Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Differential Fertility Behavior And Values In Rural And Semi-Urban Costa Rica
(USC Thesis Other)
Differential Fertility Behavior And Values In Rural And Semi-Urban Costa Rica
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
DIFFERENTIAL FERTILITY BEHAVIOR AND VALUES IN RURAL AND SEMI-URBAN COSTA RICA by Jose-Maria Blanch A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (Sociology) February 1973 INFORMATION TO USERS This dissertation was produced from a microfilm copy of the original document. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of the original submitted. The following explanation of techniques is provided to help you understand markings or patterns which may appear on this reproduction. 1. The sign or "target" for pages apparently lacking from the document photographed is "Missing Page(s)". If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting thru an image and duplicating adjacent pages to insure you complete continuity. 2. When an image on the film is obliterated with a large round black mark, it is an indication that the photographer suspected that the copy may have moved during exposure and thus cause a blurred image. You will find a good image of the page in the adjacent frame. 3. When a map, drawing or chart, etc., was part of the material being photographed the photographer followed a definite method in "sectioning" the material. It is customary to begin photoing at the upper left hand corner of a large sheet and to continue photoing from left to right in equal sections with a small overlap. If necessary, sectioning is continued again — beginning below the first row and continuing on until complete. 4. The majority of users indicate that the textual content is of greatest value, however, a somewhat higher quality reproduction could be made from "photographs" if essential to the understanding of the dissertation. Silver prints of "photographs" may be ordered at additional charge by writing the Order Department, giving the catalog number, title, author and specific pages you wish reproduced. University Microfilms 300 North Zeeb Road Ann Arbor, Michigan 48106 A Xerox Education Company 1 73-7243 I BLANCH, Jose-Maria, 1929- I DIFFERENTIAL FERTILITY BEHAVIOR AND VALUES I IN RURAL AND SEMI-URBAN COSTA RICA. f University of Southern California, Ph.D., ! 1973 Sociology, demography i I University Microfilms, A XERO X Company, Ann Arbor, Michigan | UNIVERSITY OF SOUTHERN CALIFORNIA THE GRADUATE SCHOOL UNIVERSITY PARK LOS ANGELES, CALIFORNIA 9 0 0 07 This dissertation, written by Jose-Maria Blanch under the direction of h.is... Dissertation Com mittee, and approved by all its members, has been presented to and accepted by The Graduate School, in partial fulfillm ent of requirements of the degree of D O C T O R O F P H IL O S O P H Y Da t e DISSERTATION COMMITTEE PLEASE NOTE: Some pages may have indistinct print. Filmed as received. University Microfilms, A Xerox Education Company ACKNOWLEDGMENTS I would like to acknowledge the great assistance given by Dr. Judith Friedman in the preparation of this dissertation. Her criticisms and suggestions were of great value. The support and guidance of Dr. Maurice Van Arsdol, Jr. since the very beginning of my doctoral studies have also been invaluable. My wife, Elsie, in spite of long months of suffering caused by a tragic automobile accident, has always been an encouraging presence. TABLE OF CONTENTS ACKNOWLEDGMENTS ................................... ii LIST OF TABLES..........................................vii LIST OF FIGURES.......................................xiii Chapter I. STATEMENT OF THE PROBLEM...................... 1 Salience of the Problem Pertinence of Costa Rica as a Place of Investigation Plan of the Study II. THEORY AND HYPOTHESES.......................... 20 Groups of Factors Social Background Factors I. Couple's Education II. Couple’s Residential Background Present Social Factors III. Couple's Residential History IV. Husband's Income V. Husband's Employment Status VI. Housing Conditions Biopsychological Factors VII. Coital Frequency VIII. Extended Family Living Experience IX. Attitude toward Family Planning X. Use of Contraception Measures of Fertility Chapter Page Specific Hypotheses Relevant to the Research Problem III. DATA AND METHODOLOGY....................... 51 Source of Data Temporal Sequence The Data Response Rate Accuracy of Data Sampling Errors Methods of Analysis Comparison of Means Pearsonian Product Moment Correlation Multiple Classification Analysis Summary IV. GENERAL FERTILITY DIFFERENTIALS............ 72 Fertility Level of Costa Rican Wives Differences in Three Fertility Measures Fertility Differentials by Social Background and Present Social Characteristics Education Residential Background and History Husband * s Income Husband’s Employment Status Housing Conditions Fertility Differentials by Biopsychological Characteristics Extended Family Living Experience Coital Frequency Attitude toward Family Planning Use of Contraception iv Chapter Page Summary and Conclusions V. RELATIONSHIPS AMONG INDEPENDENT VARIABLES . . 96 Relationship among Factors of Similar Character Relationship among Factors of Different Character Between Social Background and Present Social Between Biopsychological and Others Summary and Conclusions VI. SOCIO-STRUCTURAL CORRELATES OF FERTILITY . . 106 Past Fertility Social Background Characteristics Discussion Present Social Characteristics Discussion Desired Fertility Social Background Characteristics Present Characteristics Discussion Ideal Fertility Social Background Characteristics Present Social Characteristics Discussion VII. THE ROLE OF THE BIOPSYCHOLOGICAL CHARACTERISTICS ................... 149 Past Fertility Discussion Desired Fertility v Chapter Discussion Ideal Fertility Discussion VIII. SUMMARY AND IMPLICATIONS . . . . Summary Implications Fertility Control Evaluation of the Present Study Causal Order Other Variables Additivity Assumption Conclusion BIBLIOGRAPHY ........................... APPENDICES ............................. LIST OF TABLES Table Page 1.1 Costa Rica: Population, Crude Birth Rates, Crude Death Rates, Total Fertility Rates, and Rates of Natural Increase, 1940-1968 . 12 1.2 Costa Rica: Age-Specific Fertility Rates, 1960-1968 ......................... 13 1.3 Percent Change in Age-Specific Fertility Rates, Costa Rica, 1960-1968 ............ 13 1.4 Demographic and Socioeconomic Indicators, Central American Countries, 1960 and 1970. 15 1.5 Social and Economic Development in Costa R i c a .................... 15 4.1 Gross Reproduction Rates of Costa Rica and Selected Countries............ 75 4.2 Crude Birth Rates of Costa Rica and Selected Countries ....................... 75 4.3 Number of Live Births, Wanted-Expected Family Size, and Ideal Fertility Size for Costa Rican Wives, Taiwanese Wives, and White American Wives by two Broad Age Groups................ 78 4.4 Mean Number of Live Births, Wanted Family Size, and Ideal Family Size by Couple's Education, for Wives Age 20-29 and 30-49 . 84 4.5 Mean Number of Live Births, Wanted Family Size, and Ideal Family Size by Couple's Residential Background and Residential History, for Wives Age 20-29 and 30-49 . . 84 vii Table Page 4.6 Mean Number of Live Births, Wanted Family Size, and Ideal Family Size by Husband's Income and Employment Status, for Wives Age 20-29 and 30-49............. 87 4.7 Mean Number of Live Births, Wanted Family Size, and Ideal Family Size by Housing Conditions, for Wives Age 20-29 and 30-49. 87 4.8 Mean Number of Live Births, Wanted Family Size, and Ideal Family Size by Extended Family Living Experience and Coital Frequency, for Wives Age 29-29 and 30-49 . 91 4.9 Mean Number of Live Births, Wanted Family Size, and Ideal Family Size by Attitude toward Family Planning and Use of Contraception, for Wives Age 20-29 and 30-49 91 5.1 Zero-Order Correlation Coefficients among Social Background, Present Social, and Biopsychological Characteristics, for Women Age 30-49 99 6.1 Multiple Correlation Coefficients (R) between Three Fertility Measures and Groups of Socio-structural Indicators, Rural Costa Rican Couples with Wives 30-49 .......... 113 6.2 Multiple Classification Analysis Relating Number of Live Births to Social Back ground and Present Social Characteristics, Rural Costa Rican Couples with Wives 30-49 115 6.3 Multiple Classification Analysis Relating Number of Live Births to Social Back ground Characteristics by Groups of Indicators, Rural Costa Rican Couples with Wives 30-49 .................... 116 6.4 Number of Live Births by Wife's and Husband's Education and Residential Back ground, Rural Costa Rican Couples with Wives 30-49 118 viii Table Page 6.5 Number of Live Births by Couple's Education and Residential Background and Present Social Characteristics, Rural Costa Rican Couples with Wives 30-49 . ............... 119 6.6 Multiple Classification Analysis Relating Number of Live Births to Present Social Characteristics by Groups of Indicators, Rural Costa Rican Couples with Wives 30-49 123 6.7 Number of Live Births by Couple's Residential History and Housing Conditions and Other Present Social Characteristics, Rural Costa Rican Couples with Wives 30-49 129 6.8 Number of Live Births by Husband's Income and Employment Status, Rural Costa Rican Couples with Wives 30-49 ................... 129 6.9 Multiple Classification Analysis Relating Wanted Family Size to Social Background and Present Social Characteristics, Rural Costa Rican Couples with Wives 30-49 . . . 133 6.10 Multiple Classification Analysis Relating Wanted Family Size to Social Background Characteristics by Groups of Indicators, Rural Costa Rican Couples with Wives 30-49 135 6.11 Wanted Family Size by Wife's and Husband's Education and Residential Background, Rural Costa Rican Couples with Wives 30-49 135 6.12 Wanted Family Size by Couple's Education and Residential Background and Present Social Characteristics, Rural Costa Rican Couples with Wives 30-49 ......................... 136 6.13 Multiple Classification Analysis Relating Wanted Family Size to Present Social Characteristics by Groups of Indicators, Rural Costa Rican Couples with Wives 30-49 138 ix Table Page 6.14 Multiple Classification Analysis Relating Ideal Family Size to Social Background and Present Social Characteristics, Rural Costa Rican Couples with Wives 30-49 142 6.15 Multiple Classification Analysis Relating Ideal Family Size to Social Background Characteristics by Groups of Indicators, Rural Costa Rican Couples with Wives 30-49 142 6.16 Ideal Family Size by Wife's and Husband's Education and Residential Background, ' Rural Costa Rican Couples with Wives 30-49 144 6.17 Ideal Family Size by Couple's Education and Residential Background and Present Social Characteristics, Rural Costan Rican Couples with Wives 30-49 ................ 144 6.18 Multiple Classification Analysis Relating Ideal Family Size to Present Social Characteristics by Groups of Indicators, Rural Costa Rican Couples with Wives 30-49 146 6.19 Ideal Family Size by Couple's Residential History and Housing Conditions and Other Present Social Characteristics, Rural Costa Rican Couples with Wives 30-49 . . . 146 6.20 Ideal Family Size by Husband's Income and Employment Status, Rural Costa Rican Couples with Wives 30-49 ................. 146 7.1 Multiple Classification Analysis Relating Number of Live Births to Biopsychological Characteristics, Rural Costa Rican Couples with Wives 30-49 ................. 153 7.2 Multiple Classification Analysis Relating Number of Live Births to Social Background and Present Social Characteristics by Family Planning Variables, Rural Costa Rican Couples with Wives 30-49 ........ 154 x Table Page 7.4 7.5 7.6 7.7 7.8 7.9 7.10 Number of Live Births by Couple's Education and Residential Background and Biopsy chological Characteristics, Rural Costa Rican Couples with Wives 30-49 .......... Number of Live Births by .Couple's Residen tial History and Housing Conditions and Biopsychological Characteristics, Rural Costa Rican Couples with Wives 30-49 . . . Number of Live Births by Husband's Income and Employment Status and Biopsychological Characteristics, Rural Costa Rican Couples with Wives 30-49 ................. Multiple Classification Analysis Relating Three Fertility Measures of Fertility to Family Planning Variables by Various Socioeconomic Indicators, Rural Costa Rican Couples with Wives 30-49 .......... Multiple Classification Analysis Relating Wanted Family Size to Biopsychological Characteristics, Rural Costa Rican Couples with Wives .30-49 ................. Multiple Classification Analysis Relating Wanted Family Size to Social Background and Present Social Characteristics by Family Planning Variables, Rural Costa Rican Couples with Wives 30-49 . . . Wanted Family Size by Couple's Education and Residential Background and Family Planning Variables, Rural Costa Rican Couples with Wives 30-49 ................. Wanted Family Size by Family Planning Variables and Present Social Characteristics, Rural Costa Rican Couples with Wives 30-49 ................. 154 155 155 156 160 160 161 162 xi Table Page 7.11 Multiple Classification Analysis Relating Ideal Family Size to Biopsychological Characteristics, Rural Costa Rican Couples with Wives 30-49 ................ 167 7.12 Multiple Classification Analysis Relating Ideal Family Size to Social Background and Present Social Characteristics by Family Planning Variables, Rural Costa Rican Couples with Wives 30-49 ................ 168 7.13 Ideal Family Size by Couple's Education and Residential Background and Biopsychologi cal Characteristics, Rural Costa Rican Couples with Wives 30-49 ................ 170 7.14 Ideal Fertility by Couple's Residential History and Housing Conditions and Bio psychological Characteristics, Rural Costa Rican Couples with Wives 30-49 . . . 171 7.15 Ideal Family Size by Husband's Income and Employment Status and Biopsychological Characteristics, Rural Costa Rican Couples with Wives 30-49 ................ 172 7.16 Percent by which the Effect of Various Social Background and Present Social Characteristics on Ideal Family Size is Reduced when Adjusted for the Effect of a Family Planning Variable . . . 174 8.1 Summary of Hypotheses and their Confirmation Status ..................... 182 xii LIST OF FIGURES Figure Page 5.1 Associations among Pairs of Variables . . . 105 CHAPTER I STATEMENT OF THE PROBLEM I This study intends to explore the relationship between social structure and fertility level in rural Costa Rica. Fertility in this country is still high but signifi cant changes both in demographic and social structure have been taking place in recent years. The main analysis is focussed on the problem of how the demographic and social characteristics influence family planning and therefore fertility in rural Costa Rica today. Salience of the Problem Fertility is a complex phenomenon. The factors that affect human fertility are numerous and complex. Human fertility is a biological fact, but it is also, to a large extent, a social phenomenon. Social factors play a significant role in the determination of the biological result. This seems to be so even in primitive and developing countries where, in the absence of modern con traceptive devices, biological factors probably are more important in determining fertility levels, than they are in developed countries. Davis and Blake (1956:211) remark: A striking feature of underdeveloped areas is that virtually all of them exhibit a much higher fertility than do urban-industrial societies. This well-documen ted but insufficiently analyzed fact is known to be connected with profound differences in social organiza tion as between the two types of society, and is, therefore, significant for the comparative sociology of reproduction. The clarity and importance of the con trast, however, should not be allowed to obscure the equally important fact that underdeveloped areas them selves differ markedly in social organization, and that these differences appear to bring about variations in fertility. Sociologists usually approach the analysis of fertility from two major perspectives--social psychologi cal and structural. In a social psychological perspective emphasis is upon individual characteristics which affect fertility. In a structural perspective analysis links changes in the properties of entire systems to changes in fertility. Social and economic factors seem to play a very important role in producing fertility differentials. A study conducted by the United Nations (1965) reveals that a large array of measures of economic and social development markedly distinguishes the high from the low fertility countries in the world; when all countries are considered, a strong inverse association between fertility and each of the social and economic indicators is found. If we turn our attention to what happens if individuals and not nations are the subject of research, many studies indicate that a strong relationship exists between fertility level and both area characteristics and the socioeconomic status of individuals in a society (cf. Freedman et al., 1959; Whelpton et al.. 1966; Westoff et al.. 1961; Heer and Turned, 1965; Takeshita, 1962; Collver et al.. 1967; Sun, 1969). The direction and strength of association between fertility and socioeconomic variables reported in these studies varies from society to society and from time to time in a society. The overwhelming majority of studies show an inverse relationship between socioeconomic status variables and fertility, although some exceptions to the inverse pattern have been observed both in relatively high and relatively low fertility societies. (Examples include: high fertility societies: Hawley, 1955; Zikry, 1965; low fertility societies: Edin and Hutchinson, 1935; Freedman and Slesinger, 1960; Goldberg, 1960; Kiser and Whelpton, 1949). This seems in agreement with the two competitive models relating fertility to economic variables. One approach treats economic development as an inhibiting factor in fertility. The Malthusian tradition maintains that economic development raises fertility because it allows to support larger families. Many efforts have been made to relate psychological factors to fertility differences, with different success. Kiser and Whelpton posed the problem (1953:108): The Indianapolis study presents a challenge to learn the reasons for the overriding influence of socio economic status. There is good reason to believe that it is not socioeconomic status per se but rather the underlying attitudes and psychological characteristics of these classes that account for the fertility behavior. And yet, whereas characteristic patterns of fertility differentials are formed consistently in classifications by socioeconomic status, most classifi cations by psychological characteristics within socio economic groups fail to show such patterns. Studies conducted since the Indianapolis study usually have discovered low correlations between psycholog ical variables and fertility, even when fertility is measured as "desired size of family," a measure which would seem to be at a level more relevant to a psychological analysis. (Some instances are: Westoff et al., 1958; Westoff et al.. 1963; Goldberg, 1960; Westoff et al., 1961). More recently, however, large scale studies of fertility and family planning in Malaysia (Palmore, 1969) and Korea (Bom et al., 1972) among others have found that psychological factors and processes significantly affect the fertility behavior of the individual. The fact that so many studies failed to discover significant associations between psychological variables and fertility seems rather perplexing. It is not education as such that prompts an individual to act in an educated way, but it is "something" that being educated for a cer tain length of time has created in the individual (ways of perceiving, feeling, thinking, and acting). Similarly, suddenly transplanting a country man or woman into an urban area does not make him or her behave like an urbanite overnight. Urban behavior is determined by that "some thing" in the individual created through living in an urban area for a certain length of time as well as by the environmental stimuli. Perhaps the findings of insignifi cant relationships between psychological factors and fertility behavior in past studies may be attributed to either inappropriate conceptualization of the psychological variables, or inadequate measurement of psychological variables, or to inappropriate sampling of the subjects, a sampling which has too little variation because it is restricted to a narrow segment of the population. During development, societies change from being largely rural, traditional, nonindustrial, and kinship oriented to being largely urban, industrial, market- oriented. A significant fertility decline during this period is expected as an adjustment made in the society to the new life conditions. It has been noted that the family size required or preferred in a modern society is much smaller than that in a traditional society. The crucial point, however, is that, in the process, of adjustment significant intra-societal differences appear. People in certain social strata tend to react to the change in life conditions earlier and faster and those in different strata react later and more slowly or differently. It is hypothesized that, basically, it is the social structure that has a determinative differential impact in adjusting 6 to the new situation. Family limitation per se does not determine fertility levels but there are more ultimate causes which affect fertility perhaps indirectly through family limitation. Several studies have argued that such "causes" are attitudinal (Kiser and Whelpton, 1949; Goldberg, 1960; Freedman, 1961-62). This study will argue that fertility levels are to be explained by social fac tors, which may act through biological and psychological factors. Kitagawa and Hauser (1964) note the importance of the social in contrast with the biological and the consequent conjunction of demography with sociology. The study of fertility transcends matters of immediate demographic concern, since fertility is a function of the social milieu in which it occurs, even though it possesses important biological components. Hauser also remarks: The failure of the above mentioned studies (Indianapolis Study) to explain a greater proportion of variance in fertility is attributable to the fact that they have ignored the culturalogical in the sense in which Durkheim treated this subject. Durkheim’s insistence that the 'social fact’ is anterior and ex terior to the individual while exerting constraint upon his behavior may contain the clue to an important missing ingredient in fertility research. That is, the proposed design of research (which Hauser outlines in his article) is based on the assumption that fertility behavior is in large measure dependent upon the social milieu, and that changes in fertility behavior neces sarily involve social change. Or, put in another way, knowledge of the person's attitudes, values, and motivation cannot be expected to account for differ ences in fertility behavior out of their cultural con text; and, consequently, changes in fertility behavior cannot be produced through efforts to change attitudes, values, or motivation, except in the context of changes in thd social order. (1960:464-5). It seems legitimate to assume that fertility behavior is largely a social phenomenon. It is recognized, however, that within a sociological framework, the biological, psychological, and attitudinal can be viewed as "intermediate" variables with respect to. their influence in fertility. In this research, the analysis of fertility behavior will be within the realm of the sociological in an attempt to clarify in which way the variations within the social structure of a rural area of a developing country with high fertility lead, to differential fertility attitudes and behavior. Some biopsychological variables will be included as "intermediate" to explore whether or not they impinge upon reproductive behavior. This approach has been taken in a few instances in the study of fertility in less developed areas, such as Korea, Taiwan, Puerto Rico, and Malaysia. Most of the independent variables in fertility studies are overlapping in content and in their effect on fertility, but few attempts have been made to isolate the net effect of each variable. Recent studies reveal that the relationships between status variables and fertility become weaker or even change direction if the effects of other variables are held constant (cf. Freedman and Slesinger, 1960; Freedman et al.. 1959; Kuntz, 1965). For example, Goldberg (i960), after examination of the Indianapolis data, demonstrated that the original inverse relationships were found among only the farm migrants. Others have replicated these findings at the national level (cf. Duncan, 1965; Freedman and Slesinger, 1960). Analysis of the direct observed relationship between any one variable and fertility without controlling for the effect of other variables may lead to serious misinterpretations and contribute little to our understanding of how fertility differentials relate to socio-structural factors. Most of the studies in the less developed countries have considered one variable at a time, even though in recent years some have taken the multivariate approach (cf. Yankey, 1961; Takeshita, 1963; Rick, 1959; Sun, 1969). Comparatively little attention has been devoted to measur ing the amount of variance in fertility explained by one or a group of variables (for example, Freedman, 1963; Freedman et al., 1964; Freedman and Takeshita, 1965; Freedman and Coombs, 1966; Sun, 1969; Bom et al. . 1972). What has been done, however, is important on both theoretical and practical grounds. Our investigation will deal with a group of rural and semi-urban Costa Rican couples with wives aged 30 to 49. The data were obtained from the Costa Rican Rural Fertility Survey of 1969 which included women aged 15 to 49. Judging from the data available from the Urban Fertility Survey of 1963, conducted in the Metropolitan Area of San Jose, and from the fact that the demographic and social development of the Metropolitan Area is more advanced than in the rural part of the country, we can expect that fertility differentials by socio-structural factors will be somehow smaller for the rural areas of the country. The problems to be examined are the following: 1. What are the general fertility differentials by socio-structural factors such as education, residential background, residential history, income, employment status, and housing conditions, in rural and semi-urban Costa Rica at this point of time? These differentials are expected to be rather large, as many studies conducted in other countries have proved. 2. How are those socio-structural indicators related to one another? Some of these relationships are expected to be strong. 3. What are the fertility differentials by those socio-structural factors when one or a group of other socio-structural factors is taken into account? It is expected that the net fertility differentials in most of the cases will be smaller than the gross or general differentials. The analysis will discuss the independent effect of single factors, and the implications of these 10 findings. 4. How much of the variance in fertility is accounted for by each factor and by groups of factors considered in this study? 5. Do the socioeconomic characteristics of couples influence fertility behavior and values through biological and psychological factors, such as coital frequency, extended family living experience, attitude toward family planning, and use of contraception? If so, what are the channels through which this transferal of effects operates? It is expected that these biological and psychological factors will play a greater role in transferring the influence of socio-structural indicators to fertility values than to fertility behavior. The theoretical background, the scheme of analysis, and the hypotheses are discussed in detail in Chapter II. Pertinence of Costa Rica as a Place of Investigation The population trend in Costa Rica in the last thirty years can be summarized saying that it has shown a persistent high fertility until recent years with increasingly low mortality. Fertility began a slow but consistent decline in 1960, reaching 36.2 births per thousand in 1968. These trends produced a rate of natural increase almost in the shape of a normal curve peaking in 1959. Table 1.1 shows this trend clearly. The declining trend of fertility since 1960 is evident from Table 1.2, although such a trend does not necessarily guarantee a continued future decline. Past experience, however, suggests that certain general condi tions are favorable for a continuing fertility decline during the early .stages. Freedman (1965) has suggested six such pre-conditions under which fertility in a high fertility country might be expected to decline first and most rapidly. These include existence of significant social development, relatively low mortality from some time past, evidence of desire for a moderate family size and attempts at limiting family size by a part of the popula tion, effective social networks transcending local commun ities for the dissemination of family planning ideas and other modernizing influences, large-scale effective organized efforts for communicating family planning infor mation and availability of modern contraception devices to the general public. The first four of these conditions perhaps were relevant in the past as they probably are at present, while the last two have become relevant only very recently in a few countries, including Costa Rica, with the institution of an official family planning program and large scale diffusion of modern contraceptives. In the last 15 years, Costa Rica has undergone a steady social and economic development. A general 12 TABLE 1.1 COSTA RICA: POPULATION, CRUDE BIRTH RATES, CRUDE DEATH RATES, TOTAL FERTILITY RATES, AND RATES OF NATURAL INCREASE, 1940-1968 Year Population as of July 1 (in 1,000) TFR CBR CDR Natural In crease 1940 654 43.7 18.8 24.9 1945 740 44.9 16.0 26.9 1950 859 45.2 12.5 32.7 1955 1,028 48.9 10.7 38.2 1956 1,069 48.5 9.8 33.7 1957 1,110 47.8 10.4 37.5 1958 1,153 47.2 9.2 38.0 1959 1,200 48.3 9.3 39.0 1960 1,254 7335 47.5 8.9 38.6 1961 1,298 7425 46.9 8.2 38.7 1962 1,343 7240 45.4 8.9 36.5 1963 1,391 7120 45.3 9.0 36.3 1964 . . 1,493 6650 43.0 9.4 33.6 1965 1,490 6545 42.3 8.6 33.7 1966 1,541 6310 40.9 7.4 33.5 1967 1,590 5970 39.0 7.1 32.8 1968 1,634 5525 36.2 6.5 29.7 S / Source: Ricardo Jimenez, Estadisticas Demoqraficas Basicas de Costa Rica, 1970. San Jose, Costa Rica: Asociacion Demografica Costarricense, 1970. 13 TABLE 1.2 COSTA RICA: AGE-SPECIFIC FERTILITY RATES, 1960-1968 Age Groups 1960 1961 1962 1963 1964 1965 1966 1967 1968 15-19 115 118 116 118 112 112 111 110 107 20-24 353 348 336 330 307 302 290 278 261 25-29 371 368 360 350 325 320 302 281 255 30-34 304 308 300 289 268 258 246 230 216 35-39 218 226 222 225 213 212 211 195 174 40-44 90 101 98 96 89 89 87 86 80 45-49 16 16 16 16 16 16 15 14 12 Ibid. TABLE 1.3 PERCENT CHANGE IN COSTA AGE-SPECIFIC FERTILITY RICA, 1960-1968 RATES, Age Percent Chanqe Groups 1960-1966 1960-1968 15-19 -3.5 -7.0 20-24 -17.8 -26.1 25-29 -18.6 -31.3 30-34 -19.1 -28.9 35-39 -3.2 -20.2 40-44 -3.3 -11.1 45-49 -6.2 -25.0 14 overview of the demographic and socioeconomic conditions of the five Central American countries, as Table 1.4 offers, suggests that Costa Rica is in a more advanced stage of socioeconomic change and development than any of the other Central American countries. The social and economic development can be measured in terms of urbanization, literacy and education, propor tion of the labor force in non-primary sectors, and the like. Between 1960 and 1968, the per capita national income increased from 2,206 colones to 3,327, an increase of 33 percent.* This overall increase in national income has occurred along with structural changes in economy, such as improvement in agricultural technology, development of private industry, and increased rate of investment. Agriculture continues to be the dominant sector (Table 1.5). By 1967, 53 percent of the labor force was employed in the non-primary sectors. This economic development has been also accompanied by several other changes in the society, such as urbaniza tion, rising female employment, an increase in educational level, development of mass media, and improvement in sanitation, and health conditions. Such rapid changes *No major devaluation of the national currency has taken place in the last decade. The rate of exchange has been a steady 6.65 colones to one US$. The purchasing power of the colon, however, has slightly decreased. TABLE 1.4 DEMOGRAPHIC AND SOCIOECONOMIC INDICATORS, CENTRAL AMERICAN COUNTRIES, 1960 and 1970 Indicator Year Costa Rica El Salvador Guatemala Honduras Nicarac Percent 1960 32 32 29 21 35 Urban 1970 33.6 38 31 28 40 Density 1960 26 120 39 17 11 (km2) 1970 35 164 48 23 14 Dependence 1960 104 92 105 101 105 Ratio 1970 104 100 94 96 101 Infant Mortality 1960 70 70 88 52 54 (per 1,000) 1970 70 63 92b 44c 55b Child-Woman 1960 875 739 781 857 824 Ratio 1970 907 899 798 837 837 Life 1960 64 50 49 44 38 Expectancy 1970 67 56 49. 50 50 Literacy Rate 1960 84.4 49 38 45 50 (over 15) 1970 ------ ------ — — — Retention Coeffi cient (elementary 1960 35 21 16 14 10 education) 1970 40 25 18 15 12b Physicians 1960 5 2 2 2 4 (per 10,000) 1970° 5 2 2b 2 4b Per capita GNP 1958 353 238 268 196 255 (in US$) 1967 423 280 300 236 359 Retention Coeffi cient (elementary 1960 35 21 16 ’ 14 10 education) 1970 40 25 18 15 12b Physicians 1960 5 2 2 2 4 (per 10,000) 1970° 5 2 2b 2 4b Per capita GNP 1958 353 238 268 196 255 (in US$) 1967 423 280 300 236 359 Active Population 1960 50 61 66 70 60 (primary activities)1970 47b 59b 62 70b 58b a = 1965 b = 1966 c = 1967 Source: Datos Basicos de Poblacion en America Latina 1970. Washington, D. C.: Organizacion de Estados Americanos (OEA), 1970 Note: The reliability of the figures above cannot be verified. TABLE 1.5 SOCIAL AND ECONOMIC DEVELOPMENT IN COSTA RICA Indicator 1963 Later Year Living in all cities 32% 33.6%a Labor force in non-primary sectors 50% 53%b New industries 83 130 c Per Capita national income (in colones) 2.490 3.327 a a = 1970 b = 1967 c = 1968 16 probably have strong influences on the demographic situation. The improvement in health, largely attributable to the use of antibiotics and powerful insecticides and to the creation of a nationwide Social Security Service, has brought the death rate down to a very low level (6.5 per thousand in 1968). The reduction of the death rate with out, however, a corresponding reduction of the birth rate caused an increase in population, from a rate of natural increase of 26.9 in 1845 to a high 39 in 1959. Such rapid population growth in the 50’s and early part of the 60*s seemed to influence people's fertility behavior; the birth rate began to decline in 1960. The Total Fertility Rate decreased 24 percent between 1960 and 1968, and the decrease has been observed also in 1969. Marked declines have occurred for women not only in the ages over 30, but also for those in the younger ages, contrary to the pattern of decline observed in a number of Western countries at the beginning of their fertility decline (cf. Freedman and Adiakha, 1968). Before the Urban Fertility Survey in 1963, there had been only a moderate amount of family planning activi ties in Costa Rica, and these activities were restricted private groups. Family planning activities were officially sponsored by the Costa Rican government in the middle of the sixties. The fertility decline seems to have been both a process of innovation and adjustment to a specific 17 situation of high population pressure or extreme poverty. This fertility decline started before any systematic efforts in family planning came to existence. This natural situation, on the one hand, and the fact that family planning programs were developed in the second half of the sixties, on the other, makes it appropriate for the analysis of the influence of social, structural, and bio- psychological characteristics on fertility behavior and values. The fertility decline started once the country had reached a moderately high economic and social development and is a part of a whole process of modernization and social change which is affecting Costa Rican society at large. Family planning programs and availability of con traceptives may have helped to increase the rate of decrease. (See Reynolds, 1972).* The rural and semi-urban areas of the country share the general characteristics described above. The literacy rate, for example, even if below the national average, is still above 70 percent. The health services have been expanded in the last few years through a network of hospi tals, clinics, and emergency centers and medical mobile *Reynolds gives the following information: the estimated number of women who used family planning services was 3,074 out of 324,885 aged 15-49 (.95 percent) in 1966, and 17,175 out of 359,816 (4.77 percent) in 1969. In 1971, the percent of users had reached almost 10 percent of the total female population in the 15 to 49 age bracket. 18 units, as to cover the greater part of the country. The largest proportion of the rural and semi-urban population are still employed in agricultural or agriculture related jobs, even though an increase in service related occupa tions is detectable and small industries have been developed in semi-urban areas with good transportation services. In sum, Costa Rica is a high fertility area where the beginning of a possible secular fertility decline is in evidence. Mortality has been low for some time, and con siderable social and economic development relevant to fertility decline has taken place since the late 50's. Costa Rica can be considered as located at the second stage of the "demographic transition," a critical period for fertility change. As such, Costa Rica, and specifically its rural and semi-urban areas, provide us with an appro priate area to do the kind of study proposed here. Plan of the Study In the following chapters the theoretical frame work of this study will be presented and the specific hypotheses to be tested will be spelled out. The data and their accuracy will be examined next, and the methods of analysis to be used will be discussed. In Chapter IV through VI, the general fertility differentials, the relationships among the independent variables, and the 19 socio-structural correlates of fertility behavior and values will be examined in detail. Finally, in Chapter VII the analysis will deal with the role of the bio- psychological characteristics. A summary and a discussion of the implications of this study will be the theme of the last chapter. CHAPTER II THEORY AND HYPOTHESES To better approach the analysis of the problems enumerated in Chapter I, the writer will present in this chapter a conceptual scheme taking into account the avail ability of data and the social structure of Costa Rica. It is a well supported fact that fertility behavior is highly correlated with socioeconomic status. The relationship, however, differs among societies at various stages of development and at different times. Within the social and economic structure of a society there probably are mechanisms through which socio-structural factors influence fertility behavior and values. This study is an attempt to specify the structure of the social, economic, and psychological factors suggested in several models in regard to their influence on fertility behavior (see Freedman, 1961-1962). The basic hypothesis underlying the scheme is that the social and structural characteristics have marked differential effects on fertility and that their main influence is exerted directly, even if there may also be indirect influence through biological and psycho logical characteristics. 21 In order to classify the socio-structural factors, the writer has considered both the stability of the given factor over time and the time sequence in which a charac teristic is acquired, and in which it is likely to affect fertility. Those characteristics acquired before marriage and stable over time are taken as "background." They are also differentiated from those characteristics which are acquired either before or during marriage and are less stable over time. These will be called "present" characteristics. The biopsychological factors constitute a third group; these may rest on the same time base as the preceding characteristics or may be temporarily posterior. According to our scheme the effect of the social background factors on fertility is exerted: (l) directly, (2) indirectly through the present factors, (3) indirectly through the biopsychological factors. The present factors act as intervening variables; they transfer the effect of the social background characteristics to fertility either directly or through the biopsychological factors. The present factors may also act as source variables; they have an effect on fertility independent of the background variables and exert this effect in part through bio psychological factors. The biopsychological factors also act as intervening variables, even if they are also expected to have a direct effect on fertility. 22 The model can be depicted in a diagram as given below. Present Factors Fertility Social Background Factors Biopsycho logical Factors Each of these groups contains some specific variables. The specific factors selected depend in part on the nature of the study and the availability of data. In this analysis only women currently married are taken into account. Most of the "intermediate variables" sug gested by Davis and Blake are not included in this study either because of lack of available information or for the sake of simplicity. Thus variables related to the forma tion and dissolution of unions--such as permanent celibacy and amount of time spent after or between unions— are excluded. Age at union is included as a control variable, despite the fact that an earlier study found it unrelated to the fertility decline in Costa Rica (Gomez, 1970). The variables related to voluntary control of marital fertility are family limitation and coital frequency. In this context, the fertility behavior is largely the result of the couple’s decision about family size and the availability of means of fertility control. 23 The decision regarding the size of the family is very com plex. Consciously or unconsciously the pressure of social environment, the norms about family size, their own economic capacity, affect the couple's decision. Several factors are included in this analysis which seem relevant to this problem, such as wanted family size, family structure, income, husband's employment status, living con ditions, educational background, residential history and background. On the other hand, the access to modern methods of family limitation depends on several factors such as educational level and source of supply.* The following factors are, thus, included (abbre viations for each factor follow); I. Social Background Factors 1. Couple's Education (CE) 2. Couple's Residential Background (CRB), e.g., both husband and wife rural, one rural and one from village, etc. *In regard to contraceptives' availability, there have been two external sources: the official family planning program and the private (commercial) sector. The latter contributed the bulk of the contraceptives until 1966. Between 1959-1969, the private sector contributed about 55 percent compared to 45 percent for the family planning program (73 percent of the oral pills). The statistics for the oral pills distribution by the family planning program are as follows: 5,388 cycles in 1966; 27,351 in 1967; 65,963 in 1968; and 134,926 in 1969 (Reynolds, 1972). II. Present Social Factors 3. Residential History (RH) 4. Husband's Employment Status (HES) 5. Husband's Income (Hi) 6. Housing Conditions (HC), e.g., very poor, comfortable, etc. III. Biopsychological Factors 7. Coital Frequency (CF) 8. Extended Family Living Experience (EF) 9. Attitude toward Family Planning (AP) 10. Use of Contraception (UC) IV. Fertility Variables 11. Number of Live Births (LB) 12. Wanted Family Size (WFS) 13. Ideal Family Size (IFS) V. Control Variables 14. Current Age (CA) 15. Age at Union (AU) 16. Duration of Union (DU) This scheme seems adequate for Costa Rica's current stage of development. It is assumed that there are fer tility differentials in Costa Rica (Gomez, 1968; Blanch, 1972), and that these differences are accounted for mainly by the practice of family limitation among certain groups of couples. Costa Rica has been in a modernization trend which implies, among other things, the adoption of norms of 25 smaller family size than before and of modern methods of family living. The question that has to be answered is who is more likely to adopt those new norms and means of achieving these norms. Many studies have tried to answer this question and the prevalent conclusion is that those who have a higher socioeconomic status are more likely than those with lower status to do so (cf. Stycos, 1963; Stycos, 1963; Vazquez, 1968). Here, an attempt is made to group factors according to the causal relationship they have in influencing a couple's fertility behavior. One possible answer is that those with a background that includes, for example, a higher level of education tend to reach a higher status, indicated by a better occupation and income, and are more likely to adopt modern norms of family size. At higher income levels people may be culturally constrained to expend more on children than they are at lower levels, and therefore they can afford fewer children. They are, consequently, the most likely to limit family size and have smaller families. Other possible channels can be followed with the combinations of factors listed above. This kind of analysis may be important both on theoretical grounds and in terms of practical guidance in family planning programs in countries in a similar stage of development as Costa Rica. It may lead to conclusions as to what social background factors are more likely to affect 26 fertility behavior and through what other factors. Furthermore, the estimation of the individual effect of the factors located at the end of a channel will indicate what can be expected from any effort to improve these factors without improving the more remote factors. Groups of Factors The socio-structural characteristics are divided into two classes: social background and present social factors. The third class is formed by the biopsychological factors. Social Background Factors The social background factors are those already attained at the time of marriage and are less likely to be modified later. In this sense these factors can be consi dered as fixed characteristics. In our scheme these factors are those which basically influence fertility. What is going to be explored is whether or not they do influence fertility and, if they do, how much their influ ence is exerted through the other factors. Two variables are included in this group, namely, couple’s education and residential background. I. Couple's Education Not only the literacy rate in Costa Rica has been 27 for years much higher than that of other developing countries, both in Latin America and in other geographic areas of the world, but the average educational level has been rising in recent years. At present the younger people are better educated than the older ones and the men than the women. Educational level will be measured by the final level reached. It is well known and supported by many studies (cf. Grabill et al., 1958; Duncan, 1965 for the United States; Edin and Hutchinson, 1935; Moberg, 1950 for Sweden; Stycos, 1968; Miro and Mertens, 1968 for Latin America) that level of education is inversely related to fertility, even in the period of transition from high to low fertility. Bogue (1969) correlated the point at which a given country was at in making the demographic transition with a series of other variables, including education and urbanization, measured by percent urban. His measures of education consistently correlated more highly with this index than did other variables, including urbanization. Thus, percent illiterate correlated -.77, percent educated at "level one" or higher +.75, and percent in school +.99. Urbanization, on the other hand, correlated +.55. Further more, when education at "level one" was controlled, the correlation with other variables and fertility was either greatly reduced or disappeared. These data led Bogue to conclude that, "a major driving force behind fertility control appears to be education” (1969:676). A possible explanation is that an increase in education tends to lead to or be associated with residence in urban or semi-urban environments, higher levels of aspirations, more rational attitudes toward family size,, and also to better access to knowledge of the idea and methods of family limitation, and serves as a basis for more effective practice of family limitation. It is not clear, however, how education affects other factors that are also related to fertility differences. Education seems to represent one of the best measures of moderniza tion, by virtue of its direct relationship with the urbanization-industrialization process of a society through the development of skills for filling the positions in the new occupations that are created by this process. Con comitantly, the more educated are differentiated from the less educated by their urban residence and involvement in a market economy, more varied sources of information and ideas, higher levels of aspirations, and rational rather than indifferent or sentimental attitudes toward fertility behavior. Freedman argues that the role of education is fundamentally basic (1963). He suggests that with increased education and literacy, the population tends to become involved with the ideas and institutions of a larger modern culture and begins to find rewards in social relationships for which larger number of children may be 29 irrelevant. This suggests the following possible channels of influence: (l) husband’s (wife's) education ---— > employment status fertility; (2) wife's (husband's) education > residential history > fertility. Education may also operate through other channels. Both the idea of smaller families and the use of contraception were relatively new in Costa Rica in 1969, when the Rural Survey was taken. The official family planning program had been in existence only since 1966. At that time, thus, it seems obvious that the acceptance of the new idea and methods probably were largely limited to those with more education, that is with at least elementary school completed. In fact, of all those who have used any means of contraception, only 16 percent were in the lower educational category (no schooling). Therefore, it is expected that at this stage of development in rural and semi-urban Costa Rica, the effect of education per se will be significant once the elementary school completion level has been reached. II. Couple's Residential Background Residential background denotes a relationship between previous residence milieu, particularly the milieu of the formative years, and current residence environment. It assumes that fertility behavior is a pro duct of social and cultural influences at both points of time. Several studies have examined aspects of the relationship of residential background and fertility, with the emphasis being placed on the influence of the area of origin. Zarate (1967) found an inverse relationship between size of birthplace and average number of children in Monterrey, Mexico. Similar findings reported Tabah and Samuel (1962) for a sample of immigrants to Greater Santiago, Chile. Other research reporting higher fertility among immigrants of rural origin includes Hutchinson's (1961) analysis of data from eight cities in southern Brasil, and Goldberg's (1959), Freedman and Slesinger (1961), and Duncan's (1965) studies of United States data, and more recently Ritchey and Stokes (1971;1972). Studies of residential background have focussed almost exclusively on the urban population and have examined the differential fertility of urban residents according to the type of their place of origin. The present study deals only with a rural and semi-urban popu lation. The place where residence was for most of the time established before age fifteen is classified in three categories: open country or farm, village, and town. It is expected, as the above mentioned studies suggest, that the more rural the place of residence before age fifteen for both the women interviewed and their husbands, the higher one comple's fertility. The influence of exposure 31 to rural milieu on fertility may be direct and also through education and socioeconomic status. Those who live in villages and towns have better educational opportunities, on the one hand, and on the other their socioeconomic attainments are likely to be higher than among those with a farm residential background. Present Social Factors Present social factors are those mostly attained after marriage and subject to change after marriage. Since their attainment depends largely on the social background factors, they may be regarded as derived. Due to their changeability, their influence on fertility cannot be measured as precisely as that of the fixed characteristics. They are usually shown as status at the time of the survey, so they may not be relevant to fertility behavior before the survey and perhaps before attainment of the given status. This limitation also applies to the group of bio psychological variables. Nevertheless, under the assump tion that changes in status and attitudes are usually slow and are likely to follow a certain pattern, they are useful in explaining fertility differentials, especially in places where the use of modern family limitation methods is new and where use occurs most often for limitation than for spacing, as it seems to be the case in Costa Rica. This group of factors intervenes between social background 32 factors and fertility, transferring the effect of the former to the latter. These present factors can also be considered as source variables that exert their effect on fertility through biopsychological factors. The variables considered here are, couple's residential history, hus band's income and employment status, and housing conditions. III. Couple's Residential History It is assumed that migration has an independent effect on fertility in the sense that either the expecta tion of the necessity for a change in residence acts as a determinant force on fertility or family size affects social and economic factors influencing migration or other socio-structural characteristics influence both migration and family size. The distinction between residential back ground and migration has not been kept clear in most of the studies on the subject (Macisco, 1965;1968; Miro and Rath, 1965). The majority of migration takes place between urban areas (or semi-urban, in the particular situation analyzed in this study); thus the proportion of migrants with urban or semi-urban background is greater than the proportion of nonmigrants with such background. This leads to the conclusion that the fertility of migrants is influenced to a greater extent by urban or semi-urban patterns of child bearing, and the fertility of nonmigrants to a greater 33 extent by rural patterns of fertility. Ritchey and Stokes (1972) have examined the problem with United States data, testing the proposition that migrants have lower fertility than nonmigrants, controlling for size of place of origin and destination. Their conclusion is that the data strongly suggested higher fertility among migrants than among nonmigrants. In the present study, the independent effect of migration or residential history on fertility is analyzed in the case of rural or semi-urban couples. The residen tial history variable is a composite index of the separate information for wife and husband available in the survey. Three categories of residential history are formed: no migration, some moves with no residence in places of over 20,000 population, and some moves with residence in one or more places of over 20,000 population. IV. Husband1s Income Husband's income is mainly an economic variable, although its association with education cannot be denied. It stands between the basic factors and the biopsychologi- cal variables, on the assumption that income may provide the resources for the realization of a more modern living that may be inconsistent with high fertility. The rela tionship between income and fertility has been one of the most explored in the study of fertility differentials. 34 Becker (i960), who favors a purely economic approach to the analysis of fertility, states that economic development is positively correlated with ferti lity. Heer and Turner (1965) were among the first to examine the implications of Becker's model for countries at different stages of development. In a study of Latin America they observed that countries in which fertility was higher that would have been anticipated on the basis of various indexes of economic development had also experienc ed recent and large increases in per head income. Coun tries in which lower than expected fertility was observed tended to have experienced unusually low increases in income. These data led Heer (1966) to test Becker's hypothesis that while the direct effect of income increase is to raise fertility, the net result may be a decrease in fertility. This is because an increase in income usually is also associated with factors whose long term effects in decreasing fertility may be stronger, unless per capita income is increasing at a very rapid rate. Heer examined this hypothesis at the aggregate level with data from forty-one nations. He found a zero-order correlation coefficient of -.45 between income and fertility, while the partial correlation with education, infant mortality, and density held constant, was +.10, a small but consistent relation. Also, the partial correlations between fertility 35 and education, infant mortality, and density were moderate but in the anticipated direction, -.24, +.42, and -.40, respectively. Other studies provide data consistent with Becker's model (Weintraub, 1962; Adelman, 1963). Gendell (1967) showed that economic development did not appreciably affect fertility in Brasil between 1920 and 1960. Fertility continued high in the face of constant growth in real per capita income, decrease in illiteracy, and a steady dis placement of her industrial structure from a primary base to a secondary-tertiary base. Friedlander and Silver (1967) working with data from countries at different levels of development found that various indexes of income tended to have positive signs in developed countries, negative in developing. All these studies deal with aggregate data. At the individual level, Bom, Palmore, Sang and Sung (1972) show that in Korea, measures of socioeconomic status are inversely correlated with fertility. Kiser, Grabill, and Campbell (1968) in United States found that for white women, married and husband present, the relationship between income and fertility was direct at ages under 25, absent at ages 25-29, and inverse at ages 30 and over. Their interpretation of this finding is that young people of high income are more likely than people of low income to concentrate their childbearing into a relatively short 36 period after marriage. It is recognized that the measure of income has to be taken with some reservations. The peasant may be reluctant to tell an outsider the family income, sometime for fear that the information might be used as a basis for taxing him. Furthermore, it is difficult to calculate the real income, since some of it may include what he consumes from his own planting and livestock. Other income may be made from occasional day labor, of which he keeps no record. Finally, the peasant's income varies considerably from season to season and from year to year. One year he may benefit from a shortage of the particular crop he plants; another year he may plant the same crop and find the market glutted, the prices a fraction of what they were; the next year, bad weather, crop disease, or insects may destroy most of his planting. For the day laborer also, income varies considerably with both the labor market, the season, and the types of crop being planted. Thus, an accurate measure of peasant economy for any one period might be considerably misleading as to the average income, either overly high or too low. We shall be interested in learning whether for rural and semi-urban Costa Rican women, income has any independent effect on fertility and whether this effect is positive or negative. It is expected that couples with higher incomes have smaller families than couples with 37 lower incomes. V. Husband's Employment Status A substantial amount of data is available which indicate that fertility varies among occupational groups, at least in developed countries. Nevertheless, whether or not these differences exist independent of the effects of income, education, and age at marriage, is open to question. Data for developing countries relating occupation to fertility are scarce and the results contradictory. Stycos (1963;1968) drew a sample of about 2,000 currently married women from the population of Lima, Peru. Inter viewers categorized respondents according to social class, a principal determinant of which was occupation. There was a positive relation between preferred number of children and class, but a negative relationship between these two variables and actual number of children. Stycos inter preted the positive relation in economic terms; he showed that lower class respondents are aware of the deletorious economic effects of additional children. They are, how ever, unable to implement their desire for smaller families for a number of reasons, including lack of information about birth control. No systematic analysis of the relationship of this socioeconomic characteristics and fertility has been 38 performed for exclusively rural and semi-urban areas. This will be attempted in this study. Husband's employment status is classified according to traditional or modern working relationship. One of the significant features of modern society is the increase in employment outside the family institutions. This new type of employment not only reduces people's tie with kin but also increases their contact with the outside world. Modern nonfamilial institutions are places of interaction and an important source of information. It is expected that those with better education are more likely to be engaged in modern type of employment, and those who are engaged in modern type of employment are more likely to adopt a more modern outlook towards family size and family limitation practices. This factor is expected to act as an important intervening variable between education and fertility and also exert a part of its effect through the biopsychological variables. The categories of employment status are: low farm labor, unskilled nonagricultural labor, and high agricultural occupation or skilled or pro fessional labor. VI. Housing Conditions Indicators of housing conditions have not been extensively used in fertility studies. Freedman and Takeshita (1969), and Sun (1969), in their analysis of 39 Taiwan fertility differentials, report a consistent inverse relationship between ownership of modern objects and several measures of fertility. This suggests that housing conditions, roughly equivalent to ownership of modern objects, may be important net of income and residence. Housing conditions can be considered as a crude index of the extent of the desire to improve the living conditions, which suggests a modern outlook on standard of living, and of the ability to afford such living conditions which is a purely economic matter. Instead of creating an index of the number of modern objects owned, the data for this variable were taken from the report given by the interviewer as to the couple's general housing conditions. The five original categories included in the survey— very poor, poor, some comforts, comfortable, and luxurious--were reduced to four by collapsing the last two categories into one. Housing conditions separated from income is indicated by (i) the independent relationship of housing conditions with fertility, when income is held constant, and (II) the correlation of housing conditions with educa tion and husband's employment status with income held constant. Biopsycholoqical Factors In the general scheme presented in Chapter I, 40 biopsychological characteristics were hypothesized to mediate the influence of the various social background and current factors on fertility behavior. Four possible sets of biological and psychological factors may account for the fact that some subgroups of a population may be more fertile than others: (a) differential fecundity. Some subgroups may have greater difficulty in conceiving, or the prevalence of sterility may be higher. Possible biological factors related to this will be; breastfeeding, which lengthens post partum amenorrhea; involuntary abortion; diet and nutrition; temperature; humidity; diseases; especifically venereal diseases; and age of mother; (b) differential nuptiality. Some subgroups may delay marriage to a later age and thereby have a shorter duration of exposure to childbearing. Age at union would be the sociobiological factor related to this; (c) differential sexual activity. Some subgroups may have sexual relations less frequently, and some may abstain for long periods of time for religious, cultural or personal reasons. The sociobiological factor related to this would be coital frequency; and 41 (d) differential practice of contraception. Some subgroups may have a better knowledge of con traception, may have more ready access to medical or other services needed to practice contraception, may have the income with which to purchase the necessary drugs or appliances, and may lack those factors which would inhibit the use of contraceptive devices. Use of con traception would be the sociobiological indicator related to this. From the Rural Survey data is available for some of the variables mentioned, especifically, age of mother and age at union, which will be used as controls, coital frequency and use of contraception. Some information from other sources is available for some of the other variables, such as incidence of specific diseases and nutrition by area. They may be of some use in explaining area differ entials if these happen to be substantial. Besides the two variables already mentioned, coital frequency and use of contraception, two other psychological indicators will be used, namely, extended family living experience and attitude toward family planning. VII. Coital Frequency The significance level of coital frequency for the actual level of fertility is a high problematic and complex 42 question. Lopez (1967) uses this variable in association with the issue of the relation between conctraceptive practice and speed to conceive in a semi-urban Colombian community. He concludes that, given the low mean frequency of intercourse of the group under study, and its high fertility performance, the common belief that high fertil ity requires very frequent sexual relations does not seem to be supported. In this study the direct relationship of coital frequency with fertility will be examined, and also its mediating effect between the other groups of variables and fertility. VIII. Extended Family Living Experience Family structure refers to the family composition in terms of extended family system and nuclear family. An extended family is defined as one which includes other married persons who are related to the sample couple, such as parents, married brothers, sisters, cousins, or other relatives. One aspect of modernization is the increase in number of nuclear families. Modern couples are likely to leave their old homes and establish new families when they marry. Living in an extended family after marriage probably has some influence on fertility behavior. First of all, couples are more intensively exposed to the 43 influence of older relatives, such as parents, elder brothers and others who are more conservative about the traditional norms as to family size. The stronger family ties are likely to bind them to the traditional norms about the family. Secondly, the stronger relationship with their relatives is likely to give them a feeling of economic security, which may lead them to worry less about feeding a large number of children or prevent the development of a strong motivation toward family planning. They are likely, therefore, to have more children than couples in nuclear units who must be economically more selfreliant (Chandrasekar, 1955; Davis, 1955; Matheu, 1962). Another way of reasoning leads to the opposite conclusion, namely, that extended family living will reduce fertility rather than increase it because the housing situation in an extended family may inhibit the couple’s sexual behavior and also because they are expected to spend more time and energy for other family members. A single unit affords greater privacy and less adherence to the traditional periods of abstinence. Therefore, coital frequency is higher and the probability of conception greater (Nag, 1967). Freedman and Takeshita (1969) report lower fertility for nuclear families than for those living in an extended unit, while Stoeckel and Choudhury (1969), Nag (1967) working with data from West Bengal, and Pakrasi and 44 Malaker (1967), found a higher fertility in single families than in joint families. In assessing the two positions, it would seem that the latter fits better with Latin American sexual idiosincracy. The limitations on coital frequency imposed by the extended family living arrangements would lead to smaller families than in a nuclear family. The remarks made above in reference to the preceding variable, though, force some reservations upon this hypothesis. IX. Attitude toward Family Planning It is almost tautological to state that if a woman holds attitudes favoring family planning, she is more likely to control her fertility. This is not always the case, though. While Bom et al. (1972) found a correlation ratio of .36 between family planning attitude and the number of living children in Korea, Saw (1968), with data from a KAP Survey conducted in Malaya in 1966-1967, shows an insignificant relationship between approval of family planning and actual fertility measured as number of living children. We would expect a strong mediating influence of this factor in the effect of social background and present characteristics on either prospective or ideal fertility. In the case of actual fertility, the influence of attitude toward family planning should be much weaker. This would 45 be explained in the sense that the official family planning program did not start until 1966, and that this program is assumed to have been the major factor in influencing women's attitudes toward family planning but not their actual fertility, which was in most of the cases practical ly completed by that time. X. Use of Contraception The term contraception refers to any method used with the intention of preventing conception, other than a sterilizing operation. The methods included in the term are all of the mechanical and chemical methods used to prevent intercourse from resulting in conception; in addition they include abstinence,rhythm, and withdrawal. In other words, contraception is any deliberate pattern of sexual behavior intended to prevent conception. In class ifying couples as users or nonusers of contraception, no account was taken of the effectiveness of the method used. It is expected that use of contraception creates differentials in terminal fertility. The more educated women and those with a higher socioeconomic status will not only have a smaller family ideal but they are also likely to have stronger motivation for using contraception for that purpose. Therefore, we shall expect to find a socio economic differential in the use of contraception. 46 Measures of Fertility Fertility is measured by past fertility or the number of live births to date, by prospective fertility or wanted family size, and by ideal fertility or ideal family size. The number of live births the wife ever had shows effective fertility up to the date of the survey. By considering live births, problems of fetal loss are avoided; these could be a result of the couple's effort to limit their family size. Another advantage of using the number of live births is its comparability with other studies. Wanted family size is the sum of the number of children alive at the time of the survey and the number of additional children the wife wants. It represents the family size rural and semi-urban Costa Rican wives are likely to end with. Ideal family size shows the fertility values of rural Costa Rican women 30 to 49 years old. It represents the family size rural and semi-urban Costa Rican wives would have had, had they the opportunity to start again their family formation period. Past fertility differentials by socioeconomic factors are expected to be larger than prospective and ideal fertility differentials, because past fertility 47 involves more personal variations caused by intermediate variables. Specific Hypotheses Relevant to the Research Problem We have discussed in discursive detail the bases for the expected relationship between social background factors, present social factors, biopsychological factors and fertility. We shall now state these expected relation ships in the form of more formal specific hypotheses. These propositions apply to women 30. years of age and older, currently married, to whom this study is limited. I. The Contribution of the Three Groups of Factors to Fertility. 1. The three groups of factors considered will each be strongly associated with fertility. 2. The effect of social background character istics on fertility will be exerted partly through present social factors and partly through biopsychological factors. 3. The influence of present social factors on fertility will be partly exerted through the biopsychological factors. 4. Social background characteristics will have independent effects on fertility that are 48 not exerted through the other two groups of factors. 5. Present social factors will also have independent effects on fertility. 6. The intervening effect of the biopsycho logical characteristics between the social background factors, the present social factors, and fertility will be stronger in the case of ideal fertility than in the case of either prospective or actual fertility, and stronger in the case of prospective fertility than in the case of actual fertility. II. Socio-structural Differential in Fertility. 1. Each factor from the social background characteristics and present social char acteristics groups will be strongly associated with the three fertility measures. This association will be negative. 2. The effect of education on fertility will be partly exerted through husband’s income and employment status, housing conditions, and use of contraception. 3. Husband's income will influence fertility partly through attitude toward family 49 planning and use of contraception. III. Relationship among Socio-Structural Factors. The following pairs of factors will be strongly associated with each other: (a) Within the same group: couple’s education and residential background (CExCRB), husband's income and employment status (HIxHES), husband's income and housing conditions (HIxHC), husband's employment status and housing conditions (HESxHC), and attitude toward family planning and use of contracep tion (APxUC). (b) Between different groups: couple's education and husband's income, employment status, and housing conditions (CExHI, HES, HC), couple's education and attitude toward family planning and use of contraception (CExAP, UC), husband's income and attitude toward family planning and use of contraception (HIxAP, UC), and housing conditions and attitude toward family planning and use of contraception (HCxAP, UC). IV. Adjusted Fertility Differentials. 1. Income will have a negative net effect on fertility when the effects of other factors are held constant. Husband's employment status will have a weak independent effect on fertility when other factors are taken into account. Housing conditions will have a strong effect on fertility when the effect of income is removed, but will be weaker when the other factors such as education and biopsychological factors are taken into account. The influence of use of contraception will remain fairly strong even after the effect of education and income are taken into account. CHAPTER III DATA AND METHODOLOGY Source of Data The data for this study were obtained from the 1969 fertility survey of women aged 15-49 in the rural and semi- urban areas of Costa Rica. In 1963, an Urban Fertility Survey was conducted in the Metropolitan area of San Jose, capital city of the country. A rural survey program was promoted by CELADE (Centro Latioamericano de Demografia), institution which has also been responsible for the urban surveys conducted in seven Latin American capital cities, among them San Jose. The purpose of the rural surveys was to generate knowledge about fertility behavior and atti tudes in the rural and semi-urban*areas of several Latin American countries in order to complement what was already known about the urban areas. The Costa Rican Rural Fertility Survey was con- ✓ * ducted by the Direccion General de Estadistica v Censos of the Costa Rican Government, with the technical advice of *Less than 20,000 people. 51 52 CELADE. A probability sample of 1,973 women between the ages of 15 and 49 were interviewed between June and November of 1969. Only those women between 30 and 49 years of age and being currently married are included in this study. The sample size is, thus, reduced to 683 cases. The reason for confining the analysis to women 30 years of age or older is that older women are more suitable for the purpose of the investigation. First, the older women are the closest to the end of the childbearing cycle. Most of them have been married for a long enough period to have had 5 or 6 children, the most frequently reported ideal family size. Older women are trying to limit family size and have recently shown a remarkable reduction in fertility, even if this last remark could be repeated of the 20-29 age group, which also shows a large fertility reduction. In both cases they can be expected to demonstrate, by their large fertility differentials, great sensitivity to social and economic change. Temporal Sequence Since experimentation is impossible, any attempt to establish relationships and causal chains about fertility behavior must rely upon longitudinal and cross-sectional observations. In the former, naturally occurring changes and their apparent consequences are observed over time. In the latter, observations are made at roughly one point in 53 time and it is assumed that some units have undergone a change which others have not. That is, an attempt is made to infer relations from the observation of different magnitudes of variables, and it is presumed that these differences are comparable to application of different magnitudes of an independent variable. With cross-sectional data, dependence between observations and error terms, a common problem in the case of longitudinal data, is usually avoided. This fact allows to obtain more accurate estimates of correlation and regression coefficients than when using longitudinal data. A cross-sectional study, on the other hand, provides the investigator with less information about exact time sequences than in the case of longitudinal data. Only covariations and correlations can be obtained, plus know ledge of some of the temporal sequences. Another problem with cross-sectional data occurs because of the spatial continguity of the units. Due to the lack of independence between adjacent units, patterns may have merely diffused from one unit to the next (Blalock, 1968). The relationship between two variables X and Y, thus, might be spurious because of this phenomenon. Primarily because of accessibility of data, cross- sectional data are used in this study. The Data 54 Extensive information on various demographic, socio-structural, and attitudinal characteristics, as well as on use of birth control methods are available: only part is used in the present study. This data includes the usual background and fertility information as well as information on family planning and coital frequency. The sample for the Costa Rican Rural Fertility Survey was drawn from a universe of 220,695 women in the age range of 15 to 49, and living in districts of less than 20,000 inhabitants.* The basic design was a one stage stratified sample (or cluster sample) of sampling units roughly equivalent to the intermediate administrative division called ’canton' or part of it,** previously class ified in two groups: those cantones with at least a town of 20,000 inhabitants or more, and the cantones with no such towns. The 'cantones' were selected with equal probability of selection within the two strata thus formed. Each selected district was then subdivided into "segments" or clusters of 13 households, following as much as feasible *The original base number was 223,240 women, but 12,545, that is, 5.4 percent, were excluded because they resided in areas either hardly accessible or of low demo graphic density. **The largest administrative division is a 'provin- cia' and the smallest a 'distrito.' The total number of cantones is 69. 55 the natural limits. Once the "segments" were selected with equal probability of selection, every eligible woman within each "segment" was included in the sample. The number of interviews within each sampling unit or 'canton1 was determined according to its proportion of rural (less than 2,500) and semi-urban (more than 2,500 but less than 20,000) population. Even in a scientifically selected sample, relia bility of data may be affected by errors and biases arising from three major sources: response rate, accuracy of data, and sampling errors. Response Rate If the response rate is low, and individuals with certain characteristics tend to be consistently eliminated from the sample due to non-response, then a serious bias may occur in the data. In our case, out of the 2,200 originally selected households which contained eligible women, the percentage of interviews totally completed was 89.7. This rate compares favorably with other Fertility Surveys. In the Urban Fertility Survey conducted in Santiago, Chile, in 1959, the response rate was 79.7 per cent, and in the Growth of American Families Study, con ducted in the United States in 1960, it was 88 percent. Out of the 228 non-responses, only 12 cases were due to refusal to respond. The remaining 216 non-response 56 cases were due to sickness, absence, and most of all inaccessibility of the "segment” selected (146 cases).* In other words, 92 percent of the total non-response cases, was attributable to non-voluntary refusal. Accuracy of Data In any survey the respondent’s reports on his or her behavior and attitudes will include "errors," especial ly in the case of reports on the events of the past. The errors arise from a variety of sources: the respondents may not know the facts; their memories may be faulty; they may intentionally misreport on sensitive subjects, especially if others are present at the interview; the interviewer may bias the response in various ways; the form of the questionnaire, and above all the number of questions and their detail, may affect the nature of the response. It seems that at least on factual matters, the Rural Fertility Survey was done under such circumstances as to maximize the probability of valid responses, although errors of unknown magnitude undoubtedly remain despite all the precautions taken. These are some of the factors that may have increased the validity of the data: *The inaccessibility was due to natural causes, such as rain, rough terrain. Of the nine sample units, the non-response rate was high in only one, namely, 45.5 percent. 57 1. The quality, training, and supervision of the interviewers. The interviewers were carefully trained. A training period of three weeks comprised both a detailed explanation of the tools to be used and field practice of 176 interviews. The interviewers were in a great majority teachers, and teachers have a favorable reception in the community. They worked full-time during the period of the survey, which increased both their commitment and their single-minded attention to the project. The interviewing staff reported daily to the supervisors who, then, edited the interviews daily. This made it possible to talk over with the inter viewers any problems as they occurred. 2. The design of the questionnaire. The questionnaire was designed so that it pro ceeded from the more general questions to the more intimate, so that any fear that the interviewee might have felt, would have the time to dissipate. The questionnaire was especifically detailed in the areas of birth control of the couple. In some areas of questioning, questions were asked in several ways at different parts of the question naire. This allowed to check the consistency of the answers, which was found satisfactory. This 58 increases the completeness and accurateness of the answers. On the other hand, a factor that may decrease the accuracy of the information given was the fact that in 42 percent of the cases another member of the household was present during the interview (in 7.4 percent of the cases it was the husband, in 5.8 the interviewee's mothers, in 14 percent small children, and in 14.8 percent other adults). 3. Checks to improve the data. Most of the questionnaires were checked for internal errors either at the end of the interview or at the end of the day. In most cases, where errors or omissions were found, they were corrected by sending the interviewer back to the respondent to obtain a consistent set of responses on factual items. The tape used for this study was free of illegal punches. Missing data for the specific variables used are included in the tabulations. No external checks, though, have been made, for example, with vital registrations. The vital registrations system, on the other hand, is acknow ledged to contain errors, particularly in the case of infant deaths. In regard not to factual items but to attitudes or values, the question of validation is more 59 difficult. Even repeated questions in different sections of the survey (or, for that matter, in another survey) do not help, because attitudes change and it is impossible to separate unreliable responses from genuine changes. It can be said that the answers to attitudinal questions should be regarded not as exact statement of facts but as central tendencies in an area of normative think ing. Deviations from the central tendency for different women are likely to be compensating in direction, so that the total picture obtained for the whole group may not be tco far in error. This does not mean that there can be no check on the strength of the attitudes. The fact that a high correlation is found between actual fertility and both desired and ideal fertility indicates that there is a substantial correspondence between attitudes, values, and behavior. A perfect corre lation cannot be expected. It would be to hold the unrealistic view that this one attitude is so over riding as to eliminate all other considerations. Few attitudes have such dominance over behavior for any length of time. Sampling Errors The sampling error is a measure of the expected 6° variation of the sample statistic from the correspondent universe value. The object of measuring sampling error is simply to determine what the probability is that the results may vary by specific amounts, purely as a result of the fact that one selects a sample of a certain size and design rather than working with population values. This measure leads to the statement of a confidence interval, which is likely to contain the population parameter with a given amount of probability. It would have been relatively easy to calculate sampling errors, had the sample been a simple random sample. In the case of the Costa Rican Fertility Survey there would be a clustering effect on the sampling error, due to the fact that, as it was described above, probabil ity methods were used to select "segments." Once a "segment" was selected, every eligible woman in it was in the sample. This procedure would give results identical to those of the simple random selection procedure if the "segments" were not internally homogeneous— that is, if the women in each "segment" did not resemble each other more closely that they resembled women in other "segments." But this internal homogenity does exist. In rural and semi- urban Costa Rica, as in other rural or urban areas, neighborhoods tend to mold people in a similar way or attract people of similar characteristics. To the extent that this is true for the characteristics being considered, 61 the sampling error will exceed that of a simple random sampling procedure. How much the clustering effect increases the sampling error varies with the characteris tics considered, since segregation is not the same for all characteristics. Since many of the statistics used in this analysis, mainly the means, are based on rather small groups, which would involve large sampling errors, and on the other hand, many of the comparisons between small subgroups would not yield statistically significant results., in order to draw any conclusions, the writer has relied mainly on the per sistence or absence of systematic patterns of variations in the means of the dependent variable in relation to the explanatory variables, rather than on the statistical significance of the differences between pairs of means. Methods of Analysis This study has two principal goals: (l) describing the relationships between fertility and certain socio- structural and attitudinal characteristics; and (2) evaluation of hypotheses derived from theoretical premises. This first goal focuses upon the magnitude of relationships and questions of cause are unimportant. For example, from this point of view it does not matter whether housing con- dtions causes low fertility or vice versa. Such empirical generalizations are important because they are an essential 62 prerequisite to "causal” theory. The second goal requires concern with more than the degree of association. Ideally, causal hypotheses would be evaluated with so-called "structural models"— a set of equations which correspond to actual causal processes. Examples of such models include "path analysis" (see, for example: Duncan, 1966) and Blalock's "causal models" (1961). The assumptions required by such models cannot be met with available data and theory. In particular, it is not possible to unequivocally establish all causal orderings. Consequently, concern here is with evaluation of "incomplete" causal systems. The theory outlined in Chapter II includes hypotheses about intervening mechanisms and statements about "causal" relations. The concept of "cause" is one of the most ambiguous in sociology and requires a theoretical rationale which orders variables and specifies their relations to one another (cf. Theodorson, 1967). This rationale ought to suggest answers to the following questions: does X lead to Y directly, or is its effect through some other variable? Is there a direct and an indirect effect? Furthermore, this rationale should indicate which controls are important since it is not possible or meaningful to control for every variable. A goal which is related to the two primary goals is evaluation of the over-all predictive efficiency of any 63 given set of variables and their relative importance. This is partially a predictive and partially a causal analysis. To the extent that focus is upon selection of the set of indices which combine to best predict fertility, then the concern is predictive. If an attempt is made to evaluate the causal implications of this set of variables, and their relative importance, then the concern is "theoretical." This phase of analysis is largely inductive. The discussion of hypotheses in Chapter II includes statements about predictive combinations or relative importance of variables. The mathematical models utilized are variants of multiple correlation and regression, and are described below. Basically, three statistical techniques are used in this study: comparison of means, Pearsonian product moment correlation, and multiple classification analysis. Comparison of Means The means of the dependent variables--number of live births, wanted family size, and ideal family size-- have been calculated and compared for various subcategories of the sample. In doing so, the investigator has not used any statistical test of significance, as it was observed before. Two reasons have led to this decision. In the first place, in order to perform statistical tests it is necessary to calculate sampling errors for all the 64 variables used, for various sample sizes, with appropriate consideration of the clustering effect, which would be difficult. Secondly, when the classifying characteristics are continuous variables, such as income, or can be ordered, as in the case of husband’s employment status, significant differences between successive categories can not always be expected. Another criterion for evaluating compared means exists and has been used in this study, that is, monotonicity or, in other words, the presence or ab sence of systematic patterns of variations in the means. A monotonic pattern of variation among categories of the classifying variable is meaningful. The patterns of dif ferences are sometimes so uniform that they provide some confidence in the results. When such a consistent pattern of variation exists, the variations will be regarded as significant, even when, considerable sampling error may be involved in the individual comparisons or these individual differences may not be significant. Fearsonian Product Moment Correlation In addition to comparison of means, zero-order and multiple correlations have been used, the former to measure the strength of association between pairs of independent variables, and the latter to find out how much of the variation of a given dependent variable is explained by various groups of independent variables. There are a 65 number of interpretations of r. One is the amount of variance in a dependent variable ' ’explained” by an indepen dent variable. However, as noted by Labovitz and Hagedorn (1971), the magnitude of a given bivariate correlation is not entirely satisfactory as a criterion of cause given the concept of multiple causation. They suggest that two variables may be causally related even when the magnitude of the association is very low. Consequently, since a major assumption of this study is that fertility is deter mined by several factors, emphasis is given to the "rate of change” interpretation of r; that is, in standard score form, the value of r is interpreted as the amount of change produced in a dependent variable by a unit change in an independent variable (McNemar, 1966). The tables presenting the correlations also include information on whether the correlations are statistically significant at various levels. In a sample as large as ours, though, even a small amount of correlation is, of course, found to be statistically significant. The interpretation of the Pearsonian product moment correlation coefficient assumes a linear relation ship between any pair of given variables. The inspection of the pertinent scatter diagrams indicates that an approx imate linear relationship is present in most of the pairs of variables included in this study. On the other hand, since those coefficients are not used in order to make 66 strict interpretations of the mechanisms of fertility behavior and values but only in a general way to explore possible associations between independent variables, it seems adequate to use this measure to summarize those relationships even if the assumption of linearity is not always strictly met. The multiple correlation coefficient used is the one provided by the multiple classification analysis program discussed below. It is equivalent to the usual multiple correlation coefficient but adjusted for the degrees of freedom. This coefficient, when squared, indicates the proportion of variance in the dependent vari able explained by all predictors together (after adjusting for degrees of freedom).* Multiple Classification Analysis The third statistical technique in this study is the multiple classification analysis (MCA) as described by Andrews et al. (1969). It is used instead of the more common multiple regression because regression assumes a number of interval-scale variables, an assumption which in this case cannot be met. The multiple classification *The actual proportion of variance explained would be equivalent to the squared value of the multiple corre lation prior to making any allowance for degrees of freedom (cf. Andrews, 1969). 67 analysis is useful in examining the relationship of each of several predictors to a dependent variable when the effects of the other predictors are held constant and within the context of an additive model. The operation of multiple classification analysis is analogous to that of a multi variate regression analysis for situations in which the independent variables may be continuous variables divided into subclasses, or discontinuous variables. This method uses the technique of the "dummy" variable, that is, each category of a predictor is treated as a separate charac teristics by itself with values of "1" or "O", depending on whether or not an individual belongs to that category or not. This method combines certain features of both multiple linear regression analysis and analysis of variance, but has some advantages over each of these to make it particularly suitable for the analysis of social data. One of the advantages of the multiple classification analysis over the conventional multiple regression analysis is that, as it was suggested above, it can handle nominal predictors, although it requires that the dependent variable be an interval-scale variable. Furthermore, this technique makes no assumptions as to the form of relation ship of the dependent variable to any of the classificatory variables. The order of categories in a predictor, then, is not used in obtaining the solution, although it is 68 considered in interpreting the results. The advantage over the conventional analysis of variance technique is that it accepts unequal number of cases in the cells formed by cross-classification of pairs of predictor variables. In summary, it can be said that this procedure is analogous to standardization, and indicates the extent to which the original subcategory means are affected by the intercorre lations with other independent variables. There are, however, some limitations in the multi ple classification analysis program both in terms of the analytical model and the procedure to solve the equations implied by the model: (a) The distribution of the dependent variable should not be badly skewed. (b) The effects of the explanatory variables should be reasonably additive, that is, each explana tory variable should affect the explicandum in an independent manner. (c) The predictor variables should be, to some degree, associated with one another. If they are not correlated, the technique discussed is unnecessary. (d) No two independent variables should be perfect ly correlated. Assumption (a) is met, as a perousal of the one-way frequency distribution shows. As to assumption (b), to the 69 extent that there is interaction among the independent variables, this method of analysis would be inappropriate. Interactions will consequently have to be handled through separate consideration of the categories of known inter action variables. But, since.it is virtually impossible to take account of all interactions, this inadequacy of the model will have to be kept in mind. The last two limitations (c) and (d) arise from the process of solving the equations. The program arrives at a solution by a series of successive approximations, that is, by iteration. When the predictors are all uncorrelated, it would not be necessary to take this approach; but when two or more predictor subclasses are correlated perfectly, there is no solution to the equations. The predictors used in this study are correlated to some extent, but there is no perfect correlation between any two predictor sub classes . Several statistical measures are derived from the analysis, two of which will be used in this study: (l) The mean value of fertility measures for each subclass of each independent variable; this mean is expressed as a deviation from the grand mean (the "gross effect"). The "net effect" of the predictor is expressed as the deviation of the category mean from the grand mean after adjusting for the effect of other predictors. 70 The comparison of these two kinds of effects, both the pattern and the magnitude of the differentials, will indicate how the effect of one predictor on fertility is influenced by other predictors. The relationship between a predictor and fertility is expected to change when the effects of other predictors are taken into account. Our main interest is how and to what extent it changes when other predictors are controlled. (2) A measure of the relative importance of each predictor taken singly. This is the correla tion ratio or eta. In other words, this coef ficient indicates the ability of the predictor, using the categories given, to explain varia tion in the dependent variable. The proportion of variance explained is eta square. This measure of correlation requires no assumptions about linearity or ordering of variables. Summary The data used for this study are taken from the 1969 Rural Fertility Survey of Costa Rica. The accuracy of these data has been found to be satisfactory. The quality, training, and supervision of the interviewers, the design of the questionnaire, and the checks conducted to improve the data, account for that accuracy. The methods of analysis to be used in the remaining chapters are the comparison of means, employed for the description of the general fertility differentials and for the discovery of possible interaction effects among pairs of predictors on fertility; the Pearsonian product moment correlation for the description of the relationships among independent variables; and the multiple classification analysis and the several measures this program provides, for the evaluation of the specific hypotheses presented in Chapter II. CHAPTER IV GENERAL FERTILITY DIFFERENTIALS Before proceeding to a more detailed analysis of the way in which socio-structural differentials in fertil ity are brought about, examination will be made regarding some of the patterns of differential fertility in Costa Rica indicated by this and other surveys' data. Goimez (1968), in his analysis of the Urban Fertility Survey of the Metropolitan Area of San Jose, the Costa Rican capital, reports fertility differentials by socioeconomic status. He indicates also that San Jose couples practice birth control mainly for limitation of family size and begin that practice relatively late in their married lives. No systematic studies have been previously con ducted on the fertility behavior and values of rural and semi-urban Costa Rican couples. If we assume that the socioeconomic factors are already acting among the rural and semi-urban Costa Rican couples as they were among the urban couples to produce fertility differentials, it is expected that the actual fertility of those rural and semi- urban couples will show clear differentials by socio- structural indicators. According to the theoretical 73 framework presented in Chapter II, it is not expected that the differentials in fertility values (i.e., wanted and ideal family size) will be as large as in the case of actual fertility. Thus, as an introduction to the main analysis, the writer will examine in this chapter the general fertility differentials of the rural and semi-urban area population. The analysis will begin by locating the fertility level of rural and semi-urban Costa Rican wives among those of developed and developing countries. The differentials in fertility will be shown, measured at the three levels of actual, prospective, and ideal fertility, by the various socio-structural and biopsychological factors used in this study. The analysis is restricted to the older rural and semi-urban Costa Rican couples because, as it was noted earlier, they are near the completion of, or have already completed, the family building process. With an illustra tive purpose, the same differentials are shown for the younger couples (20-29). This is being done only in this chapter. Fertility Level of Costa Rican Wives The fertility level in Costa Rica is higher than the fertility level of developed countries, but it is relatively low if compared to other developing countries (Table 4.1). 74 Costa Rican fertility has been falling continuously since around 1960. The crude birth rate decreased by 11.5 percent just in two years, from 1966 to 1968, one year before the Rural Fertility Survey was taken. The corres ponding rates of decline for other developed and many developing countries, except for the case of Chile, are much lower (see Table 4.2). This fact, combined with the already very low mortality* and the socioeconomic develop ment of the recent years, suggests that conditions are favorable for a continuing fertility decline. As it was mentioned in Chapter I, the recent fertility decline is pronounced among women of all age brackets, not only among women over 30, the typical pattern of many Western countries at the beginning of their fertility decline. In other words, even if the fertility level of Costa Rican wives, mainly those in the rural and semi-urban areas, was relatively high at the time of the Rural Fertility Survey', a modern pattern of fertility decline seems to have already started. It is expected that fertility differentials among the rural and semi-urban Costa Rican wives, specially the older wives, at the time of this survey will be large enough to show the different reactions of couples to the forces of modernization and other social pressures that induce fertility declines. *The crude death rate in the period 1960 to 1965 was between 8 and 9, and in 1968 was down to 6.5. TABLE 4.1 GROSS REPRODUCTION RATES OF COSTA RICA AND SELECTED COUNTRIES Country Date GRR Costa Rica 1967 2,910 Canada 1967 1,261 El Salvador 1966 3,166 Mexico 1966 3,190 Panama 1967 2,548 United States 1967 1,255 Chile 1967 2,336 Source: UN Demographic Yearbook 1969, Table 31. TABLE 4.2 CRUDE BIRTH RATES OF COSTA RICA AND SELECTED COUNTRIES CBR Percent Change Country 1966 1968 1966-1968 Costa Rica Canada 40.9 19.3 36.2 17.5 11.5 9.4 jranama United States Chile iy&/ 1967 1967 1,255 2,336 Source: UN Demographic Yearbook 1969, Table 31. CRUDE BIRTH TABLE 4.2 RATES OF COSTA RICA AND SELECTED COUNTRIES Country 1966 CBR 1968 Percent Change 1966-1968 Costa Rica 40.9 36.2 11.5 Canada 19.3 17.5 9.4 El Salvador 45.4 43.2 4.9 Mexico 44.3 43.5 1.9 Panama 38.4 38.3 .3 United States 18.4 17.5 4.9 Chile 31.6 26.6 15.9 Taiwan 32.4 29.3 9.6 Source: UN Demographic Yearbook 1970, Table 13. 76 Differences in Three Fertility Measures In this analysis, fertility is measured by three indices which range from actual fertility to expressions of fertility values. They are the number of live births to date, the wanted family size, and the ideal family size. We have already located Costa Rican fertility level among the developed and developing countries in the preceding section. Now it will be pertinent to compare those three fertility measures with those of the United States, where fertility has been relatively high among the developed countries and those of Taiwan, which has a relatively low fertility level among developing countries'.* Table 4.3 provides a crude but interesting comparison of the ferti lity level of Costa Rican wives, both urban and rural and semi-urban, and taiwanese and white American wives. The urban Costa Rican wives aged 20 to 49 have had 3.3 live births on the average. This is .3 less than the average Taiwanese wife had in 1962 and 1.0 more than the average white American wife had at the time of the 1960 Growth of the American Family Study. The rural and semi-urban Costa Rican wives aged 20 to 49 have 2.0 live births more than *It would have been preferable to establish these comparisons with other Latin American countries. Regret fully, the necessary information as to number of live births, wanted family size, and ideal family size for the two age groups 20-29 and 30-49 has not been available. 77 their urban counterparts. The differences among the three populations remain the same for wanted family size if we compare rural and semi-urban Costa Rican wives and Taiwanese wives (1.7) and almost the same (2.8) if the comparison is established between rural and semi-urban Costa Rican wives and white American (measured as expect ed family size). Those differences narrow down if we com pare ideal size. The largest difference is of 1.5 more children for rural and semi-urban Costa Rican wives as compared with white American. This would seem to suggest that although the Costa Rican wives, mainly those in rural and semi-urban areas, have had the higher fertility in the past and eventually expect to have more children, a modern norm appears to be emerging. Thus, it is expected that Costa Rican fertility will continue to decline in the near future. When we look at these three fertility measures by the two broad age groups, we find that the differences in fertility levels between the different populations are always greater for the older wives. The older rural and semi-urban Costa Rican wives have had nearly twice as many live births as their urban counterparts, while the younger wives have only had one third more. The same patterns shows if the comparison is made between both groups of Costa Rican wives on the one hand and either Taiwanese or American wives on the other. This indicates that the TABLE 4.3 NUMBER OF LIVE BIRTHS, WANTED-EXPECTED FAMILY SIZE, AND IDEAL FAMILY SIZE FOR COSTA RICAN WIVES, TAIWANESE WIVES AND WHITE AMERICAN WIVES BY TWO BROAD AGE GROUPS Fertility Measure Age 20--29a Aqe 30-•39b Total CR Tai US CR Tai US CR Tai US U R U R U R Live Births 2.4 3.4 2.4 1.8 4.3 7.1 4.8 2.7 3.3 5.3 3.6 2.3 W-E FSC -- 5.3 3.8 3.2 -- 6.6 4.6 3.0 -- 5.9 4.2 3.1 IFS 3.9 4.2 3.8 3.2 4.1 5.3 4.1 3.4 4.0 4.8 3.9 3.3 No. of Wives ’ 709 495 1165 1113 683 683 1278 1313 1392 1178 2443 2414 aWhite American wives aged 18-29. ^Rural and semi-urban Costa Rican wives aged 30-49. cThis measure for Costa Rican and Taiwanese wives is wanted family size, while that for White American Wives is expected fsimilv size._______________________ Source: T. H. Sun, Socio-Structural Analysis of Fertility Differentials in Taiwsm, unpublished doctoral dissertation, 1969; P. K. Whelpton et al., Fertility and Family Planning in the United States (Princeton: Princeton University Press, 1966). 03 79 difference in the level of past fertility among those populations is mainly due to a sustained high fertility among older Costa Rican wives. The recent fertility de cline in this age group is expected to narrow the differ ences in the near future. There is not much difference between the two age groups in ideal family size, though younger wives prefer a slightly smaller family. It should be noted that the ideal family size for the older Costa Rican wives is smaller than their wanted family size and that the number of live births is greater than the wanted family size. This presumably shows the effect of mortality on family size. Live births are counted regardless of subsequent deaths, but wanted size takes into account only those children alive at the time of the survey. This relationship among fertility measures, namely, "number of live births greater than want ed family size greater than ideal family size" is a phenomenon which probably appears in a period of rapid demographic transition from high to low fertility as a result of the combined effects of high mortality in the past, lack of means to control high fertility and recent modernization leading to preferences for smaller families. Some of these wives were probably having more children than they wanted in order to make sure that the desired number survived. Since recent modernization has brought down the mortality level, many wives have found themselves with more 80 children than they preferred. The recognition of mortality change lagged behind the reality of the situation. It is this group of women who are highly motivated to take up family planning. Couples who have ever used any family limitation methods have approximately the same number of live births but a smaller ideal size than those who have never used anything. Fertility Differentials by Social Background and Present Social Characteristics Generally speaking, there is a strong negative correlation, as it was expected, between the fertility of the rural and semi-rurban Costa Rican wives and their socio-structural characteristics. (Tables 4.4 to 4.7). This is consistent with the argument that in the period of demographic transition, those with higher social status have better knowledge of fertility control, use means of family limitation more often and more effectively and therefore have lower fertility. This negative association stands in both age groups, even if for the younger rural and semi-urban couples the relationship is somehow weaker mainly for the actual fertility measure. In the case of the normative fertility measures, the inverse association with indices of socioeconomic status is always evident. 81 • Education All three fertility measures are inversely related to the couple’s education level- in both age groups, 20-29 and 30-49 (Table 4.4). The negative association decreases r when one moves from actual fertility to ideal fertility, showing that even for those couples with almost no or little formal education normative fertility is approaching a more modern level. The central questions, however remain: what are the channels through which education brings about this change? What are the direct causes of this change? We note that the difference in past fertility, that is, number of live births, is especially large between those with and without a complete elementary school education. The existence of this gap in past fer tility between two educational levels suggests an inter esting relationship between level of education and change in fertility behavior. It has been suggested that there is a minimum level of education at which people in a society with a given degree of development will be more likely to be influenced by modernization and consequently to change their fertility behavior. Table 4.4 indicates that in rural and semi-urban Costa Rica, at the time when, wives in our sample were of school age, the essential dis tinction lies in the fact of being exposed to the 82 educational system until at least graduation from elemen tary school. This finding is consistent with at least two interpretations: (1) education differentiates styles of life which are relevant for fertility behavior, and (2) the impact of education is a function of the relative exposure of the population to the educational system. This can be explained in the sense that Heer (1971) suggests. He sub mits that increases in educational attainment may reduce the preference for children (a) through forcing an acquain tance with new types of goals which furnish more satisfac tion than having children, (b) by facilitating the acquisi tion of methods of child care which allow a greater propor tion of existing children to survive, and (c) by reducing the rewards parents obtain from the gainful labor of their children. Residential Background and History The Costa Rican wives are classified according to the couple's residential background, before age 15, and to the couple's residential or migration history, which is roughly equivalent to rural experience after marriage. Residential background seems to have differential influences on fertility behavior (Table 4.5). In general, for older wives, the amount of the couple's rural back ground is positively correlated to the three fertility measures. This holds without exception. The patterns in 83 the case of the younger wives is less clear and regular except for ideal fertility. The greater the "rurality" of their background, the larger the family they consider as ideal. When the couple’s residential history is consider ed, older wives have a lower past fertility only if either husband or wife or both have lived in a large town (Table 4.5). The general pattern is that of a positive associa tion between rural experience after marriage and all three fertility measures. Younger wives in every single category want a similar family than their older counterparts and their ideal fertility is also considerably lower than that of older wives. Husband's Income Past fertility or number of live births to date of the older rural and semi-urban Costa Rican wives shows a negative linear relationship with husband's income (Table 4.6). Those with husbands who do not work, are paid in kind or earn less than 100 colones a month, and those with husbands in the lowest income group have the highest past fertility (7.9 to 7.5 live births). Those in the highest income category have more than 1.5 fewer live births (5.9). The same pattern holds when fertility is measured by wanted and ideal family size. Those in the lowest TABLE 4.4 MEAN NUMBER OF LIVE BIRTHS (LB), WANTED FAMILY SIZE (WFS), AND IDEAL FAMILY SIZE (IFS) BY COUPLE’S EDUCATION, FOR WIVES AGE 20-29 AND 30-49 Couple’s Education LB WFS IFS No. % Amount of Elementary School: 30-49 Both had less than 1 year 7.6 6.8 5.6 570 8 One incomplete 7.7 6.7 5.6 171 25 Both incomplete 7.3 6.6 5.2 279 41 One completed 6.0 6.7 5.0 87 13 Both completed 4.4 5.6 4.4 44 6 Total ------ ------ ------ 683 100 Amount of Elementary School: 20-29 Both had less than 1 year 4.9 5.6 4.5 17 3 One incomplete 3.5 5.5 4.5 79 16 Both incomplete 3.8 5.6 4.7 193 39 One completed 3.2 5.0 4.4 107 22 Both completed 2.3 4.5 3.5 57 12 Total — - ------ — - 495 100 TABLE 4.5 MEAN NUMBER OF LIVE BIRTHS (LB) , WANTED FAMILY SIZE (WFS), AND IDEAL FAMILY SIZE (IFS) BY COUPLE'S RESIDENTIAL BACKGROUND AND RESIDENTIAL HISTORY, FOR WIVES AGE 20-29 AND 30-49 Characteristic LB WFS IFS No. % Residential Background: 30-49 n A a. * " 7 c A o o n n s r TABLE 4.5 MEAN NUMBER OF LIVE BIRTHS (LB), WANTED FAMILY SIZE (WFS), AND IDEAL FAMILY SIZE (IFS) BY COUPLE'S RESIDENTIAL BACKGROUND AND RESIDENTIAL HISTORY, FOR WIVES AGE 20-29 AND 30-49 Characteristic LB WFS IFS No. % Residential Background; Both rural 7.4 30-49 6.7 5.4 380 56 At least one from village 7.4 6.6 5.2 95 14 Both from village 7.2 6.7 4.7 39 6 At least one from town 6.3 6.5 5.0 107 16 Both from town 4.7 6.0 4.2 29 4 Total ------ ------ ------ 683 100 Both rural 3.6 20-29 5.7 4.9 281 57 At least one from village 3.0 5.2 4.3 68 14 Both from village 2.7 4.1 4.0 20 4 At least one from town 3.4 5.1 3.8 85 17 Both from town 2.6 3.7 3.3 23 5 Total --- --- --- 495 100 Residential Historv: No moves 7.4 30-49 6.9 5.6 131 19 Some moves, no large town 7.4 6.6 5.4 361 53 Some moves, some large towns 6.1 6.4 4.7 181 26 Total --- — — _ , ------ 683 100 No moves 3.6 20-29 5.5 ' 4.8 96 19 Some moves, no large town 3.5 5.6 4.2 266 54 Some moves, some large towns 3.0 4.4 3.7 121 24 Total — _ _ _ _ _ _ — — 495 100 85 income group want and prefer the largest size family, and those in the highest income group want and prefer the smallest size family. But the differentials in those two fertility measures are much smaller than those found in past fertility. The differences between the two extreme groups is 2.0 children in past fertility while only .6 children for wanted family size and 1.2 in ideal family size. For the younger wives the same negative correlation with husband’s income is found whether we consider past, prospective, or ideal fertility. The main difference in the relationship of income and fertility for the older wives and the younger wives is that the latter want a much smaller family than their older counterparts, mainly in the middle income and highest income groups.* *The cases with an income of 600 colones or more are subdivided in the following way: LB_____WFS IFS_____N 600- 900 colones 6.2 30-49 6.2 4.8 41 900-1200 colones 6.1 6.5 4.4 14 1200 colones or more 5.3 6.1 4.5 32 600- 900 colones 2.8 20-29 4.7 3.7 40 900-1200 colones 2.4 5.2 3.3 11 1200 colones or more 2.2 3.4 3.8 20 86 Husband's Employment Status For the three groups which comprise husband's employment status, namely tenant farmers, unskilled labor, and farm owners and administrators, and skilled laborers, and professionals, there is a negative association with past fertility (Table 4.6). For the older wives the dif ferential is very strong for past fertility but it dis appears when fertility is measured as wanted family size. In the case of ideal family size there is a .7 differential between the two extreme groups. For the younger wives the differentials are sus tained in all three fertility measures. The association between employment status and wanted family size appears to be curvilinear and very pronounced. Housing Conditions Of the three measures of fertility, number of live « births shows a strong negative correlation with housing conditions both for older and younger wives. Prospective and ideal fertility show the same pattern, even if the differentials between the extreme groups are smaller than those found in past fertility (Table 407). This negative relationship between fertility and housing conditions is an consistent as the relationship found between fertility and income. It has been often TABLE 4.6 MEAN NUMBER OF LIVE BIRTHS (LB), WANTED FAMILY SIZE (WFS), AND IDEAL FAMILY SIZE (IFS), BY HUSBAND’S INCOME AND EMPLOYMENT STATUS, FOR WIVES AGE 20-29 AND 30-49 Characteristic LB WFS IFS No. % Husband1s Income: 30-49 No work, up to 100 colones 7.9 6.8 5.8 45 7 100-300 colones 7.5 6.7 5.6 282 41 300-600 colones 7.1 6.6 5.1 132 19 600 colones or more 5.9 6.2 4.6 87 13 Total _ — — — — — —-- 683 100 20-29 No work, up to 100 colones --- --- --- 16 3 100-300 colones 3.8 5.8 4.9 200 40 300-600 colones 3.4 5.0 4.3 117 24 600 colones or more 2.6 4.4 3.7 71 14 Total --- --- --- 495 100 Husband1s Employment Status: 30-49 Tenant farmers 7.3 6.6 5.4 516 76 Unskilled labor 6.9 6.3 5.0 50 7 Farmers or skilled labor 5.6 6.5 4.7 82 12 Total --- --- --- 683 100 20-29 ' - Tenant farmers 3.6 5.7 4.8 353 71 • Unskilled labor 3.2 3.9 3.6 48 10 Farmers or skilled labor 2.7 4.4 4.0 76 15 Total 495 100 Unskilled labor 6.9 6.3 5.0 50 7 Farmers or skilled labor 5.6 6.5 4.7 82 12 Total -- --- --- 683 100 Tenant farmers 3.6 Unskilled labor 3.2 Farmers or skilled labor 2.7 20-29 5.7 3.9 4.4 4.8 3.6 4.0 353 48 76 71 10 15 Total -- --- --- 495 100 TABLE 4.7 MEAN NUMBER OF LIVE BIRTHS (LB) , WANTED FAMILY SIZE (WFS) , IDEAL FAMILY SIZE (IFS), BY HOUSING CONDITIONS, FOR WIVES AGE 20-29 AND 30-49 AND Housinq Conditions LB WFS IFS No. % Very poor 7.9 30-49 6.8 5.7 149 22 Poor 7.4 6.7 5.3 305 45 Some comforts 6.4 6.5 4.0 147 22 Comfortable 5.5 6.1 4.7 76 11 Total --- --- --- 683 100 Very poor 4.2 20-29 5.5 4.3 104 21 Poor 3.5 5.5 4.2 219 44 Some comforts 3.0 5.1 4.2 119 24 Comfortable 2.4 4.4 3.8 50 10 Total ------- — — — — 495 100 88 argued that housing conditions (or ownership of modern objects) seems to indicate differential orientation toward consumption better than income does. The conclusion drawn is that housing conditions or ownership of modern objects has a stronger differential influence on fertility than income because orientation toward consumption is more meaningful for explaining fertility than income per se. Our findings do not support this reasoning. Each housing conditions and income show an equally strong negative association with fertility. Fertility Differentials by Biopsycho- logical Characteristics Four biopsychological indicators have been selected as attitudinal and intervening between the socio-structural variables and fertility: extended family living experience, coital frequency, attitude toward family planning, and use of contraception. Generally speaking, no strong pattern of associa tion can be found between any of these indicators and fertility both past and prospective. In the two variables related to family planning, a certain trend seems to be developing in the direction of a differential relationship with prospective and ideal fertility. 89 Extended Family Living Experience Fertility is lower by all three fertility measures for couples who are living in an extended family unit than for those who are not, in both age groups (Table 4.8). There are two exceptions. In the case of older wives and their prospective fertility, the differential disappears, and for the same age group the ideal family size is higher when they are living with their parents than when they are not. The limitations that the questionnaire imposes on the analysis have to be acknowledged. The questions about living arrangements deal only with the current situation at the time of the survey. They do not contain specific information about the duration of the extended family living experience nor about whether or not the couple ever had lived in that situation. Within this limitation, though, it seems fair to conclude that in rural and semi-urban Costa Rica an extend ed family living experience has a negative effect on fertility, which is opposite to what has often been hypothesized and verified. This fact could be explained as due either to consequent inhibitions in the couple’s sexual behavior or to the incompatibility of a large family with shared croweded living arrangements. 90 Coital Frequency This biopsychological indicator has seldom been used in fertility studies. As Table 4.8 shows, coital frequency bears a slight positive systematic relationship with number of live births in the older age group. No differentials appear in the two middle groups, but between the two extreme groups the difference is of .9 more chil dren for those with the highest frequency of intercourse. For the younger wives, there is no such systematic varia tion between actual fertility and coital frequency. In the case of both prospective and ideal fertil ity, two remarks can be made. First, the differences between groups are very small, never more than .5 more, children for one group than for another; second, these variations are sometimes curvilinear, sometimes bimodal. Thus, it can be concluded that basically there is no con sistent relationship between coital frequency and normative fertility. In other words, rural and semi-urban Costa Rican couples do not associate their frequency rate of sex ual intercourse with wanted family size or fertility values. Attitude toward Family Planning As Table 4.9 shows, in both groups of wives, there is no relationship between past fertility and attitude TABLE 4.8 MEAN NUMBER OF LIVE BIRTHS (LB), WANTED FAMILY SIZE (WFS) , AND IDEAL FAMILY SIZE (IFS), BY EXTENDED FAMILY LIVING EXPERIENCE AND COITAL FREQUENCY, FOR WIVES AGE 20-29 AND 30-49 Characteristic LB WFS IFS No. % Extended Family Living Experience; No 7.1 30-49 6.6 5.2 619 91 Yes 6.4 6.6 5.8 60 9 Total ------ ------ ------ 683 100 No 3.5 20-29 5.4 4.3 415 84 Yes 2.9 4.9 3.7 73 15 Total ------ - — ------ 495 100 Coital Frequency: 30-49 None or less than 1 a month 6.5 6.4 5.2 32 5 1-3 a month 7.0 6.7 5.1 171 25 1-2 a week 7.0 6.5 5.2 282 41 3 or more a week 7.4 6.7 5.4 137 20 Total --- --- -- 683 100 20-29 None or less than 1 a month --- --- --- 13 3 1-3 a month 3.4 5.4 4.5 94 19 1-2 a week 3.4 5.3 4.0 234 47 3 or more a week 3.4 5.2 4.3 127 26 Total — — — --- --— 495 100 - • “X **# J iy 1-2 a week 3.4 5.3 4.0 234 47 3 or more a week 3.4 5.2 4.3 127 26 Total ------ ------ ------ 495 100 TABLE 4.9 MEAN NUMBER OF LIVE BIRTHS (LB), WANTED IDEAL FAMILY SIZE (IFS), BY ATTITUDE AND USE OF CONTRACEPTION, FOR WIVES FAMILY SIZE (WFS), AND TOWARD FAMILY PLANNING AGE 20-29 AND 30-49 Characteristic LB WFS IFS No. % Attitude toward Family Planning: Against Favor 7.1 7.1 30-49 6.9 6.3 5.9 4.7 305 359 45 55 Total ------ ------ ------ 683 100 Against Favor 3.4 3.4 20-29 6.2 4.8 4.7 3.9 166 312 34 66 Total ------ ------ - : — 495 100 Use of Contraception: No Yes 7.1 7.0 30-49 6.7 6.4 5.5 4.8 454 228 66 34 Total --- — — — — — “ 683 100 No Yes 3.4 3.3 20-29 5.6 4.7 4.3 3.9 312 179 63 37 Total — — — — — — — — — 495 100 92 toward family planning. But when fertility is measured as wanted or ideal family size, those who favor family planning want a smaller family and consider as ideal a smaller number of children than those who are opposed to family planning. The differentials for both measures are larger in the younger wives group. This would indicate: (l) that couples do not plan their families in terms of spacing, and (2) that future fertility is likely to be lower than it is currently, if the attitudes that the younger couples have toward family planning are taken as indicative of future trends in actual fertility. Use of Contraception This variable follows very closely the pattern of association with fertility that was found for the preceding indicator. While there is no concomitant variation between past fertility and use of contraception, a certain negative pattern develops when prospective and ideal fertility are taken into consideration. From Table 4.8 nothing cam be said about the effectiveness of the use of contraception on family limita tion since the wives in the older age group have already exceeded in actual fertility their own wanted family size and the younger wives have not yet reached the family size they want. It can only be said that current users of contraceptive methods hardly differ in their past fertility 93 from non users in both age groups, but given late use of contraception this does not necessarily mean they will not differ at the end of the childbearing period. Summary and Conclusions The level of fertility in Costa Rica is high when compared to that prevalent in the developed world, but relatively low among the developing countries. If only rural and semi-urban Costa Rica is taken into account, the level of fertility continues to be high among the latter group of countries. Among rural and semi-urban wives, those of the younger group prefer smaller families than the older cohort. For the older group the ideal family size is smaller than the wanted size, which is in turn smaller than the number of live births they have had. This pattern suggests a process of fertility change within a moderniza tion framework. As mortality declines and aspirations change, a gap opens between the actual and the ideal fertility. Generally speaking, the whole set of relationships between fertility and socio-structural indicators holds in the expected direction. For the older group, those who are more modern in characteristics have relatively low past fertility and prefer a smaller family. Modern characteris tics in this study are: higher level of education, less rural experience and more mobility, higher income, 94 employment in modern sectors, better housing conditions. These results are consistent with those reported in various studies in other developing countries and in the earlier stages of demographic transition in the now developed countries. For the older age group, the fertility differen tials are usually larger when measured by past fertility than by wanted or ideal family size. This indicates large personal differences in implementing ideas about family size. The socio-structural variables indicate differen tials in regard to not only preferred family size but also ability to take effective action. Fertility differentials are larger among older couples than among younger couples, mainly for past fertility. This is probably associated with the fact that birth control methods are used mainly to limit family size rather than to space childbirths. It would seem that the rural and semi-urban Costa Rican wives prefer to have the number of children wanted as early as possible in their married life. This would be based on their desire to com plete the family cycle relatively soon so that when older they can rely on their grown children. For this reason very few begin using birth control methods before they have the number of children they want. The differentials in ideal fertility among younger couples are as large as those among older couples, although 95 younger couples prefer a smaller family than older couples. The relationships between ideal fertility and socio- structural factors for the younger couples are very similar to those observed among older couples, implying that the actual fertility differentials of the younger group will also be inversely related to socio-structural factors in the future, although the relationships may be somewhat weaker. CHAPTER V RELATIONSHIPS AMONG INDEPENDENT VARIABLES Previously, the manner in which fertility in rural Costa Rica varies with different indicators of socio- structural and biops-ychological characteristics has been observed. The relationships between fertility and these factors vary in strength and direction, and for the most part they are consistent with the findings of many other studies on fertility differentials of developing societies. The question of the mechanism by which these factors influ ence fertility still remains. To study this, it is useful to examine the relationships among those socio-structural and biopsychological characteristics. It is probably that these characteristics are elements of a social structure which act as dependent and independent variables in the process of social change in relation to fertility decline.* While some variables have rather direct influence on fertility, others may have only *For example, Bernard (1971) suggests that under some circumstances education can be explained in terms of birth rates, and not vice versa. The same can be said of other variables in their relation to fertility or-among themselves. 96 97 indirect influence through certain other variables. The influence of some variables may be to some extent over lapping. In this section, the writer examines these inde pent variables to make explicit the relationship among factors in the same group, the relationships among factors classified in different categories, and the possible channel of influence of these factors on fertility. The analysis henceforth is based entirely on the older Costa Rican group. One assumption is made when examining associations between independent variables, that is, a strong association between two independent variables indicates that they measure the same underlying characteristics in their effect on the dependent variable. The concept of two variables measuring the same underlying characteristics is important. When two independent variables which belong each to a different group form a causal relationship in influencing fertility, if they measure the same underlying characteristic, a strong channel of influence is suggested. The technique of analysis used to examine these relationships among independent variables is the Pearsonian product moment correlation. The correlation coefficient summarizes in an easily understandable way the strength of association between pairs of variables. 98 Relationship among Factors of Similar Character The independent variables are grouped in three categories, according to their character and effects on fertility. The first category comprises those character istics which constitute the social background of the inter viewees, namely, residential background and education of the couple. The second category, that of present social characteristics, consists of socioeconomic factors, which can vary over time, namely, couple’s residential history, husband's income and employment status, and housing conditions. The group of biopsychological indicators con sists of the following variables: extended family living experience as partly psychqlogical partly biological, coital frequency as purely biological, and the two family planning variables, namely, attitude toward family planning and use of cpntraception. The two social background variables are strongly related, as expected; the correlation coefficient is .41 (Table 5.1). Those who lived mostly in towns before marriage have a higher education than those who lived in villages or in a rural area because, (l) towns have more and better educational facilities than villages and villages have more and better educational facilities than rural areas; (2) people's motivation to attain higher TABLE 5.1 ZERO-ORDER CORRELATION COEFFICIENTS AMONG SOCIAL BACKGROUND, PRESENT SOCIAL AND BIOPSYCHOLOGICAL CHARACTERISTICS, FOR WOMEN AGE 30-49 2 3 4 5 6 7 8 9 10 11 12 13 1. No. Live Births .39 .22 -.24 -.17 -.14 -.10 -.16 -.23 -.96 .04 .01. -.01 2. Wanted Fam ily Size .17 -.14 -.09 -.10 -.06 -.93 -.13 -.01 • O to i to o -.07 3. Ideal Fam ily Size -.09 -.14 -.14 -.13 -.09 -.13 .06 .05 -.22 -.11 4. Couple1s Education .41 .20 .33 .33 .44 .01 -.03 .15 .28 5. Couple1s Res. Background .28 .27 .28 .27 .00 -.05 .15 .22 6. Couple1s Res. History .21 .13 .12 .06 -.06 .15 .12 7. Husband's Income .25 .34 -.03 -.08 .07 .16 8. Husband’s Employment Status .24 -.02 .03 .03 .16 9. Housing Conditions -.01 -.05 .13 .27 10. Extended Family Living Experience .01 -.07 -.01 11. Coital Frequency .06 .00 12. Attitude Toward Family .30 13. Use of Contraception ----- With N=683, absolute values greater than .07 are statistically significant at the 5 percent level, and those greater than .09 at the 1 percent level. Variables 1 to 3 are fertility variables; variables 4 and 5 are social background variables; variables 6 to 9 are present social variables; and variables 10 to 13 are biopsychological variables. VO vo 100 education is greater in towns than in villages and still greater than in rural areas. Probably this is due to the fact that, while it is true that literacy is not saleable in a peasant economy, the hope of obtaining jobs outside the village induces people to force their sons and daugh ters to continue their schooling so that they may have the chance to achieve a better life in nonagricultural work. In the second group of variables, those which show present social characteristics, hypothesis Ilia predicted a strong association between husbarrLfe income and both husband's employment status and housing conditions, and among husband's employment status and housing conditions. These are precisely the strongest associations found as measured by the product moment correlation coefficient. The association between husband's income and housing con ditions is .34, between husband's income and employment status .25, and between husband's employment status and housing condition .24. There is also a fairly strong association between husband's income and couple's residen tial history (.21). Probably those more mobile and with residence in large towns have better jobs opportunities with higher salaries. They move in search of those oppor tunities. The association between husband's employment status and residential history is lower (.13). The third group of variables are those that are biological and attitudinal in nature. Hypothesis Ilia 101 predicted only a strong association between attitude toward family planning and use of contraception. The hypothesis is confirmed. Those who have a positive attitude toward family planning are more likely to use available contra ceptive methods; the correlation is .30. Correlations among other pairs of biological and attitudinal factors are each below .10. Relationships among Factors of Different Character Between Social Background and Present Social The associations hypothesized (Illb) are between education and husband's income, husband's employment status, and housing conditions. The three associations are confirmed. The higher the educational level people attain, the better jobs they hold, the more income they earn, and the more likely they will live in towns. For example, about 67 percent of husbands with High School graduation are employed in skilled labor or better, while only 30 percent of those without High School graduation have the same kind of jobs. Almost 75 percent of husbands with High School graduation have monthly earnings of 900 colones or more, whereas only 35 percent of those without High School graduation earn the same amount. About 75 percent of the couples whose husbands have not graduated from Elementary 102 School have never lived in towns of 20,000 people. This percentage lowers to 55 percent for the couples whose husbands have some High School education, and to near 25 percent if the husbands have graduated from High School or have some University education. Nothing was hypothesized regarding the association between couple’s residential background and present factors. Table 5.1 shows that those associations are also fairly strong: an r of .28 with residential history and husband's employment status, and an r of .27 with husband's income and housing conditions. The more rural the couple's residential background has been, the more likely the husband will be employed in either agricultural low or un skilled jobs, the lower the monthly earnings will be, the poorer their living conditions will be. Regarding the association with residential history, it must be noted that among husbands with less rural background, almost 50 per cent have moved more times and more often to large towns, while only 11 percent have had more rural experience after marriage than before. Between Biopsychological and Others Hypothsis Illb predicts strong associations between the following: couple's education and both atti tude toward family planning and use of contraception, husband's income and the same two family planning variables 103 and housing conditions and those two indicators. Extended family living experience does not show any strong association with any of the social background and present characteristics. The fact that only slightly under 10 percent of the wives in the age group under considera tion were currently living with their parents at the time of the interview, reduces in any event the weight of this variable in our analysis. If we assume that the recent fertility decline has resulted mainly from voluntary fertility control, any variables that influence fertility should have a fairly strong association with family planning variables. The last two colums of Table 5.1 do not fully substantiate this statement and the hypotheses stated above. Education correlates .15 with attitude toward family planning and .28 with use of contraception. In the case of housing conditions, the hypothesis is also confirmed; it correlates .13 with attitude toward family planning, and .27 with use of contraception. Husband's income correlates only weakly with use of contraception (.16). The same coefficient is found between husband's employment status and use of con traception, which had not been hypothesized. The associations of coital frequency and the various social background and present indicators are con sistently low. This variable, thus, does not seem to have much explanatory value in our scheme. 104 Many channels of influence on fertility are possible--channels going from social background character istics to fertility through present social factors and bio- psychological indicators. The main possible channels that emerge seem to be those shown in the following diagram (Figure 5.1). Only those channels suggested by a Pearsonian product moment correlation coefficient of .10 or more are shown here. The values shown beside the lines connecting the various variables are the Pearsonian r's. These channels will be examined in detail in the following chapters. Summary and Conclusions The variables selected in this study, key variables in social change as well as in fertility change, are related to one another as different measures of the same phenomena or as cause and effect in the process of social and fertility change. The above analysis suggests that in some cases there appears to be some overlapping in content and effects on fertility among some variables, while there also probably exists some strong channel of influence of these variables on fertility. The possible channels of influence of these variables on fertility are shown in figure 5.1. Figure 5.1. Associations among pairs of variables (figures in parentheses are Pearsonian R1s). EfRH [ (.15) (-.30) List of Variables: CE = Couple1s education CRB = Couple's RH = Couple's Resid. history; HES = Husband's Employment Status; HI = Husband's Income; AP = Attitude toward Family Planning; UC = Use of Contraception. 105 CHAPTER VI SOC10-STRUCTURAL CORRELATES OF FERTILITY It has been established in the preceding chapter that strong associations exist among certain predictors and that these associations may indicate that certain predic tors measure to some extent the same underlying character istic. This fact can be interpreted in two different ways, depending on the character of the predictors involved. The predictors can belong to either of two groups: the social background characteristics group, which includes couple's education and residential background, and present social characteristics group, which comprises couple's residential history, husband's income and employment status, and housing conditions. If the two predictors are from the same group, a strong association means that both are partially sharing the explanatory value about fertility differentials. On the other hand, if the two predictors are from different groups, a strong association can mean that the preceding predictor exerts much of its influence on fertility through the following predictor. The latter could, thus, be considered as "intervening" or "mediating" the influence of the preceding and fertility. We will be 106 107 observing each possibility by looking at the change in the differential fertility pattern when one variable is controlled for the effect of another. This will tell us not only whether or not this mediation is present and how large it is, but also whether or not there is a change in the character of fertility differential patterns. A negative relation may be changed to a curvilinear pattern, no pattern, or even to a positive relation when the effect of another predictor is removed. The change of pattern itself may be significant in explaining relationships between social structure and fertility. Two statistical techniques are used for this analysis. One is multiple classification analysis (MCA) and the other simple comparison of means. The multiple classification analysis technique provides the gross and net effects or unadjusted and adjusted deviations from the grand mean; by comparing both we can see the extent to which the unadjusted deviations are affected by the inter correlations with other independent variables. Further more, from the value of the correlation ratio or eta we can see the ability of the predictor to explain variation in the dependent variable. At the same time, we can examine the net fertility pattern to see how if differs from the original pattern. Multiple classification analysis can show inter action effects, but in the way it is used in the present 108 study, it does not. This can be accomplished by the comparison of means technique. It is analogous to cross tabulations where each mean summarizes the distribution of a complete row or column of a contingency table. It will reveal the different patterns within specific categories of a predictor. A shortcoming of this technique, though, is that often cells contain very few cases. As the tables show, the means for cells with less than 10 cases are not reported, and those for cells with between 10 and 19 cases are reported in parentheses. The questions asked are, thus: (l) if a strong association is found between two predictors, what can be said of their independent explanatory value in relation to fertility. In a more specific way we want to know (a) whether or not each of them keeps its independent explana tory value, if they are from the same group or character, and (b) whether or not the channels of influence diagrammed above seem to hold. For example, if we find a strong association between couple's education and residential background, we want to know to which extent each keeps its independent effect on fertility where the other is con trolled. If what we find is a strong association between couple's education and housing conditions, the question to be answered is whether the former influences fertility mainly through the latter or it maintains an influence.of its own. (2) How the gross fertility differentials change 109 when one predictor is controlled for the effect of another predictor. (3) Whether or not interaction effects exist between two predictors. Question (la) will be answered by comparing the gross and net effects of two predictors of the same character, that is, both from either social background or present characteristics. For question (lb), the same com parison will be used between two predictors of different character, that is, one social background characteristic and one present. Question (2) will be answered by studying the change in patterns between the gross and net effects and also fertility differentials for specific categories of a predictor as they are shown in the comparison of means tables. Finally, question (3) can be answered by examining all comparison of means tables and trying to find any major patterns of interaction between pairs of predictors. A further problem to be examined is what happens if several predictors are considered together. More specifi cally the following questions can and must be raised: (l) Does a social background characteristic exert its effect completely through a group of our specific intervening variables from the present characteristics group, or does it have an additional independent effect? (2) Does the effect of the "intervening" present characteristics depend completely on a group of social background characteristics 110 in explaining fertility differentials? (3) How does the net effect of a predictor differ from its gross effect when the concurrent effects of a group of predictors are removed? These questions will also be answered by probing the multiple classification analysis program outputs presented in the pertinent tables. The predictors will be considered by groups in the following way: all other pre dictors from the same group, the same adding age at union, the same adding duration of union, all predictors from the other group, the same adding age at union, the same adding duration of union, other possible combinations of predictors. The three measures of fertility, namely, number of live births or past fertility, wanted family size or desir ed fertility, and ideal family size or ideal fertility, will be examined separately. They do not necessarily have the same relationships with the independent variables. The intercorrelations among the three measures are as follows: past fertility correlates .39 with desired fertility, and .22 with ideal fertility; desired fertility correlates .17 with ideal fertility. The fact that the three indicators of fertility do not completely overlap confirms what on theoretical grounds is assumed, that is, that they reflect different dimensions of fertility. Consequently, their relationship with the independent Ill variables cannot be expected to follow the same patterns. For each of the three fertility measures, each of both the two social background and the four present characteristics will be considered. The analyses are based only on older rural and semi-urban Costa Rican couples with wives 30 to 49. Past Fertility Social Background Characteristics Hypothesis 1-1 predicts a strong association of these group of characteristics and fertility. It is seen in Table 6.1 that the two background characteristics taken together show a multiple correlation coefficient of .26 with past fertility, which is consistent with the hypothe sis. The association, however, is not strong. According to hypothesis 1-2, it is expected that part of the effect of these two indicators on fertility will be exerted through the present factors, and part will be, according to hypothesis 1-4, independent of this group of factors. Table 6.3 confirms the two hypotheses.* When *It may be useful to explain the set-up of the tables presenting the unadjusted and adjusted deviations obtained from the multiple classification analysis used in this study. For example, in Table 6.3, the dependent var iable is number of live births, indicated in the title of the table, and the independent variables and their sub classes are given under the left hand side heading. The means for each subclass of each independent variable are 112 either the effect of couple's education or residential background on past fertility is adjusted for the effects of the four present characteristics taken together, the devia tions are reduced but not altered. This indicates that the two social background characteristics have an effect on past fertility which is partly independent, partly depend ent on the intervening action of the present characteris tics . On the individual level, hypothesis II-l predicts a strong association of the two social background indicators on past fertility, and hypothesis II-2 that part of the effect of couple's education on past fertility will be exerted through husband's income and employment status, and housing conditions. The correlation ratios of .27 between couple's education and past fertility, and of .19 between couple's residential background and past fertility are given in the first column. The unadjusted deviations from the grand mean for each category of the independent vari ables are in the second column. Thus, the mean number of live births for the couples with no formal education is .48 more than the mean for all women (Grand Mean) given at the bottom of the table. The six columns to the right of the unadjusted deviations present the adjusted deviations for the categories of each variable when all other variables being considered in a given column are controlled. For in stance, in the fifth column from the left, the variables considered in the analysis are couple's residential his tory, husband's income and employment status, and housing conditions. The effect of the four variables other than couple's education on the number of live births for couples with no formal education is to reduce the deviation from .48 to .05. TABLE 6.1 MULTIPLE CORRELATION COEFFICIENTS (R)* BETWEEN THREE FERTILITY MEASURES AND GROUPS OF SOCIO-STRUCTURAL INDICATORS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Groups of Indicators Live Births Wanted Family Size Ideal Family Size Social Background (Couple's Education + Couple's Residential Background) .26 .15 .16 Social Background + Age at Union .41 .21 .21 Social Background + Duration of Union .36 .17 .22 Present Social (Residential History + Husband's Income + Husband's Employment Status + Housing Conditions) .28 .10 .20 Present Social + Age at Union .41 .19 .22 Present Social + Duration of Union .37 .15 .23 *The multiple correlation coefficients used in this table are from the multiple classification analysis program (adjusted for degrees of freedom). 113 114 consistent with hypothesis II-l (Table 6.2). Hypothesis II-2 is only partly confirmed. The net effect of couple's education on past fertility when either husband's income or employment status is controlled is almost comparable to the gross effect, indicating that neither of these two present factors mediate the effect of couple's education on past fertility. In the case of housing conditions, the hypo thesis is confirmed. Mainly in the case of couples with no formal education at all or less than one year of elementary school, the effect of illiteracy on actual fertility is almost all accounted for by the attitude they have toward consumption, represented here as housing conditions. It was not hypoth©sized that the effect of couple's residential background on pasft fertility would be partially mediated by either couple's education or housing conditions. From Table 5.2 it may be ©ncluded that this is the case. Thus, the mechanisms that the results shown in the tables suggest are: couple's education ---- > present factors > past fertility couple's residential background --- > present factors > fertility couple's education ---- > housing conditions > fertility couple's residential background > couple's education > fertility TABLE 6.2 MULTIPLE CLASSIFICATION ANALYSIS RELATING NUMBER OF LIVE BIRTHS TO SOCIAL BACKGROUND AND PRESENT SOCIAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 ____________ Number of Live Births________________ Deviation From No. Grand Adiusted Deviation of Characteristic Mean Mean CE CRB RH HI HES HC Cases Couple's Education (CE) Amount of Elementarv School: Both had less than 1 year 7.6 .48 .41 .39 .39 .40 .16 57 One incomplete 7.7 . 66 .60 .60 .60 .59 .54 171 Both incomplete 7.3 .20 .14 .16 .18 .18 .21 279 One completed 6.0 -1.1 -.96 -.96 -.96 -.95 -.89 87 Both completed 4.4 -2.7 -2.3 -2.4 -2.4 -2.4 -2.1 44 Correlation Ratio .27 Couple1s Residential Backq round (CRB) Both Rural 7.4 .32 .12 .22 .23 .26 .21 380 One from Village 7.4 .32 .23 .34 .32 .28 .39 95 Both from Village 7.2 .10 .50 .14 .32 .25 .31 39 One from Town 6.3 -.73 -.33 -.55 -.55 -.57 -.54 107 Both from Town 4.7 -2.3 -1.3 -2.0 -2.0 -2.1 -1.7 29 Correlation Ratio .19 Couple's Residential History (RH) No moves 7.4 .34 .18 .22 .18 to 00 .28 131 Some moves, no large town 7.4 .36 .27 .30 .34 .35 .31 361 Some moves, some large towns 6.1 -.93 -.61 -.72 -.77 -.82 i • 00 181 Correlation Ratio .17 H l l c h a n H • c T n n n m o I T - T T \ . J-U .ou .14 .32 .25 .31 39 One from Town 6.3 -.73 -.33 -.55 -.55 -.57 -.54 107 Both from Town 4.7 -2.3 -1.3 -2.0 -2.0 -2.1 -1.7 29 Correlation Ratio .19 Couple*s Residential History (RH) No moves 7.4 .34 .18 .22 00 H • .28 CO C M • 131 Some moves, no large town 7.4 .36 .27 .30 .34 .35 .31 361 Some moves, some large towns 6.1 -.93 -.61 -.72 -.77 -.82 -.78 181 Correlation Ratio .17 Husband * s Income (HI) No work— 100 colones 7.9 .81 .70 .85 .81 .65 .67 45 100-300 colones 7.5 .41 .19 .31 .30 .34 .15 282 300-600 colones 7.1 .01 .00 .00 .05 -.03 .09 132 600 colones or more 5.9 -1.1 -.44 -.79 -1.0 -.92 -.66 87 Correlation Ratio .17 Husband’s Employment Status (HES) Tenant farmers 7.3 .19 .08 .10 .12 .13 .11 516 Unskilled labor 6.9 -.16 .19 .24 -.04 .09 .13 50 Farmers or skilled labor 5.6 -1.5 -.86 H • H 1 -1.3 -1.1 H • H 1 82 Correlation Ratio .17 • Housing Conditions (HC) Very Poor 7.9 .80 .51 .74 .76 .68 .69 149 Poor 7.4 .36 .26 .31 .32 .31 .32 305 Some Comforts 6.4 -.63 -.47 -.62 -.58 -.56 -.58 147 Comfortable 5.5 -1.5 -.86 -1.3 -1.4 -1.3 -1.3 76 Correlation Ratio .24 Grand Mean 7.1 TABLE 6.3 MULTIPLE CLASSIFICATION ANALYSIS RELATING NUMBER OF LIVE BIRTHS TO SOCIAL BACKGROUND CHARACTERISTICS BY GROUPS OF INDICATORS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Social Background Characteristic Deviation from Grand Mean Mean A Adjusted Deviation* B C D E F No of Cases Couple’s Education Amount of Elementary School: Both had less than 1 year 7.6 .48 .42 .53 .05 .10 .09 .20 57 One incomplete 7.7 .66 .66 .52 .44 .48 .53 .39 171 Both incomplete 7.3 .20 .11 .11 .17 .16 .13 .13 279 One completed 6.0 -1.1 -1.0 -.88 -.73 -.82 -.89 -.72 87 Both completed 4.4 -2.7 -2.0 -2.0 -1.7 -1.8 -1.4 -1.4 44. Couple1s Residential Backqround Both Rural 7.4 .32 -.02 .08 .09 .09 -.05 .05 380 One from Village 7.4 .32 .47 .30 .39 .30 .56 .37 95 Both from Village 7.2 .10 .60 .32 .50 .53 .62 .37 39 One from Town 6.3 -.73 -.23 -.33 -.30 -.29 -.18 -.29 107 Both from Town 4.7 -2.3 -1.3 -1.2 -1.4 -1.1 -1.1 -1.0 29 Grand Mean 7.1 *Predictor Variables in each group: A - Other social background factor adding age at union B - Other social background factor adding duration of union from Adjusted Deviation* No Social Background Grand of Characteristic Mean Mean A B C D E P Cases Couple's Education Amount of Elementary School: Both had less than 1 year 7.6 .48 .42 .53 .05 .10 .09 < .20 57 One incomplete 7.7 .66 . 66 .52 .44 .48 .53 .39 171 Both incomplete 7.3 .20 .11 .11 .17 .16 .13 .13 279 One completed 6.0 -1.1 -1.0 -.88 -.73 -.82 -.89 -.72 87 Both completed 4.4 -2.7 -2.0 -2.0 -1.7 -1.8 -1.4 -1.4 44 Couple1s Residential Background Both Rural 7.4 .32 -.02 .08 .09 .09 -.05 .05 380 One from Village 7.4 .32 .47 .30 .39 .30 .56 .37 95 Both from Village 7.2 .10 .60 .32 .50 .53 .62 .37 39 One from Town 6.3 -.73 -.23 -.33 -.30 -.29 -.18 -.29 107 Both from Town 4.7 -2.3 -1.3 -1.2 -1.4 -1.1 -1.1 -1.0 29 Grand Mean 7.1 *Predictor Variables in each group: A - Other social background factor adding age at union B - Other social background factor adding duration of union C - All present characteristics D - Couple's education/residential background adding housing conditions E - Column D adding age at union F - Column D adding duration of union 117 couple's residential background ----> housing conditions > fertility Discussion This seems to lead to some clear conclusions. (i) Couple's education has exerted a distinctive effect in the differential fertility performance of rural and semi-urban Costa Rican couples, as expected. By itself, this indicator is a good predictor of fertility. (II) Couple's education has had a differential effect on past fertility of rural and semi-urban Costa Rican couples insofar as it has allowed the husband to be employed in a higher level job, earn a higher salary, and also because it has allowed the couple to live in better housing conditions. These characteristics, most of the time, have proved to be incom patible with a large family. Couple's education has, thus, negatively influenced fertility in an indirect way in per formance .' This conclusion is reinforced when the comparison of means tables are examined (Tables 6.4 and 6.5). Those tables show in a more detailed manner in which way couple's education affects fertility within each specific sub category of the other variables under consideration. They clearly indicate (a) that among rural and semi-urban Costa Rican couples, their education has been operating in a negative way in producing fertility differentials; (b) that TABLE 6.4 NUMBER OF LIVE BIRTHS BY WIFE'S AND HUSBAND'S EDUCATION AND RESIDENTIAL BACKGROUND, RURAL COSTA RICA COUPLES WITH WIVES 30-49 Number of Live Births Residential Wife's Education Husband's Education Rackaround 1 2 3 T 1 2 3 T N Wife's Rural 7.4 7.6 5.2 6.7 8.0 7.3 6.1 7.1 478 Village (7.9) 6.9 6.7 6.8 (7.7) 6.7 (6.9) 7.1 106 Town A z ) 6.3 4.0 5.2 (-) 6.8 3.9 5.3 97 Total 7.5 7.3 5.3 6.4 7.9 6.9 5.0 6.6 Husband's Rural 7.5 7.4 5.6 6.8 8.0 7.1 5.9 7.0 478 Village (8.4) 7.3 4.9 6.8 (-) 7.1 5.2 6.1 100 Town (7.4) 7.1 4.6 6.5 (-)__ 7.2 4.4 5.8 73 Total 7.6 7.1 5.0 6.6 (-) 7.1 5.2 6.3 N 170 426 86 144 402 92 683 (-) Base number less than 10 cases ( ) Less than 20 but at least 10 cases Code for Education: 1 = Less than 1 year of elementary school 2 = Elementary school incomplete £ 3 = Elementary school completed co TABLE 6.5 NUMBER OP LIVE BIRTHS BY COUPLE’S EDUCATION AND RESIDENTIAL BACKGROUND AND PRESENT SOCIAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Present Social Number of Live Births Character Couple's Education Couple1s Resid. Back. istics 1 2 3 4 5 T 1 2 3 4 5 T N Couple's Residen tial History No moves Some moves, no large town Some moves, some large towns (7.7) 7.7 * 7.7 8.0 8.1 8.0 7.5 6.1 (6.1) 6.3 5.6 * 7.1 (4.7)6.8 4.2 6.0 7.2 7.5 7.2 * 7.7 6.3 (8.5) * (7.5)6.9 (5.3)5.8 * 7.8 * 7.4 3.9 5.7 131 361 181 Total 7.7 7.6 7.2 6.0 4.4 6.6 7.3 7.0 7.1 6.3 3.9 6.9 Husband's Income No Work-100 colones * * 8.1 * * 8.1 7.9 * * * * 7.9 45 100-300 colones 7.9 7.8 7.3 6.4 * 7.3 7.5 7.9 (7.2)6.9 * 7.4 282 300-600 colones * 7.7 7.2 (6.0) . * 7.0 6.9 7.4 * 7.2 * 7.1 132 600 colones or More * * 6.7 5.7 4.3 5.6 (7.7)(6.0) * 5.5 (4.0)5.8 87 Total 7.9 7.7 7.3 6.0 4.3 7.0 7.5 7.1 7.2 6.5 4.0 7.0 Husband's Employment Status Tenant Farmers 7.5 7.7 7.5 5.6 5.3 6.7 7.4 7.6 7.1 6.7 * 7.2 516 Unskilled Labor * * 7.3 * * 7.3 7.7 * * (6.9) * 7.3 50 Farmers or Skilled Labor * * 5.7 6.2 3.9 5.3 5.9 * * 5.3 (4.2)5.1 82 Total 7.5 7.7 6.8 5.9 4.6 6.4 7.0 7.6 7.1 6.3 4.2 6.5 Housinq Conditions Very Poor 7.7 8.0 7.8 * * 7.8 7.9 (7.8) * (9.5) * 8.4 148 Poor 7.3 8.2 7.4 6.1 * 7.5 7.4 8.5 (7.0)6.8 * 7.4 305 Some Comforts * 7.1 6.7 5.9 (5.4) 6.3 6.9 6.4 * 5.4 * 6.2 147 « - . . . . / ? ___1_1 _ M. t Z / e c . \ a r\ e n c. n - V- ( n i W/i k U q i \ e a n & Husband’s Income No Work-100 colones 100-300 colones 300-600 colones 600 colones or More * 7.9 * * * 7.8 7.7 * 8.1 7.3 7.2 6.7 * 6.4 (6.0) 5.7 * * * 4.3 8.1 7.3 7.0 5.6 7.9 * 7.5 7.9 6.9 7.4 (7.7)(6.0) * * * 7.9 (7.2)6.9 * 7.4 * 7.2 * 7.1 * 5.5 (4.0)5.8 45 282 132 87 Total 7.9 7.7 7.3 6.0 4.3 7.0 7.5 ,7.1 7.2 6.5 4.0 7.0 Husband’s Employment Status Tenant Farmers 7.5 7.7 7.5 5.6' 5.3 6.7 7.4 7.6 7.1 6.7 * 7.2 516 Unskilled Labor * * 7.3 * * 7.3 7.7 * * (6.9) * 7.3 50 Farmers or Skilled Labor * * 5.7 6.2 3.9 5.3 5.9 * * 5.3 (4.2)5.1 82 Total 7.5 7.7 6.8 5.9 4.6 6.4 7.0 7.6 7.1 6.3 4.2 6.5 Housinq Conditions Very Poor 7.7 8.0 7.8 * * 7.8 7.9 (7.8) * (9.5) * 8.4 148 Poor 7.3 8.2 7.4 6.1 * 7.5 7.4 8.5 (7.0)6.8 * 7.4 305 Some Comforts * 7.1 6.7 5.9 (5.4) 6.3 6.9 6.4 * 5.4 * 6.2 147 Comfortable * * 7.5 (5.6) 4.0 5.7 6.9 * (7.1)(4.5)(3.1)5.4 76 Total 7.5 7.8 7.3 5.9 4.7 6.8 7.3 7.6 7.0 6.5 3.1 6.8 N 57 171 279 87 44 380 95 39 107 29 683 *Base number less than 10 cases ( ) Less than 20 but at least 10 cases Code for Education: 1 = Both had less than 1 year of elementary school 2 = One elementary school incomplete 3 = Both elementary school incomplete 4 = One elementary school completed 5 = Both elementary School completed Code for Residential Background: 1 = Both rural 2 = One from village 3 = Both from village 4 = One from town 5 = Both from town 120 this negative effect starts operating when at least either husband or wife has completed elementary school; and (c) that this negative effect has been at its highest when both have completed elementary school and both have had a residential background mainly in towns, or their residen tial history includes at least one town of 20,000 people, or the husband's monthly salary is at least of 600 colones, or the husband's employment status is at an intermediate level or in a skilled job, or they live in comfortable housing conditions. (ill) Couple's education has had a differential negative effect on fertility performance in part due to the mediating effect of housing conditions. The more educated tend to live in better housing conditions (roomier and with more comforts), which, as it was noted before, is most of the time not compatible with a large family. (IV) Couple's residential background by itself has had a clear differential effect on reducing past fertility; on the other hand, this effect is explained partly in the sense that those who have resided in towns for most of their 15 first years of age have had better educational opportunities, and better education, all of which, as noted above, leads to a decrease in fertility. Further more, another possible channel of influence suggested by the results is couple's residential background --- > couple's education --- > housing conditions > 121 fertility, that is, the mediating influence of couple's education between residential background and fertility would be partially exerted through housing conditions, for the reasons given above. The comparison of means (Table 6.4) of couple's residential background in combination with any of the present factors suggest the same interactive effect that we found between these factors and couple's education. The combination of residential background in a town with either residential history that includes at least one town of 20,000 people, or a husband's monthly income of 600 colones or more, or a husband's employment status in an intermediate agricultural job or in skilled labor, or living conditions which are comfortable, have produced a fertility performance lower than average. These conclusions are in agreement with the find ings of the majority of studies that have analyzed the differential fertility of developing countries. Education and residential background have a powerful influence in shaping attitudes and aspirations which in turn lead to a reduction in fertility performance. At this respect housing conditions would seem a good indicator of those attitudes and aspirations. The way a couple lives indi cates not only the ability to afford a larger and more comfortable house, but also a willingness to either main tain or improve a way of living which, most of the time, 122 could not be sustained if they were to have and support a large family. Present Social Characteristics This group of indicators is expected to be strongly associated with past fertility, as hypothesis 1-1 predicts. Table 6.1 shows that the multiple correlation coefficient is .28. The hypothesis is, thus, confirmed, even if the association is not strong. It was also anticipated that present social factors would have independent effect on past fertility, according to hypothesis 1-5. This is also confirmed. When the effect of other social background or- present social variables, it keeps its own independent effect on past fertility (Tables 6.2 and 6.6). The prediction of a strong association of the present social characteristics with past fertility is con firmed (hypothesis II-2). The correlation ratio of each couple's residential history, husband's income and employ ment status with number of live births is .17, and that of housing conditions is .24. The associations, however, are not strong. Income, as it was anticipated (hypothesis IV-l), retains its negative effect on past fertility when the effects of other social background and present social characteristics are taken into consideration. TABLE 6.6 MULTIPLE CLASSIFICATION ANALYSIS RELATING NUMBER OF LIVE BIRTHS TO PRESENT SOCIAL CHARACTERISTICS BY GROUPS OF INDICATORS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Present Social Characteristic Mean Deviation from Grand Mean A Adjusted B C Deviation D E F No. of Cases Couple's Residen tial Historv No Moves 7.4 .34 .21 .11 .10 .13 .08 .05 131 Some Moves, no large town 7.4 .36 .32 .13 .31 .25 .06 .25 361 Some Moves, some large towns 6.1 -.93 -.71 -.31 -.62 -.55 .16 -.46 181 Husband ' s Income No Work- 100 Colones 7.9 .81 .42 .41 .39 .76 .54 .62 45 100-300 Colones 7.5 .41 .05 .09 .02 .16 .17 .13 282 300-600 Colones 7.1 .01 .06 -.04 -.01 .00 .03 -.05 132 600 Colones or More 5.9 -1.1 -.37 -.29 -.20 -.30 .21 -.16 87 Husband's Emplov- ment Status Tenant Farmers 7.3 .19 .04 .04 .03 .04 .01 .02 516 Unskilled Labor 6.9 -.16 -.17 -.13 -.18 .40 .56 .43 50 Farmers or Skilled Labor 5.6 -1.5 -.83 -.75 -.68 -.71 .67 -.62 82 Housinq Conditions Very Poor 7.9 .80 .65 .72 .74 .54 .62 .60 149 Poor 7.4 .36 .27 .26 .23 .25 .23 .21 305 Some Comforts 6.4 -.63 -.53 -.59 -.51 -.50 .55 -.46 147 Comfortable 5.5 -1.5 -1.1 -1.2 -1.2 -.82 .88 -.95 76 Grand Mean 7.1 Predictor Variables in each group: No Work- 100 Colones 7.9 .81 .42 .41 .39 .76 .54 .62 45 100-300 Colones 7.5 .41 .05 .09 .02 .16 .17 .13 282 300-600 Colones 7.1 .01 .06 -.04 -.01 .00 -.03 -.05 132 600 Colones or More 5.9 -1.1 -.37 -.29 -.20 -.30 -.21 -.16 87 Husband's Employ ment Status Tenant Farmers 7.3 .19 .04 .04 .03 .04 .01 .02 516 Unskilled Labor 6.9 -.16 -.17 -.13 -.18 .40 .56 .43 50 Farmers or Skilled Labor 5.6 -1.5 -.83 -.75 -.68 -.71 -.67 -.62 82 Housinq Conditions Very Poor 7.9 .80 .65 .72 .74 .54 .62 .60 149 Poor 7.4 .36 .27 .26 .23 .25 .23 .21 305 Some Comforts 6.4 -.63 -.53 -.59 -.51 -.50 -.55 -.46 147 Comfortable 5.5 -1.5 -1.1 -1.2 -1.2 -.82 -.88 -.95 76 Grand Mean 7.1 Predictor Variables in each group: A - Other present social characteristics B - Column A adding age at union C - Column A adding duration of union D - Social Background characteristics E - Column D adding age at union F - Column D adding duration of union H to to 124 Husband's employment status retains in most of the cases its independent effect on number of live births when controls are introduced. It had been expected a weakening of that independent effect (hypothesis IV-2). Only when either couple's education or the other three present social characteristics together are considered, the net effects of husband's employment status is considerably lower than its gross effects. It was expected that housing conditions would keep a strong effect on past fertility when income is controlled but that its effect would weaken when couple's education is taken into account (hypothesis IV-3). Our findings confirm these anticipated mechanisms of influence. The tables reveal also other interesting relation ships. The effect of couple's residential history on actual fertility is weakened when either couple's educa tion, or both couple's education and residential back ground, or these two social background variables plus either age at union or duration of union are controlled. The effect of income on number of live births remains fairly strong, as it was mentioned above, but there is a partial weakening of its effect when any of the following controls are used: couple's education, couple's residential background, both together, all the other present social characteristics taken at the same time. The addition of any of the two demographic controls, namely, 125 age at union and duration of union, to the social back ground or the present social indicators, further reduces the independent effect of income on past fertility. Finally, in the case of husband's employment status, the added control of either age at union or dura tion of union to the social background or the present social variables, weakens still more its effect on past fertility. Discussion A careful observation of the summary above allows the following conclusions. (i) Each one of the four present characteristics has exerted a distinctive effect in producing a differen tial fertility performance among rural and semi-urban Costa Rican couples than in other developing countries, in agreement with the theoretical scheme proposed in Chapter II. (II) Couple's education explains in part the effect that every one of the present factors has on fertility. It is the one single factor which plays a role in the relationship of each one of the variables so far considered. (ill) Both residential history and housing condi tions operate independent of the other two present factors, namely, husband's income and employment status, in their 126 effect on fertility performance. Where the couple has resided and in which housing conditions they are living has had a marked differential effect on fertility performance. Perhaps these variables are suggesting an attitude incompa tible with a large family. Those who move and go to live in cities, do so usually in search of better job opportuni ties, better salaries, more comfortable surroundings, all of which indicates a more modern attitude toward life, regardless of what they achieve. Besides, in large towns they have a better chance to assimilate modernizing norms leading to a lower actual fertility. On the other hand, those who live in better housing conditions show also, as it was suggested above, a modern attitude toward consump tion, an effort to improve their conditions. In other words, it is not as much how much they earn but how they sprend their money that influences also their fertility performance. The fact that either couple's education alone or in combination with couple's residential background weakens to a considerable extent the effect of either residential history and housing conditions on past fertility indicates that it is not totally residential history or housing conditions by themselves that explain fertility differen tials, but in part it is due to the couple's education and residential background. (IV) Husband's income has been greatly dependent on 127 other characteristics in differently affecting fertility behavior. It is true that it always keeps a certain explanatory value, and the pattern of association with past fertility is always negative. Its negative effect on fertility, though, is largely dependent on couple’s education, housing conditions, couple's residential back ground; the two social background factors together, or the other present factors. This indicates that income is the less independent of all the socioeconomic indicators used in this study. In other words, it is not as much income per se what explains differential fertility, but the couple's other socioeconomic characteristics. (V) The effect of husband's employment status on past fertility is to a certain extent independent of the effect of any other indicator and partly dependent. The association of this factor with past fertility weakens con siderably, but not as much as in the case of husband's income, when the effects of other socioeconomic factors are taken into account. (VI) The two control variables, age at union and duration of union, when applied, have a marked effect in reducing the fertility differentials in several cases. The effect of residential history on past fertility practically disappears when age at union is included as a control over and above either social background factors or present factors. Also in the case of husband's income, the 128 adjustment for any of the two controls further decreases the effect of that variable on past fertility. (VII) The comparison of means Tables 6.7 and 6.8 suggest that there are interaction effects between any pair of indicators from this group. A reinforcing effect in lowering fertility performance is found when couples who have resided in at least one town of 20,000 population find themselves in the highest income bracket, or in the highest employment status, or living in comfortable housing conditions; also in the case of couples whose husbands are in the highest occupational level and earn the highest monthly income or live in comfortable housing conditions; and finally in the case of couples who live in comfortable housing conditions and whose husbands are in the highest income bracket. These data suggest important socio-structural differences among rural and semi-urban Costa Rican couples. Although complete interpretations of these differences are not justified, certain implications can be explored. Education is the most important factor affecting fertility. Residential background, housing conditions, and residential history are also important, but less than education. Husband's employment status and husband's income have the lower independent effect on past fertility. Education acts in a rural or semi-urban society in the same way that it has been operating in urban areas. TABLE 6.7 NUMBER OF LIVE BIRTHS BY COUPLE'S RESIDENTIAL HISTORY AND HOUSING CONDITIONS AND OTHER PRESENT SOCIAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Current Character istics Number of Live Births Couple's Residential History Housinq Conditions N 1 2 3 T 1 2 3 4 T Housinq Conditions 1 Very poor 8.1 8.0 7.4 7.8 148 2 Poor 7.6 7.7 6.7 7.3 305 3 Some Comforts 6.4 6.8 5.9 6.3 147 4 Comfortable (7.0) 6.3 4.0 5.8 76 Total 7.3 7.2 6.0 6.8 Husband’s Income No Work- 100 Colones (8.6) 7.7 (7.5) 7.9 (8.9) 7.4 * * 8.1 45 100-300 Colones 7.8 7.6 6.7 7.4 7.8 7.5 6.8 * 7.4 282 300-600 Colones 6.5 7.9 6.1 6.8 (8.1) 7.9 6.1 (4.9) 6.7 132 600-Colones or More * 6.4 5.2 5.8 * 7.1 5.7 5.0 5.9 87 Total 7.6 7.2 6.0 6.9 8.2 7.5 6.2 5.0 7.0 Husband's Employ ment Status Tenant Farmers 7.4 7.7 6.1 7.1 7.9 7.5 6.5 6.1 7.0 516 Unskilled Labor * 6.9 6.8 6.8 * 7.5 (6.9) (5.1) 6.5 50 Farmers or Skilled Labor (6.4) 5.5 5.4 5.8 * 6.2 5.9 4.6 5.6 82 Total 6.9 6.7 6.1 6.6 * 7.1 6.4 5.3 6.4 N 131 361 181 148 305 147 76 *Base number less than 10 ( )Less than 20 but at least 10 or More *_______ b.4 5. 4. b.ts * /.jl s. /____ s. u o. y «'/ Total 7.6 7.2 6.0 6.9 8.2 7.5 6.2 5.0 7.0 Husband's Emplov- ment Status Tenant Farmers 7.4 7.7 6.1 7.1 7.9 7.5 6.5 6.1 7.0 516 Unskilled Labor * 6.9 6.8 6.8 * 7.5 (6.9) (5.1) 6.5 50 Farmers or Skilled Labor (6.4) 5.5 5.4 5.8 * 6.2 5.9 4.6 5.6 82 Total 6.9 6.7 6.1 6.6 * 7.1 6.4 5.3 6.4 N 131 361 181 148 305 147 76 *Base number less than 10 ( )Less than 20 but at least 10 Code for Couple's Residential History: 1 = No Moves 2 = Some Moves, no large towns 3 = Some Moves, some large towns NUMBER TABLE 6.8 OF LIVE BIRTHS BY HUSBAND'S INCOME RURAL COSTA RICAN COUPLES WITH AND EMPLOYMENT WIVES 30-49 STATUS, Husband's Employment Number of Live Births Status Husband's Income No Work-100 100-300 300-600 600 Colones Total N Colones Colones Colones or More Tenant Farmers 8.3 7.6 7.3 5.8 7.2 516 Unskilled Labor * * 6.7 (6.5) 6.6 50 Farmers or Skilled Labor * (6.2) (6.5) 5.4 6.0 82 Total * 6.9 6.8 5.9 6.6 N 45 282 132 87 683 *Base number less than 10 ( )Less than 20 but at least 10 130 Even if literacy is not saleable in a peasant economy, school attendance seems to be a function of hopes for jobs outside the farm or the village, for better life. The shortage of land and the increasing population are undoubtedly factors which turn the farmer's and villager's attention to the city, to new industries nearby which demand a higher level of literacy. This in turn leads to new aspirations and attitudes non compatible with a large family. Regardless of whatever classificatory position the couple occupies in the other sociostructural factors considered in this study, education has always had a decreasing effect on fertility. The fact that the country is intensifying its efforts to retain children until grad uation from elementary school should lead to a greater decrease in fertility in the next decades. The semi-urban and urban population of the country has been increasing in the recent years, and the internal migration into cities is in the upsurge. This should also contribute to a reduction in fertility based on the find ings about the effects of both residential background and residential history. A smaller proportion of the popula tion will be raised in farms and or small villages, and more people will be moving into cities and towns or among them. Husband's employment status and husband's income emerge as relatively unimportant in their ability to 131 independently explain fertility differentials. Desired Fertility It is recognized that attitudes and desires do not seem of prime importance in determining average family size in a country like Costa Rica, mainly in its rural and semi- urban areas, where family planning has not been a common practice in married life. Nevertheless, it is of some value to know views on ideal family size and the number of children they actually want. In dealing with the socio-structural differentials in desired fertility, the purpose will be mainly to point out the differences from the patterns found in past fertility differentials. It should not be assumed, of course, that the number of children a wife wants remains constant throughout the childbearing years of her married life. Even if there were no major economic and social changes in the society as a whole, as in fact it was the case during the sixties in Costa Rica, the complex multitude of factors that affect the daily lives of individual couples would undoubtedly cause some wives to revise their personal family size pre ferences upward or downward from time to time. Our data reflect the desired family size of rural and semi-urban Costa Rican wives at one point in time. Generally speaking, it can be said that the desired 132 fertility differentials are much smaller than past fertility differentials, and that this occurs because those wives who have a higher than average number of children desire a family closer in size to the average and those who have a much smaller than average number of live births desire a family larger in size than they actually have and hence closer to the average. Social Background Characteristics The multiple correlation coefficient between couple's education and residential background taken together and desired fertility is .15, which indicates that hypothesis 1-1 even if confirmed does not prove to be strongly supported (Table 6.1). Each of the two variables in this group, couple’s education and residential background, as hypothesis II-l predicts, is associated with desired fertility, education more than residential background. Hypothesis 1-2 is only partly confirmed. It predicts that each of these two factors will exert their effect on desired fertility through the present factors, and it proves to be so only in the case of residential background (Table 6.10). Hypothesis II-2 is not confirmed. Education exerts its influence on desired fertility independent of any of the present factors. Nothing was predicted as to the way in which residential background exerts its effect on TABLE 6.9 MULTIPLE CLASSIFICATION ANALYSIS RELATING WANTED FAMILY SIZE TO SOCIAL BACKGROUND AND PRESENT SOCIAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Wanted Family Size Deviation from No. Grand Adlusted Deviation of Characteristic Mean Mean CE CRB RH HI HES HC Cases Couple's Education (CE) Amount of Ele mentary School: Both had less than 1 year 6.8 .18 .17 .16 .16 .19 .12 57 One incomplete 6.7 .12 .11 .12 .11 .12 .05 171 Both incomplete 6.6 .01 .00 .01 .00 .01 .01 279 One completed 6.7 .05 .06 .07 .08 .05 .09 87 Both completed 5.6 -1.0 -1.0 -.99 -.98 -1.0 -.89 44 Correlation Ratio .18 Couple's Residential Backqround (CRB) Both Rural 6.7 .09 .04 .08 .07 .09 .06 380 One from Village 6.6 -.02 -.07 .02 -.01 -.04 -.03 95 Both from Village 6.7 .04 .16 .02 .06 .06 .13 39 One from Town 6.5 -.10 .02 -.04 -.05 -.08 -.05 107 Both from Town 6.0 -.62 -.30 -.54 -.49 -.57 -.41 29 Correlation Ratio .10 Couple's Residential History (RH) No moves 6.9 .27 .22 .23 .23 .26 .26 131 Some moves, no large town 6.6 -.02 -.04 -.03 -.02 -.02 -.04 361 Some moves, some large towns 6.4 -.19 -.09 -.14 -.15 -.17 -.15 181 Correlation Ratio .10 Husband's Income (Hi) Both from Village 6.7 .04 .16 .02 .06 .06 .13 39 One from Town 6.5 -.10 .02 -.04 -.05 -.08 -.05 107 Both from Town 6.0 -.62 -.30 -.54 -.49 -.57 -.41 29 Correlation Ratio •10 Couple1s Residential History (RH) No moves 6.9 .27 .22 .23 .23 .26 .26 131 Some moves, no large town 6.6 C\J O • i i • O >12. -.03 -.02 -.02 -.04 361 Some moves, some larqe towns 6.4 -.19 -.09 -.14 -.15 i • H < 1 -.15 181 Correlation Ratio» .10 Husband's Income (Hi) No work- 100 Colones 6.8 .20 .15 .21 .19 .17 .16 45 100-300 Colones 6.7 .06 .00 .04 .03 .05 -.01 282 300-600 Colones 6.6 -.05 -.06 -.06 -.04 -.04 -.04 132 600 Colones or more 6.2 -.38 -.16 -.29 -.31 -.36 -.23 87 Correlation Ratio> 10 Husband's Employment Status (HES) Tenant Farmers 6.6 .03 -.01 .00 .00 .00 .00 516 Unskilled Labor 6.3 -.32 -.20 -.22 -.26 -.24 -.24 50 Farmers or skilled Labor 6.5 -.10 .09 -.02 -.05 .00 .03 82 Correlation Ratio ,07 Housinq Conditions (HC) Very poor 6.8 .16 .09 .15 .13 .15 .15 149 Poor 6.7 .08 .05 .07 .09 .07 .08 305 Some Comforts 6.6 -.06 -.05 -.05 -.04 -.04 -.05 147 Comfortable 6.1 -.54 -.28 -.49 -.54 -.51 -.53 76 Correlation Ratio .12 Grand Mean 6.6 134 desired fertility, but the tables show that it is partly due to the effect of education and housing conditions (Table 6.9). The following conclusions seem, thus, in order. (i) The social background characteristics have less predictive value for desired fertility than they proved to have for past fertility. (II) Couple's education has a differential effect on desired fertility completely independent of any other socio-structural factor. (III) Couple's residential background has a signi ficant differential effect on desired fertility, but it is smaller than that of education and partly due to the intervening influence of either education or the present characteristics, mainly housing conditions. The comparison of means in Tables (6.11 and 6.12; se.e also Appendix B) show that between pairs of social background and present factors, the reinforcing effect, so pronounced in the case of past fertility, is much smaller for desired fertility. Present Characteristics The same pattern found in the case of the relation ship of these variables to past fertility is visible when they are associated with desired fertility. The relation ships are present but weaker than in the case of past TABLE 6.10 MULTIPLE CLASSIFICATION ANALYSIS RELATING WANTED FAMILY SIZE TO SOCIAL BACKGROUND CHARACTERISTICS BY GROUPS OF INDICATORS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Social Background Characteristic Mean Deviation from Grand Mean A Adlusted Deviation B C D E F No. of Cases Couple's Education Amount of Ele- mentarv School: Both had less than 1 year 6.8 .18 .16 .12 .11 .11 .10 .06 57 One incomplete 6.7 .12 .10 .09 .10 .08 .07 .06 171 Both incomplete 6.6 .01 -.01 .01 .00 .00 .01 .01 279 One completed 6.7 .05 .06 .06 .09 .09 .09 .09 87 Both completed 5.6 -1.0 -.90 -.94 -.82 -.85 .74 -.78 44 Couple's Residential Backqround Both Rural 6.7 .09 .02 .03 .04 .04 .01 .02 380 One from Village 6.6 -.02 -.04 -.05 .03 -.07 -.03 -.04 95 Both from Village 6.7 .04 .21 .15 .09 .18 .24 .18 39 One from Town 6.5 -.10 .03 .02 .01 .02 .04 .03 107 Both from Town 6.0 -.62 -.29 -.22 -.30 -.23 .22 -.15 29 Grand Mean 6.6 Predictor Variables in each group: A - Other social background factor adding age at union B - Other social background factor adding duration of union C - All present social characteristics D - Couple's education/residential background and housing conditions E - Column D adding age at union F - Column D adding duration of union________________________________ TABLE 6.11 Grand Mean 6.6 Predictor Variables in each group: A - Other social background factor adding age at union B - Other social background factor adding duration of union C - All present social characteristics D - Couple's education/residential background and housing conditions E - Column D adding age at union F - Column D adding duration of union________________________________ TABLE 6.11 WANTED FAMILY SIZE BY WIFE'S AND HUSBAND'S EDUCATION AND RESIDENTIAL BACKGROUND, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Wanted Family Size Residential Wife's Education Husband's Education Backqround 1 2 3 T 1 2 3 T N Wife1s Rural 6.6 6.8 6.1 6.5 6.8 6.7 6.7 6.7 478 Village (7.0) 6.7 6.5 6.7 (7.0) 6.5 (6.7) 6.7 106 Town (-) 6.6 5.4 6.0 (-) 6.4 5.8 6.1 97 Total 6.8 6.7 6.0 6.5 6.9 6.5 5.2 6.5 Husband1s Rural 6.7 6.7 6.7 6.7 6.8 6.7 6.8 6.8 478 Village (6.9) 6.4 5.4 6.2 (-) 6.1 6.2 6.1 100 Town (6.9) 7.1 5.4 6.5 (-) 6.9 5.8 6.4 73 Total 6.8 6.7 5.8 6.4 (-) 6.6 6.3 6.4 N 170 426 86 ..144 402 92 683 (-) Base number less than 10 cases ( ) Less than 20 but at least 10 cases Code for Education: 1 = Less than 1 year of elementary school 2 = Elementary school incomplete 3 = s Elementary school completed TABLE 6.12 WANTED FAMILY SIZE BY COUPLE’S EDUCATION AND RESIDENTIAL BACKGROUND AND PRESENT SOCIAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Present Wanted Family Size Character istic Couple's Education Couple's Res id. Back. 1 2 3 4 5 T 1 2 3 4 5 T N Couple's Residen tial History No moves (6.7) 6.8 7.1 (6.8) * 6.8 6.9 * (6.8) * * 6.8 131 Some moves, no large town 6.7 6.7 6.5 6.6 (6.1) 6.5 6.7 6.4 (6.5) 6.6 * 6.5 361 Some moves, some large towns * 6.7 6.5 6.7 5.3 6.3 6.4 6.7 (6.7) 6.5 5.7 6.4 181 Total 6.7 6.7 6.7 6.7 5.7 6.5 6.7 6.5 6.7 6.5 5.7 6.6 Husband1s Income No work-100 Colones * * l'<3 * * . 7.3 7.0 * * * * 7.0 45 100-300 Colones 6.8 6.8 6.6 6.8 * 6.7 6.7 6.9 (6.2) 6.5 * 6.6 282 300-600 Colones * 6.5 6.8 (6.6) * 6.6 6.5 6.8 * 6.7 * 6.7 132 600 Colones or more * * 6.0 6.8 5.6 6.1 (6.7)(5.3) * 6.4(5.9)6.1 87 Total 6.8 6.6 6.7 6.7 5.6 6.6 6.7 6.3 6.2 6.5 5.9 6.6 Husband's Employment Status Tenant Farmers 6.8 6.8 6.7 6.6 (5.4) 6.5 6.7 6.8 6.4 6.4 * 6.6 516 Unskilled Labor * * 6.5 * * 6.5 6.3 * * (7.3) * 6.8 50 Farmers or Skilled Labor * * 6.3 7.0 (6.2) 6.5 6.8 * * 6.5(6.1)6.5 82 Total 6.8 6.8 6.5 6.8 5.8 6.5 6.6 6.8 6.4 6.7 6.1 6.6 Housinq Conditions Very Poor 6.7 6.9 6.7 * * 6.8 6.7 (7.1) * (7.1) * 7.0 148 Poor 6.9 6.8 6.6 6.8 * 6.8 6.7 6.9 (6.5) 6.5 * 6.6 305 Some Comforts * 6.5 6.6 6.6 (6.4) 6.5 6.7 6.3 * 6.6 * 6.5 147 Comfortable * * 6.7 (6.7) 5.5 6.3 6.4 * (7.0) (5.9) (5.5)6.2 76 Total 6.8 6.7 6.6 6.7 6.0 6.6 6.6 6.8 6.7 6.5 5.5 6.6 N ______________ 57 171 279 87 44 380 95 39 107 29 683 #Ra c e > nnmhor 1 occ +Vi=»n TO large towns * 6.7 6.5 6.7 5.3 6.3 6.4 6.7 (6.7) 6.5 5.7 6.4 181 Total Husband’s Income 6.7 6.7 6.7 6.7 5.7 6.5 6.7 6.5 6.7 6.5 5.7 6.6 No work-100 Colones * * , 7;3 * * . 7.3 7.0 * * * * 7.0 45 100-300 Colones 6.8 6.8 6.6 6.8 * 6.7 6.7 6.9 (6.2) 6.5 * 6.6 282 300-600 Colones 600 Colones * 6.5 6.8 (6.6) * 6.6 6.5 6.8 * 6.7 * 6.7 132 or more * * 6.0 6.8 5.6 6.1 (6.7)(5.3) * 6.4(5.9)6.1 87 Total 6.8 6.6 6.7 6.7 5.6 6.6 6.7 6.3 6.2 6.5 5.9 6.6 Husband * s Employment Status Tenant Farmers 6.8 6.8 6.7 6.6 (5.4) 6.5 6.7 6.8 6.4 6.4 * 6.6 516 Unskilled Labor * * 6.5 * * 6.5 6.3 * * (7.3) * 6.8 50 Farmers or Skilled Labor * * 6.3 7.0 (6.2) 6.5 6.8 * * 6.5(6.1)6.5 82 Total 6.8 6.8 6.5 6.8 5.8 6.5 6.6 6.8 6.4 6.7 6.1 6.6 Housinq Conditions Very Poor 6.7 6.9 6.7 * * 6.8 6.7 (7.1) * (7.1) * 7.0 148 Poor 6.9 6.8 6.6 6.8 * 6.8 6.7 6.9 (6.5) 6.5 * 6.6 305 Some Comforts * 6.5 6.6 6.6 (6.4) 6.5 6.7 6.3 * 6.6 * 6.5 147 Comfortable * * 6.7 (6.7) 5.5 6.3 6.4 * (7.0)(5.9)(5.5)6.2 76 Total 6.8 6.7 6.6 6.7 6.0 6.6 6.6 6.8 6.7 6.5 5.5 6.6 N 57 171 : 279 87 44 380 95 39 107 29 683 *Base number less than 10 ( ) Less than 20 but at least 10 cases Code for Education: 1 = Both had less than 1 year of elementary school 2 = One elementary school incomplete 3 = Both elementary school incomplete 4 = One elementary school completed 5 = Both elementary school completed Code for Residential Background: 1 = Both Rural 2 = One from Village 3 = Both from Village 4 = One from Town ' _____ 5 = Both from Town 137 fertility with the exception of husband's employment status which practically does not have any differential effect on fertility (Tables 6.9 and 6.13). Education proves once again to explain part of the effect of any of the present factors on desired fertility. Only in one other case, namely, when the effect of hus band's income on desired fertility is controlled for hous ing conditions, a significant reduction in differential desired fertility occurs. Discussion Few, if any, studies have used desired fertility as a dependent variable in the way that this study does. Most of the time, when used, it is an intervening variable employed to show the attitudes toward family size, which are supposed to have influenced fertility performance. Given the assumption that any major changes in attitudes toward family planning in the rural and semi-urban areas of Costa Rica probably are attributable to the official family planning program, and that this program started scarcely three years before the Rural Fertility Survey was taken, it was preferred to use desired fertility as a dependent and not as an intervening variable between socio-structural factors and fertility performance. In this way, the dif ferential effects of the various socio-structural indica tors on attitudes toward family size can be assessed and TABLE 6.13 MULTIPLE CLASSIFICATION ANALYSIS RELATING WANTED FAMILY SIZE TO PRESENT SOCIAL CHARACTERISTICS BY GROUPS OF INDICATORS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Deviation from Adjusted Deviation* No. Present Grand of Characteristic Mean Mean A B C D E F Cases Couple’s Residential Historv No Moves 6.9 .27 .25 .21 .21 .20 .18 .18 131 Some Moves, no large town 6.6 -.02 -.03 -.07 -.03 -.05 -.08 -.04 361 Some Moves, some large towns 6.4 -.19 -.15 -.04 — 12 -.08 -.01 -.07 181 Husband’s Income No work-100 Colones 6.8 .20 .17 .14 .15 .17 .10 .14 45 100-300 Colones 6.7 .06 -.04 -.01 -.03 .01 -.04 .01 282 300-600 Colones 6.6 -.05 -.03 -.07 -.04 -.06 -.07 -.07 132 600 Colones or more 6.2 -.38 -.16 -.14 -.14 -.15 -.12 -.14 87 Husband's Employment Status Tenant Farmers 6.6 .03 .00 -.01 -.01 -.01 -.02 -.02 516 Unskilled Labor 6.3 -.32 -.17 -.13 -.18 -.16 -.13 -.19 50 Farmers or Skilled Labor 6.5 -.10 .10 .13 .16 .10 .13 .14 82 Housinq Conditions Very Poor 6.8 .16 .13 .15 .14 .10 .12 .13 149 Poor 6.7 .07 .04 .03 .01 .05 .04 .03 305 Some Comforts 6.5 -.06 -.03 -.03 -.04 -.05 -.06 -.05 147 Comfortable 6.1 -.54 -.53 -.54 -.52 -.29 -.32 -.30 76 Grand M e a n ______________________6.6 *Predictor Variables in each group: A = Other present characteristics B = Column A adding age at union C = Column A adding duration of union D = Social background characteristics E = Column D adding age at union F = Column D addina duration of uni on v • u — • — • wo - .0 / - .Uffc - .Ufc> - .07 -.07 132 600 Colones or more 6.2- -.38 -.16 -.14 -.14 -.15 -.12 -.14 87 Husband's Employment Status Tenant Farmers 6.6 .03 .00 -.01 -.01 -.01 -.02 -.02 516 Unskilled Labor 6.3 -.32 -.17 -.13 -.18 -.16 -.13 -.19 50 Farmers or Skilled Labor 6.5 -.10 .10 .13 .16 .10 .13 .14 82 Housinq Conditions Very Poor 6.8 .16 .13 .15 .14 .10 .12 .13 149 Poor 6.7 .07 .04 .03 .01 .05 .04 .03 305 Some Comforts 6.5 -.06 -.03 -.03 -.04 -.05 -.06 -.05 147 Comfortable 6.1 -.54 -.53 -.54 -.52 -.29 -.32 -.30 76 Grand Mean 6.6 *Predictor Variables in each group: A = Other present characteristics B = Column A adding age at union C = Column A adding duration of union D = Social background characteristics E = Column D adding age at union _____F = Column D adding duration of union 138 139 the implications of these differential attitudes evaluated. The data presented suggest small socio-structural differences in wanted family size. For the United States, Whelpton et al. (1966) found also small differences in desired fertility by education, income, husband’s occupa tion, and place of residence. Data for developing countries are not available. In our case, only education has been significantly affecting desired fertility, either by itself or through other factors. Thus, regardless of the possible impact of the family planning programs in lowering desired family size, education, if increased, will play a major role in producing that lowering effect. On the average, however, a wanted family size of 6.6 is extremely high when thinking in terms of attitudes that eventually would lead to a much smaller actual fertility than the present. Ideal Fertility In many areas of life discrepancies are found between ideals and realities. The fertility ideals of Latin American men and women have proved to be considerably lower than their actual fertility. Hill et al. (1959), in their study of population control in Puerto Rico, found that over 75 percent of the couples interviewed considered as ideal a family of two or three children, while their average family size was in the vicinity of five children. 140 Kahl (1967), working with data from Brasil and Mexico, found a mean fertility ideal of 2.6 for the former and 3.9 for the latter, while the crude birth rates of the two countries were at the time between 43 and 47 per thousand, which obviously indicates a higher mean number of live births than they considered ideal to have. This seems to be the case in rural and semi-urban Costa Rica. In all the categories of the six socio- structural characteristics included in this analysis, the mean number considered as ideal is smaller than the corresponding mean number of live births.* Furthermore, on balance the wives as a group think as ideal a smaller number of children than they actually want. The differentials between the extreme categories of the same indicator are smaller than in the case of past fertility, due to the fact that wives in the lower categor ies of each indicator considered as ideal a much smaller number of children than they already had. Here again, only those differential ideal fertility patterns will be remarked and commented upon that differ from what was found in the analysis of past fertility. *This is not always the case in the developed coun tries. In the United States, as Whelpton et al. (1966) re port, the ideals of many wives call for more children than they feel it is prudent for them to have (for example, of the 2,414 wives interviewed, those with college education considered as ideal 3.4 but wanted only 3.3; those with grade school education considered as ideal 3.7 but wanted 3.5). 141 Social Background Characteristics The two factors in this group show a multiple correlation coefficient of .16, which confirms hypothesis 1-1, even if the association is not substantial (Table 6.1). Hypothesis 1-2, which predicts that the effect of each one of these factors on ideal family size will be due to the intervening effect of present social factors is also con firmed (Table 6.15, column C). The peculiarity in this case is that the differential effect of couple's education on ideal fertility almost disappears when we control for the effects of the present factors, while couple's resid ential background retains its effect to a considerable extent regardless of the controls used. Hypothesis II-l is confirmed. Each of the two factors are associated with ideal fertility (Table 6.14), and hypothesis II-2 is only partly confirmed. The effect of couple's education on ideal fertility is exerted in part through husband's income and housing conditions, but not through husband's employment status. Thus, these are the suggested channels: couple's education --- > present factors --- > ideal fertility; couple's residential background --- > present factors > ideal fertility; couple's education ---- > husband's income --- > ideal fertility; and couple's education > housing conditions --- > ideal fertility. TABLE 6.14 MULTIPLE CLASSIFICATION ANALYSIS RELATING IDEAL* FAMILY SIZE TO SOCIAL BACKGROUND AND PRESENT SOCIAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Ideal Family Size Deviation from No Grand ______Adjusted Deviation_____ of Characteristic Mean Mean CE CRB RH HI HES HC Cases Couple's Education (CE) Amount of Elementary School: Both had less than 1 year 5.6 .33 .25 .23 .19 .28 .16 57 One incomplete 5.6 .37 .33 .34 .29 .33 .29 171 Both incomplete 5.2 -.08 -.08 -.08 -.10 -.09 -.07 279 One completed 5.0 -.28 -.15 -.20 -.14 -.22 -.16 87 Both completed 4.4 -.83 -.60 -.59 -.49 -.70 -.55 44 Correlation Ratio .14 Couple's Residential Background (CRB) Both Rural 5.4 .11 .04 .05 .05 .08 .07 380 One from Village 5.2 -.04 -.05 .02 -.05 -.06 -.01 95 Both from Village 4.7 -.52 -.37 -.53 -.39 -.49 -.44 39 One from Town 5.0 -.23 -.13 -.10 -.10 -.18 -.17 107 Both from Town 4.2 -1.1 -.77 -.82 -.81 -.89 -.85 29 Correlation Ratio .17 Couple's Residential History (RH) No moves 5.6 .38 .35 .38 .27 .36 .35 131 Some moves, no large town 5.4 .09 .06 .07 .07 .07 .07 361 Some moves, some large towns 4.7 -.54 -.45 -.46 -.41 -.48 -.48 181 Correlation Ratio .17 Husband's Income (HI) No work-100 Colones 5.8 .56 .55 .47 .57 .61 .52 45 100-300 Colones 5.6 .38 .32 .34 .32 .34 .30 282 300-600 Colones 5.1 -.21 -.22 -.21 -.19 -.23 -.18 132 600 Colones or more 4.6 -.69 -.52 -.50 -.53 -.62 -.54 87 Correlation Ratio .19 Husband's Employment Status (HES) T o n a n f Parmor c ; a i r\ r\^t r\Q AO rv r - 600 Colones or more 4.6 -.69 -.52 -.50 -.53 -.62 -.54 87 Correlation Ratio .19 Husband's Employment Status (HES) Tenant Farmer Unskilled Labor Farmers or Skilled Labor 5.4 5.0 4.7 .10 -.22 -.58 .07 -.11 -.36 .08 -.13 -.38 .08 -.07 -.41 .05 -.03 -.29 .07 -.08 -.38 516 50 82 Correlation Ratio .10 Housinq Conditions (HC) Very Poor 5.7 .45 .34 .32 .38 .28 .40 149 Poor 5.3 .07 .05 .07 .07 .03 .06 305 Some Comforts 5.0 -.29 -.24 -.26 -.24 -.18 -.26 147 Comfortable 4.7 -.61 -.40 -.40 -.56 -.33 -.52 76 Correlation Ratio .15 Grand Mean 5.3 TABLE 6.15 MULTIPLE CLASSIFICATION ANALYSIS RELATING IDEAL FAMILY SIZE TO SOCIAL BACKGROUND CHARACTERISTICS BY GROUPS OF INDICATORS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Deviation Social Background from Grand Adjusted Deviation* No. of Characteristic Mean Mean A B C D E F Cases Couple1s Education Amount of Elementary School: Both had less than 1 year 5.6 .33 .26 .26 .01 .13 .13 .13 57 One incomplete 5.6 .37 .39 .35 .26 .28 .33 .29 171 Both incomplete 5.2 -.08 -.10 -.11 -.10 -.07 -.10 -.10 279 One completed 5.0 -.27 -.15 -.09 -.04 -.07 -.07 -.01 87 Both completed 4.4 -.83 -.62 -.53 -.20 -.41 -.42 -.31 44 Couple's Residential Backqround Both Rural 5.4 .11 .01 .03 -.03 .03 -.01 .02 380 <Dne from Village 5.2 -.04 .02 .02 .04 -.02 .06 .04 95 Both from Village 4.7 -.52 -.34 -.42 -.38 -.35 -.32 -.39 39 One from Town 5.0 -.24 -.14 -.19 .01 -.11 -.12 -.17 107 Both from Town 4.2 -1.1 .77 -.82 -.62 -.71 -.71 -.74 29 Grand Mean 5.3 *Predictor Variables in each group: A = Other social background factor adding age at union. B = Other social background factor adding duration of union. C = All present characteristics. D = Couple’s education/residential background and housing conditions. E = Col. D adding age at union. F = Coly. D add. duration • J . W Housing Conditions (HC) Very Poor 5.7 .45 .34 .32 .38 . .28 .40 149 Poor 5.3 .07 .05 .07 .07 .03 .06 305 Some Comforts 5.0 -.29 -.24 -.26 -.24 -.18 -.26 147 Comfortable 4.7 -.61 -.40 -.40 -.56 -.33 -.52 76 Correlation Ratio .15 Grand Mean 5.3 TABLE 6.15 MULTIPLE CLASSIFICATION ANALYSIS RELATING IDEAL SOCIAL BACKGROUND CHARACTERISTICS BY GROUPS OF RURAL COSTA RICAN COUPLES WITH WIVES FAMILY SIZE INDICATORS 30-49 TO » Social^Background Characteristic Mean Deviation from Grand i Mean A Adjusted Deviation* B C D E F No. of Cases Couple's Education Amount of Elementary School: Both had less than 1 year 5.6 .33 .26 .26 .01 .13 .13 .13 57 One incomplete 5.6 .37 .39 .35 .26 .28 .33 .29 171 Both incomplete 5.2 -.08 -.10 -.11 -.10 -.07 -.10 -.10 279 One completed 5.0 -.27 -.15 -.09 -.04 -.07 -.07 -.01 87 Both completed 4.4 -.83 -.62 -.53 -.20 -.41 -.42 -.31 44 Couple's Residential Background Both Rural 5.4 .11 .01 .03 -.03 .03 -.01 .02 380 0ne from Village 5.2 -.04 .02 .02 .04 -.02 .06 .04 95 Both from Village 4.7 -.52 -.34 -.42 -.38 -.35 -.32 -.39 39 One from Town 5.0 -.24 -.14 -.19 .01 -.11 -.12 -.17 107 Both from Town 4.2 -1.1 -.77 -.82 -.62 -.71 -.71 -.74 29 Grand Mean 5.3 *Predictor Variables in each group: A = Other social background factor adding age at union. B = Other social background factor adding duration of union. C = All present characteristics. D = Couple's education/residential background and housing conditions. E = Col. D adding age at union. F = Coly. D add. duration H to 143 These findings are partly unexpected. (I) Couple's education plays a role in molding differential fertility ideals, but it is mostly through the socioeconomic charac teristics of the couple that it exerts its influence. (II) Couple's residential background is a powerful indicator of fertility ideals. Its influence in shaping ideals remains strong, independently of the socioeconomic status of the couple. As tables 6.16 and 6.17 show, some strong inter actions occur between education and residential background, and between each education and residential background on the one hand, and each of the present social indicators on the other. The lack of information for many of the cells, mainly in Table 6.17, prevents from reaching further conclusions. Present Social Characteristics This group of factors is associated with ideal fertility, as hypothesis 1-1 anticipates. The multiple correlation coefficient is .20 (Table 6.1). Each one of the four factors included in this group have independent effects on ideal fertility, confirming hypothesis 1-5 (Table 6.14). On the individual level, each of the four factors has a fairly strong association with ideal fertility, except husband's employment status, which shows a correlation TABLE 6.16 IDEAL FAMILY SIZE BY WIFE’S AND HUSBAND'S EDUCATION AND RESIDENTIAL BACKGROUND, RURAL COSTA RICAN COUPLES WITH WIVES, 30-49 Ideal Family Size Residential Wife's Education Husband '.s Education Background 1 2 3 T 1 2 3 T N Wife’s Rural 5.6 5.4 5.2 5.5 5.6 5.4 5.3 5.4 478 Village (5.8) 4.7 4.8 5.1 (5.3) 4.7 (5.0) 5.0 106 Town (-) . 4.7 4.3 4.5 (-) 5.1 4.0 4.6 97 Total 5.7 4.9 CO • 5.1 5.5 5.1 4.7 5.0 Husband's Rural 5.5 5.2 4.8 5.2 5.4 5.3 4.7 5.1 478 Village (5.8) 5.0 4.8 5.2 (-) 5.2 4.9 5.1 100 Town (5.9) 5.0 4.5 5.1 (-) 5.2 4.4 4.8 73 Total 5.7 5.1 4.7 5.2 (-) 5.2 4.7 5.0 N 170 426 86 144 402 92 683 (-) Base number less than 10 cases ( ) Less than 20 but at least 10 cases Code for Education: 1 = Less than 1 year of elementary school 2 = Elementary school incomplete _________________________ 3 = Elementary school completed_________ TABLE 6.17 IDEAL FAMILY SIZE BY COUPLE'S EDUCATION AND RESIDENTIAL BACKGROUND AND PRESENT SOCIAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Present • Ideal Family Size Social Char Couple's Education Couple's Resid. Back. acteristic 1 2 3 4 5 T 1 2 3 4 5 T N Couple * s Resid. History No moves (6.3) 6.0 5.2 (5.6) * 5.8 5.7 * (4.5) * * 5.1 131 Some moves, Social Char J.U' Couple's Education ccxjl r eui i ijl- l y Couple's Resid. Back. acteristic 1 2 3 4 5 T 1 2 3 4 5 T N Couple's Resid. Historv No moves (6.3) 6.0 5.2 (5.6) * 5.8 5.7 * (4.5) * * 5.1 131 Some moves, no large town 5.2 5.7 5.3 5.3 (4.9) 5.3 5.4 5.4'(5.1) 5,1 * 5.2 361 Some moves, some lg. towns * 5.0 4.9 4.3 4.2 4.6 4.7 4.g (4.6) 4.9 3.9 4.6 181 Total 5.7 5.6 5.1 5.1. 4.5 5.2 5.3 5.1 4.7 5.0 3.95.0 Husband's Income No work- 100 Colones * * 5.4 * * 5.4 5.6 * * * * 5.6 45 100-300 Colones 5.9 5.7 5.5 5.8 * 5.7 5.7 5.9 (3.7) 5.1 * 5.1 282 300-600 Colones * 5.6 5.2 (4.5) * 5.1 5.2 4.6 * 4.9 * 4.9 132 600 Colones or more * * 4.3 4.2 (4.5) 4.3 (4.3)(4.0) * .5.0 (3.9H.3 878 Total 5.9 5.6 5.1 4.8 4.5 5.1 5.2 4.8 3.7 5.0 3.9 5.0 Husband's Emplov. Status Tenant Farmers 5.6 5.6 5.3 5.3 (4.3) 5.2 5.4 5.2 4.4 5.2 * 5.0 516 Unskilled Labor * * 5.2 * * 5.2 5.9 * * (4.4) * 5.1 501 Farmers or Skilled Labor * * 4.4 4.8 (4.4) 4.5 4.7 * * 4.6 (3.7)4.3 82 Total 5.6 5.6 5.0 5.0 4.3 5.0 5.3 5.2 4.4 4.7 3.74.8 Housinq Conditions Very poor 6.1 5.7 5.3 * * 5.7 5.6 (5.6) * * * 5.6 149 Poor 5.3 5.7 5.4 4.9 * 5.3 5.2 5.7(5.2) 5.0 * 5.3 305 Some Comforts * 5.4 4.8 5.1 * 5.0 5.4 4.3 * 5.4 * 5.0 147 Comfortable * * 5.0 (4.0) 4.2 4.4 5.4 * (4.9) (3.7) (3.9)4.5 76 Total 5.7 5.6 5.1 4.7 4.2 5.1 5.4 5.2 5.0 4.7 3.9 5.1 N 57 171 279 87 44 380 95 39 107 29 683 *Base number less than 10 cases ( ) Less than 20 but at least 10 cases Code for Education: 1 = Both had less than 1 year of elementary school. 2 = One elementary school incomplete. 3 = Both elementary school incomplete. 4 = One elementary school completed. __________________________5 = Both elementary school completed.______________ 145 ratio of only .10 (Table 6.14). This confirms in part hypothesis I1-2. As to the specific channels anticipated in hypotheses IV-1 through IV-3, as Table 6.14 shows, hus band's income keeps its negative effect when controlled for the other single indicators (IV-1, confirmed), husband's employment status sees its effect weakened when adjusted for the effects of other variables, especially when the effect of income is taken into consideration or the effects of the other three present social indicators or of the social background variables are considered (Table 6.18); housing conditions has its effect on ideal fertility decreased to a greater extent when adjusted for the effects of husband's income than for the effects of couple's educa tion (hypothesis IV-3 reversed). The comparison of means tables (6.19 and 6.20) show some interaction between pairs of present social factors when individuals are classified as "high" in both. They also reveal that wives with husbands in the highest income bracket consistently have lower fertility ideals, regard less of the position they occupy in the other indicators. In summary, husband's income is the most important of the present factors in explaining ideal fertility both for its independent effect and for its mediating effect in the association between any of the other factors and ideal fertility. Husband's employment status, on the other hand, is the factor with less differential effect on ideal TABLE 6.18 MULTIPLE CLASSIFICATION ANALYSIS RELATING IDEAL FAMILY SIZE TO PRESENT SOCIAL CHARACTERISTICS BY GROUPS OF INDICATORS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Present Social Characteristic Mean Deviation from Grand Mean A Ad-jus ted Deviation* B C D E F No. of Cases Couple's Residential Backqround No moves 5.6 .38 .27 .24 .25 .37 .35 .35 131 Some moves, no large town 5.4 .09 .05 .05 .03 .06 .03 .05 361 Some moves, some lg. towns 4.7 -.54 .40 -.34 -.37 -.42 -.36 -.38 181 Husband's Income No work-100 Colones 5.8 .56 .72 .71 .71 .48 .45 .48 45 100-300 Colones 5.6 .38 .23 .23 .21 .30 .30 .28 282 300-600 Colones 5.0 -.21 .20 -.22 -.19 -.21 -.27 -.20 132 600 Colones or more 4.6 -.69 .38 -.36 -.34 -.43 -.39 -.37 87 Husband's Employment Status Tenant Farmers 5.4 .10 .03 .02 .03 .06 .05 .05 516 Unskilled Labor 5.0 -.22 .18 .17 .16 -.08 -.07 -.08 50 Farmers or Skilled Labor 4.7 -.58 .13 -.10 -.12 -.27 -.23 -.25 82 Housinq Conditions Very poor 5.7 .45 .25 .30 .29 .25 .29 .29 149 Poor 5.3 .07 .04 .03 .01 .05 .04 .03 305 Some comforts 5.0 -.29 .16 -.17 -.13 -.22 -.22 -.17 147 Comfortable 4.7 -.61 .34 -•.39 -.41 -.28 -.30 -.35 76 Grand Mean 5.3 *Predictor Variables in each group: A = Other present social characteristics B = Column A adding age at union C = Column A adding duration of union D = Social background characteristics E = Column D adding age at union F = Column D adding duration of union TABLE 6.19 IDEAL FAMILY SIZE BY COUPLE'S RESIDENTIAL HISTORY AND HOUSING TABLE 6.19 IDEAL FAMILY SIZE BY COUPLE'S RESIDENTIAL HISTORY AND HOUSING CONDITIONS AND OTHER PRESENT SOCIAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Ideal Family Size Present Social Couple's Resident.History Housing Conditions Characteristic 1 2 3 T 1 2 3 4 N Housinq Conditions 1 Very poor 6.2 5.6 5.2 5.6 148 2 Poor 5.3 5.4 5.1 5.3 305 3 Some Comforts 5.7 5.0 4.5 5.1 147 4 Comfortable (5.6)_ 5.0 3.8 4.8 76 Total 5.6 5.2 4.6 5.2 Husband•s Income No work-100 Colones (6.1) 5.6 (5.9) 5.9 (5.2) 6.3 * * 5.8 45 100-300 Colones 5.8 5.7 5.1 5.5 5.8 5.6 5.3 * 5.6 282 300-600 Colones 5.3 5.1 4.6 5.0 (5.4) 5.1 4.9 (4.8) 5.1 132 600- Colones or more * 4.8 4.2 4.5 * 4.7 4.5 4.4 4.5 83 Total 5.7 5.3 4.9 5.3 5.5 5.4 4.9 4.6 5.1 Husband's Employment Status Agric. low 5.8 5.4 4.8 5.3 5.7 5.4 4.9 4.9 5.2 516 Unskilled * 5.2 4.7 4.9 * 5.1 (5.9)(3.9) 5.0 50 Skilled— Agr. Int. (.3.7) 5.3 4.4 4.5 * 4.7 4.6 4.6 4.6 86 Total 4.8 5.3 4.6 4.9 * 5.1 5.1 4.5 4.9 N 131 361 181 148 305 147 76 683 *Base number less than 10 ( ) Less than 20 but at least 10 Code for Couple's Residential History: 1 = No moves 2 = Some moves, no large towns ________________________________________________3 = Some moves, some large towns TABLE 6.20— IDEAL FAMILY SIZE BY HUSBAND'S INCOME AND EMPLOYMENT STATUS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 ______________ Ideal Family Size_________________________ Husband' s Employment _______________ Husband' s Income_______________ Status__________________________No work-100 C. 100-300 C. 300-600 C. 600 C. + T N Tenant Farmers 6.1 5.7 5.0 4.4 5.3 510 Unskilled Labor * * 5.1 (4.8) 5.0 50 Farmers or Skilled Labor *________ (4.6) (5.2) 4.8 4.9 86 Ideal Family Size Present Social Couple's Resident.History Housing Conditions Characteristic 1 2 3 T 1 2 3 4 N Housinq Conditions 1 Very poor 6.2 5.6 5.2 5.6 148 2 Poor 5.3 5.4 5.1 5.3 305 3 Some Comforts 5.7 5.0 4.5 5.1 147 4 Comfortable (5.61 5.0 3.8 4.8 76 Total 5.6 5.2 4.6 5.2 Husband's Income No work-100 Colones (6.1) 5.6 (5.9) 5.9 (5.2) 6.3 * * 5.8 45 100-300 Colones 5.8 5.7 5.1 5.5 5.8 5.6 5.3 * 5.6 282 300-600 Colones 5.3 5.1 4.6 5.0 (5.4) 5.1 4.9 (4.8) 5.1 132 600- Colones or more * 4.8 4.2 4.5 * 4.7 4.5 4.4 4.5 83 Total 5.7 5.3 4.9 5.3 5.5 5.4 4.9 4.6 5.1 Husband's Employment Status Agric. low 00 • 5.4 4.8 5.3 5.7 5.4 4.9 4.9 5.2 516 Unskilled * 5.2 4.7 4.9 * 5.1 (5.9)(3.9) 5.0 50 Skilled— Agr. Int. (3.7) 5.3 4.4 4.5 * 4.7 4.6 4.6 4.6 86 Total 4.8 5.3 4.6 4.9 * 5.1 5.1 4.5 4.9 N 131 361 181 148 305 147 76 683 *Base number less than 10 ( ) Less than 20 but at least 10 Code for Couple's Residential History: 1 = No moves 2 = Some moves, no large towns __________________________________________ 3 = Some moves, some large towns TABLE 6.20— IDEAL FAMILY SIZE BY HUSBAND'S INCOME AND EMPLOYMENT STATUS, COSTA RICAN COUPLES WITH WIVES 30-49 RURAL Ideal Family Size Husband's Employment Husband's Income Status No work-100 C. 100-300 C. 300-600 C. 600 C. + T N Tenant Farmers 6.1 5.7 5.0 4.4 5.3 510 Unskilled Labor * * 5.1 (4.8) 5.0 50 Farmers or Skilled Labor * (4.6) (5.2) 4.8 4.9 86 Total * 5.2 5.1 4.7 5.0 N 45 282 132 87 *Base number less than 10 ( ) Less than 20 but at least 10 H o\ 147 fertility. Discussion It should be noted that simple statements of ideal family size are deceptive in a context where attitudes may be uncrystallized or ambivalent, which seems to be the case in rural and semi-urban Costa Rica. The field of fertility ideals i:s particularly vulnerable to conflicting opinions among relatives and peer groups. The fact that they con sider as ideal a smaller family than they say they want and than they have already had may represent a transitional stage of attitude development in the society between an unequivocal preference for large families and an unequivo cal preference for small. During such a period, the individual may be subject to two opposing value systems, both of which he or she can agree with. The numerical statement of ideal family size may represent not so much a fixed ideal as a kind of compromise between more extreme values current in the culture, which are simultaneously accepted by the individual. At this respect early family size attitudes and especially their intensity may be a more useful tool for the prediction of fertility and fertility control behavior than current ideals, which are subject to a number of influences, among which fertility itself must be counted. Of all the factors included in this study, two 148 prove to be specially meaningful in explaining ideal fertility: couple's residential background and husband's income, while education, which was expected to have a larger independent effect than any other factor, shows to be largely dependent, in its effect on ideal fertility, on socioeconomic characteristics. Hill et al. (1959) found also that for Puerto Rican couples differences in ideal family size due to education were not large. Given the strong influence of income on ideal fertility, it would seem that the small family size ideal is a rationalization which supersedes the education influence. This rationalization is dependent on the current socioeconomic status, mainly income, and on the actual fertility achieved. While modes of thought and the culture may be held in common in rural and semi-urban areas, significant variations in socioeconomic class influ ence important differences in the formation of attitudes and values, such as ideal fertility. In other words, socioeconomic status, as indicated by income, and personal life experience, in this case measured as residential background, produce (in combination with other influences not included in this study) specific attitudes about fertility. CHAPTER VII THE ROLE OF THE BIOPSYCHOLOGICAL CHARACTERISTICS Socio-structural factors may affect fertility behavior and ideals directly and indirectly through inter vening biopsychological variables. The biological and psychological indicators used in this study are coital fre quency, extended family living experience, attitude toward family planning, and use of contraception. In the preced ing chapter it was seen that socio-structural variables are related to fertility behavior and ideals in different ways. Knowledge and specific attitudes pertaining to family planning are included as dependent variables in relation to socioeconomic factors under the assumption that different social setting impose different environmental conditions and constraints upon the individual whose psychological characteristics are molded as they interact with the surroundings. A detailed study of these biological and psycholog ical factors would require, an analysis of such questions as couple's attitudes toward contraception, the proportion of couples who use contraception, when use of contraception began, socioeconomic differentials in the use of 149 150 contraception, family planning patterns and their relation ship to various demographic and socioeconomic variables, the pregnancy planning status of couples, changes in planning status within marriage, and also an inquiry of the specific methods used and their effectiveness. Obviously, such a detailed exploration is not possible within the limits of this study. In this analysis some simple questions are explored: do the biopsychological variables included here have any significant relationship to fertility behavior and ideals? Is there any substantial association between demographic and socio-structural characteristics and these biopsychological variables? To what extent biopsychologi cal indicators increase our ability to understand and predict fertility levels and fertility ideals? The biopsychological variables may increase our understanding in different ways: (1) They may be one of the mechanisms through which the socio-structural indicators operate to affect fertility behavior and ideals. In other words, they may be one of the underlying reasons for the recurrent finding that socio-structural variables have substantial effects on fertility behavior and ideals. (2) The psychological indicators may be independ ently important and not simply underlying mechanisms for earlier findings, as a new dimension of measurement. 151 (3) Also, the psychological indicators may interact with the socio-structural variables to influence fertility behavior and ideals. In some socio-structural conditions, they may be important determinants of behavior, and under different socio-structural conditions, they may play a less crucial role. In this chapter, an attempt will be made to assess in which way these biological and psychological variables affect, if at all, the fertility behavior and ideals of rural and semi-urban Costa Rican couples with wives 30 to 49 years of age. As in the preceding chapter, two methods of analysis will be used: multiple classification analysis and comparison of means. The role of the biopsychological indicators will be examined separately for past, desired, and ideal fertility. Past Fertility The hypotheses formulated in Chapter II regarding the role of the biopsychological variables and their con firmation status are as follows: Hypothesis 1-1 predicted a strong association of this group of variables with past fertility, and is not confirmed. The multiple regression coefficient is an insignificant .06 (Table 6.1). Hypothssis 1-2 anticipated that the effects of the social background factors would be exerted partly through 152 these biopsychological characteristics. It is not con firmed. The same happens with hypothesis 1-3. The data show that the influence of present social factors on fertility is not exerted even partly through the biopsy chological indicators, as it was predicted (Table 7.2). The hypotheses related to adjusted fertility dif ferentials are not confirmed either. Neither couple's education nor husband's income nor housing conditions exert part of their effect through use of contraception (contrary to what was hypothesized in II-2, II-3, and IV-3). In summary, the role of the biopsychological factors in relation to past fertility is practically non existent. Only coital frequency and extended family living experience have had a slight effect on number of live births (Table 7.1). None of the four, though, operate as underlying mechanisms. There is a certain cumulative ef fect when either attitude toward family planning is favor able or they have used contraception on the one hand, and, on the other hand, they are in the highest category of any of the social background or present factors. This indicates that the two family planning variables operate partly in interaction with the socio-structural character istics included in this study (Tables 7.3 to 7.6). TABLE 7.1 MULTIPLE CLASSIFICATION ANALYSIS RELATING- NUMBER OF LIVE BIRTHS TO BIOPSYCHOLOGICAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Number of Live Births Deviation from No. Biopsychological Grand Ad-justed Deviation of Characteristics Mean Mean CF EF AP UC Cases Coital Frequency (CF) Less than 1 a month 6.5 -.58 -.59 -.61 -.57 32 1-3 a month 7.0 -.07 -.09 -.04 -.07 171 1-2 a week 7.0 -.11 -.11 -.14 -.11 282 3 or more a week 7.4 .28 .29 .26 .28 137 Correlation Ratio .07 Extended Family Livino Experience (EF) No 7.1 .06 .06 .06 .06 619 Yes 6.4 - .66 -.69 -. 66 -.65 60 Correlation Ratio .07 Attitude toward Family Planninq (AP) Against 7.1 .02 .01 .03 .00 305 Favor 7.1 .05 .06 .04 .08 359 Correlation Ratio .09 Use of Contraception (UC) No 7.1 .02 .01 .02 .04 454 Yes 7.0 -.07 -.04 -.07 -.09 228 Correlation Ratio .05 Grand Mean 7.1 153 TABLE 7.2 MULTIPLE CLASSIFICATION ANALYSIS RELATING NUMBER OF LIVE BIRTHS TO SOCIAL BACKGROUND AND PRESENT SOCIAL CHARACTERISTICS BY FAMILY PLANNING VARIABLES, RURAL COSTA RICAN COUPLES WITH WIVES. 30-49 Social Background. and Present Characteristics } Mean Number of Live Births Deviation from Adiusted Deviation ielnd Fain I PI. Use of C No. of Cases Couple1s Education Amount of Elementary School: Both had less than 1 year 7.6 .48 .47 .56 57 One incomplete 7.7 . 66 .67 .72 171 Both incomplete 7.3 .20 .21 .20 279 One completed 6.0 -1.1 -1.1 -1.1 87 Both completed 4.4 -2.7 -2.7 -2.8 44 Couple's Residential Background Both rural 7.4 .32 .32 .33 380 One from village 7.4 .32 .33 .33 95 Both from village 7.2 .10 .11 .09 39 One from town 6.3 -.73 -.72 -.75 107 Both from town 4.7 -2.3 -2.4 -2.4 29 Couple's Residential History No moves 7.4 .34 .36 .35 131 Some moves, no large town 7.4 .36 .36 .37 361 Some moves, some large towns 6.1 -.93 -.94 -.92 181 Husband's Income No work-100 Colones 7.9 .81 .85 .86 45 100-300 Colones 7.5 .41 .42 .44 282 300-600 Colones 7.1 .01 -.02 .02 132 600-Colones 'or more 5.9 -1.1 -1.3 -1.3 87 Husband's Employment Status Tenant Farmers 7.3 .19 .19 .20 516 Unskilled Labor 6.9 -.16 -.20 -.17 50 Farmers or Skilled Labor 5.6 -1.5 -1.4 -1.5 82 Housing Conditions Very poor 7.9 .80 .80 .86 148 Poor 7.4 . •a« one 50 82 148 305 147 76 Grand Mean 7.1 TABLE 7.3 NUMBER OF LIVE BIRTHS BY COUPLE’S EDUCATION AND RESIDENTIAL BACKGROUND AND BIOPSYCHOLOGICAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Farmers or Skilled Labor Housing Conditions Very poor Poor Some comforts Comfortable 5.6 7.9 7.4 6.4 5.5 -1.5 -1.4 - . j-t -1.5 .80 .36 -.63 ■1.5 .80 .35 -.61 - 1.6 .86 .36 -.64 - 1.6 Number of Live Births Biopsych. Couple’s Education Couple1s Residen. Back. Characteristic 1 2 3 4 5 T 1 2 3 4 5 T N Coital Freguencv Less than 1 a month * * (5.9) * * 5.9 (7.8) * * * * 7.8 32 1-3 a month (6.8) 7.5 7.3 6.3(4.3) 6.4 6.9 7.4 (8.9) 6.9 * 7.5 171 1-2 a week 7.2 8.0 7.1 5.5(4.9) 6.7 7.3 6.8 (6.0) 6.2 (4.6) 6.2 282 3 or more week 18.7) 7.1 7.6 7.3 * 7.7 7.6 7.8 * (6.7) * 7.4 137 Total 7.6 7.5 7.0 6.6 4.6 6.7 7.4 7.3 7.4 6.6 4.6 7.2 Attitude toward Family Planning Against 7.3 7.4 7.1 6.3(5.4) 6.7 7.0 8.2 (8.0) 6.7 * 7.5 305 Favor 8.0 8.2 7.5 5.8 4.1 6.7 7.9 7.1 6.7 6.2 4.4 6.5 359 Total 7.6 7.8 7.3 6.0 4.7 6.7 7.4 7.6 7.4 6.5 4.4 7.0 Use of Contrac. No 7.3 7.5 7.1 6.0 4.6 6.5 7.2 7.7 7.6 6.1 (4.6) 6.6 454 Yes (9.0) 8.5 7.6 6.0 4.3 7.1 7.8 6.8 (6.6) 6.6 (4.8) 6.5 228 Total 8.1 8.0 7.3 6.0 4.5 6.7 7.5 7.2 7.1 6.3 4.7 6.6 N 57 171 279 87 44 380 95 39 107 29 683 * Base number less than 10. ( ) Less than 20 but at least 10. Codes: For Couple’s Education: 1 = Both had less than 1 year. 2 = One elementary school incomplete. 3 = Both elementary school incomplete. 4 = One elementary school completed. 5 = Both elementary school completed. Codes: For Couple’s Residential Background: 1 = Both rural; 2 = One from village; _______3 = Both from village; 4 = One from town; 5 = Both from town.______________ TABLE 7.4 NUMBER OF LIVE BIRTHS BY COUPLE'S RESIDENTIAL HISTORY AND HOUSING CONDITIONS AND BIOPSYCHOLOGICAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Biopsychological Characteristics Number of Live Births Resid. History 1 2 3 Housinq Conditions T 1 2 3 4 T N Coital Frequency Less than 1 a month * * * * * * * * * 32 1-3 a month 6.8 7.4 6.5 6.9 7.7 7.5 6.2 6.2 6.9 171 1-2 a week 7.0 7.4 5.7 6.7 7.7 7.1 6.9 4.7 6.6 282 3 or more a week 7.8 7.6 6.6 7.3 8.0 7.6 6.8 * 7.3 137 Total 7.2 7.5 6.3 7.0 7.8 7.4 6.6 5.5 6.9 Extended Family Livinq Experience No 7.4 7.5 6.2 7.0 6.9 7.6 6.5 5.6 6.9 619 Yes * 7.0 5.5 6.6 * 6.0 * * * 60 Total * 7.2 5.8 6.8 * 6.8 * * * Attitude toward Fam. Planninq Against 7.3 7.5 5.8 6.9 7.5 7.2 7.2 4.8 6.7 305* Favor 7.5 7.5 6.3 7.1 8.5 7.7 5.8 5.8 6.9 359 Total 7.4 7.5 6.0 7.0 8.0 7.5 6.4 5.3 6.8 Use of Contraception No 7.3 7.4 6.3 7.0 7.7 7.4 6.3 5.4 6.7 454 Yes 7.7 7.6 5.9 7.1 8.5 7.6 6.6 5.6 7.1 228 Total 7.5 7.5 6.1 7.1 8.1 7.5 6.4 5.5 6.9 N 131 361 181 149 : 305 147 76 683 * Base number less ■ than 10. ( ) Less than 20 but at ; Least 10. Code: For Couple's Residential History: 1 = No moves. 3 = Some! moves, no large town. 3 = Some moves, some large t. For Couple's Housing Conditions: 1 = Very poor: 2 = Poor; 3 = = Some comforts; 4 = Comfortable. TABLE 7.5 — NUMBER OF LIVE ] BIRTHS BY HUSBAND'S INCOME AND EMPLOYMENT STATUS AND BIO- PSYCHOLOGICAL CHARACTERISTICS. RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Number of Live Births Biopsychological Husband1s Income Hus. Employ.Status Characteristics 1 2 3 4 T 1 2 3 T N Coital Frequency T o c? t ? +V» n 1 « ' iji .3&X isi_________ 149 305 147 76 683 * Base number less than 10. ( ) Less than 20 but at least 10. Code: For Couple's Residential History: 1 = No moves. 3 = Some moves, no large town. 3 = Some moves, some large t. For Couple's Housing Conditions: 1 = Very poor; 2 = Poor; 3 = Some comforts; 4 = Comfortable. TABLE 7.5 — NUMBER OF LIVE BIRTHS BY HUSBAND'S INCOME AND EMPLOYMENT STATUS AND BIO- PSYCHOLOGICAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Number of Live Births____________________ Biopsychological Characteristics Husband's; Income Hus. Employ.Status 1 2 3 4 T 1 2 3 T N Coital Frequency Less than 1 a month * * * * * 6.2 * * * 32 1-3 a month * 7.5 6.4 6.2 6.7 7.2 * 5.8 6.5 171 1-2 a week * 7.1 7.7 5.4 6.6 7.1 7.6 5.7 6.8 282 3 or more a week * 7.8 7.2 5.9 7.0 7.7 * * * 137 Total * 7.5 7.1 5.8 6.8 7.1 * 5.7 6.5 Extended Family Living Exper. No 7.9 7.7 7.2 5.9 7.2 7.3 6.7 5.6 6.5 619 Yes * 6.0 * * * 6.4 * * * 606 Total * 6.8 * * * 6.9 * * * Attitude toward Fam. Planning Against 8.0 7.1 6.9 * 7.3 7.1 7.0 6.2 6.8 305 Favor * 8.0 7.2 5.7 7.1 7.4 6.9 5.4 6.6 359 Total * 7.5 7.0 * 7.2 7.2 7.0 5.8 6.7 Use of Contraception No 7.8 7.5 6.8 5.5 6.9 7.2 7.1 5.6 6.7 454 Yes * 7.5 7.7 6.1 7.1 7.4 6.7 5.6 6.5 228 Total * 7.5 7.1 5.8 7.0 7.3 6.9 5.6 6.6 N 45 282 132 87 516 50 82 683 * Base number less than 10. ( ) Less than 20 but at least 10. Code: For Husband's Income: 1 = No work-100 Colones; 2 = 100-300 Colones; 3 = 300-600 Colones; 4 = 600 Colones or more. For Husband's Employment Status: 1 = Tenant Farmers; 2 = Unskilled Labor; _____________ 3 = Farmers or Skilled Labor TABLE 7.6 — MULTIPLE CLASSIFICATION ANALYSIS RELATING THREE MEASURES OF FERTILITY TO FAMILY PLANNING VARIABLES BY VARIOUS SOCIOECONOMIC INDICA TORS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Family Planning Variables Mean Deviation from Grand Mean Adjusted Deviation* CE CRB HI HC No. of Cases Attitude toward Family Planning Number of Live Births Against 7.1 .02 -.11 -.06 -.14 -.10 305 Favor 7.1 .05 .16 .12 .19 .14 359 Use of Contraception No 7.1 .02 -.16 -.06 -.08 -.11 454 Yes 7.0 -.07 .29 .10 .14 .21 228 Grand Mean 7.1 Attitude toward Family Planning Wanted Family Size Against 6.9 .31 .29 .30 .30 .28 305 Favor 6.3 -.29 -.26 -.28 -.27 -.27 359 Use of Contraception No 6.7 .07 .26 .06 .06 .04 454 Yes 6.5 -.15 -.06 -.13 -.12 -.08 228 Grand Mean 6.6 Attitude toward Family Planning Ideal Family Size Against 5.9 .64 .60 .61 .58 .60 305 Favor 4.7 -.53 -.50 -.50 -.48 -.50 359 Use of Contraception No 5.4 .22 .17 .18 .16 .17 454 Yes 4.8 -.43 -.34 -.34 -.32 -.33 228 Grand Mean 5.3 *CE = Couple's education; CRB = Couple's Residential Background; HI = Husband's Income; HC = Housing Conditions. 156 157 Discussion Attitude toward family planning, as explored in the survey, refers only to current dispositions, which most likely have developed in the case of the great majority of wives in the rural and semi-urban areas of the country after the official family planning efforts were started. This seems to be the reason why this factor did not, dif ferentially affect the fertility performance of those rural and semi-urban Costa Rican couples. The fact that use of contraception does not appear to have differentially influenced fertility could be attributed to the same reason. The fact that there is a cumulative effect between each of the two family planning variables and various indicators of social structure and socioeconomic status leads to partially contradictory conclusions to what was said sibove. There has been for a certain time a favorable attitude toward family planning, even if not widespread, and contraception has been used. Both have affected family size. The couples with smaller families are those who have such a favorable attitude and used contraception and, at the same time are better educated, or have been raised mainly in towns, or are in higher than average socioeconomic status. Nothing can be said regarding the "causal" relationship between those factors, whether education, or 158 residential background, or socioeconomic status has led to a favorable attitude toward family planning or to an effec tive use of contraception, nor regarding the time when these attitudes developed or the use of contraception started. Due to the lack of strong social pressures in favor of a small family, family planning probably has been a matter of individual or family decision, and its use has been most likely casual. Institutional patterns and social sanctions have done little in the past to determine contra ceptive use. Actual performance, thus, depends in great part on individual interest and concern by the particular family. Even if not markedly so, the families most likely to have used contraception are those with the character istics described above. The statistical analysis, though, does not warrant the inclusion of this group of variables as having signi ficantly affected the fertility performance of rural and semi-urban Costa Rican couples. A better understanding of the relevance of this group of variables for past fertility would have been gained by investigating intensity and consistency of attitudes, and perception of family size as a problem. The data available and the scheme of analysis do not allow such investigation. Desired Fertility No outstanding differences are found in the way 159 these variables relate to desired fertility if compared with their relation to past fertility, but some new trends appear. The explanatory value of this group of variables on desired fertility increases to a multiple correlation coefficient of .17 (hypothesis 1-1 confirmed). The two family planning variables do present a slight differential effect on desired fertility, larger in the case of attitude toward family planning than in the case of use of contra ception (Table 7.7). It appears that attitude toward family planning is more important than use of contraception in explaining desired fertility. As Table 7.7 shows, the effect of attitude toward family planning on desired fertility is not exerted through use of contraception, while the effect of use of contraception on desired ferti lity is explained by the intervention of attitude toward family planning. None of the other hypotheses regarding the influ ence of education and husband's income on desired fertility through any of the biopsychological indicators is confirm ed. Only in the case of husband's income is a trend noticed. Part of the fact that husband's income has a dif ferential effect on desired fertility is due to the differ ential attitude toward family planning of those in differ ent income brackets. This channel seems to be operating: husband's income --- > attitude toward family planning TABLE 7.7 — MULTIPLE CLASSIFICATION ANALYSIS RELATING WANTED FAMILY SIZE TO BIOPSYCHOLOGICAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Wanted Family Size Deviation Biopsychological Characteristics Mean from Grand Mean Adiusted Deviation CF EF AP UC Cases Coital Frequency (CF) Less than 1 a month 6.4 -.27 -.25 -.24 -.24 32 1-3 a month 6.7 .09 .08 .05 .08 171 1-2 a week 6.5 -.06 -.06 -.05 -.06 282 3 or more a week 6.7 .05 .05 .07 .04 137 Correlation Ratio .06 Extended Family Livinq Experience(EF) No 6.6 . .00 .00 .00 .00 619 Yes 6.6 -.09 -.08 -.14 -.07 60 Correlation Ratio .04 Attitude toward Family Planninq (AP) Against 6.9 .29 .31 .32 .31 305 Favor 6.3 -.26 -.29 -.29 -.28 359 Correlation Ratio .19 Use of Contraception (UC) No 6.7 .07 .07 .08 .01 454 Yes 6.4 -.15 -.15 -.15 -.03 228 Correlation Ratio .18 Grand Mean 6.6 TABLE 7.8 -- MULTIPLE CLASSIFICATION ANALYSIS RELATING WANTED FAMILY SIZE TO SOCIAL BACKGROUND AND PRESENT SOCIAL CHARACTERISTICS BY FAMILY PLANNING VARIABLES, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Deviation Adiusted Deviation No. Social Background and from Attitude toward Use of Con- of Present Characteristics Mean Grand Mean Family Planning traception Cases Couple1s Education Amount of Elementary School: Both had less than 1 year 6.8 .18 .14 .17 57 One incomplete 6.7 .12 .06 .11 171 VARIABLES, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Deviation Adiusted Deviation No. Social Background and from Attitude toward Use of Con- of Present Characteristics Mean Grand Mean Family Planning traception Cases Couple's Education Amount of Elementary School: Both had less than 1 year 6.8 .18 .14 .17 57 One incomplete 6.7 .12 .06 .11 171 Both incomplete 6.6 .01 .03 .01 279 One completed 6.7 .05 .08 .06 87 Both completed 5.6 -1.0 -.95 -1.0 44 Couple's Residential Backqround Both Rural 6.7 .09 .06 .08 380 One from village 6.6 -.02 .03 -.03 . 95 Both from village 6.7 .04 .08 .06 39 One from town 6.5 -.10 -.04 -.06 107 Both from town 6.0 -.62 -.52 -.57 29 Couple's Residential Historv No moves 6.9 .27 .20 .26 131 Some moves, no large town 6.6 -.02 -.02 -.02 361 Some moves, some large towns 6.4 -.19 -.13 -.17 181 Husband's Income No work-100 Colones 6.8 .20 .08 .17 45 100-300 Colones 6.7 .06 .02 .05 282 300-600 Colones 6.6 -.05 -.01 -.05 132 600 Colones or more 6.2 -.38 -.24 -.33 87 Husband's Emplovment Status -Tenant Farmers 6.6 .03 .02 .02 510 Unskilled Labor 6.3 -.32 -.28 -.29 50 Farmers or Skilled Labor 6.5 -.10 -.07 -.06 82 Housinq Conditions Very poor 6.8 .16 .11 .14 149 Poor 6.7 .08 .09 .08 305 Some comforts 6.5 -.06 -.06 -.05 147 Comfortable 6.1 -.54 -.45 -.50 76 Grand Mean 6.6 TABLE 7.9 -- WANTED FAMILY SIZE BY COUPLE'S EDUCATION AND RESIDENTIAL BACKGROUND AND FAMILY PLANNING VARIABLE, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Wanted Family Size Biopsych. Couple's Education Couple's Residential Back. Characteristic 1 2 3 4 5 T 1 2 3 4 5 T N Attitude toward Family Planning Against 7.4 6.9 6.8 7.2(6.7) 7.0 6.9 7.1 (7.1) 7.4 * 7.1 305 Favor 6.1 6.5 6.5 6.3 5.0 6.0 6."5 6.4 6.3 6.1 5.8 6.2 359 Total 6.7 6.7 6.6 6.7 5.9 6.5 6.7 6.7 6.7 6.7 5.8 6.6 Use of Contrac. No 6.7 6.7 6.6 6.8(6.6) 6.7 6.7 6.8 6.6 6.6 (6.9) 6.7 454 Yes (7.4) 6.8 6.7 6.5 5.2 6.5 6.7 6.2 (6.8) 6.5 (5.3) 6.3 228 Total 7.0 6.7 6.6 6.8 5.9 6.6 6.7 6.5 6.7 6.6 6.1 6.5 N 57 171 279 97 44 380 95 39 107 29 683 * Base number less than 10. ( ) Less than 20 but at least 10. Codes: For Couple's Education: 1 = Both had less than 1 year. 2 = One elementary school incomplete. 3 = Both elementary school incomplete. 4 = One elementary school .completed. 5 = Both elementary school completed. For Couple's Residential Background: 1 = Both rural; 2 = One from village; 3 = Both from village; 4 = One from town; ____________________ 5 = Both from town. TABLE 7.10 — WANTED FAMILY SIZE BY FAMILY PLANNING VARIABLES AND PRESENT SOCIAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Wanted Family Size Att.toward Fam.Pl. Present Characteristic Anainst Ravnr t Use No of Contracep. Yes T N Couple’s Residential History No moves 7.1 6.5 6.8 6.9 6.8 6.8 131 Some moves, no large town 6.9 6.3 6.6 6.6 6.6 6.6 361 Some moves, no large towns 6.7 6.3 6.5 6.6 6.2 6.4 181 Total 6.9 6.4 6.6 6.7 6.5 6.6 Husband’s Income No work-100 Colones 7.0 * 7.0 6.8 * 6.8 45 100-300 Colones 6.8 6.5 6.6 6.7 6.6 6.6 282 300-600 Colones 6.8 6.4 6.6 6.6 6.4 6.5 132 600 Colones or more * 5.9 5.9 6.1 6.3 6.2 87 Total 6.9 6.3 6.6 6.6 6.4 6.5 Husband1s Employment Status Tenant Farmers 7.0 6.4 6.7 6.7 6.5 6.6 516 Unskilled Labor 6.5 6.2 6.4 6.2 6.4 6.3 50 Farmers or Skilled Labor 7.0 6.2 6.6 6.7 6.3 6.5 82 Total 6.8 6.3 6.6 6.5 6.4 6.5 Housing Conditions Very poor 7.0 6.5 6.7 6.7 7.1 6.9 148 Poor 6.9 6.5 6.7 6.7 6.6 6.7 305 Some comforts 6.9 6.3 6.6 6.7 6.3 6.5 147 Comfortable 6.9 5.7 6.3 6.1 6.0 6.1 76 Total 6.9 6.3 6.6 6.6 6.5 6.6 N 305 359 454 : 228 683 *Base number less than 10 162 > desired fertility. Cumulative effects seem to be produced when those who favor family planning or use contraception are also better educated couples or have had a residential back ground mainly in towns. Discussion A greater role of the biopsychological indicators was expected both in their direct influence on desired fertility and as a channel through which the effect of the social background and present characteristics on desired fertility would run. The results reported above clearly indicate that fertility, if measured as wanted family size, is not significantly affected by the biopsychological variables included in this study, nor do these factors meaningfully intervene as channels through which the socio- structural characteristics affect fertility. It could be argued, in trying to explain the lack of differential effect of these psychological indicators, mainly the two family planning variables, on desired fertility, that since high fertility is still normative in the rural and semi-urban areas of Costa Rica, this norm is still operative on the level of fertility desires, with no significant differences among couples from different social background and socioeconomic status, and regardless of their attitude toward family planning or of their use of 164 contraception. This perhaps reflects the still meaningful importance of familial and kinship ties in community life, prevalent in a rural society. The fact that this norm is not operative at the level of actual fertility in all the cases could be explained as due to the fact that perform ance is more the result of individual's reactions to sets of circumstances than of a strict adherence to a specific fertility norm. Ideal Fertility The trends that were found in the preceding section on desired fertility and the role of psychological variables are more noticeable when dealing with ideal fertility, and some new ones develop. The multiple correlation coefficient of this group of variables with ideal fertility is .26, which confirms hypothesis 1-1 insofar it predicted a strong relationship of this group of indicators with ideal fertility. As a matter of fact, it is stronger than the relationship with ideal fertility of either the social background or the present factors. Hypotheses 1-2 and 1-3, which predict that both social background and present indicators will exert part of their influence on fertility through this group of factors, are partially confirmed. It is confirmed in the case of education, husband's income and housing conditions. In the 165 case of the other three indicators--residential background, residential history, and husband’s employment status--, as Table 7.12 shows, there seems to be present an intervening effect of the family planning variables but only to a small degree. . . Regarding the effect of education being mediated through use of contraception, and the effect of husband's, income through attitude toward family planning and use of contraception, both hypotheses (I1-2 and I1-3) are confirmed. It must be added that in the case .of education, attitude toward family planning has also had an intervening effect (Table 7.12). As to the adjusted fertility differentials, hypothesis IV-3 predicts that the effect of housing condi tions on ideal fertility will be weaker when the bio psychological factors are taken into account. This hypothesis is confirmed, even if the independent effect of housing conditions remains fairly strong. Hypothesis IV-4, which postulates that the influence of use of contra ception on ideal fertility will remain fairly strong after the effect of education and income are taken into account, is also confirmed (Tables 7.12 and 7.6). It must be added that the effect of attitude toward family planning is also strongly independent of the various socio-structural indicators. Hypothesis 1-6 predicts that the intervening effect 166 of the biopsychological indicators between social background factors and present factors on the one hand and fertility on the other', will be stronger for ideal fertil ity than in the case of the other two measures of fertility and stronger for desired fertility than for past fertility. This is confirmed (Tables 7.2; 7.8 and 7.12). Some non-hypothesized channels of influence must be noticed. (a) The effect of attitude toward family planning on ideal fertility is largely independent of the effect of use of contraception, but not vice versa (Table 7.11). (b) The differential effect of extended family living exper ience on ideal fertility is in part due to the effect of attitude toward family planning, even if it retains a fairly strong independent effect. Some cumulative effects are noticeable, mainly between couple's residential background and both family planning variables, and housing conditions and use of con traception. When couples are in the highest subclass in both variables, the ideal fertility is considerably lower than average (Tables 7.13 and 7.14). Discussion In the case of ideal fertility, it appears that the trends operating in Costa Rica, not only agree with what was hypothesized but with what studies in other developing countries have found. The biopsychological variables seem TABLE 7.11 — MULTIPLE CLASSIFICATION ANALYSIS RELATING IDEAL FAMILY SIZE TO BIOPSYCHOLOGICAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Ideal Family Size Deviation No - from Ad-justed Deviation of Biopsychological Characteristics Mean Grand Mean CF . EF AP UC Cases Coital Frequency (CF) Less than 1 a month 5.2 -.01 .00 -.03 .01 32 1-3 a month 5.1 -.14 -.11 -.19 -.15 171 1-3 a week 5.2 -.04 -.03 .00 -.01 282 3 or more a week 5.4 .18 .15 .21 .16 137 Correlation Ratio .05 Extended Family Livinq Experience(EF) No 5.2 -.06 -.06 -.04 -.06 619 Yes 5.8 .52 .50 .39 .51 60 Correlation Ratio .09 Attitude toward Family Planninq(AP) Against 5.9 .64 .64 .62 ‘ .59 305 Favor 4.7 -.53 -.54 -.52 -.49 359 Correlation Ratio .26 Use of Contraception(UC) No 5.5 .22 .21 .22 .11 454 Yes 4.8 -.43 -.42 -.43 -.21 228 Correlation Ratio .14 Grand Mean 5.3 167 TABLE 7.12 — MULTIPLE CLASSIFICATION ANALYSIS RELATING IDEAL FAMILY SIZE TO SOCIAL BACK GROUND AND PRESENT SOCIAL CHARACTERISTICS BY FAMILY PLANNING VARIABLES, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Social Background and Present Characteristics Mean Deviation from Grand Mean Adiusted Deviation Att.toward Use of Fam.Plan. Contrac. No. Of Cases Couple1s Education Amount of Elementary School: Both had/less than 1 year 5.6 .33 .24 .25 57 One incomplete 5.6 .37 .23 .31 171 Both incomplete 5.2 -.08 -.02 -.07 279 One completed 5.0 -.28 -.23 -.22 87 Both completed 4.4 -.83 -.64 -.62 44 Couple's Residential Background Both Rural 5.4 .11 .04 .07 380 One from village 5.2 -.04 .05 -.06 95 Both from village 4.7 -.52 -.45 -.47 39 One from town 5.0 -.24 -.11 -.14 107 Both from town 4.2 -1.1 -.90 -.93 29 Couple's Residential History No moves 5.6 .38 .25 .35 131 Some moves, no large town 5.4 .09 .08 .07 361 Some moves, some large towns 4.7 -.54 -.42 -.49 181 Husband's Income No work-100 Colones 5.8 .58 .31 .47 45 100-300 Colones 5.6 .38 .31 .34 282 300-600 Colones 5.1 -.21 -.15 -.22 132 600 Colones or more 4.6 -.69 -.44 -.55 87 Husband's Employment Status Tenant Farmers 5.4 .10 .09 .08 510 Unskilled Labor 5.0 -.22 -.16 -.14 50 Farmers or Skilled Labor 4.7 -.58 -.49 -.45 82 168 Social Background and Present Characteristics TABLE 7. Mean 12— (Cont.) Deviation from Grand Mean Att.toward Fam. Plan. Use of Contrac. No. of Cases Housinq Conditions Very poor 5.7 .45 .35 .36 149 Poor 5.3 .07 .07 .07 305 Some Comforts 5.0 -.29 -.29 -.27 147 Comfortable 4.7 -.61 -.39 -.45 76 Grand Mean 5.3 169 TABLE 7.13 — IDEAL FAMILY SIZE BY COUPLE'S EDUCATION AND RESIDENTIAL BACKGROUND AND BIOPSYCHOLOGICAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Ideal Family Size Biopsychol. Characteristic Couple's Education Couple's Resid. Background 1 2 3 4 5 T 1 2 3 4 5 T N Coital Frequency Less than 1 month * * (4.8) * * 4.8 (5.8) * * * * 5.8 32 1-3 a month (5.6) 5.5 4.9 (5.2)(4.9) 5.2 5.1 4.8 (5.9) 4.9 * 5.2 171 1-2 a week 5.2 5.3 5.3 5.1 (4.2) 5.0 5.3 5.4 (3.9) 5.0 (4.2) 4.8 282 3 or more a wk. (7.4) 5.9 5.5 (4.5) * 5.8 5.6 5.5 . * ..(5,4). * 5.5 137 Total 6.1 5.6 5.1 4.9 4.5 5.3 5.4 5.2 4.9 5.1 4.2 5.3 Extended Family Living Experience No 5.6 5.5 5.1 5.0 4.4 5.1 5.3 5.2 4.7 5.0 4.2 4.9 619 Yes * (6,_5) 6.0 * * 6.2 5.9 * * * * 5.9 60 Total 5.6 6.0 5.5 5.0 4.4 5.4 5.6 5.2 4.7 5.0 4.2 5.3 Attitude toward Family Planning Against 6.1 6.2 5.8 5.6 (4.7) 5.7 5.9 6.0 (4.9) 5.9 * 5.7 305 Favor 4.9 4.8 4.8 4.6 4.4 4.7 4.8 4.7 4.6 4.6 3.7 4.5 359 Use of Contrac. No 5.7 5.6 5.4 5.4 (4.4) 5.3 5.5 5.3 4.8 5.4 (4.6) 5.1 454 Yes (5.3) 5.7 4.8 4.5 4.5 5.0 4.9 5.1 (4.7) 4.7 (3.9) 4.7 228 Total 5.5 5.6 5.1 5.0 4.4 5.1 5.2 5.2 4.7 5.0 4.2 4.9 N 57 171 279 87 44 380 95 39 107 29 683 * Base number less than 10. I ) Less than 20 but at least 10. Codes: For Couple's Education: 1 = Both less than 1 year; 2 = One elementary school incomplete; 3 = Both elementary school incomplete; 4 = One elementary school completed; 5 = Both elementary school completed. For Couple's Residential Background: 1 = Both rural; 2 = One from village; 3 = Both from village; 4 = One from town; 5 = Both from town. O TABLE 7.14 — IDEAL FERTILITY BY COUPLE'S RESIDENTIAL HISTORY AND HOUSING CONDITIONS AND BIOPSYCHOLOGICAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Ideal Family Size Biopsychological Characteristics Resid. History Housing Conditions 1 2 3 T 1 2 3 4 T N Coital Frequency Less than 1 a month * * * * * * * * * 32 1-3 a month 5.1 5.5 4.4 5.0 5.1 5.5 4.4 5.2 5.0 170 1-2 a week 5.5 5.2 4.9 5.2 5.7 5.2 5.2 4.5 5.1 282 3 or more a month 6.5 5.6 4.5 5.5 5.8 5.1 5.4 * 5.4 137 Total 5.7 5.4 4.6 5.2 5.5 5.3 5.0 4.9 5.2 Extend. Fam. Livinq Exper. No 5.6 5.3 4.7 5.2 5.7 5.3 4.9 4.5 5.1 619 Yes * 6.2 5.0 5.6 * 5.9 * * * 60 Total * 5.8 4.8 5.4 * 5.6 * * * Attitude toward Fam. Planninq Against 5.9 6.1 5.2 5.7 6.0 6.2 5.4 5.3 5.7 305 Favor 5.2 4.7 4.5 4.8 5.3 4.7 4.6 4.3 4.7 359 Total 5.5 5.4 4.7 5.2 5.7 5.5 5.0 4.8 5.2 Use of Contraception No 5.9 5.5 4.8 5.4 5.7 5.6 5.0 5.3 5.4 454 Yes 5.0 4.9 4.6 4.8 5.8 4.8 5.0 4.3 5.0 228 Total 5.5 5.2 4.7 5.1 5.7 5.2 5.0 4.8 5.2 N 131 361 181 149 305 147 76 683 * Base number less than 10. ( ) Less than 20 but at least 10. Code: For Couple's Residential History: 1 : = No moves; 2 = Some moves, no large town; 3 : = Some moves, some large towns. For Couple's Housing Conditions: 1 : 4 : = Very poor; 2 = = Comfortable. Poor; 3 = Some < comforts; 171 TABLE 7.15 -- IDEAL FAMILY SIZE BY HUSBAND'S INCOME AND EMPLOYMENT STATUS AND BIOPSYCHO LOGICAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Ideal Family Size Biopsychological Characteristics Husband's Income Hus.Employ.Status 1 2 3 4 T 1 2 3 T N Coital Frequency Less than 1 a month * * * * * 5.5 * * * 32 1-3 a month * 5.5 4.8 4.7 5.0 5.1 * 4.9 5.0 170 1-2 a week * 5.6 5.1 4.2 4.9 5.4 4.9 4.4 4.9 282 3 or more a week • * 5.9 5.3 5.0 5.4 5.6 * * * 137 Total * 5.7 5.1 4.5 5.0 5.4 * 4.6 5.0 Extend. Family Livinq Exper. No 5.7 5.6 5.0 4.5 5.2 5.3 4.9 4.7 5.0 619 Yes * 5.9 * * * 5.8 ' * * * 60 * 5.7 * * * 5.5 * * * Attitude toward Fam. Planning Against 6.2 6.2 5.8 * 6.1 6.1 5.4 5.0 5.5 305 Favor * 5.0 4.5 4.5 4.7 4.8 4.8 4.5 4.7 359 Total * 5.6 5.1 * 5.4 5.4 5.1 4.7 5.1 Use of Contraception No 6.0 5.9 5.2 4.5 5.4 5.6 5.3 4.9 5.3 454 Yes * 4.9 4.7 4.6 4.7 4.9 4.8 4.5 4.7 228 Total * 5.4 5.0 4.6 5.1 5.3 5.0 4.7 5.0 N 45 282 132 87 516 50 . 82 683 * Base number less than 10. ( ) Less than 20 but at least 10. Code: For Husband1s Income • • 1 = No work-■100 Colones; 2 = 100-■300 Colones; 3 = 300-600 Colones; 4 = 600 Colones or more. For Husband1s Employment Status : 1 = Tenant Farmers; 2 == Unskilled Labor; 3 = Farmers or Skilled Labor. 172 to be one of the reasons that account for the fact that socio-structural variables have substantial effects on fertility ideals. Education, husband's income, and housing conditions are the three socio-structural factors which exert part of their influence on ideal fertility through family planning variables. This implies that those couples with better education have a more favorable attitude toward family planning and make more use of contraception, and this in part explains why they have smaller family ideals than their counterparts with less education. The same process seems to operate with those couples with better incomes or living in better housing conditions (Table 7.16). In Puerto Rico (see Hill et al., 1959), Taiwan (see Freddman and Takeshita, 1969), and Korea (Bom et al., 1972), the patterns are roughly the same. Attitude toward family planning and use of contra ception emerge as fairly important determinants of ideal fertility, regardless of the effects of other variables. The case could be made, though, for a reverse relationship. Ideal fertility could be the determinant of a favorable attitude toward family planning and lead the couple to use contraception, which in turn would differentially affect fertility behavior. On theoretical grounds, the scheme is sound, even if not utilized in this study. Another possi bility would be to conceive of the two variables--attitude toward family planning and ideal fertility--as simultaneous 174 TABLE 7.16 PERCENT BY WHICH THE EFFECT OF VARIOUS SOCIAL BACKGROUND AND PRESENT SOCIAL CHARACTERISTICS ON IDEAL FAMILY SIZE IS REDUCED WHEN ADJUSTED FOR THE EFFECT OF A FAMILY PLANNING VARIABLE Social Background and Present Social % Reduction when Adjusted for Characteristic Att.t.Fam.PI. Use of Contrac. Both Couple’s Education 24 24 34 Couple’s Residential Background 22 17 26 Couple’s Residential History 27 9 28 Husband's Income 40 18 44 Husband’s Employment Status 15 22 24 Housing Conditions 30 24 41 175 and both differentially determining use of contraception, and thus actual fertility. The fact remains, though, that both family planning variables are fairly strongly associated with ideal fertility, in whatever direction this relationship be conceived. Since we are dealing at the level of attitudes when analyzing ideal fertility and attitude toward family planning, the fact that their association is stronger than the association found between ideal fertility and use of contraception, is foreseeable. The latter measures action, and attitudes not always lead to congruent action. This explains also the findings that attitude toward family planning differentially affect ideal fertility independent ly from the couple's use of contraception, but not the opposite. Use of contraception differentially affects ideal fertility, but most of this effect is attributable to the influence of attitude toward family planning. CHAPTER VIII SUMMARY AND IMPLICATIONS Summary A principal goal of this dissertation is to explore and analyze the determinants of fertility in rural and semi-urban Costa Rica. Fertility has been traditionally high in this country, but both demographic and socio- structural changes are taking place. An implicit assump tion is that these changes lead to a differential adjust ment in fertility levels. The main questions explored are the following: (i) What are the observed fertility differentials by indicators of social structure and socioeconomic status in rural and semi-urban Costa Rica at the present stage of development? (II) To which extent these determinants of differential fertility are independent of the intervening effect of other indicators? (III) Is there a systematic cumulative effect on fertility between pairs of indicators? (IV) Do the attitudinal and biological factors play a role in determining fertility differentials? Is this 176 177 role an independent one or is it merely a transferral of effects from socio-structural determinants to fertility? (V) Does the scheme analyzed operate in the same way for past fertility than for prospective and ideal fertility? It was hypothesized that the social background characteristics affect fertility in a differential way (a) directly, (b) through present characteristics, (c) through attitudinal factors, and that the present characteristics influence fertility (a) directly, and (b) through attitudinal factors. The specific factors included in each category are: 1. Social background factors: couple’s education couple’s residential background 2. Present factors: couple’s residential history husband's income husband’s employment status housing conditions 3. Biopsychological factors: coital frequency extended family living experience attitude toward family planning use of contraception 178 4. Control variables: age at union duration of union The theoretical background of the scheme was dis cussed in Chapter II with reference to relevant findings for other developed and developing countries. A specific set of hypotheses was spelled out. The sources of data and the statistical methods used were examined in Chapter III. With Chapter IV the analysis was introduced by discussing the general fertility differentials of rural and semi-urban Costa Rican couples, after a previous comparison of fertility levels in Costa Rica and in other countries. In Chapter V we probed the relationships among the various independent factors in regard to fertility behavior in Costa Rica. In Chapter VI we proceeded to discuss the socio-structural correlates of fertility, and the mechan isms of influence between various indicators and fertility. In the last chapter of analysis, Chapter VII, the role of the biological and psychological factors was examined. Throughout, the analysis was sustained at the three levels of fertility--past, prospective, and ideal--and was mainly based on the older rural and semi-urban Costa Rican couples. The hypotheses outlined in Chapter II were in general supported. Variables included in these hypotheses were examined individually and in groups; they suggested a 179 set of "causes" of rural and semi-urban fertility. It appears from these data that each one of the six social background and present indicators is of some importance and makes a unique contribution to fertility differentials. Over all, the two social background factors and husband's income and housing conditions from the group of present indicators, seem to be most important. Among the four bio psychological variables, the only strictly biological, that is, coital frequency, and extended family living experience appear to have no importance. The other two, which have been considered as family planning variables, proved to be important only in relation to ideal fertility. In a more specific way, the older and semi-urban couples with the following characteristics have remarkably lower past fertility and consider as ideal smaller families: those who have completed elementary school, have had a residential background mostly in towns, have resided in at least one town in their adult life, the husbands have a salary of 600 colones or more, the husbands work in skilled jobs or agricultural jobs of at least an inter mediate level, live in comfortable housing conditions. These results are consistent with those reported in various studies of other developing countries. The differentials in fertility among the younger rural and semi-urban couples are also considerably high. This, combined with the fact that they want smaller 180 families than the older couples, leads to the conclusion that they will have a consistent differential fertility behavior in the future and that they will have smaller families than their older counterparts. The most inter esting difference is that younger couples show more pronounced differentials in prospective fertility than the older couples. The ideal family size is smaller than the wanted size, which in turn is smaller than the number of live births. This pattern, which probably is the result of the combined effects of high mortality in the past, lack of means to control fertility, and recent rapid modernization leading to preference for smaller families, implies a process of fertility change. Some of the socio-structural indicators are strongly associated among them. Because of this, the net effects--adjusted for the effect of another indicator--of a predictor are generally weaker than the gross or unadjusted effects. Most of the hypotheses about adjusted fertility differentials have proved to be acceptable in the case of past and ideal fertility, and some non-hypothesized channels have been found. Table 9.1 presents a summary of all the hypotheses and their level of confirmation for the three fertility measures. Treated as groups, the social background and present characteristics tend to explain more 181 variance in actual fertility, while the attitudinal factors explained more variance in ideal fertility. Thus, the fertility measures of rural and semi- urban Costa Rican couples vary significantly by the socio- structural and socioeconomic factors considered. The ultimate level of fertility is the outcome of a complex interplay of variables. The effect of remote causes is exerted through more recent ones, and recent causes also interplay among them to affect fertility. For example, it was shown that housing conditions is part of the path through which education affects fertility, and it also appears to have indirect effects through its association with income. Housing conditions also has direct effects upon fertility, effects which are partly masked by its association with residential history. Implications The scheme"tested in this study is consistent with the concept of multiple causation. Although the causal ordering of many of the variables used is not clear and could sometimes be reversed, fertility is clearly a dependent variable. This variable is caused by a large number of independent variables, none of which produces a large amount of change. For example, the correlation coefficient between income and the number of live births is -.10. This correlation, when interpreted as a slope, 182 TABLE 8.1 SUMMARY OF HYPOTHESES AND THEIR CONFIRMATION STATUS Confirmation Status Past Desired Ideal Hypotheses____________________Fertility Fertility Fertility I - Contribution of the three groups of factors to fertility la Social Background factors yes weak yes b Present factors yes weak yes c Biopsychological factors no yes yes 2a Social background partly through present social yes only CRB yes b Social background partly through biopsychological no no only CE 3 Present social factors partly through biopsychological no no only Hi 4 Social background have independent effects yes no only CRB 5 Present social factors have independent effects yes RH,HI,HC yes 6 Intervening effects of bio psychological factors stronger for ideal than for desired fertility, stronger for desired than for past fertility yes-------- II- Socio-structural Differen tials in Fertility 1 Each social background and present social factor strongly associated with except fertility yes only CE,RH,HC CE,HES 2 CE partly through HI no no yes CE partly through HES no no no CE partly through HC yes no yes CE through UC no no yes 3 HI partly through AP no no yes HI partly through UC no no yes 183 TABLE 8.1--(Cont.) Confirmation Status Hypotheses Past Desired Fertility Fertility Ideal Fertility IV- Adjusted Fertility differentials 1 HI strong negative effect when adjusted for other factors yes no(except yes 2 HES weaker net effect HC) when adjusted for other factors Yes(CE, or all present) no only HI 3a HC strong net effect when adjusted for HI yes weak no b HC weaker net effect when adjusted for CE no no yes c HC weaker net effect when adjusted for biopsychological no no yes 4a UC strong net effect when adjusted for CE no no yes b UC strong net effect when adjusted for HI no no yes Code for Variables: CE - couple's education CRB - couple's residential background RH - couple's residential history HI - husband's income HES - husband's employment status HC - housing conditions AP - attitude toward family planning UC - use of contraception____________ 184 indicates that it takes 10 units of change in income to produce one more live birth*. If this correlation is interpreted in isolation from other variables in the rural and semi-urban setting, it might be regarded as unimport ant. But, when all socio-structural variables considered in this study, including income, are analyzed together in a multiple correlation equation, the resultant correlation coefficient is .30. This indicates that proportion of variance in past fertility accounted for has increased. Consequently, an implication of this study is that 1 1 the1 1 cause of fertility cannot be isolated. There do not appear to be two or three variables which act to produce fertility even if some are more important than others, specifically education for past fertility and income for ideal fertility according to this study. The concept of multiple causa tion, thus, seems appropriate. On the other hand, the values of explained variance obtained seem to indicate that even a set of crucial socio- structural variables leaves most of the variance unex plained. It is possible that the addition of other avail able variables would add little statistically, perhaps because the most important factors (idiosincratic?) operating in relation to fertility cannot be isolated. It is true, however, that some distinct and consistent patterns are present and operating in rural and semi-urban Costa Rica. It has been found that when couples 185 are classified in the highest category in any pair of soc ial background, present social, and family planning indicators, their fertility performance and ideals are much lower than average. This would suggest that there are several possible combinations of characteristics that have a strong effect in lowering fertility. A final remark at this respect. The fact that the factors differentially affecting rural and semi-urban fertility in Costa Rica are roughly the same that have been found in urban settings, leads to the conclusion that certain classes of variables are important determinants of fertility wherever they are found. It is possible, though, that the type of environment, rural versus urban, may interact to produce different relationships among variables. For example, rural environments may select people who are oriented toward large families. Such persons, even when employed in skilled jobs or when earning better salaries, may still have many children. In other words, the relation between occupational status or income may not be the same in rural as in urban environments. Fertility Control This study has implications for public policy aimed at fertility control. The most general finding is that not a single factor would be very effective in decreasing fertility, but that a whole set of factors 186 should be approached simultaneously. Thus, the findings of this analysis seem to support the "economic development" approach to population growth, that is, that decreases in family size are the long range resultant of a socioeconomic transformation and moderniza tion which, in turn, leads to a desire for fewer children. They also somewhat support the family planning approach which assumes that education and communication regarding birth control will eventually reduce births to a level in keeping with low mortality. In practical terms, however, population growth in rural and semi-urban Costa Rica is still fast; thus, it impedes the achievement of an economic development that would lead to a curtailment of that same population growth. Furthermore, the advantages of the family group in the modern world cannot be discounted. One must also take into account the strong social controls which isolate the individual in modernized societies from alternative roles and satisfactions and, hence, bolster their intense feelings of dependency on the family, and a fortiori on having children. It would, thus, appear that, concomitantly with the furthering of the modernization process and the intensifi cation of the family planning programs, the family institution has to be modified to make the small family a way of life. The motivational framework of reproduction could be affected through various policies that would 187 undermine the utilities found in offspring (thereby allowing a sense of increased costs to prevail), and would stress crucial foci of change in the society. Such would be female labor-force participation, a rigidly compulsory education of children which would remove them as potential economic assets and put intellectual barriers between them and past generations, and some others. Evaluation of the Present Study There are a number of problems and limitations with this study. They will become apparent when evaluating both the conceptual framework and the methods of analysis. Causal Order One of the most frequent problems encountered in the analysis of hypotheses, even if not always specifically mentioned, was the determination of the causal ordering of variables. Sometimes these problems could be partially dealt with theoretically. For example, it was argued that education precedes residential history, and therefore the former can legitimately be considered a cause of the latter. On other occasions a certain ambiguity in the logic of the conceptual scheme could not be prevented. The interpretations of the results would have been different if certain variables had been ordered in different sequences, 188 which theoretically would have been acceptable. In the body of the analysis the case was mentioned of the differ ent options in sequentially ordering attitude toward family planning and ideal number of children. There is some ambiguity in the causal relationship between present characteristics and attitudes. Actually, these two classes of variables rest on the same time basis. The actual ordering of present factors leading to attitudes was simply based on the preference for and also on the somehow greater likelihood of this sequence than any other. Such questions are critical. The actual ordering of variables determines whether or not a given relationship with fertility is "spurious*' or is "interpreted." For example, if income is temporally prior to both residential history and fertility, then the association between the latter may be spurious. This would result in the conclu sion that residential history is an unimportant variable in the determination of fertility. If income intervenes temporally between residential history and fertility, then an important path through which residential history affects fertility has been isolated. A related problem is mutual causation. It is possible that income and housing conditions have reciprocal effects upon one-another. As with determination of tempor al sequence, a clear theoretical rationale is essential. However, to evaluate hypotheses of this kind, certain kinds of data are also required. One statistical technique that handles this problem is "lag correlation" (of. Pelz and Andrews, 1964; and Bohrnstedt, 1969). This type of correlation requires data on each of two variables used in the study for two differ ent points on time. With such data it would be possible to correlate one variable at the prior point of time with another variable in the posterior point of time, and vice versa. If one correlation is substantially larger than the other, then insight into the proper causal ordering is obtained. If both are of approximately equal magnitude, then support is given to hypotheses about reciprocal effects. There is also a certain theoretical weakness in using present social characteristics to account for life time fertility behavior because most of the present indicators are expressed in terms of the situation at the time of the survey. This was tolerated under the assumption that changes in status are usually slow and are likely to follow a certain pattern. In order for present factors to be used properly in the study of fertility, they should be measured as the situation at an earlier period, prior to family building and the eventual use of family limitation. Finally, if all the independent variables had been recoded in a numerical scale, other stronger statistical 190 measures and techniques might have been used. This recoding would be difficult but not impossible. Further more, some predictors might have been more useful if they had been expressed differently from the way they were in this study. For example, income could be expressed as monthly per capita income. Other Variables This study is concerned with evaluation in a specific rural and semi-urban context of a developing country, of hypotheses relating independent variables of known importance to fertility. The goal is the further understanding of the determinants of rural and semi-urban fertility. Many other sociological, demographic, and attitudinal variables were not included because data were not available for their measurement and because of the need to keep the analysis within manageable proportions. There are also important methodological reasons for including other variables. An implicit assumption underlying any multiple regression analysis, such as the multiple classification analysis used in this study, and also the test of bivariate hypotheses is that all relevant variables are included. This is probably not the case. A serious problem arises if relevant variables are excluded because the magnitudes of the coefficients depend specifically upon the variables included in the equation 191 (cf. Gordon, 1968). If variables which are correlated with a given set of independent variables are excluded spurious ly high coefficients may result. Consequently, it may be erroneously concluded that a particular variable makes a large independent contribution to fertility. The interpretation-of bivariate hypotheses implies analogous dangers. If an unanalyzed variable is correlated with two variables included in an hypothesis, the predicted relationship may be spuriously high or spuriously low. Of course, these problems are always present. Blalock (1961) remarks that the researcher must at some time consider his system closed. It would be advisable, though, to extend the theoretical and methodological rationale of the analysis of fertility to include more pertinent variables. Additivity Assumption Both the perspective developed above and the multiple classification analysis technique used require an additive model. In other words the assumption is made that unit increments in two variables lead to twice the change in a dependent variable as a unit change in just .one of these variables. However, it may be that the coeffi cients vary according to the level of some third variable. For example, it may be that income and fertility are re lated in different ways under the condition of high 192 education than they are under the condition of low education. Such hypotheses require a more sophisticated theory than proposed here. This and similar studies, though, might be extended to permit hypotheses about inter action effects. These studies can utilize techniques for the analysis of interaction now being developed (see, for example, Sunquist et al.. 1964; Blalock, 1965). Conclusion Because of the limitations summarized in the section above, the findings of this study are somewhat tentative. The contribution of this study to the field may be seen as twofold. First, because of the systematic way in which it analyzes the gross and net fertility differ entials, confirming or rejecting specific hypotheses regarding the association between three different levels of fertility and socio-structural indicators that are consi dered to be operating in a developing society. Secondly, insofar it specifies the channels of influence through which the various socio-structural factors affect fertility. These factors affect fertility behavior and values through t complex processes. Not all of them influence fertility simultaneously and with the same intensity. They form certain sequences, which are tentatively specified. A further contribution of this study is seen in the fact that 193 deals exclusively with a rural and semi-urban setting of a developing country, which has seldom been done. It may be that more sophisticated methods of analysis would lead to alterations in the conclusions reached. Nonetheless, it is hoped that the basic scheme and its analysis have proved to be sound. The fertility level of individual couples, even in rural and semi-urban environments, has started to be, and will increasingly be, a function of the decision to take or not effective action in planning fertility. This decision is affected by one's access to the idea and methods of family planning and by one's attitudes toward new values, which in turn are a function of the environment, both social and physical. BIBLIOGRAPHY REFERENCES Adelman, Irma. 1963 "An Economic Analysis of Population Growth." American Economic Review 53: 314-339. Andrews, Frank, James N. Morgan, and John A. Sonquist. 1969 Multiple Classification Analysis. Ann Arbor: Survey Research Center, the University of Michigan. Becker, Gary. 1960 "An Economic Analysis of Fertility." Pp. 209-231 in National Bureau Committee for Economic Research (ed.), Demographic and Economic Change in Developed Countries. Princeton: Princeton University Press. Bernard, Jessie. 1971 "Education as a Demographic Variable." Pp. 1876- 1889 in International Population Conference, London 1969. Liege: The International Union for the Scientific Study of the Population. Vol. III. Blalock, Hubert. 1961 Causal Inferences in Non-experimental Research. Chapel Hill: University of North Carolina Press. 1965 "Theory Building and the Statistical Concept of Interaction." American Sociological Review 29: 374-380. 1968 "Theory Building and Causal Inferences." Pp. 155- 198 in Hubert M. Blalock and Ann Blalock (eds.), Methodology in Social Research. New York: McGraw-Hill. / / Blanch, Jose-Marxa. 1972 "Factores Estructurales y Ecologicos en la Fecundidad de Centroamerica y Panama." In Conferencia Regional Latinoamericana de Poblacion. Mexico, D.F. 1970. 195 196 Bogue, Donald. 1969 Principles of Demography. New York: John Wiley and Sons. Bohrns tedt, George. 1969 "Observations on the Measurement of Change." Pp. 113-133 in Edgar F. Borgatta (ed.), Sociological Methodology. San Francisco: Jossey-Bass. Bom, Mo Chung, et al. 1972 Psychological Perspectives: Family Planning in Korea. Seoul: Hollym Corporation. Chandras ekar, S. 1955 "Cultural Barriers to Family Planning in Under developed Countries." Pp. 64-70 in Proceedings of the 5th International Conference on Planned Parenthood. London. Coliver, Andrew, et al. 1967 "Local Variations of Fertility in Taiwan." Population Studies 20: 329-343. Davi s, King s1ey. 1955 "Institutional Patterns Favoring High Fertility in Underdeveloped Areas." Eugenics Quarterly 21: 33-39. Davis, Kingsley, and Judith Blake. 1956 "Social Structure and Fertility: An Analytical Framework." Economic Development and Cultural Change 4: 211-235. Duncan, Otis. 1965 "Farm Background and Differential Fertility." Demography 2: 240-249. 1966 "Path Analysis: Sociological Examples." American Journal of Sociology 72: 1-16. Edin, K. E., and E. P. Hutchinson. 1935 Studies of Differential Fertility in Sweden. London: P. S. King and Son. 196 197 Freedman, 1961 1963 1965 Freedman, 1963 1967 Freedman, 1960 Freedman ■ , 1959 Freedman, 1964 Freedman, 1965 Freedman, 1966 Freedman, 1968 Ronald. "The Sociology of American Fertility: A Trend Report and Bibliography." Current Sociology 10- 11 (2): 35-121. "Norms for Family Size in Underdeveloped Countries." Pp. 220-245 in Proceedings of the Royal Society 159. "The Transition from High to Low Fertility: Challenge to Demographers." Population Index 31: 417-429. Deborah. "The Relation of Economic Status to Fertility." American Economic Review 53: 414-426. "The Role of Consumption of Modern Durables in a Developing Economy. The Case of Taiwan." Ph.D. Dissertation. The University of Michigan (unpublished). Ronald, and D. P. Slesinger. "Fertility Differentials for the Indigenous Nonfarm Population of the United States." Population Studies 15: 161-173. Ronald, and Morris Silver. Family Planning, Sterility, and Population Growth. New York: McGraw-Hill. Ronald, et al. "Fertility and Family Planning in Taiwan: A Case Study of the Demographic Transition." American Journal of Sociology 70: 16-27. Ronald, and J. Y. Takeshita. "Studies of Fertility and Family Limitation in Taiwan." Eugenics Quarterly 12: 233-250. Ronald, and Lolangene Coombs. "Economic Considerations in Family Growth Decisions." Population Studies 20: 197-222. Ronald, and A. L. Adiakha. "Recent Fertility Declines in Hong Kong: The Role of the Changing Age Structure." Population Studies 22: 186-198. 198 Freedman, Ronald, and John Y. Takeshita. 1969 Family Planning in Taiwan. Princeton: Princeton University Press. Friedlander, Stanley, and Morris Silver. 1967 "A Quantitative Study of the Determinants of Fertility Behavior." Demography 4: 30-70. Goldberg, David. 1958 "Family Role Structure and Fertility." Pp. 595- 596 in Dissertation Abstracts 19 (3). 1959 "The Fertility of Two-generation Urbanites." Population Studies 17: 214-222. 1960a "Another Look at the Indianapolis Fertility Data." Milbank Memorial Fund Quarterly 38: 23- 36. 1960b "Some Recent Developments in American Fertility Research." Pp. 137-151 in National Bureau of Economic Research (ed.), Demographic and Economic Change in Developed Countries. Princeton: Princeton University Press. Gendell, Murray. 1967 "Fertility and Economic Development in Brazil." Demography 4: 143-157. Gomez, Miguel. 1968 Informe de la Encuesta de Fecundidad en el Area Metropolitana. San Jose, Costa Rica: Universidad de Costa Rica. 1970 "El Rapido Descenso de la Fecundidad en Costa Rica." Pp. 271-308 in Quinto Seminario Nacional de Demografia-Informe. San Jose, Costa Rica. Gordon, Robert. 1968 "Issues in Multiple Regression." American Journal of Sociology 73: 592-616. Hauser, Philip. 1960 "On Design for Experiment and Research in Fertility Control." In C. V. Kiser (ed.), Research in Family Planning. Princeton: Princeton University Press. 199 Hawley, A. H. 1955 "Rural Fertility in Central Luzon." American Journal of Sociology 20: 21-27. Heer, David. 1966 "Economic Development and Fertility." Demography 3: 423-444. 1971 "Educational Advance and Fertility Change." Pp. 1903-1914 in International Population Conference, London, 1968, Vol. III. Liege: The Inter national Union for the Scientific Study of Population. Heer, David, and Elsa Turner. 1965 "Areal Differences in Latin American Fertility." Population Studies 18: 279-292. Hill, Reuben, J. Mayone Stycos, and Kurt W. Back. 1959 The Family and Population Control. A Puerto Rican Experiment in Social Change. Chapel Hill: The University of North Carolina Press. Hutchinson, B. 1961 "Fertility, Social Mobility, and Urban Migration in Brazil." Population Studies 14i 182-189. Kahl, Joseph. 1967 "Modern Values and Fertility Ideals in Brazil and Mexico." Journal of Social Issues 23: 99-114. Kiser, C. V., and P. K. Whelpton. 1949 "Family Planning and Fertility of Socio- Economic Status." Milbank Memorial Fund Quarterly 27: 188-244. 1953 "Resume of the Indianapolis Study of Social and Psychological Factors Affecting Fertility." Population Studies 7: 95-110. Kiser, C. V., and P. K. Whelpton (eds.). 1949 Social arid Psychological Factors Affecting Fertility. New York: Milbank Memorial Fund, 5 vols. Kiser, C. V., Wilson H. Grabill, and Arthur A. Campbell. 1968 Trends and Variations in Fertility in the United States. Cambridge: Harvard University Press. 200 Kitagawa, Evelyn, and Philip M. Hauser. 1964 "Trends in Differential Fertility and Mortality in a Metropolis: Chicago." In E. Burgess and D. Bogue (eds.), Contributions to Urban Sociology. Chicago: University of Chicago Press. Kuntz, P. R. 1965 "The Relation of Income to Fertility." Journal of Marriage and Family Living, 27: 509-513. Labovitz, Sandford, and Robert Hagedorn. 1971 Introduction to Social Research. New York: McGraw-Hill. Lopez, Alvaro. 1967 "Some Notes on Fertility Problems in a Colombian Semi-urban Community." Demography 4: 453-463. Macisco, J. J., Jr. 1965 "Fertility in Puerto Rico: An Ecological Study." Sociological Analysis 26: 157-164. 1968 "Fertility of White Migrant Women, United States, 1960: A Stream Analysis." Rural Sociology 33: 474-479. Matheu, A. K. 1962 "Preliminary Lessons Learned from the Rural Population Control Study of Singapore." In C. V. Kiser (ed.), Research in Family Planning. Princeton: Princeton University Press. McNemar, Quinn. 1966 Psychological Statistics. New York: John Wiley and Sons. Miro, Carmen, and Ferdinand Rath. 1965 "Preliminary Findings of Comparative Fertility Surveys in Three Latin American Countries." Milbank Memorial Fund Quarterly 43: 36-62. Miro, Carmen, and Walter Mertens. 1968 "Influences Affecting Fertility in Urban and Rural Latin America." Milbank Memorial Fund Quarterly 46: 97-128. 201 Moberg, Sven. 1950 "Marital Status and Family Size among Matriculat ed Persons in Sweden." Population Studies 4: 115-127. Nag, M. 1967 "Family Type and Fertility.” Pp. 160-163 in Proceedings of the World Population Conference, Belgrade. New York: United Nations. Pakrasi, K., and C. Malaker. 1967 "The Relationship between Family Type and Fertility.” Milbank Memorial Fund Quarterly 45: 458-472. Palmore, James. 1969 "Malaysia: The West Malaysian Family Survey, 1966-1967." Studies in Family Planning 40: 11-20. Pelz, Donald, and Frank Andrews. 1964 "Detecting Causal Priorities in Panel Study Data." American Sociological Review 29: 836-848. Reynolds, Jack. 1972 "Costa Rica: Measuring the Demographic Impact of Family Planning Programs." Paper presented at the Population Association of American Meetings, Toronto, Canada. Rick, H. 1959 Fertility Patterns in Selected Areas in Egypt. Ph.D. Dissertation. Princeton University (unpublished). Ritchey, P. Neal, and C. Shannon Stokes. 1971 "Residence Background, Socioeconomic Status, and Fertility." Demography 8: 369-377. 1972 "Residence Background, Migration, and Fertility." Demography 9: 217-230. Saw Swee-Hock 1968 "Family Planning Knowledge, Attitudes, and Practice in Malaya." Demography 5: 702-709. 202 Sonquist, John, et al. 1964 The Detection of Interaction Effects. Ann Arbor: Institute for Social Research. Stoeckel, John, and Moqbul A. Choundhury. 1969 "Differential Fertility in a Rural Area of East Pakistan." Milbank Memorial Fund Quarterly 47: 189-198. Stycos, J. Mayone. 1963a "Social Class and Differential fertility in Peru." Pp. 123-128 in Proceedings of the Inter national Population Conference, London. Vol. II. 1963b "Culture and Differential Fertility in Perii." Population Studies 16: 257-270. 1968 Human Fertility in Latin America: Sociological Perspectives. Ithaca: Cornell University Press. Sun, T. H. 1969 Socio-Structural Analysis of Fertility Differentials in Taiwan. Ph.D. Dissertation. The University of Michigan (unpublished). Tabah, L., and R. Samuel. 1962 "Preliminary Findings of a Survey on Fertility and Attitudes toward Family Formation in Santiago, Chile." In C. V. Kiser (ed.), Research in Family Planning. Princeton: Princeton University Press. Takeshita, J. Y. 1962 Socioeconomic Correlates of Urban Fertility in Japan. Ph.D. Dissertation. The University of Michigan (unpublished). 1963 "Population Control in Japan: A Miracle or Secular Trend." Journal of Marriage and Family Living 25: 44-52. Theodorson, George.. 1967 "The Uses of Causation in Sociology." Pp. 131- 152 in Llewellyn Gross (ed.), Sociological Theory. New York: Harper and Row. United Nations. 1965 Population Bulletin 7. New York: United Nations. 203 Vazquez, Jose L. 1968 "Fertility Decline in Puerto Rico: Extent and Causes." Demography 5: 855-865. Weintraub, Robert. 1962 "The Birth Rate and Economic Development: An Empirical Study." Econometrics 40: 812-817. Westoff, C. F., et al. 1958 "Fertility Through Twenty Years of Marriage." American Sociological Review 23: 549-556. 1961 Family Growth in the Metropolitan America. Princeton: Princeton University Press. 1963 The Third Child. Princeton: Princeton University Press. Whelpton, P. K., et al. 1966 Fertility and Family Planning in the United States. Princeton: Princeton University Press. Yankey, D. 1961 Fertility Differences in a Modernizing Country. Princeton: Princeton University Press. Zarate, Alvan. 1967 "Differential Fertility in Monterrey, Mexico: Prelude to Transition?" Milbank Memorial Fund Quarterly 45: 93-108. Zikry, A. M. 1965 "Fertility Differentials of the United Arab Republic Women." Pp. 179 in United Nations, Proceedings of the World Population Conference. Vol. II. APPENDICES 204 205 APPENDIX A This appendix contains the questions from which the data for the present study were taken. The Rural Fertility Survey interview schedule was prepared by CELADE (Centro ✓ Latinoamericano de Demografia) for the use by all the countries which participate in the program of rural fertility surveys. First the questions are presented in its original language, Spanish; then, the writer's trans lation follows. Interview Schedule I. Ideales de Fecundidad 2. iCuantos hijos es bueno (es mejor) (esta bien) que una mujer tenga? (insista en que la entrevistada le diga un numero especifico. Si la mujer contesta, "Los que Dios manda," "Los que vengan,1 1 pregunte:) iCuantos hijos es bueno (es mejor) (esta bien) que Dios mande (que vengan)? II. Motivos para una Alta y una Baja Fecundidad III. Historia de Embarazos ^The complete interview schedule can be obtained from CELADE, Apartado 5249, San Jose, Costa Rica, C. A. IV. V. 206 14. ^Tiene Ud. hijos vivos? 15. iHa tenido hijos que nacieron vivos y murieron despues? 17. Muchas mujeres tienen embarazos que terminan en nacidos muertos o abortos. iHa tenido algun nacido muerto o aborto? (La entrevistadora prepara el resumen siguiente. Una vez terminado este resumen lealo en voz alta a la entrevistada). (a) Hijos nacidos vivos actualmente vivos. (b) Hijos nacidos vivos actualmente muertos. (c) Numero total de hijos nacidos vivos (a+b). Algunos Aspectos Explicativos de la Fecundidad Algunos Datos Personales de la Entrevistada 30. A. iCuando nacio usted? (su fecha de nacimiento) B. iCuantos anos tiene? (icual es su edad?) 32. Antes de tener 15 anos, £donde vivio la mayor parte del tiempo, en el campo (en una hacienda) (finca) (monte), en un pueblo, o en una ciudad? VI. VII. VIII. 207 33. Aparte de su lugar de nacimientotha vivido usted en otros lugares m£s de 6 meses? A. <LEn cuales? 35. tFue usted a la escuela? A. tQue ano (curso) (grado) (nivel) de escuela (de educacion) termino usted? Historia de Uniones 38. <LEs usted soltera, casada, conviviente, separada, viuda, divorciada (anulada)? (tCual es su estado conyugal (matrimonial) actual?) Conciencia Sobre la Posibilidad de Planificar la Familia y Actitudes Hacia Esta Posibilidad. 51. <j,Le parece a usted bien o no usar cosas (metodos) para no quedar embarazada (no encargar hijos) y tener los hijos solo cuando lo quiera? Conocimiento y Uso de Metodos de Planificacion Familiar 59. Usted y su marido (companero) than usado esta cosa (este (metodo) alguna vez? condon - retiro - esterelizaci<5n ✓ - ritmo - pildora - lavado interno - 208 anillo - supositorios - jalea - diafragma de la mujer. 61. iCada cu^nto tiempo tiene relaciones sexuales (se acuesta con su marido (esposo) (companero))? IX. Deseos de Tener Mas Hijos 66. iQuiere tener mas hijos o no, o le da lo mismo? A. iCuantos hijos mas quiere tener? X. Algunos Datos Personales del Marido (Companero) 73. tEtonde vivio/ su marido la mayor parte del tiempo antes de tener 15 anos, en el campo (una hacienda) (monte) (finca), en pueblo o en una ciudad? 74. iSiempre ha vivido su marido (esposo) (companero) aqui ? A. iSu marido (esposo) (companero) ha vivido alguna vez en una ciudad durante 6 o mas meses? 75. ^Que ano (curso) (grado) (nivel) de escuela (de educacion) termino su marido (esposo) (companero)? 76. A. iEn que trabaja su marido (esposo) (companero) la mayor parte del tiempo? 209 B. iQu^ es lo que el hace en ese (su) trabajo? H. iCu^nto gana en dinero su marido (esposo) (companero) por todos los trabajos que el hace (diariamente, semanalmente o mensualmente? 84. iViven con usted en este hogar (1) sus padres? (2) hermanos (as) no casados (as) o convivientes ? (3) hermanos (as) casados (as) o convivientes? Si es casada o conviviente, (4) los padres de su marido (companero)? (5) cunados (as) no casados (as) o convivientes? (6) cunados (as) casados (as) o convivientes? Informe de la Entrevistadora 2. Tipo de vivienda Vivienda muy pobre Vivienda pobre Vivienda modesta con algunas comodidades Vivienda comoda Vivienda de lujo Imposible de clasificar II. III. IV. 210 English Version Fertility Ideals 2. How many children is it good (is better) (is ok) for a woman to have? (insist in obtaining a specific number. If the interviewee answers, "As many as God will," "As many as they come," ask:) How many children is it good (better) (ok) that God will (they come)? Motives for a High and Low Fertility Pregnancy History 14. Do you have any children alive? 15. Did you ever have any children born alive and who died afterwards? 17. Many women have pregnancies which end in stillbirths or abortions. Have you ever had any stillbirths or abortions? (The interviewer prepares the following summary.' Once concluded, read it loud to to the interviewee). (a) Children born alive still alive. (b) Children born alive now dead. (c) Total number of children born alive (a+b). Some Explanatory Aspects of Fertility VI. VII. VIII. 211 Personal Information about the Interviewee 30. A. When were you born? (birthdate) B. How old are you? (which is your age?) 32. Before being 15, where did you live most of the time, in a farm, in a village, or in a town? 33. Besides the place where you were born, have you lived in other places for more than six months? 35. Did you go to school? A. Which year (grade) (level) of school (education) did you complete? History of Unions 38. Which is your current marital status? Are you single, married, living in common law union, separated, widow, divorced (annulled)? Awareness as to the Possibility of Planning the Family and Attitudes toward that Possibility 51. Do you think that it is all right to use things (methods) not to become pregnant and so to have children whenever one wants? Knowledge and Use of Methods of Family Planning 59. Have you and your husband ever used this thing (method)? condom - withdrawal - IX. X. 212 sterilization - rhythm - pill - vaginal wash - ring - suppository - cream, jelly - diaphragm - woman's sterilization. 61. C. How often do you have sexual inter course (go to bed with your husband or companion)? Desire for More Children 66. Do you want more children or not, or are you indifferent? A. How many more children do you want? Personal Information Regarding the Husband (Mate) 73. Where did he live most of the time before being 15, in a farm, village or town? 74. Has your husband always lived here? A. Has your husband (mate)ever lived in a city for six months or more? 75. Which grade (level) of school (education) did your husband (mate) complete? 76. A. Which is your husband's (mate's) occupation most of the time? B. What does he do in that occupation? H. How much does he earn for all the kinds of work he does? (daily, weekly, monthly). 84. Do they live with you in this house < 213 (1) your parents? (2) your brothers (sisters) not married or living in common law unions? (3) brothers (sisters) married or in common law unions? If she is married or living in common law law unions? (4) your husband's (mate's) parents? (5) brothers (sisters)-in-law not married or living in common law unions? (6) brothers (sisters)-in-law married or living in common law unions? Interviewer's Report 2. Type of housing very poor poor some comforts comfortable luxurious impossible to classify APPENDIX B This appendix includes some tables related to the contents of Chapters VI and VII suad not included in the text. Tables A-l and A-2 are related to Chapter VI and Tables B-l through B-8 to Chapter VII. 214 TABLE A-l WANTED FAMILY SIZE BY COUPLE'S RESIDENTIAL HISTORY AND HOUSING CONDITIONS AND OTHER PRESENT SOCIAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Wanted Family Size Current Couple' s Resident. H. Housing Conditions Characteristics 1 2 3 T 1 2 3 4 T Housina Conditions 1 Very Poor 7.0 6.7 6.6 6.8 2 Poor 6.8 6.7 6.6 6.7 3 Some Comforts 7.0 6.6 6.3 6.6 4 Comfortable (6.5) 6.0 6.0 6.2 Total 6.8 6.5 6.4 6.6 Husband's Income No work-100 Colones (6.7) 7.0 (6.7) 6.8 (6.5)7.4 * * 6.9 100-300 Colones 7.0 6.7 6.1 6.6 6.8 6.6 6.6 * 6.7 300-600 Colones 6.5 6.6 6.7 6.6 (6.9)6.6 6.7 (5.4) 6.5 600 Colones or More * 6.2 6.2 6.2 * 6.6 *6.4 5.9 6.3 Total 6.7 6.6 6.4 6.6 6.7 6.8 6.6 5.7 6.6 Husband’s Employment Status Tenant Farmers 6.9 6.6 6.4 6.6 6.8 6.7 6.5 6.1 6.5 Unskilled Labor * 6.0 6.4 6.2 * 6.4 (6.1)(6.4) 6.3 Farmers or Skilled Labor (6.5) 6.7 6.3 6.5 * 6.5 7.0 6.0 6.5 Total 6.7 6.4 6.4 6.5 * 6.5 6.5 6.2 6.4 * Base number less than 10 • ( ) Less than 20 but at least 10. Code for Couple's Residential History: 1 = No Moves; 2 = Some Moves, no larqe Towns; 3 = Some moves, some large Towns. 215 TABLE A-2 DESIRED FERTILITY BY HUSBAND’S INCOME AND EMPLOYMENT STATUS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Wanted Family Size Husband1s Husband1s Income Employment No Work-100 Status Colones 100-300 Colones 300-600 600 Colones Colones or More Total Tenant Farmers 7.1 Unskilled Labor * Farmers or Skilled Labor * 6.7 * (6.1) 6.6 6.2 (6.8) 6.2 (5.6) 6.5 6.7 5.9 6.5 Total * 6.5 6.5 6.1 6.4 * Base number less than 10. ( )Less than 20 but at least 10. 216 TABLE B-l NUMBER OF LIVE BIRTHS BY FREQUENCY OF INTERCOURSE AND EXTENDED FAMILY LIVING EXPERIENCE AND OTHER BIOPSYCHOLOGICAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Number of Live Births Frequency of Intercourse Extended Family Living Biopsychological Experience Characteristics 1 2 3 4 T No Yes T Extended Family Living Experience No 6.7 7.1 7.0 7.5 7.1 Yes * (6.3) 6.4 (6.1) 6.3 Total * 6.8 6.7 6.8 6.7 Attitude toward Family Planninq Against (6.0) 6.8 7.2 7.2 6.8 7.2 6.1 6.6 Favor (6.9) 7.4 6.8 7.5 7.1 7.1 6.9 7.0 Total (6.5) 7.1 7.0 7.3 6.9 7.1 6.4 CO • VO Use of Contraception No 6.3 7.1 6.9 7.4 6.9 7.2 6.4 6.8 Yes (6.9) 6.8 7.1 7.2 7.0 7.1 6.5 6.8 Total 6.6 6.9 7.0 7.3 7.0 7.1 6.4 6.8 * Base number less than 10. ( ) Less than 20 but at least 10. Code for Frequency of Intercourse: 1 = Less than 1 a Month; 2 = 1-3 a Month; 3 = 1-2 a Week; 4 = 3 or more a Week. 217 TABLE B-2 NUMBER OF LIVE BIRTHS BY ATTITUDE TOWARD FAMILY PLANNING AND USE OF CONTRACEPTION, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Use of Contraception Number of live Births Attitude Toward Family Planninq Against Favor Total No 7.1 7.2 7.2 Yes 7.0 7.0 7.0 Total 7.0 7.1 7.1 218 TABLE B-3 — WANTED FAMILY SIZE BY HUSBAND'S INCOME AND EMPLOYMENT STATUS AND BIO PSYCHOLOGICAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Wanted Family Size Biopsychological Characteristics Husband's Income H' s Empl. Status 1 2 3 4 T 1 2 3 T Frequencv of Intercourse Less than 1 a Month * * * * * 6.3 * * * 1-3 a Month * 7.0 6.5 6.2 6.6 6.7 * 6.9 6.8 1-3 a Week * 6.6 6.7 5.9 6.4 6.6 6.4 6.2 6.4 3 or more a Week * 6.7 6.5 6.5 6.6 6.8 * * * Total * 6.8 6.6 6.2 6.5 6.6 * 6.5 6.6 Extended Familv Livina Experience No 6.9 6.7 6.6 6.2 6.6 6.7 6.2 6.5 6.5 Yes * 6.1 * * * 6.5 * * * Total * 6.4 * * * 6.6 * * * * Base number less than 10. ( ) Less than 20 but at least 10. Code for Husband's Income: 1 = No Work-100 Colones; 2 = 100-300 Colones; 3 = 300-600 Colones; 4 - 600 Colones or More. Code for Husband's Employment Status: 1 = Tenant Farmers; 2 = Unskilled Labor; ________________________________________ 3 - Farmers or Skilled Labor. __ 219 TABLE B-4 — WANTED FAMILY SIZE BY COUPLE'S RESIDENTIAL HISTORY AND HOUSING CONDITIONS AND BIOPSYCHOLOGICAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Wanted Family Size Biopsychological Characteristics Residential History 1 2 3 T Housing Conditions 1 2 3 4 T Frequency of Intercourse Less than 1 a Month * * * * * * * * * 1-3 a Month 6.6 6.8 6.5 6.6 7.0 6.9 6.6 6.1 6.8 1-2 a Week 6.9 6.4 6.5 6.6 6.4 6.6 6.8 6.1 6.5 3 or More a Month 7.2 6.7 6.3 6.7 7.0 6.8 6.2 * • 6.7 Total 6.9 6.6 6.4 6.7 6.8 6.8 6.5 6.1 6.6 Extended Family Livinq Experience No 6.9 6.6 6.4 6.7 6.8 6.7 6.5 6.1 6.6 Yes * 6.5 6.4 6.5 * 6.5 * * * Total * 6.6 6.4 6.6 * 6.6 * * * *Base number less than 10. ( ) Less than 20 but at least 10. Code for Residential History: 1 = No Moves; 2 = Some Moves no large Town; 3 = Some Moves, Some large Towns. Code for Housing Conditions: 1 = Very Poor; 2 = Poor; 3 = Some Comforts; 4 = Comfortable. 220 TABLE B-5 -- WANTED FAMILY SIZE BY FREQUENCY OF INTERCOURSE AND EXTENDED FAMILY LIVING EXPERIENCE AND OTHER BIOPSYCHOLOGICAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Wanted Family Size Biopsychological Characteristics Frequency of Intercourse Extended Familv Livinq Experience 1 2 3 4 T 1 2 T Extended Family Livino Experience No 6.5 6.6 6.6 6.8 6.6 Yes * (7.9) 6.4 (5.9) 6.7 Total * 7.2 6.5 6.3 6.6 Attitude toward Familv Planninq Against (5.8) 7.0 7.0 7.1 6.7 7.0 6.6 6.8 Favor (6.9) 6.4 6.2 6.3 6.5 6.3 6.4 6.4 Total (6.4) 6.7 6.6 6.7 6.6 6.6 6.5 6.6 Use of Contraception No 6.2 6.8 6.5 6.8 6.6 6.7 6.6 6.6 Yes (6.6) 6.4 6.6 6.3 6.5 6.5 6.4 6.5 Total 6.4 6.6 6.6 6.6 6.6 6.6 6.5 6.6 * Base number less than 10. ( ) Less than 20 but at least 10. Code for Frequency of Intercourse: 1 : = Less than 1 a Month; 2 = 1-3 a Month; 3 = = 1-2 a week; 4 = 3 or More a Week. to to TABLE B-6 WANTED FAMILY SIZE BY ATTITUDE TOWARD FAMILY PLANNING AND USE OF CONTRACEPTION, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Wanted Family Size Attitude Toward Family Planninq Use of Contraception Against Favor Total No 6.9 6.3 6.6 Yes 7.0 6.3 6 . 6 Total s 6.9 6.3 6 . 6 f l 222 TABLE B-7 -- IDEAL FAMILY SIZE BY FREQUENCY OF INTERCOURSE AND EXTENDED FAMILY LIVING EXPERIENCE AND OTHER BIOPSYCHOLOGICAL CHARACTERISTICS, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Ideal Familv Size Extended Familv Biopsychological Frequencv of Intercourse Livinq Experience Characteristics 1 2 3 4 T No Yes Total No 5.2 5.1 5.1 5.4 5.2 Yes * (5.1) 5.1 (5.9) 5.4 Total * 5.1 5.1 5.6 5.3 Attitude toward Familv Plannina Against (6.2) 5.7 5.9 6.3 6.1 5.9 6.1 6.0 Favor (4.4) 4.5 4.8 4.9 4.7 4.7 5.3 5.0 Total (5.3) 5.1 5.3 5.6 5.3 5.3 5.7 5.5 Use of Contraception No 5.6 5.3 5.4 5.8 5.6 5.4 5.9 5.7 Yes (4.6) 4.8 4.8 4.7 4.7 4.8 5.5 5.1 Total 5.1 5.0 5.1 5.3 5.1 5.1 5.7 5.4 *Base number less than 10. ( ) Less than 20 but at least 10. Code for Frequency of Intercourse: 1 = Less than 1 a Month; 2 = 1-3 a Month; 3 = 1-3 a Week; 4 = 3 or More a Week. 223 TABLE B-8 IDEAL FAMILY SIZE BY ATTITUDE TOWARD FAMILY PLANNING AND USE OF CONTRACEPTION, RURAL COSTA RICAN COUPLES WITH WIVES 30-49 Ideal Familv Size Use of Contraception Attitude toward Familv Planninq Against Favor Total No 6.0 4.8 5.4 Yes 5,6 4.6 5.1 Total 5.8 4.7 5.3 224
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
The Construction And Empirical Test Of A Theory Based On Selected Variables In Small-Group Interaction
PDF
Referential Dissociation And Response To Stress
PDF
The Effects Of Social Security Programs On Fertility Levels
PDF
Determinants Of Intercounty Migration: California, 1970-1973
PDF
Economic Differentiation And Social Organization Of Standard Metropolitanareas In The United States: 1950
PDF
Status Consistency Among The Clergy
PDF
Familism, Suburbanization, And Residential Mobility In A Metropolis
PDF
Determinants Of Neonatal And Postneonatal Mortality: The Case Of Malawi, 1992
PDF
The Impact Of The Local Metropolitan Environment On The Patterning Of Social Contact
PDF
A Longitudinal Model Of Residence Change
PDF
Consensus Of Role Perceptions In A Welfare Planning Council
PDF
Prediction Of Overt Behaviors In Hospitalized Psychiatric Patients
PDF
Transformational Processing Of Sentences Containing Adjectival Modifiers
PDF
Employment Distribution, Income, And City Size: A Statistical Analysis
PDF
Passage Through The Expressive Student Subculture: The Initiation And Cessation Of Marijuana Use
PDF
Social Components Of Housing Cost In The Western Metropolis
PDF
Narrative And Lyric Originality In The Old French Versions Of "La Vie De Saint Eustache"
PDF
Ecology Of Negro Communities In Los Angeles County: 1940-1959
PDF
An Empirical Examination Of The Relationship Of Vertical Occupational Mobility And Horizontal Residential Mobility
PDF
Conflict And Generalized Conflict In Verbal And Motor Responses
Asset Metadata
Creator
Blanch, Jose-Maria
(author)
Core Title
Differential Fertility Behavior And Values In Rural And Semi-Urban Costa Rica
Degree
Doctor of Philosophy
Degree Program
Sociology
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
OAI-PMH Harvest,sociology, demography
Language
English
Contributor
Digitized by ProQuest
(provenance)
Advisor
Friedman, Judith J. (
committee chair
), Morgner, Aurelius (
committee member
), Van Arsdol, Maurice D., Jr. (
committee member
)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c18-780955
Unique identifier
UC11363439
Identifier
7307243.pdf (filename),usctheses-c18-780955 (legacy record id)
Legacy Identifier
7307243
Dmrecord
780955
Document Type
Dissertation
Rights
Blanch, Jose-Maria
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA
Tags
sociology, demography