University of Southern California Dissertations and Theses

An Economic Analysis Of Nurse Mobility Patterns

(USC Thesis Other)

An Economic Analysis Of Nurse Mobility Patterns

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content 71-7732 PAYNE, Richard Dee, 1937- AN ECONOMIC ANALYSIS OF NURSE MOBILITY PATTERNS. University of Southern California, Ph.D., 1970 Economics, general University Microfilms, Inc., Ann Arbor, Michigan THIS DISSERTATION HAS BEEN MICROFILMED EXACTLY AS RECEIVED AN ECONOMIC ANALYSIS OF NURSE MOBILITY PATTERNS by Richard Dee Payne 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 (Economics) August 1970 UNIVERSITY OF SOUTHERN CALIFORNIA TH E G RA D U A TE S C H O O L U N IV E R SIT Y PA RK L O S A N G E L E S , C A L IF O R N IA 9 0 0 0 7 This dissertation, written by RICHARD[..DEE ..PAYNE.............. under the direction of hXS... Dissertation C o m mittee, and a p p ro v e d by all its members, has been presented to and accepted by T h e G r a d u ate School, in partial fulfillment of require ments of the degree of D O C T O R O F P H I L O S O P H Y t Dean D a te A u g u s t 1.97.0. DISSERTATION COMMITTEE * r - v C h ^ 9 * / 7 ACKNOWLEDGEMENTS As is so often the case, this study could not have been carried out without the cooperation and help of several individuals and organizations. I am particu larly indebted to Professor Donald E. Yett, Director of the Human Resources Research Center for the research faci lities, data, and helpful suggestions which he made available to me. I also gratefully acknowledge the many very relevant suggestions and criticisms offered by my colleague Robert T. Deane. His advice was sought on innumerable occasions and he always stood ready to help. Financial support was provided through a grant from the United States Public Health Service, Division of Nursing. The present study is a part of a larger study in progress. Complete services were provided by the U.S.C. Computer Sciences Laboratory and the computer facility in the Graduate School of Business Administration. In this respect, I want to express heartful thanks to Bert Steece and Ron Schechtman for their expertise in computer programming. Theirs was the unlucky lot to inter pret, often from rather vague instructions, the infor mation to be extracted from the data. Somehow they were ii iii always able to determine what I wanted and to make it available. I would be amiss not to express my thanks to Professors E. Bryant Phillips, Spencer D. Pollard, Aurelius Morgner and Edward H. Barker for their guidance as mem bers of my committee. Professor Phillips who was the chairman deserves special thanks for the considerable effort he expended on my behalf. For her understanding and patience during our stay here at U.S.C. , I will ever appreciate my good wife Pat. TABLE OF CONTENTS Chapter Page ACKNOWLEDGEMENTS ........................... ii TABLE OF CONTENTS........................... iv LIST OF TABLES............................. vii LIST OF FIGURES............................. x I. INTRODUCTION ............................... 1 Statement of the Problem ................... 1 Importance of the Problem................... 13 Current Mobility Literature ................. 15 Definition of Terms and Concepts ...... 15 Organization of the Remainder of the Dissertation ............................. 29 II. LITERATURE REVIEW ........................... 32 Turnover — Voluntary versus Involuntary . . 33 Turnover — Turnover Rate.................. 39 Turnover — Relative Frequency ............ ^9 Turnover — Instability R a t e .............. 50 Turnover — Expectancy of Service .......... 51 Mobility.................................... 56 Voluntary versus Involuntary Mobility .... 57 Frictional Mobility ......................... 58 iv V Chapter Page Labor Force Mobility....................... 58 Intensity of Work Status Mobility .......... 6l Interfirm Mobility ......................... 6l Geographic Mobility ......................... 62 Industrial Mobility ......................... 68 Interoccupatlonal Mobility ................. 70 Mobility Flows ............................. 72 Personal Characteristics of Job Changers . . 74 Reasons for Job Change..................... 78 Other Aspects of Nurse Turnover ............. 86 III. STATEMENT OF RELATIONSHIPS TO BE INVESTIGATED AND DESIGN OF SURVEY .......... 90 Statement of Relationships to be Investigated in Chapters IV and V .......... 90 Survey Design and Data Description ........ 97 Problems Encountered ....................... 104 IV. INTERPRETATION OF TURNOVER DATA............... 107 Hospital Turnover Rates ..................... 107 Refined Turnover Rates ..................... 113 Other Turnover Measures ..................... 118 Terminations by Reason Groups ............... 124 Average Length of S t a y ........................142 Regression on A L O S ............................146 Vi Chapter Page Seasonal Turnover ........................... 150 Summary........................................ 153 V. INTERPRETATION OF MOBILITY DATA...............157 Mobility Flows ............................. 157 Mobility Types and Reason Groups .......... 163 Geographic Mobility by Reason and Demographic Characteristic ................. 175 Reason Distributions of Chapter IV and V Compared...................................... 198 Movement to Areas of Net Economic Advantage . 201 Summary........................................ 205 VI. SUMMARY AND CONCLUSIONS....................... 210 Turnover Findings ........................... 212 Mobility Findings ........................... 216 BIBLIOGRAPHY ...................................... 222 APPENDIX A .......................................... 231 APPENDIX B .......................................... 231 * APPENDIX C .......................................... 236 APPENDIX D .......................................... 242 APPENDIX E .......................................... 245 APPENDIX .......................................... 248 LIST OF TABLES Table Page 1. Mobility Flows................................ 30 2. Turnover Rates for Worker Groups .............. 45 3. Nurse Turnover Studies ........................ 47 4. Three Turnover Rates Compared ............... 52 5. Reason for Turnover in the Seven Nursing Studies...................................... 84 6. Hospital Turnover Rates ...................... 109 7. Predicted and Actual Signs of the Coeffi cients in the Hospital Regressions.............Ill 8. Refined Turnover Rates for Four Hospitals . . 114 9. Three Turnover Rates for the 18 Hospitals . . 119 10. Comparison of Expectancy of Service Rates (Nurses)..................................122 11. Percentage of Terminations by Reason Group . . 125 12. Reason Group by Race - Terminations...........129 13. Reason Group by Race - Terminations Percentage Distribution ..................... 130 14. Reason Group by Age - Terminations.............132 15. Reason Group by Age - Terminations Percentage Distribution ..................... 13^ 16. Reason Group by Marital Status - Terminations. 136 17. Reason Group by Marital Status - Termina tions - Percentage Distribution ............ 137 vii viii Table Page 18. Reason Group by Highest Educational Level - Terminations ................................. 139 19. Reason Group by Highest Educational Level - Terminations - PercentageDistribution .... 140 20. Average Length of Stay by Hospital..............143 21. Average Length of Stay by Race, Marital Status, Shift and Education (18 Hospitals).......... 1*14 22. Predicted and Actual Signs of the Coefficients in the Length of Stay Regressions.............148 23. Hires and Terminations by Month in the 18 Hospitals...................................... 151 24. Mobility Flows versus Reason (Hires) ......... 159 25. Eight Mobility Types versus Reason Groups . . 164 26. Column Percentages for Eight Mobility Types versus Reason Groups ......................... 166 27. Number of Geographic Moves by Race..............177 28. Percentage of Geographic Moves by Race .... 177 29. Number of Geographic Moves by Race and Reason G r o u p .......................................... 179 30. Percent of Geographic Moves by Race and Reason G r o u p .......................................... 180 31. Geographic Type Mobility by Reason Group With and Without Filipinos ....................... 182 32. Number of Geographic Moves by Marital Status . 183 33. Percentage of Geographic Moves by Marital Status.......................................... 183 34. Number of Geographic Moves by Marital Status and Reason Groups..............................185 35. Per Cent of Geographic Moves by Marital Status and Reason Groups..............................186 ix Table Page 36. Number of Geographic Moves by Age Group . . . 188 37. Per Cent of Geographic Moves by Age Group . . 188 38. Number of Geographic Moves by Age and Reason Group ................................. 190 39. Per Cent of Geographic Moves by Age and Reason Group ................................. 191 40. Number of Geographic Moves by Education . . . 194 41. Per Cent of Geographic Moves by Education . . 194 42. Number of Geographic Moves by Reason and Education...................................... 195 43. Per Cent of Geographic Moves by Reason and Education..................................196 44. Number and Per Cent of Terminations and Hires by Reason Groups..........................199 45. Count of Nurses by Financial Situation by Move Type.................................... 202 46. Percentage of Nurses by Financial Situa tion by Move T y p e ..............................203 LIST OF FIGURES Figure Page 1. Monopsony Model ............................. 7 2. Three Turnover Rates Compared ............... 120 x CHAPTER I INTRODUCTION Statement of the Problem In recent years concern has been expressed over the possibility that there exists a shortage of nurses. Such concern has been most prevalent within the hospital setting. As a consequence, considerable effort has been devoted to learning the possible causes of the nurse shortage(s). One of the most fully developed models of the nurse shortage has relied heavily upon the geographic mobility patterns of registered nurses (RNs). This monopsony-oligopsony model of hospital nurses forms the framework for this dissertation, and suggests several hypotheses which will be investigated. One of the pos sible shortcomings of the model, however, is the fact that data on the mobility of nurses has been extremely limited. One economist has gone so far that he suggested: ^"Donald E. Yett, "An Economic Analysis of the Hospital Nursing Shortage," (Unpublished Doctor of Philosophy Dissertation, The University of California, Berkeley, 1968), p. 521. 2 Data on the geographic mobility patterns of professional nurses and the factors associated with their mobility are non existent. These data are necessary to appraise the probable effects of programs which are designed to increase the output from nursing programs in states which ^ have relatively few professional nurses. While there are no nurse mobility studies that we have been able to find, there are a number of nurse turnover studies. Unfortunately, none of these studies were conducted by economists, and therefore they do not lend themselves well to the study of mobility. Never theless, high turnover rates are associated with high costs, and are therefore of considerable interest to hos pital administrators and other employers of nurses. Turnover statistics, however, shed little light upon the underlying motivational forces bringing about a change of job, and do not indicate the situation of the RN both before and after her job change. Such information is critical to the analysis of the nurse shortage and to any policy aimed at its correction. While turnover studies are of very limited use, the more important of those found in the literature will be reviewed briefly in Chapter II. 2Jeffrey H. Weiss, "Nursing Manpower Program Analysis" (a study for the Department of Health, Education and Welfare, Washington, D.C., 1967), p. 7. (Processed.) 3 One of the goals of this dissertation is to obtain data on the mobility patterns of registered nurses entering and leaving general hospitals in the Los Angeles area. Such data, it is hoped, will make it possible to determine whether or not there is a shortage of nurses in an economic sense, and, if such a shortage exists, might provide use ful insights pertaining to its correction.3 Labor Mobility and Wage Theory as Applied to Nursing. The primary conclusion of wage theory which is relevant to this study is that wages (and other differen tials) cause the reallocation of labor in a market which is competitive. Such reallocation would include the geographic movement of nurses and movement within and between the various areas of nursing such as hospital, industrial, private duty, public health, school nursing, etc. Theoretically, each individual nurse will seek employment which is advantageous to her, and avoid employ ment which is not. The end result of acting according to one's preferences would lead to an equality of jobs. Rottenberg points out: It is of primordial importance to understand ^It might be noted, that while our primary objec tive is to obtain mobility data, statistics on nurse turnover will also be forthcoming as a sub-set or by product of the mobility data. 4 that the early economists said that is was "the whole of the advantages and disadvan tages' * in all employments that would be equal ... They did not say wages are equal in all employments ... Occupations equal in other respects would tend to be equal in price, but occupations unequal in other respects would be unequal in price.^ These "other respects" referred to by Rottenberg in the case of nurses might include such considerations as hours worked, holidays, weekends, fringe benefits, etc. The mobility of workers (nurses), then, is the mechanism by which the conclusions of wage theory are fulfilled. To the extent that mobility is impeded, the conclusions of wage theory are frustrated. (See Appendix A for a brief discussion of state licensing procedures and unions.) The Monopsony-Oligopsony Model of the Nursing Pro fession. Contrary to the rather widely held view that the market for RNs is competitive and also national in scope, the Monopsony-Oligopsony model suggests that the market is, in fact, segmented into imperfectly competitive local markets. These local markets can be divided into the hospital and non-hospital sectors. The hospital sector ^Simon Rottenberg, "On Choice in Labor Markets," Industrial and Labor Relations Review. IX (January, 1956), p. lbl». 5 accounts for approximately 70 per cent of the active RNs. Roughly two-thirds of the remaining RNs find their wages tied quite closely to those of nurses employed by hospi tals. The other one-third (or roughly 10 per cent of total RNs) find their wages to be somewhat higher than those of hospital nurses, but, due to their relatively few numbers, they exert only minor influence on the 5 general level of nurse salaries. This latter point is significant, because it allows us to concentrate on the hospital sector without introducing such a great bias. The model suggests that in many instances (par ticularly in rural areas) there may be only one hospital within a given geographical area, in which case, such a hospital would be the only employer of registered nurses, and hence, would behave as a monopsonist. In the more urban regions, where there might be three or four hospitals in a given area, the model proposes that such hospitals, in effect, would act as though they were oligopsonists. In metropolitan areas with many hospitals, some kind of oligopsonistic or cartel arrangements would prevail. The significant fact of these three arrangements is that, ^Donald E. Yett, "Yes, Virginia, there is a short age of nurses — but it's not as simple as all that ..." (paper presented at the Second Conference on the Economics of Health, Baltimore, Maryland, December 5-7, 1968), pp. 31-3^. in each case, the hospital (or hospitals) is confronted by the market supply curve of nurses which is upward- sloping. Given such a supply curve, the marginal cost curve is also upward-sloping and lies above the supply curve. The significance of this fact is that imperfectly competitive employers of nurses attempting to maximize profits would usually be expected to report vacancies at the prevailing wage. These vacancies would indicate the existence of a nurse "shortage" in economic welfare theory terms in the sense that fewer nurses are employed in hospitals than would be the case if their wage was equal to the value of their services. This kind of shortage could be expected to persist indefinitely since it repre sents an equilibrium market position and, therefore, no endogenous adjustment process would be generated to "cor rect" it. The above ideas are illustrated below for the case of a monopsonist employer of nurses, the other cases, i.e., oligopsony and collusion, are similar. In the fol lowing diagram we see that the employer finds it profitable to employ quantity OA and pay them a wage OW which is below that which would have prevailed in a competitive market. He reports vacancies of quantity AB which quan tity represents the additional number of RNs he would have been willing to employ at the wage OW. 7 MC X w nurses B A 0 Figure 1 Monopsony Model From the above discussion and earlier sections of this dissertation, at least three different concepts of nurse shortage can be distinguished. The first of these which was mentioned above, is a shortage due to the imper fection of the market. This might be termed the "monop sony shortage." The second concept, and the one most commonly referred to by economists, is an "economic" or "dynamic shortage" caused by a situation where the quan tity demanded of labor is greater than the quantity sup plied at some prevailing wage. This shortage is due to a disequilibrium situation in the market, and can be corrected from within the market. The third concept is that used by most nurse employers. Such a shortage is based on a comparison of the number of "active" nurses relative to the number "required" to provide some desired quality or level of patient care. This concept might well be termed a "need" shortage. Unlike the former two con cepts, the quantity of labor desired is not inversely related to the wage paid. Of the three concepts, only the first two have relevance in strict economic discus sions. Further, from a standpoint of policy, it becomes vitally important to distinguish between these three kinds of shortages. If it could be demonstrated that the mar ket for nurses is in equilibrium, then all observed vacan cies will represent an economic or dynamic shortage. While these two kinds of shortages can easily be separated conceptually, serious problems are confronted when we attempt to do so empirically. There are at least three elements which are essen tial to the validity of the Monopsony-Oligopsony model. Each of these was mentioned either directly or indirectly in the discussion above. The first concerns the assump tion that hospitals attempt to maximize profits. In view of the fact that many hospitals are operated on a non profit basis, some doubt is cast as to how accurately the monopsony-oligopsony model describes the nursing market. However, such doubts can be mitigated if it can be shown that hospital administrators "behave" as though they were maximizing profits. For example, if they seek to maximize output (rather than profits) subject to the constraint that they must break even, then they, in effect, maximize profits. Such a policy would be accomplished by utilizing the factors of production in the least-cost combination for whatever output is produced, and this in turn could be accomplished by combining inputs in such proportions that the ratios of their marginal products to their mar ginal costs are equal. Under these conditions, "any hos pital which is a monopsonist or oligopsonist in the nurse market will hire fewer RNs than it would have as a com petitive employer. Such a hospital will substitute other factors (i.e., Licensed \ocational Nurses (LVNs), Aides, etc.) for nurses while, at the same time, reporting un filled nursing positions at the going salary scale. Such does appear to be the case, and therefore, on these grounds at least, the monopsony-oligopsony hypothesis of the nurse market seems to be acceptable. Further, it might be noted that a definite trend toward corporate ownership of hospitals is becoming apparent. Clearly, corporations will attempt to maximize profits. ^Yett, op. clt. , "An Economic Analysis of the Hospital Nursing Shortage," pp. 69-70. 10 Even more crucial to the acceptance of the monop sony-oligopsony hypothesis are the following conditions: (1) that there be few employers of nurses, and (2) that there be little wage induced geographic mobility of nurses. There is considerable evidence suggesting that in many areas, there are few employers of nurses. For example, it is noted that over 10 per cent of the hospi tals surveyed in a Hill-Burton survey were the only hos pitals in their service area. Approximately 30 per cent were located in areas with one or two hospitals, ^5 per cent were in areas with less than four hospitals, and over 60 per cent were in areas with less than six hospi- 7 tals. Even in many of the larger cities where there are several dozen hospitals, the market for nurses is far less competitive than one would imagine. In a survey of the 31 largest metropolitan hospital associations, all but one of the 15 that replied reported they had been successful in establishing and operating a "wage-standar- O dization" program among the hospitals in the association. Such collusive action on the part of hospitals in a large metropolitan area creates, in effect, a situation some what similar to that of a single employer of nurses in a 7Ibid., p. 57 and no. 139. ®Ibid., p. 57 and no. 1^0. 11 rural region. As to the second condition, that nurse geographic mobility be low, Professor Yett has suggested some circum stantial evidence which seems to support this hypothesis. He contends that nurses are generally secondary wage earners and, as such, are not in a position to respond to better job possibilities in other geographical areas. This implies that although married nurses may move, they do so in response to their family plans rather than in response to differences in wage or working conditions, and hence, are not mobile in the sense that such movement could remove the effects of monopsony or oligopsony. Single RNs who change Jobs often do so in response to factors other than those of job betterment. Such moves can come about as a result of travel plans, or they may be due to the RN's desire to join family and friends. In either case, although a move was made, it was not made in response to wage or working condition differences, and, therefore, need not negate the effects of monopsony and oligopsony. The hypothesis that the geographic mobility of nurses or the inter-area movement of nurses is "low," implies that the relevant nursing market is the local 12 g rather than the national market. If the mobility of nurses were sufficiently high, then there would be, in effect, only one nurse market, the national market, and the monopsony-oligopsony theory would break down. If it could be shown that the relevant nurse market is local rather than national in scope, we would expect to find some action taken by employers to reduce intra-area mobility. There is some evidence which suggests that this is actually the case. This writer observed that several of the hospitals in this area have tacit anti-pirating agreements which discourage movement within the area. In addition to this, many hospitals have wage policies which penalize an RN who changes jobs in the local market. This may occur by dropping a newly employed RN back one or more pay steps at the time of job change. This can also occur when hospitals fail to give pay credit for exper ience and training. The above ideas can be summarized by simply stating that the monopsony-oligopsony power of hospitals can be maintained when the geographic mobility of nurses is low and, by exerting this monopsony power, hospitals are also better able to reduce intra-city or intra-area movement. ^An obvious problem here is to determine what is "low,” or low relative to what. It is hoped that the literature can provide us bench marks with which compari son can be made. 13 In order that the monopsony-oligopsony hypothesis be relevant to the nursing profession then, it is critical that the wage or wage-related geographic mobility of nurses be low. If mobility does exist (and it need not be perfect mobility), nurses would migrate to areas where their productivity (wages) would be higher, or at least to areas where the working conditions were more favorable. Such migration would force monopsonistic or oligopsonistic employers to raise their wages and/or improve their working conditions. In other words, monopsony and oligopsony could not exist when nurses are mobile. Of the three requisites for the validity of the monopsony-oligopsony model (namely, (1) profit maximizing behavior, (2) few employers of nurses, and (3) low geographic mobility of nurses), the latter point is the most crucial, and ‘ unfor tunately, is the least well-documented in Professor Yett's model. Importance of the Problem Several Important implications can be drawn from the concept of labor mobility. "If labor mobility is ’functional' in the sense of adapting the labor supply to changing labor requirements, individual workers must tend to choose employment where the need for them is 14 greatest."1^ Such a movement of labor (nurses) assures a more efficient allocation of the labor supply which is clearly advantageous from the social point of view. This movement has still further advantages, when observed from the point of view of the individual worker (nurse). By adjusting according to her advantage, the nurse’s situa tion is also improved.1" ' ' By actually noting the movement of nurses (this includes both their pre- and post-move status) and relating this to the reason for job change, it is hoped that the degree of mobility, and, hence, the nature of the nurse market will be clarified. As noted above, such informa tion would be of considerable value to hospital adminis trators, leaders in the nursing profession and anyone responsible for the enactment of policy affecting nurses. A primary objective of this dissertation is to investigate the mobility patterns of registered nurses, i.e., the extent of movement and also the type of move ment. (Possible types of movement are discussed in the section on terms and concepts which follows.) Special 1(^Herbert S. Parnes, Research on Labor Mobility, An Appraisal of Research Findings in the United States (New York: Social Sciences Research Council, 1954), p. 9. •^Mobility of nurses will not necessarily be held as a desirable objective by everyone; as noted above, em ployers of nurses will likely view mobility as costly turnover. 15 consideration will be given to the hypothesis that the market for hospital nurses is in fact characterized by imperfect competition on the buyer's side of the market. Other objectives include the following: (1) to investi gate the motivation behind the job changes of hospital nurses; (2) to describe some of the personal characteris tics of the job changes; and (3) to look at some institu tional factors which affect the turnover of nurses. Current Mobility Literature The literature dealing with the mobility of labor is extensive and diverse. For this reason it will be treated separately in Chapter II. Definition of Terms and Concepts This section contains a description of the terms and concepts to be used in Chapters IV and V. The dis tinction between voluntary, involuntary and "worker not controllable" job changes is made. Then the differences between turnover and mobility are pointed out. This is followed by a discussion of the various turnover and mobility concepts. Voluntary versus Involuntary Job Changes. In some cases it is useful to differentiate between those workers who change jobs voluntarily, and those who are discharged. This is particularly helpful when one of the objectives is to study the motivation behind job changes. Clearly, the person who is discharged is not motivated to make such a shift, and it would be erroneous to include such among statistics used to analyze worker motivations. On the other hand, a study whose primary interest is the ability to change jobs would have to include those workers who were asked to leave. The ability to change jobs is evidenced just as much by involuntary as by voluntary movement. In this dissertation, differentiation will be made between voluntary and involuntary turnover and mobility. Involuntary job changes are defined the same as they gen erally are in the literature, i.e., "fired" and "layoff." For many purposes, however, we will expand the involuntary category to include all job changes over which the nurse did not have complete control. This expanded category we will call "worker not controllable." This will include such reasons as pregnancy, husband transferred, sickness, etc. It is felt that this broadening of the involuntary category has special relevance to a study dealing with women, many of whom are secondary wage earners. 17 Turnover and Mobility Differentiated. Turnover has been defined as the change in personnel that takes place during a given period in the work force of a firm or industry. It is generally measured as a percentage: the number of workers separated (or hired) per 100 12 workers on the average payroll during the period covered. Turnover statistics do not take into consideration the status of the worker both before and after the job change. That is, a worker may go from one job to another, to the ranks of the unemployed, to another industry or occupation, etc., and these facts are simply not noted in turnover statistics. Parnes, citing W. S. Woytinsky, suggests: ... labor mobility and labor turnover are distinguishable not on the basis of the kinds of job shifts that are involved, but on the basis of the viewpoint from which these shifts are seen. Analysis of mobility requires a comparison of the worker's status before and after a labor market transaction — for example, a change of jobs, a move Into or out of employment, or into or out of the labor force. Turn over, on the other hand, is based on a count of separations or accessions. 3 While turnover statistics can be derived from mobility data, the converse is not the case. Although the primary objective of this study is to obtain and investigate l2Parnes, op. clt., p. 23. 13Ibid., pp. 23-4. 18 mobility data, because of the above, turnover statistics will also be forthcoming. Turnover. In their article "New Ways to Measure Personnel Turnover in Hospitals," Levine and Wright dis- 14 cuss several different turnover measures. The first measure, which they simply call the "turnover rate," is the measure most commonly used. It is the percentage of the average number of persons employed terminating their employment during the year. It can be found by adding the number of RNs on staff at the beginning of the year to the number on staff at the end of the year and dividing by two. This number is then divided into the number of terminations during the year. A turnover rate of 100 per cent, for example, means that the number of resignations during the year equaled the average number of persons employed by the hospital during the year. The turnover rate, it is felt, is adequate for many descriptive purposes, but when turnover rates are to be compared (as between two or more hospitals) and the statistical significance of the differences in rates is to be evaluated, then, in many instances, the turnover rate ^Eugene Levine and Stuart Wright, "New Ways to Measure Personnel Turnover in Hospitals," Hospitals, Journal of the American Hospital Association, XXXI (August, 1957), pp. 3H-V2. ---- can be misleading. This may occur due to random sampling fluctuations. Levine and Wright suggest an alternative measure which they call "the relative frequency of turn over. " This statistic is computed by dividing the number of terminations during the year by the total (as opposed to the average) number of persons on staff during the year (that is to say, by the average number employed during the year plus the number who terminated during the year). The relative frequency of turnover measure will be computed in those instances where comparisons of turnover rates are made. A possible shortcoming of the turnover rate and also the relative frequency of turnover is that neither takes into consideration any demographic characteristics of the hospitals and nurses concerned. For example, it might be expected that a hospital comprised mostly of single or young RNs would have a higher rate of turnover than that of a hospital where many are older and married. The high turnover rate in the former hospital would prob ably be due, in large part, to such things as marriage, pregnancy, etc., rather than job dissatisfaction. Conclu sions drawn from comparisons of turnover rates, where such factors are not taken into consideration, could be greatly in error. Levine and Wright suggest several ways to redefine these "crude" turnover rates, one of which is to 20 separate nurses Into similar categories for purposes of comparing turnover rates. These categories might include positions such as supervisors, head nurses, staff nurses, etc. Likewise, the turnover rates in various service units such as medical, surgical, intensive care, etc., in different hospitals could be compared. Similar groupings for marital status and age could also be devised. Where single total turnover rates are desired, they can be calculated by assuming identical age, etc., distributions in the hospitals to be compared, and the turnover rates calculated on that basis. Levine and Wright term such rates "refined” turnover rates. A further shortcoming of "crude" rates (both turn over rates and relative frequency of turnover) is that they observe differences in tenure of employees. For ex ample, two hospitals may have the same turnover rates, but the one may have five of its ten positions filled twice each for a turnover rate of 100 per cent while a second hospital may have two of its ten positions filled five times each, also for a turnover rate of 100 per cent. While both hospitals have the same turnover rates, the staff of the second is the more stable of the two because eight of its ten positions remained filled during the entire year, while only five positions in the first were filled the whole year. Two measures which emphasize the 21 length of service and thereby overcome the above-mentioned drawback of turnover rates are the "instability rate” 15 and the "average length of employment." The instability rate is found by simply dividing the number of nurses who terminated during the year and who were on payroll at the beginning of the year by the average number of nurses employed during the year. In our example above, the first hospital has an instability rate of 50 per cent, while the second has an instability rate of 20 per cent. (It might be noted that while the second hospital has more overall stability, two of its positions are extremely unstable.) The second measure of length of service, the "average length of employment" is a useful measure in that it takes into consideration an employee's total length of service, and thus offers a "longer range view of the turn over process." Such a figure could be obtained by simply adding the total number of years worked by the staff prior to a given date, and dividing this total by the total num ber of RNs on staff. The last measure of turnover discussed in the article by Levine and Wright, and one which they felt might be the best measure of turnover behavior, is that of "expectancy of service." Such a statistic is rather 15Ibid., pp. 40-41. difficult to obtain, since it takes into consideration the total turnover history of the hospital. In order to cal culate the expectancy of service, one would have to ob serve a group of new employees from the time they entered service until they terminated. A table could be used to show how many of the original group were still employed at the end of consecutive years of service. A study would be completed at the time the last member of the original group terminated. It might take as many as 30 or 40 years for this to happen, which, of course, renders such a pro cedure almost useless. An alternative course would in volve an examination of the work histories of every RN who every worked at a given hospital and noting their length of service. From such an analysis the average per centage of personnel who remained on the payroll after various numbers of years of service could be calculated. These percentages could then be applied to a new group of employees just beginning work at a hospital. Such information would be useful to an administrator who was interested to know how long his employees might remain with him. Because of the difficulties in obtaining this information, it may not be possible to include this measure of turnover in the present study. In addition to those measures of turnover mentioned above, it might be useful in some instances to compare the 23 average length of service of the whole hospital staff to the average length of stay of those RNs who are voluntarily terminating. Such a comparison would be helpful in deter mining whether or not a small number of RNs are responsible for most of the turnover within a hospital. In the present study, all of the various turnover measures will be used, some on a limited or modified basis. Mobility. The concept of labor mobility upon first glance appears to be unambiguous and to simply denote the movement of labor. However, upon attempting to measure mobility empirically it becomes readily apparent that such is not the case. Probably one of the best discussions on the concept of mobility is that of Herbert S. Parnes. (Much of what follows relies heavily upon his analysis of mobility. Parnes suggests that, in an economic sense, there are at least three basic concepts of labor mobility: (1) the ability of workers to move from one Job to another or into or out of employment; (2) their willingness or propensity to make such moves when given the opportunity; and (3) their actual movement. The first has important implications for determining the maximum potential flexi bility of the labor force. It is, however, virtually ^Parnes, op. cit., pp. 13-23• 2M impossible to devise an operational measure of this con cept. Therefore, it is not surprising to find that studies to date have been little concerned with this approach. The second concept — the propensity to change jobs — plays a central role in the traditional theory of labor allocation and wage determination. It is this propensity vnhich is presumed to produce the equalization of the terms of employment for comparable jobs. The threat of losing employees to firms offering more attractive posi tions provides the impetus for competitive adjustments which, ceteris paribus, result in a tendency toward uniform employment conditions for similar workers. Unfortunately, as in the previous case, this definition has little opera tional significance. It is clearly impossible to present even a small sample of workers with an extremely wide range of alternative job offers, and to note their responses under varying conditions. Consequently, even those economists who advocate measuring mobility in terms of ability or willingness to move have had to use data on actual movement in order to study mobility patterns. Moreover, in addition to its relative ease of measurement, the latter definition is the most objective. Only by observing what nurses actually do, as opposed to what they say they might do, will we be able to describe the existing pattern of nurse mobility. The measure of 25 mobility used in this dissertation will be actual movement. In order to investigate the mobility patterns of nurses, it becomes useful to distinguish between several classifications or types of labor movement. In this study eight possible categories will be identified: (1) inter- firm movement, (2) occupational movement, (3) industrial movement, (4) geographical movement, (5) movement from employment to unemployment (frictional mobility) and vice versa, (6) movement into and out of the labor force, (7) vertical movement, and (8) movement from full-time to part-time and vice versa (intensity). It is apparent that none of these categories are mutually exclusive, and quite often a job change will involve several of these categories. Each of these mobility classifications, as they apply to nursing, will be defined below. Inter-firm movement consists of a nurse moving from one employer to another. Such a move might also in volve several of the other categories. Occupational mobility involves a move from the nursing profession into a non-nursing profession and the reverse. An example might be a nurse leaving hospital nursing to become a secretary at a bank. In addition to being an occupational change, this move also involves an inter-firm move and possibly other types as well. An industrial change con sists of a move from one area of nursing, such as hospital 26 nursing, into another area of nursing, such as private duty or public health nursing. Such a move could involve several of the other classifications. Geographical mobility has been described as "a situation in which a worker changes his residence so as to make himself available for jobs for which he previously would not have been available because of their distance 17 from his home." ' Such a definition, however, presents several problems, because "not every change of residence nor every change in location of work, is embraced by this concept of geographical mobility. For example, workers may change the state of their residence (to say nothing of their county) without changing their place of work, or take jobs in another state or county without changing 18 their residence." Parnes suggests that these problems can be largely eliminated by defining geographic mobility as being any move in excess of a specified number of miles.19 While this recommendation appears to be optimal, we elected to define a geographic move as a change in the location of the job across the Los Angeles Standard Metro politan Statistical Area (SMSA) boundary. This boundary 17Ibid., p. 33. l8Ibld., p. 3^. 19Ibld., p. 31 *. also coincides with the Los Angeles County boundary. The reasoning for such a decision was based on the fact that most other mobility studies had also used the SMSA defini tion. If any "bench marks" as to wage-induced geographic mobility are to be obtained, the probability is higher that they would come from studies with comparable defini tions. Even with this concession, however, there are still some differences between the present study and some of those in the literature. One of these differences is the fact that some studies used a change of residence as the unit of measure. In the case of movement into and out of employment (frictional mobility) it is necessary to specify some minimum period during which the worker was out of work and actively looking for a job. The period selected is arbi trary, but a period which seems to be reasonable is 30 days . This period will constitute the minimum period of this study. Any nurse, then, who has been out of work and who has been actively looking for a job for 30 days or more will be considered to be unemployed. Upon finding a job, such a person will be considered to have moved from unemployment to employment. The key to making this deter mination, in addition to the time period, is the fact that the person is looking for a job. RNs who take extended vacations of one or two months will not be considered 28 unemployed until they begin to look for a new Job. RNs who take extended leaves of absence but return to the same hospital also will not be considered unemployed. Movement into and out of the labor force has refer ence to those nurses who leave the nursing profession to become housewives and mothers, or those who return to nursing from the home. It should be noted, however, that those nurses who leave nursing to become secretaries, etc., are not considered to have left the labor force; rather, such a move would fall into the occupational mo bility category mentioned above. As in the case of fric tional mobility, it is necessary to specify some minimum time period. For purposes of this dissertation, an RN who has been at home for more than 90 days will be con sidered to be out of the labor force. Upon return to nursing, such an RN is considered to be moving from out side the labor force back into it. Vertical mobility is the movement from one Job title to another. This, of course, could take place within a single place of employ ment, as with an RN moving up from staff RN to head nurse or supervisor. It could also occur in several of the other classifications. We decided to include only those vertical moves which accompanied a change of employer. Vertical moves within a hospital were Ignored. The last mobility category might be called the intensity of work 29 (i.e., full-time or part-time) category. Here too it was decided to include only those RNs whose intensity move was accompanied by a change of employer. As was mentioned above, it is possible that more than one of these mobility classifications could occur at the same time. Such mobility combinations will be termed "mobility flows," and the "patterns" of nurse mobility will be identified in terms of these flows. In the table below, all the possible mobility flows are identified. Organization of the Remainder of the Dissertation The last task of the present chapter is to outline briefly the remaining chapters of this dissertation. Chapter II consists of a review of the relevant mobility and turnover literature, both nursing and non nursing. One of the things which becomes readily apparent in the discussion of the nursing literature is that most of the studies were done by people outside the economics profession. The implication of this fact is that although many of these studies are called mobility studies, they are, in fact, primarily concerned with the problem of turnover rather than mobility, as defined above. Another shortcoming of the existing nursing literature is its lack of uniformity of definitions and objectives. Nevertheless, it does contain a number of studies with which the find- 30 TABLE 1 MOBILITY FLOWS Abbreviations Used: 1. Interfirm INT 2. Interoccupational 0 3. Interindustrial IND 4. Labor Force L 5. Unemployment- Employment E 6. Geographical G 7. Vertical V 8. Intensity FTPT Mobility Flows: 1. INT 16. INT-IND-FTPT-G 2. INT-G 17. L 3. INT-V 18. L-G 4. INT-V-G 19. E 5. INT-FTPT 20. E-G 6. INT-FTPT-G 21. Not certain 7. INT-FTPT-V 22. Extra Job 8. INT-FTPT-V-G 23. INT-O-IND 9. INT-0 24. INT-O-IND-G 10. INT-O-G 25. INT-0-IND-FTPT 11. INT-O-FTPT 26. INT-0-IND-G-FTPT 12. INT-O-FTPT-G 27. New Grad-L 13. INT-IND 28. New Grad-L-G 14. INT-IND-G 15. INT-IND-FTPT ings of the present study can be compared. Chapter III outlines several of the relationships to be investigated in Chapters IV and V, and includes a description of the data and survey design. In Chapter IV the turnover data is analyzed, and in Chapter V the mobil ity data is investigated. This chapter forms the real core of this dissertation. In Chapter VI the significant findings are summarized and conclusions drawn. Sugges tions for possible future research are made. CHAPTER II LITERATURE REVIEW The purpose of this chapter is to examine briefly the findings of several turnover and mobility studies with data which are comparable to the present study. First we consider various turnover concepts according to the following categories: (1) all workers, (2) female workers, (3) professional workers, (*♦) female professional workers, and (5) nurses. Next, mobility is examined with respect to the same five subdivisions. The next section treats the personal characteristics of Job changers, followed by a section on reasons for turnover. The final section deals with other factors related to turnover. As will be recalled from Chapter I, the major distinction between turnover and mobility is the point of view from which the job change is considered. Turnover is simply a count of hires or terminations with no con sideration being given to the worker's past or subsequent labor force status. This is typically the view taken by the employer. Mobility, on the other hand, concerns itself not only with the Job change, but more importantly 32 33 with an analysis of the status of the worker both before and after the Job change. Mobility, then, is analysis of Job change from the employee's point of view. Turnover — Voluntary versus Involuntary From an economic point of view, it is necessary to separate those Job changes which are voluntary from those which are not. One way to accomplish this is simply to differentiate between "quits," which are initiated by the employee, from "fires and layoffs," which are employer initiated. One of the consistent findings in the litera ture is that quits and fires-layoffs fluctuate in opposite directions during the stages of the business cycle. During prosperous times, the proportion of quits rises and fires-layoffs fall; in bad times, the reverse. Much of the general turnover literature is couched in terms of quits and fires-layoffs. Another measure of voluntary-involuntary turnover with important implications is obtained by broadening the involuntary category to include any reason for Job change over which the worker has no control. This kind of turnover could be called "employee not controllable" turnover. Although not prevalent in the literature, this concept will receive considerable attention in the present study. 34 All workers. The U. S. Department of Labor breaks down turnover Into "quits,” "layoffs” and "other" (includes transfers) per 100 manufacturing employees. Fbr 1961, 1965 and 1968 the annual "layoff" rates were 26.4, 16.8 and 10.0 per cent respectively. The total turnover rates for these years were 48, 49.2 and 55.2 per cent. From these figures it is seen that "layoff" comprised about 55 per cent of all terminations in 1961, 34.1 per cent and 18.1 per cent in 1965 and 1968 respectively.^" It should be noted that the "layoff" rate in 1961 is very high, due to the recession conditions prevailing at that time. It appears from the above that "quits" and "trans fers" accounted for 45 to 80 per cent of all terminations in manufacturing during the period studied. Bancroft and Garfinkle estimated for 1961 that 37.5 per cent of the job shifts for men were due to "lost Job" with 10.9 per cent due to "termination of temporary Job" plus an unspe- 2 cified amount due to other involuntary reasons. From their figures, voluntary turnover appears to be between 45 and 55 per cent of all turnover. In 1967 the 1U. S. Department of Labor, Bureau of Labor Sta tistics, "Establishment Data Labor Turnover," Monthly Labor Review, XCIII (February, 1970), p. 97. p Gertrude Bancroft and Stuart Garfinkle, "Job Mobility in 1961," Monthly Labor Review, (August, 1963), p. 903. Administrative Management Society found that among office workers in the United States and Canada approximately 21 per cent of the male employees left their Jobs for reasons of ’ ’dismissal" and "staff reduction." In 1965, this figure was 17 per cent. In 1964 the percentage of all office workers was 15 per cent. Retirement and military 3 service were not included in the figures cited above. All females. In the studies by the Administrative Management Society, involuntary separations by women office workers were approximately 13 per cent in both 1965 and 1967. * * Prom the Bancroft and Garfinkle figures in 1961, it is difficult to specify accurately the per cent of voluntary separations due to a very large (32.2 per cent) "other" category. On the basis of a comparison of the percentage of those women who "lost" their job (20.6 per cent) with the percentage of men who lost their jobs (37.5 per cent), it might be concluded from the Bancroft and Garfinkle study that a greater proportion of women 3Administrative Management Society, "Survey: Turnover of Office Personnel," Administrative Management (June, 1965), pp. 55-58; Administrative Management Society "Administrative Management Society Reports Results of Turn over Survey," Administrative Management (July, 1966), pp. 43-44; Administrative Management Society, "Personnel Turnover: AMS’s Findings," Administrative Management (July, 1968), pp. 43-46. i i Ibid. left for voluntary reasons than men. 36 Professional Workers. We would expect, a priori, a greater proportion of professional workers to make volun tary job changes than other occupational groups. The available literature, while not as conclusive as one might want, did seem to support this hypothesis. Bancroft and Garfinkle noted that only 25 per cent of the professional workers (compared with 37.5 per cent for all men) left for reason "lost job." It must be noted, however, that "lost Job" does not include fired. They also found that 42.4 per cent of professional job changes were made for "improvement in status" reasons. Fbr all men this was 5 33.7 per cent. Ladinsky also found that a large per centage (67 per cent) of professional workers leave for "work related" reasons.^ Professional women. As in the case of profes sional men, Bancroft and Garfinkle found that a relatively small percentage of professional women (23.3 per cent) left for the reason of "lost Job." Unfortunately, as in the above case, the unspecified category "other" was Bancroft, op. cit., p. 903. ^Jack Ladinsky, "Sources of Geographic Mobility," Demography, IV, No. 1 (1967), p. 304. 37 7 quite large. While the statistics are sparse, it probably can be concluded that a greater percentage of women and pro fessionals make voluntary job changes than do men and other workers. Nurses. Considerable attention was given in the nursing literature to the question of reason for turnover. One of the problems, however, was that almost every study categorized the reasons for turnover by a different method. For example, the study by the American Hospital Association found that 95 per cent of all turnover was voluntary and Q only five per cent was involuntary. Involuntary included only "fired" and "asked to resign." Dodge, on the other hand, defined involuntary turnover as being essentially leaving for any reason over which the RN has no control. This is the concept "employee not controllable" turnover mentioned at the beginning of this section. This concept includes husband moving to new Job, pregnancy, sickness, etc. as involuntary reasons. Dodge estimated her "involuntary turnover" to comprise 60 per cent of all 7Bancroft, op. clt. Q American Hospital Association, Division of Re search, "Survey of Personnel Turnover in Volunteer Hospi tals" (an unpublished abstraction from final report, 1962) . 38 terminations.^ Several studies viewed turnover reasons from the hospital's point of view, i.e., whether turnover was "avoidable" or not. In a study of the Department of Health, Education and Welfare (HEW) it was estimated that 75-80 per cent of turnover was unavoidable.10 The United Hospital Fund of New York suggested that approximately 70 per cent of turnover is "unavoidable.1,11 Foley and Hough attempted to determine the extent of "controllable" nurse 12 turnover. Foley found that roughly 7^ per cent of turnover could not be controlled, while Hough found that only 55 per cent was not controllable. Catania, Smith and Saleh viewed the reasons for turnover as being essen tially "individual related" and "Job related." In each study the individual reasons accounted for 55 to 75 per 9Joan S. Dodge, "Why Nurses Leave— And What to do about it," Modern Hospital, XCIV (May, i960), p. 116. 10U.S. Department of Health, Education and Wel fare, Hospital Personnel; Report of a Personnel Research Project, (Washington, D.C.: Public Health Service Bulletin No. 930-C-9, 1964). 11United Hospital Fund of New York, Analyzing and Reducing Employee Turnover in Hospitals: A Special Study in Management Practices and Problems'! (New York: United hospital Fund of New York, 196b), p. 12. 12Margaret Foley, "Can We Minimize Staff Turnover," Hospital Progress, XXXIII (March, 1952), pp. 66-67; and L. B. Hough, "What Are the Reasons for Nursing Service Turnover?" Hospital Management, LXXIX (January, 1955), pp. ^3-46. 39 13 cent of the reasons. Although the nursing studies vary greatly in the manner in which they group turnover reasons, some conclu sions can nevertheless be drawn. It appears that nurses make a considerably larger proportion of voluntary separa tions than do other worker groups. This figure might be as high as 90 to 95 per cent. Of the five to ten per cent which might be considered involuntary reasons, "fired" and "layoff" probably account for only three to six per cent. When the involuntary category is expanded and be comes "employee not controllable," the proportion of voluntary separations probably falls to 40 to 60 per cent of all terminations. Reasons for turnover will be dis cussed later. Turnover — Turnover Rate Turnover rates have usually been defined as the number of terminations per period of time divided by the average number of workers on staff during that period. ^James j. Catania, "Why Do Nurses Change Jobs," Hospital Management, XCVIII (August, 1964), pp. 93-94; Shoukry D. Saleh, Robert J. Lee and Erich P. Prien, "Why Nurses Leave Their Jobs — an Analysis of Female Turnover," Personnel Administration, XXVIII (January-February, 1965), pp. 2$-28; and Phil M. Smith, Influence of Wage Rates on Nurse Mobility, (Chicago: Graduate School, March, 1962). I MO This definition is readily operational for a single firm, and somewhat less so for industry in general. It becomes very difficult to apply when all workers are to be con sidered. In spite of this, a few rough estimates of the labor turnover of all workers can be obtained. All workers. Bancroft and Garfinkle, using U.S. Bureau of the Census data, estimated that in 1961, slightly over eight million workers changed jobs at least once.11* In a study, the Organization for Economic Coopera tion and Development (OECD) estimated that 21.M per cent of all job changers changed jobs two times per year, 5.6 per cent three times per year and 2.6 per cent four or 15 more times per year. ^ When these figures are applied to those of Bancroft and Garfinkle, a very crude estimate of the turnover rate for all workers in 1961 becomes 17.2 per cent. The OECD study estimates that the turnover rate for all men in 1961 was roughly 18 per cent.^ They felt this figure, however, to be understated. Eldridge and Wolksteln, using data from the Continuous Work History Sample of the Bureau of Old-Age and Survivors Insurance, ■^Bancroft, op. cit. , p. 888. ^OECD, Wages and Labor Mobility, (Paris: OECD, 1965), PP. 5M-55. ^Ibld. y p. 56. 41 estimate that in 1952 about 30.8 per cent of all workers 17 employed in 1952 were employed by more than one employer. 1 Bunting et. al. found that less than 30 per cent of his sample in three southern states was mobile. They defined a worker to be mobile when he had two or more employers 1 f t during the year. The Administrative Management Society found the turnover rates for a nationwide sample of office workers to be 22 per cent, 20 per cent and 2** per cent for 1961*, 1965, and 1967 respectively.^ Considerably higher turn over rates are noted for all manufacturing workers. In 1961 the annual rate was 48 per cent; in 1965 and 1968 it was 49.2 and 55.2 per cent respectively.20 From the above discussion, it is apparent that it is almost impossible to estimate the turnover rate for all workers. The figures above indicate only that the range is quite broad, and that manufacturing workers generally Eldridge and I. Wolkstein, "Incidence of Employer Change," Industrial and Labor Relations Review, X (October, 1956), pp. 101-107. ■^Robert L. Bunting, Lowell D. Ashby and Peter A. Prosper, "Labor Mobility in Three Southern States," Industrial and Labor Relations Review, XIV (April, 1961), pp. 432-445. IQ ^Administrative Management Society, op. cit. 20 U.S. Department of Labor, op. cit. have quite high annual turnover rates. All Female Turnover. As in the case of all workers, it is very difficult to determine the turnover rate for all women. However, using the Bancroft-Garfinkle data and the OECD adjustment described above, the turnover for all women in 1961 was estimated to be 1*1.5 per cent 21 per year. The OECD study estimated the turnover rate P P for all women to be 13.6 per cent. In the studies of the Administrative Management Society the turnover rate for women office workers was estimated to be 28 per cent 2*3 in 1965 and 32 per cent in 1967. Turnover rates for all women in manufacturing were 61.5 per cent in 1962, p h 59.8 per cent in 1965 and 67.5 per cent in 1968. All professional workers. Only two studies were found which had data on the rate of job changing of professional workers. Using Bancroft and Garfinkle's data in a manner similar to that in the paragraph on all workers ^Bancroft, op. cit. , p. 898 and OECD, op. cit. 22OECD, op. cit., p. 56. 23 Administrative Management Society, op. cit. 2k United States, Bureau of Labor Statistics, Employ ment and Earnings Statistics for the United States 1909- 196^» (Washington, D.C.: Government Printing Office, ISbk), p. 38; Ibid., Employment and Earnings, XII (February, 1966), p. 102;~TbTd., XV (February, 1970), p. 153- 43 gives a rough estimate of 14.4 per cent for professional 25 men. An explanation of this is provided in the footnote below.Behrend, in an English study in 1947-1952, found that m£_le grammar school teachers had a turnover rate of 27 10.3 per cent per year. Female Professional Workers. An approximation to the turnover rate for professional women can be obtained from the Bancroft and Garfinkle study. They found that the approximately 8.3 per cent of the professional women had made at least one Job change (almost the same as for nQ professional men). ° Following the same procedure as outlined in footnote 26, a turnover rate of 14.0 per cent is obtained. In the Behrend study it was found that female grammar school teachers had a turnover rate of 29 15.7 per cent. ^ In her study of nursing turnover rates, “ ^Bancroft, pp. cit., p. 903. 2^We obtained an estimate of the number of profes sional men employed in 1961 (4,935,000), multiplied this by the per cent of professional men who changed Jobs (8.5). This figure was then adjusted for multiple Job changes, and finally this adjusted figure was divided by the total number of professional men in 1961. ^Hilde Behrend, "Normative Factors in the Supply of Labor," The Manchester School of Economics and Social Studies, (January, 1955), p. 66. 2®Bancroft, op. cit., p. 903. ^Behrend, op. cit. Nash stated that the turnover rate for school teachers was on 18 per cent per year. All of the turnover rates discussed to this point are summarized In the following table. From the few observations we are not able to determine conclusively which sex has the higher turnover rate. In manufacturing and office management, the rates for women are higher, but in the Bancroft and Garfinkle study (B&G) and the OECD study the rates for women are lower. It does appear that turnover rates for manufacturing are higher than those for other industries. Professional workers appear to have lower turnover rates than persons in other occu pations. We would expect this to be the case because they are less subject to layoff than other worker groups. Turnover Rate of Nurses. In contrast to the rela tively few studies concerned with the overall turnover rate categories discussed above, the literature reveals a rather sizable number of studies dealing with the turn over of hospital nurses. A major problem, however, is encountered when comparisons among the various studies are made. Most of the nurse turnover studies were not consistent as to the characteristics of the nurses to be ■^Mary K. Nash, ’ ’Turnover of Psychiatric Staff Nurses," Nursing Outlook, XIV (August, 1966), p. 29. TABLE 2 TURNOVER RATES FOR WORKER GROUPS (All figures in per cent) Worker Category 1952 1953 1961 1962 1964 1965 1967 1968 All Workers 30.8 Less than 30.0 18.0 (OECD) 17.2 (B&G ) All Manufacturing 48.0 49.2 55.2 All Office Workers 22.0 20.0 24.0 Women (All) 13.6 (OECD) 14.5 (B&G) Women Manufacturing (All) 61.5 59.8 67.5 Women Office 28.0 32.0 Professional 10.3 14.4 (B&G) Female Professional 15.7 14.0 (B&G) 18.0 Source: Previously cited works. included in their statistics. Some dealt only with full time nurses while others included various unspecified part- time categories. The studies were not consistent in their treatment of nurses on extended vacations, pregnancy, leaves, military leaves and transfers. Some included stu dent nurses, nurses on vacation relief, nuns, etc., while others did not. Probably still more serious than the many differences in the categories of nurses used in cal culating the turnover rate, is the fact that most studies simply did not specify which registered nurses were in cluded and which were not. In spite of these problems it is instructive to look at some of the turnover rates calculated by several studies. The results of 13 of these studies are shown in the table below. Table 3» which covers approximately 20 years, shows the wide range of turnover studies which have been conducted. Even disregarding the differences among the 13 studies, no perceptible trend in turnover rates can be noted. A simple average of the rates yields a turn over rate of approximately 55 per cent. The spread in the rates extends from a low of 26 per cent to a high of 112 per cent. Such rates probably indicate that the turnover of hospital nurses is considerably higher than those in many other firms and occupations, and about the TABLE 3 NURSE TURNOVER STUDIES Study Sample Size (Hospitals) Year Turnover Rates (per cent) Hough 2 General 1950-51 112.0 (Staff RNs only) Foley 48U Catholic Gen. 1951 32.0 (All RNs) Levine 51 General 1955 66.9 (Staff RNs FT) ANA 311 General 1955 46.5 (All RNs) Dodge 6 General 1957 26.5-52.3 (FT RNs) HEW 1 General 1959 65.5 (Nurses?) 1963 54.2 (Nurses?) AHA 226 Voluntary 1962 44.3 (All RNs) Nash 1 Psychiatric 1963 69.0 (Staff RNs) Saleh, et. al. 1 Teaching Gen. 1963-64 58.0 (RNs) Bain 23 Canadian Hospital 1965 41.7 (FT Nurses) 1967 38.6 (FT Nurses) ANA 181 General (200-299 Beds) 1967(August) 5.4 (64.8)a RNs (over 300) 1967(August) 4.7 (56.4)a RNs TABLE 3 (Continued) Study Sample Size Year Turnover Rates (Hospitals) (per cent) ANA 192 General (200+ Beds) 1968(April) 3.0 (36.0)a RNs ANA 167 General 1969(January) 3.9 (46.8)a RNs aAssumes that the monthly rate is typical for the entire year. Evidence suggests that this is not realistic. Source: American Hospital Association, Division of Research, "Survey of Personnel Turnover in Voluntary Hospitals," (an unpublished abstraction from final report, 1962) American Nursing Association, Research and Statistics Department, Facts about Nursing, A Statistical Summary 1955-1956, (New York: ANA, 1956), pp. 20-29; American Nursing Association, Research and Statistics Department, "Salary Ranges and Other Employment Conditions for RN Staff Nurses in Non-federal short-term General Hospitals," (unpub lished data supplied by ANA, August, 1967); American Nursing Association, Division of Research, "Salary Ranges for RN Staff Nurses in Non-federal short-term General Hospi tals," (unpublished data supplied by ANA, April, 1968); American Nursing Association, Research and Statistics Department, "Salary Ranges for RN Staff Nurses in Non-federal short-term General Hospitals," (unpublished data supplied by ANA, January, 1969); William Bain, "Turnover of Nursing and Paramedical Staff in 23 Ontario Hospitals," Canadian Hospital, XLVI (June, 1969), p. 38; Dodge, op. cit.; Foley, op. cit.; Depart- ment of health, Education and Welfare, op. cit.; Hough, op. cit. ; Eugene Levine, "Ana lyzing Turnover Among Hospital Personnel. Part II Turnover Among Nursing Personnel in General Hospitals," Hospitals, XXXI (September, 1957), pp. 50-53; Nash, op. cit.; and Saleh, op. cit. 1 4 9 same as workers in manufacturing industries. Turnover — Relative Frequency Relative frequency of turnover was defined in Chapter I as being the total number of terminations di vided by the sum of the average number of nurses employed and the number who terminated during the year. No study relative to all workers, female workers, professional workers or female professional workers could be found which calculated the measure of turnover. There were two nursing studies whose data was in a form which allowed ■31 calculation of this measure of turnover. Nurses. In the turnover study conducted by Wright in 1954-1955, the relative frequency of turnover was calculated to be 46.3, 45.0 and 51.4 per cent for the three hospitals of his study. The corresponding turnover rates were 86.3, 81.7 and 105.9 per cent respectively. Comparison of the turnover rates with the relative fre quency of turnover rates shows that the turnover rate spread is greatly reduced when the relative frequency of ^Stuart Wright, "Analyzing Turnover Among Hospi tal Personnel. Part III Turnover and Job Satisfaction," Hospitals, XXXI (October, 1957), pp. 47-52; and American Hospital Association, op. cit. 50 turnover rates are calculated. A statistical test to determine whether the turnover differences are due to random sampling fluctuations or to differences in the hos pitals now becomes more significant. The American Hospi tal Association study calculated the turnover rate to be M.3 per cent, while the relative frequency of turnover rate was only 29.6 per cent. It is seen from the above examples, and by definition, that the relative frequency of turnover will always be smaller than the turnover rate. Turnover — Instability Rate The instability rate was defined in Chapter I as the number of nurses who terminated during the year and who were on payroll at the beginning of the year, divided by the average number of nurses employed during the year. It has an advantage over the previously mentioned rates because it emphasizes the length of service of the nurses. Again, no general turnover studies were found which used this measure of turnover. Three nursing studies, however 32 did use this rate. Nurses. Wright found the instability rate in his study to be 50.6 per cent. As mentioned above, ^Wright, op. cit. ; Levine, op. cit. ; and Health, Education and Welfare, op. cit. 51 however, he did not separate the several nursing cate gories. The Instability rate in Levine's study was 48.5 per cent. The HEW study found the instability rate in the sample hospital to be 33.4 per cent in 1958 and 28.1 per 3 3 cent in 1959* Again it should be noted that these rates apply to most of the hospital employees and not only to registered nurses. The relationship of the three turn over rates discussed to this point is summarized in the following table. As would be expected, the instability rate is smaller than the simple turnover rate. Unfortunately, there is only one study which presented all three rates. High instability rates indicate that much of the hospi tal turnover is a result of employees who have been at the hospital for a year or more now leaving the hospital. Low rates indicate that much of the turnover is occurring among new employees. Turnover — Expectancy of Service The expectancy of service measure was described in Chapter I. It is felt that this measure is probably ^It should be noted that the rates provided in the HEW study are really "stability" rates rather than "instability" rates. The figures above were obtained by subtracting the stability rates from 100. TABLE 4 THREE TURNOVER RATES COMPARED Study Turnover Rate Relative Frequency of Turnover Instability Rate Wrighta 89.7? 47.3? 50.6? AHA 44.3? 29.6? — Levine 66.9? — 48.5? HEWb 61.8J5 45-9? 1958 1959 33.455 1958 28.1? 1959 aWright's data Is for entire nursing service. bHEW data is for all hospital employees. Source: Previously cited studies in Table 3* vn ro 53 the best measure of turnover behavior, but it is also the most difficult to obtain. Several studies approxi mated this measure of turnover by either calculating rela tive frequency distributions for the length of stay of those workers who were terminating or by calculating it for those who were continuing in their employment. All Workers. In 1966, Hamel found that of all male workers who were currently employed in his sample, 49.3 per cent had been at their job five years or less. Almost 15.8 per cent had been employed over five to ten years, and 35 per cent over ten years. The median number of years on the job was 5.3 years.31 * The Administrative Management Society (AMS) found in 1965 that at the time of termination, 40 per cent of all office workers had been on the job less than one year; 42 per cent had been there one year but less than five; and 18 per cent had been there five years or more. In 1967, AMS found that 52 per cent of those terminating had been on the job less than one year; 34 per cent, one year but less than five; and 35 14 per cent over five years. 3^Harvey R. Hamel, "Job Tenure of Workers, Janu ary, 1966," (Special Labor Force Report) Monthly Labor Review, XC (January, 1967), p. 35. o tr Administrative Management Society, 1966, op. cit.; and Administrative Management Society, 1968, op. cit. 54 Women Workers. Women who were currently employed In the Hamel study In 1966 were found to have been on the job a shorter period of time than were men. Those employed five years or less comprised 63.9 per cent of the sample; 15.3 per cent had been on the job over five to ten years, while 20.8 per cent had been on the job ten years or more. 36 The median length of time on the job was 2.8 years. Professional Workers. Behrend found that only 9 per cent of the male teachers in her sample stayed less than one year. The average length of service was 5.9 37 years. Hamel found that for those male professional workers who were still on the job, 49.2 per cent had been there five years or less, 18.6 per cent over five to ten years and 32.2 per cent over ten years. The median length o Q of time on the job was 5*3 years. It will be noted that these figures differ but little from those of all males above. Hamel suggests that because professional occupa tions are expanding so rapidly, increased opportunities probably induce professional workers to change jobs after a few years. Hamel, op. cit. , p. 35. ■^Behrend, op. cit., p. 67. Hamel, op. cit. 55 Women Professional Workers. Of those women teachers who were leaving a Job, Behrend found that 13 per cent left in less than one year, while the average 39 length of service was 6.7 years. Hamel found that some 60.1 per cent of the currently employed female professional workers had been on the job five years or less. About 15.2 per cent had been employed over five to ten, and 24.6 per cent had been employed over ten years. The median length 40 of stay was 3.5 years. Nurses. Catania set up a relative frequency dis tribution of the lengths of stay for his sample of termina tions. He observed that 43 per cent of the nurses resigned after having worker one year or less. Thirty-seven per cent resigned after one year but less than three years. 41 Eighteen per cent resigned after three or more years. Frances Elder found that 45 per cent of the terminations 42 in his sample had been employed one year of less. In the study by Saleh, et. al. it was found that 15 per cent of those terminating left in six months or less, 38 per ^Behrend, pp. cit. 40 Hamel, op. cit. 41 Catania, op. cit., p. 93* 4 ? Frances Elder, "Why Nurses Don't Stay Put," Registered Nurse, XXI (May, 1958), pp. 60-63. cent in one year or less, 26 per cent were employed longer than one year but less than two, 25 per cent between two and five years, and 10 per cent over five years.^3 Null found that in her sample, 58 per cent of the nurses re- iiij signed before having completed one year. Based on the above review, it appears that women in general have a shorter expectancy of service than male workers, and that nurses tend to have an even shorter period of service. Mobility In Chapter I it was concluded that while there are at least three ways to measure mobility in a theoretical sense (i.e., ability to change jobs, propensity to change Jobs and actual Job change), in parctice nearly all of the mobility studies have used either a count of job changes or a change of residence as their measure. Utiliz ing either of these measures of mobility, we can distin guish several different kinds or categories of mobility. The first task is to distinguish between voluntary and involuntary mobility. Others are as follows: (1) fric tional mobility or movement into and out of an employed ^Saleh, op. cit., p. 25. ^Virginia M. Null, "Facts Tell A Story," Nursing Outlook, III (August, 1955), pp. ^18-^22. status within the labor force; (2) movement into and out of the labor force; (3) movement between full-time and part-time work or intensity of work; (4) interfirm move ment; (5) industry movement or movement from one industry to another; (6) occupational movement or movement from one occupation to another; (7) geographical movement; and (8) combinations of these, called mobility flows. Studies which provide insight into each of the above mobility categories will be reviewed in terms of their application to all workers, female workers, professional workers, female professional workers and nurses. In the case of several of the above categories, little or no information was found. Voluntary versus Involuntary Mobility Where voluntary and involuntary mobility have been distinguished in the mobility literature, the definitions were the same as those discussed in the section on turn over. Due to the fact that most large studies utilized aggregate national data, they were unable to distinguish between voluntary and involuntary mobility, and, hence, little more than conjecture is presently available. As mentioned in the turnover section, to distinguish the kind of Job movement is of considerable interest to the present study. We will distinguish between voluntary and involuntary mobility but will go a step further to dis tinguish between reasons for Job change which are beyond the worker's control, i.e. "worker not controllable." This broadening of the involuntary category is particularly useful to a study of the mobility patterns of groups of workers, many of whom are secondary wage earners. Frictional Mobility No mobility study was found which considered (in an empirical way) the movement of workers into and out of unemployment. Due to the favorable job market for hospi tal nurses, we expect to find few, if any, unemployed nurses (i.e., unemployed in the sense that they have been out of work for 30 days or more, and have been actively looking for work). Labor Force Mobility The mobility literature again provided few statis tics on the movement of people in general into and out of the labor force. There were, however, several studies pertaining to the movement of women. Female Labor Force Movement. Rosenfeld and Parrella estimate that 1.2 million women entered, and 59 945,000 women left the labor force in 1963.^5 When con verted into percentages of the female civilian labor force these figures become 4.8 and 3.8 per cent respectively. As percentages of the total male and female civilian labor force, they become 1.7 per cent and 1.3 per cent respectively. Altman estimates that married women with drew from the labor force at a rate of 17 per cent per year from 1959 through 1962. For men, the rate was less 46 than 2 per cent per year. Several studies indicate that women are becoming an ever larger portion of the labor force, and, hence, their rate of entry overall must exceed the rate of exit from the labor force.^ Bowen and Finegan found that the age profile revealed by the labor force participation of married women roughly resembles an inverted "U," with a peak reached at about age 20-24, ^Carl Rosenfeld and Vera C. Perrella, "Why Women Start and Stop Working: A Study in Mobility," Monthly Labor Review, XXCVIII, No. 9 (September, 1965), pp. 1077-1082. 46 Stuart Harold Altman, "Factors Affecting the Un employment of Married Women: A Study of the Dynamics of the Labor Force Behavior of Secondary Family Workers," (unpublished Ph.D. dissertation, University of California at Los Angeles, 1964), p. 37. ^Glen Cain, Married Women in the Labor Force, (Chicago: University of Chicago Press, 19&6); William G. Bowen and Aldrich T. Finegan, The Economics of Labor Force Participation (New Jersey: Prindeton University Press, 1969), p. 897; Vera C. Perrella, "Women and the Labor Force," Monthly Labor Review, (February, 1968), pp. 1-12; and OECD, op. cit. 60 and maintained until age ^5—^9 when it begins to de- cline. McNally found that the labor force participation of women reaches two peaks, one before families are begun (at 20-24 years of age), after which the' participation declines. Then it turns up after the children are in school or grown and reaches the second peak at about 40-45 49 (or earlier). All of these studies show that the family cycle is very important in explaining the movement of women into and out of the labor force. Professional Women and Nurses. No statistics were found which actually measured the movement into and out of the labor force of all professional workers or profes sional women and nurses. However, some insight into these questions can be gained by examining the labor force participation rates of the various groups. Bowen and Finegan found that these rates differed greatly (as one would expect) with marital status and training. For all married women participation rate was 37 per cent. For married professional women it was 64.4 per cent. For all ^®Bowen, op. cit., pp. 108-113. ^ g . B. McNally, "Patterns of Female Labor Force Activity," Industrial Relations, (May, 1968), pp. 210-211; and M. S. Cohen, "Participation of Married Women in the Labor Force," Monthly Labor Review, XCII (October, 1969), pp. 31-35. 61 single women the rate was 86.8 per cent and for single SO women with education it was still higher. These figures suggest that the movement out of the labor force of pro fessional women is less than that of other women, and, hence, we would also expect the exit from the labor force of nurses to be less. Intensity of Work Status Mobility No statistics were found which indicated the num ber of workers moving between full-time and part-time status. Clearly the measurement of this kind of mobility would be very difficult, particularly moves which take place within a given firm. The present study will attempt only to measure a change in the intensity of work status which occurs at the time of a Job change. Interfirm Mobility Interfirm mobility is the most general of the re maining mobility categories to be discussed. A worker is said to have made an interfirm move when he begins work for a different employer. Such moves would include a change in his industry, his occupation, the geographical location of his work, and the intensity of his work ■^Bowen, op. cit., p. 88, 255 and 128. status. They would be either voluntary or involuntary. The only two mobility categories not included are the frictional and labor force groups. Most of the mobility studies in question were concerned with a specific kind of mobility (i.e., geographical, industrial, etc.), which makes it difficult to obtain statistics on the general measure of mobility (i.e., interfirm). Actually, one rough measure of interfirm mobility is simply the given turnover rate minus any frictional or labor force movement. In other words, the Individual worker must have gone from one employer to another. Geographic Mobility It will be recalled from Chapter I that for pur poses of this dissertation a person is said to have made a geographical move if he changed his place of employment across the Los Angeles Standard Metropolitan Statistical Area boundary (SMSA). This boundary is also the Los Angeles County line. Such a move need not, but usually will have been accompanied by a change of residence. This definition was chosen because it Is the definition most commonly used in the mobility literature. It was hoped that its use would facilitate comparison of the findings of the present study with those found in the literature. It was noted, however, that several studies were based upon residence change rather than job change across the SMSA boundary. If it can be assumed that most residence changes were also accompanied by job changes, then meaningful comparison with such studies can be made. Some of the more important geographic mobility studies are reviewed below. All Workers. Saben found that about seven per cent of the male labor force moved to a different county between March, 1962 and March, 1963. Approximately 3.8 per cent moved between states and 3.1 per cent between 51 counties of the same state. The United States Depart ment of Labor estimated that in any year approximately six per cent of the nation's population could be expected to 52 move across county lines. Marsh estimates that about five to 6.8 per cent of the population move to a different 53 labor market each year. Tarver found that between 1955 and I960, 17.M per cent and 16.7 per cent of the ■^Samuel Saben, "Geographic Mobility and Employ ment Status, March 1962 - March 1963," Monthly Labor Review, XXCVII (August, 1964), pp. 873-HFT; 52U. S. Department of Labor, Manpower Report of the President (Washington, D.C.: Government Printing Office, 1962), p. 143. 53 R. E. Marsh, "Geographical Labor Mobility in the United States; Recent Findings," Social Security Bulletin, XXX (March, 1967), p. 15. residences in 212 Standard Metropolitan Statistical Areas and 213 urban areas, respectively, moved across county c ii lines. On an annual basis this would be slightly over three per cent for each sample. Raimon estimated that approximately 3.5 per cent of the nation’s population move across state lines each year.-^ Over a two year period (1955-1957) Gegan and Thompson found that workers migrated out of a labor surplus area at the rate of nine per cent per year."^ To summarize the above statistics, it appears that geographical moves of all workers ranged between five and nine per cent per year. The percentage of interstate moves lies between three and five per cent per year. Female Workers. A study of the United States D. Tarver, "Metropolitan Area Intercounty Migration Rates: Reply," Industrial Labor Relations Review, XIX (January, 1§66) ,p. T7 ^^Robert L. Raimon, "Interstate Migration and Wage Theory," Review of Economics and Statistics, XLIV (Novem ber, 1962), p. 429• eg ■ ' Vincent F. Gegan and Samuel H. Thompson, "Worker Mobility in a Labor Surplus Area," Monthly Labor Review. VIII, pt. 2 (December, 1957), p. 1453. 65 Department of Commerce revealed that in 1963-1964 the interstate mobility of all labor force females was 2.9 per cent while that of men was 3*6 per cent. It should be noted that intercounty moves within the same state were not included in these figures. Gegan and Thompson found that over a two year period approximately 15.7 per cent 58 of the women workers moved out of the county. The yearly rate would be roughly 7.8 per cent. They also found that the migration rate for the male workers was greater than that for female workers. It must be remem bered, however, that the geographic mobility of married women workers will depend to a great extent on the mobility of their husbands. Looking only at "all women," it ap pears that from the available evidence women are less geo graphically mobile than men. Professionals. Saben found that in 1962-1963 professional and technical workers accounted for 19 per cent of the migrant men, while they comprised only 12 per cent of all employed men 18 to 64 years of age. Using these figures and others in his article, an estimate of 57u. s. Department of Commerce, Bureau of the Cen sus, "Mobility of the Population of the United States, March 1963 to March 1964," Current Population Reports: Population Characteristics, Series P-20, (September, T965). 5®Gegan, op. clt., p. 1456. 66 the migration rate of professional males becomes 9.1 per cent per y e a r . ^ 9 In a study of the geographic mobility of professional workers conducted by Ladinsky, it was found that male professionals had a migration rate of ten per cent per year between 1964 and 1965. Six per cent moved between states while four per cent moved between counties within the same state.Ladinsky found that professional males were more mobile geographically than any other occu pational group except farm laborers who also had a migra tion rate of ten per cent. Several other studies indi cated that professional workers are geographically more mobile than other occupational groups.^ Professional Women and Nurses. There was no infor mation available which specifically treated the geogra phic mobility rates of professional women. It does seem reasonable, however, on the basis of the discussion above, that professional women are more geographically mobile than other women. ^^Saben, op. clt. 6°Jack Ladinsky, "The Geographic Mobility of Pro fessional and Technical Manpower," Journal of Human Re sources , II, No. 4 (Fall, 1967), p. 4d0. 6*Tarver, op. clt., p. 219. Ladinsky broke down the category professional, technical and kindred workers into 33 specific profes sional occupations with no regard to sex. He then ranked these occupations by migration rates from highest (clergy men - 54 per cent) to lowest (funeral directors - 3 per cent). These rates, however, were for the five-year period 1955-1960. Nurses ranked 17th going from low to high rates with a rate of 24 per cent. Elementary and secondary teachers ranked 21st with a rate of 28 per cent 62 and librarians were 19th with a rate of 25 per cent. When these rates are converted to yearly rates, they be come close to five per cent per year. This rate is slightly above those for all women, and there is a high probability that these rates are understated due to the fact that (1) census data were used and (2) the period 63 of time was five years. While the studies cited above have indicated that professional and technical workers are more mobile geo graphically than other occupational groups, they also indicate that professional people are less mobilt or make fewer Job changes locally. Ladinsky for example found that they ranked seventh out of ten occupational groups ^Ladinsky, op. clt., p. 488. Ladinsky, op. clt., p. 481. 68 over a one year period with a rate of 13 per cent. Over the five year period 1955-1960 they ranked ninth out of ten.6" Industrial Mobility The concepts of interindustry mobility as used in the present study is fairly straight forward. The nursing profession is treated as the occupational group, while various areas of nursing such as hospital, public health, school nurse, etc., are treated as industries. Outside the nursing profession several definitional problems are encountered.^5 jn spite of these problems, some useful data is available. All Workers. In 1961 Bancroft and Garfinkle found that about 56 per cent of the male Job shifts involved a change in the industry. R>r 1955 they found the rate of industry shifts for males to be 66 per cent.^ Palmer, in her Six City study found that between 1940 and 1949 al most 75 per cent of the Job changes involved a change of ^Ladinsky, op. clt., p. 480. ^Laurence C. Hunter and Graham L. Reid, Urban Worker Mobility (Paris: Organization for Economic Cooper ation and Development, 1968), p. 63. ^Bancroft, op. clt., pp. 905-906. 69 67 industry also. Gallaway found that 56.3 per cent of male workers whose region of major job changed between 68 1957 and I960, also changed industry. It should be noted that an interregional move would exclude intercounty moves and many interstate moves (i.e., those not crossing one of the nine regional boundaries). All interregional moves then are geographic moves, but not all geographic moves are interregional. In a small study of a labor surplus area, Gegan and Thompson found that 67 per cent of the migrants changed their industry at the time they 6Q moved. ? From the above, it would seem reasonable to infer that the interindustry mobility rate might lie some where between 50 and 60 per cent of job changes. Female Workers. Only one of the studies provided interindustry mobility rates for all female workers. Bancroft and Garfinkle found the rate to be 55*5 per cent and 60.4 per cent for 1961 and 1955 respectively. These rates were, for all practical purposes, the same as those for males. ^ g. l . Palmer, Labor Mobility In Six Cities, 1940- 1949, (New York: Social Sciences Research Council, 1954). ^®Lowell E. Gallaway, "The Effect of Geographic Labor Mobility on Income: A Brief Comment," The Journal of Human Resources, IV (Winter, 1969), p. 105. ^Gegan, op. cit., p. 1454. 70 Professional Men and Women Workers. Once again the study by Bancroft and Garfinkle provides some useful statistics. They found that 52.5 per cent of the Job changes by professional men in 1961 involved a change of industry. For professional women this figure was 31.2 70 per cent. Here a very significant difference is noted between the rate for men and women. Evidently, in this study at least, professional women are very much less likely to change industry at the time of Job change than are men. Nurses. The review of the available nursing liter ature revealed no statistics on interindustry Job changes. Interoccupatlonal Mobility Interoccupational mobility may be defined as the rate at which workers change their occupation at the time of job change. As in the case of interindustry mobility, some definitional problems are encountered when this measure of mobility is applied. Hunter and Reid provide 71 an excellent discussion of the problems. In the present study, any nursing field will be considered the "nursing ^Bancroft, op. clt., p. 905. ^Hunter, op. clt., pp. 78-81. 71 occupation" and, therefore, any move into another kind of work will be considered an occupational change. All Workers (Men and Women). In the study by Bancroft and Garfinkle, it was found that in 1961 a worker's attachment to his occupation was somewhat more stable than to his industry. For all men it was found that the interoccupational mobility rate was 49 per cent. For women, this rate was 44 per cent. In 1955 the rates for men and women were 54.2 per cent and 49.6 per cent- respectively.^ For the period 1962-1963, Saben found that approximately 34 per cent of all workers who moved 7 3 geographically also changed their occupation.'J This last figure is considerably lower than that found by Bancroft and Garfinkle, and it may indicate greater occu pational attachment of workers who make geographical moves. It appears, from the studies cited, that women may tend to make fewer occupational moves than men. Professional Men and Women. All of the studies indicate that professional workers have considerably more occupational attachment than other groups of workers. Bancroft and Garfinkle found that professional men in ^Bancroft, op. clt., p. 906. ^Saben, op. clt., p. 880. 72 1961 had an interoccupational mobility rate of 35*2 per 74 cent while women had a rate of 24.8 per cent. Saben noted in 1962-1963 that only 18.7 per cent of all pro fessional workers who migrated during the year changed 75 their occupation. This result lends credence to his finding above that occupational loyalty is greater among the geographically mobile workers. This fact is also attested to by an OECD study. Nurses. Once again no information on the occupa tional mobility of nurses was available. Mobility Flows As noted in Chapter I, seldom will a job change take place without involving more than one of the mobility categories discussed above. Any combination of these mobility categories which describes a Job change will be termed a mobility flow. This concept has received little attention in the literature. Those few studies which refer to such flows will be cited below. All Workers (Male and Female). Palmer found that ^Bancroft, op. cit., p. 906. 75 '^Saben, op. clt. 76OECD, op. clt., p. 71. 73 about half of all job shifts made during 1940-1949 by workers in the six cities involved simultaneous changes 77 of employer, occupation and industry. Bancroft and Garfinkle found that 39 per cent of the job changes for men and 34 per cent for women involve both occupation 7 f t and industrial changes. The OECD study found that con siderably more than half of the industry changes are associated with a change of occupation. In other words, roughly 25 per cent of all job changes involve both 79 industrial and occupational shifts. Professional Workers — Male and Female. The study by Bancroft and Garfinkle was the only study which provided mobility flow data for professional workers. They found that 29.4 per cent of the job changes by pro fessional men involved both occupational and industrial changes. For professional women the rate was only 21 per 80 cent. As might be expected, these rates are lower than those for other male and female workers. Nurses. No information was available on the ^Palmer, op. clt., pp. 74-75. ^Bancroft, op. clt., p. 905. 79qecd, op. clt., p. 71. 8n Bancroft, op. clt., p. 905. 74 mobility flows of nurses. We might expect, however, due to the degree of specialization of their training, that the magnitude of the flow involving interindustry and interoccupational movement of nurses will be relatively low. Personal Characteristics of Job Changers Several studies have examined the characteristics of Job changers. In addition to those characteristics which have already been discussed above, (i.e., sex and occupational status), four other characteristics warrant some attention. These are age, race, marital status and educational level. Age — All Workers. One of the most consistently treated subjects in the studies dealing with the mobility of workers is the effect of age on mobility. Bunting et. al. found that over half of the in and out migrants from Q n three southern states were under 30. The studies by Raimon, Lansing and Mueller, Eldridge and Wolkstein, Tarver, and Marsh all found that mobility was greatest ^Bunting, op. clt., p. 432. 75 Op in the younger age categories. Raimon, Lansing and Mueller and Marsh’s studies concerned the geographical movement of workers. Gegan and Thompson found that workers under 25 migrated in substantially higher proportions than did older groups.®^ Bancroft and Garfinkle found that for both males and females, the highest rate of job chang- Ojj ing occurred between the ages of 18 and 24 years old. In his study of occupational mobility, Saben found that 60 per cent of those workers who changed occupation were 85 under 35 years of age. From the literature it seems reasonable to conclude that the mobility rates of workers reach a peak at about the age of 23-26 years, and then begin to decline in the late 20’s and early 30's. Several studies found that women reach their peak two or three years before men. Parnes suggests, however, that the relationship between mobility and age is complex, and that it is interlinked with other factors which are partly a function of age. ®2Raimon, op. clt., p. 428; John B. Lansing and Eva L. Mueller, The Geographic Mobility of Labor, Insti tute for Social Research, Survey Research Center, (Ann Arbor: University of Michigan, 1967), p. 421; Eldridge, op. clt.; Tarver, op. clt.; and Marsh, op. clt. ^Gegan, op. clt., p. 1453. Oji Bancroft, op. clt.» p. 898. ®^Samuel Saben, ’ ’Occupational Mobility of Employed Workers," The Monthly Labor Review, XC (June, 1967), p. 31. 76 For example, young workers just entering the labor force are possibly still looking for the right job. The marital status of young workers will likely be different than those of older workers. Young workers will not be much affected by such matters as home ownership, job seniority, etc. Marital and family status are probably of greatest importance when the mobility of female workers is con sidered. Labor force participation studies such as those of Gain, Bowen and Finegan have found that the participation of married women is very much affected by the family cycle, rising up to the birth of the first child, roughly main taining this peak, then reaching a second peak when the last child is in school or grown.It is clear that the family cycle has an important effect on at least one mo bility category, i.e., movement into and out of the labor force. Race. The effect of race upon the mobility of workers is uncertain due to the conflicting results of the studies which have considered it. Parnes, reviewing the literature up to the early 1950's found that most of the 86 H. S. Parnes, Research on Labor Mobility: An Appraisal of Research Findings in the United States, (New York: Social Science Research Council, ), pp. 105-109. ®?Cain, pp. clt.; Bowen, op. clt.; and McNally, op. cit. 77 studies showed that Negro males made more job shifts than white males. Negro females were slightly less mobile than po g o whites. More recently, Bunting found similar results. Both the Lansing-Mueller and Saben studies came to the opposite conclusion; they found that the geographic mobil- 90 ity of Negro males was lower than that of whites. Be cause of the high probability that other variables — such as educational level, occupational and industrial dif ferences, discrimination, etc. — are intermingled with race, it is extremely difficult to draw any definite con clusions as to the effect of race on mobility. Other findings of these studies are mentioned below. All of them found that family income has a negative effect on la bor force participation, and that wages have a positive effect. The level of education had a strong positive effect while the presence of children in the home had a negative one. Negro women had a higher participation rate than did white women. Professional women had higher par ticipation rates than other women. Women who had never been married had the highest participation of any female ^R. l. Bunting, "Labor Mobility: Sex, Race and Age," Review of Economics and Statistics, XLII (May, I960), p. 231. ^°Lansing, op. cit.; and Saben, op. clt. 78 group. While the factors cited above are not directly related to mobility rates (except in the case of entry into and exit out of the labor force) they nevertheless provide useful information which can serve as a point of reference in a study of the mobility of women. Reasons for Job Change In this section, studies which have attempted to analyze the reasons for job changes will be reviewed: first, as they pertain to all workers, and then as they pertain to the nursing profession. It should be noted that relatively few non-nursing studies have dealt with the subject of reason for job change. All Workers. In five studies prior to 1951 *, Parnes, although aware of the difficulties involved, attempted to summarize and compare the reasons given for leaving a job. He found that wages were a decisive factor in only a minority of voluntary separations reported by the various studies. The percentage ranged from four per cent in one of the studies to 25 per cent in another. The wage factor accounted for approximately ten to 15 per cent of the voluntary separations in the other three studies. "Other economic factors," which included such considerations as 79 "steadiness of employment" and "chance for advancement or improvement," accounted for from zero to 26 per cent of the terminations in the five studies. After combining both "wage" and "other economic factors" for each study reviewed, the range extended from a low of about 11 per cent to a high of about 40 per cent. It should be noted that Parnes did not distinguish between reasons for geo graphic moves and reasons for non-geographic moves. Parnes also found that substantial proportions of job terminations were of the "personal" variety, resulting in 91 temporary periods of no labor force activity. Of the more recent studies, the one by Lansing and Mueller treats th'e subject most comprehensively. They found that for men, (the rates for women 22-64 years old were somewhat lower) approximately 49.5 per cent of those who moved geogra phically did so for "work related" reasons (these were broken down: (1) to take a job, 29.5 per cent, (2) to look for work, 11.9 per cent, (3) job transfer, 8.1 per cent). Another 14.6 per cent left for "marriage and family" reasons, and 35.3 per cent left for "other" reasons.^ Using a somewhat different reason classifica tion, Lansing and Mueller found that for all persons in ^Parnes, op. clt., pp. 151-153. 92Lansing, op. clt., p. 37. and out of the labor force, 58 per cent moved geogra phically for "economic reasons only," 14 per cent for both economic and non-economic reasons, 23 per cent for non economic reasons only and five per cent for no stated reason. They found that professional and technical workers moved geographically for "economic reasons only" 74 per cent of the time, while 19 per cent of the migrants who 93 were not in the labor force moved for some reason. Saben found that male professional workers moved geogra phically for "work related factors" 63.6 per cent of the time. This was broken down: 42.3 per cent, to take a job; 1.9 per cent, to look for work; 19.4 per cent, job transfer; 36.4 per cent, other.^ Ladinsky also found that professional workers make geographic moves more often 95 for job related reasons than other occupational groups. ^ In their study, Bancroft and Garfinkle found that 33 per cent of all job shifts were made to improve status, 33 per cent due to loss of Job, and the remaining 33 per cent for such reasons as the ending of a temporary job, illness or other personal reasons. For all women these percentages 93Lansing, op. clt., p. 60. S^Saben, "Geographic ... , " op. clt., p. 877. 95jack Ladinsky, "Sources of Geographic Mobility," Demography, IV, No. 1 (1967), p. 304. 81 were 30.0, 20.6 and 49.3 per cent respectively. For pro fessional women the rates were 23.3 per cent, 16.8 per a 6 cent and 59.9 per cent. No explanation was given for the latter figures (23.3 and 59-9 per cent) on profes sional women which seem to run counter to expected results. One of the most consistent findings of recent labor mo bility studies is that net geographical movement of labor is typically from areas of low net advantage to areas of 97 high net advantage as measured by income differentials. In their study entitled "Why Women Start and Stop Working: A Study in Mobility," Rosenfeld and Perrella found that 43.3 per cent of married women who left the labor force did so because of pregnancy, thirteen per cent due to illness, 14.2 per cent due to family responsibili ties, 7.2 per cent were laid off, 3.2 per cent due to unsatisfactory Job, 7.6 per cent moved, and 11.6 per cent due to "other" reasons. With respect to reasons for ^Bancroft, op. clt. , p. 903. 97 Bancroft, op. clt. ; Bunting, Ashby, Lowell and Prosper, op. clt.; Raimon, op. clt.; Lowell E. Gallaway, "Industry variations in Geographic Labor Mobility Pat terns," Journal of Human Resources, II (Fall, 1967), pp. 461-474; John B. Lansing and James N. Morgan, "The Effect of Geographic Mobility on Income," Journal of Human Re sources, II (Fall, 1967), pp. 449-460; Lansing and Mueller, op. clt.; and M. J. Greenwood, "An Analysis of the Deter minants’ of Geographic Labor Mobility in the United States," Review of Economics and Statistics, LI, No. 2 (May, 1969), pp. 189-1$'4'. ----------------------- 82 married women returning to the labor force, they found that 41.6 per cent returned due to financial necessity, 16.9 per cent wanted extra money, 18.9 per cent for per sonal satisfaction, 6.8 per cent because husband lost Job, Q8 and about 15 per cent distributed among other reasons. Family Status. Parnes suggested that: On the basis of the available evidence, the most reasonable conclusion is that there is a slight relationship between the marital sta tus and the mobility of male workers, married men making somewhat fewer shifts than non married . 99 Saben found that married men had lower rates of occupation al mobility than single men, but among women the rates varied little with marital status. Heneman found that family heads were more mobile occupationally and indus trially, while secondary family members (wives, etc.) made more shifts from unemployment to employment, and into and out of the labor force. 9®Carl Rosenfeld and Vera C. Perrella, "Why Women Start and Stop Working: A Study in Mobility," Monthly Labor Review, XXCVIII, No. 9 (September, 1965),10/8 and 1081. 99parnes, op. clt., p. 121. 100Saben, "Occupational ... op. clt., p. 32. 101Herbert G. Heneman, Jr., "Differential Short- run Labor Mobilities: St. Paul, 1941-1942," Minnesota Manpower Mobilities, (University of Minnesota: Industrial Relations Center), Bulletin No. 10, 1950, p. 40. 83 Due to the tremendous increase in the labor force participation of married women during the last 20 years, more attention has been turned to a study of the factors which determine their participation. One of the most interesting findings of these studies is the fact that although the presence of preschool children in the family had a negative affect on the participation rates, it is nevertheless in this category that one of the most signi ficant increases in married women participation has occurred.^02 Nurses. As noted in an earlier section of this chapter, the reasons for turnover of registered nurses was given considerable attention in the literature. Many of these studies have provided the actual count or per centages of the more important reasons. In the table below, we attempt to group these reasons into five major 102cain, op. clt.; Bowen, op. clt. ; Alfred Telia, "Labor Force Sensitivity to Employment by Age, Sex," Industrial Relations, IV, No. 2 (February, 1965)* pp. 69-83; Jacob Mincer, "Labor Force Participation of Married Women," Aspects of Labor Economics, (Princeton: Princeton University Press, 1962), pp. 63-97; and Thomas Dernburg and Kenneth Strand, "Hidden Unemployment 1953- 62: A Quantitative Analysis by Age and Sex," American Economic Review, LVI (March, 1966), pp. 71-95. TABLE 5 REASON FOR TURNOVER IN THE SEVEN NURSING STUDIES (Figures in percentages) _ _ _ j . _ RN Not RN Control- RN Con- RN Con- RN Not Control- lable — trol- trollable Control lable— Not Job lable- Job Rela- lable — Study Not Job Year Related Related Wage- Related ted Dis satisfac tion Fired In volun tary Not Known Total Lotspeicha 19^8 30.8 38.0 16.5 10.0 --- 4.7 100.0 Hougha 1950-51 30.6 51.7 8.0 6.5 1.6 --- 98.4 Foleya 1951 46.5 33.5 15.0 2.0 2.0 3.0 102.0 Null 1954 41.0 23.2 19.6 13.1 1.0 2.0 100.0 a Catania 1958-62 43.5 25.5 15.1 5.2 6.8 3.8 99.9 Smith 1960(Sng) 4.5 47.5 18.5 9.0 ___ 21.0 100.5 (Mar)50.0 17.0 7.0 16.0 --- 9.0 99.0 Saleha 1963-64 53.5 19.5 7.0 16.0 --- 4.0 100.0 aIn those cases where a sizable percentage (15-40 per cent) is simply noted as "moved," the percentage was divided equally between "not controllable by RN" and "controllable by RN" categories. Source: Previously cited studies. co 85 103 categories. Reasons which were included in group one were those over which the RN had no control, such as pregnancy, sick ness, husband moving, etc., but did not include fired or asked to resign. Those in group two included reasons over which the nurse had control, but were not specifically related to her work. Such reasons as marriage, desire to move out of an area, extended vacation, move to further experience, travel, etc. were included here. Group three included such things as leaving for a higher wage, a bet ter Job, etc. Group four included all reasons having to do with an RN's dissatisfaction with her job. It goes almost without saying that any attempt to rearrange the findings of others is difficult, and at best offers only a rough approximation as to the actual distribution of reasons. In spite of this, several general observations can be made from the above table. Probably the first of these is that a very sizable propor tion of nurses left their job for reasons beyond their immediate control. The table suggests that this range might extend from roughly 25 to 55 per cent. Another l^Ruth l . Lotspeich, "Why Do General Duty Nurses Resign," American Journal of Nursing, LI (July, 1951), pp. 468-569; Hough, op. cltTl Ftaley, opT clt.; Null, op. clt.; Catonia, op. clt. ; Phil M. Smith, Influence of Wage Rates on Nurse Mobility(Chicago; The Graduate School, 1962); Saleh, op. clt. 86 finding not apparent from the table is that relatively few nurses left a job specifically to obtain higher wages. Smith found that only four per cent of married women did so, and eight per cent of the single women. Hough also found that only four per cent of the terminations in his study left because they desired higher wages. A problem is encountered when category three is expanded from "higher wages" to include "better job," "promotion," etc. The range for this group is from zero to 29.6 per cent. It is quite possible that some of the reasons included in group four, the job dissatisfaction group, might well have been placed in the "wage or wage related" group, and the reverse might also have been possible. Column five indicates that fairly few RN's were fired or asked to resign (zero to 6.8 per cent). By adding columns one and two, we obtain a rough indication of the percen tage of terminations over which a hospital has no control. This percentage probably lies between 50 and 85 per cent of all terminations. Other Aspects of Nurse Turnover In this section the available literature concerning several other determinants of nurse turnover will be re viewed. Three of these variables are hospital character istics, i.e., hospital size, ownership, and RN/patient 87 ratio. Size of Hospital. Levine and Welland found that 104 turnover was highest in larger hospitals. Dodge, on the other hand, could find no relationship between hospi- 105 tal size and turnover. Two studies by the American Nurses Association (ANA), one in 1955-1956, another in 1967, offered inconclusive evidence as to the relationship of hospital size and turnover.10^ If any conclusion were drawn from the ANA studies, it would probably be that larger hospitals have lower turnover rates. Ownership of Hospital. Levine found that turnover was highest in church owned hospitals, while it was lowest 107 in government and "other" hospitals. An ANA study came to opposite conclusions. In their study, government hos pitals had the highest turnover, church owned the lowest, 108 and other nongovernment hospitals were in the middle. Prom the little available evidence, no conclusion can be 104 Levine, op. clt., and George P. Weiland, "Studying and Measuring Cursing Turnover," International Journal of Nursing Studies, VI (July, 1969), pp. 61-70. ^■^Dodge, pp. clt., p. 170. 106ANA, op. clt. 107 Levine, op. clt. 10 8,.., ANA, op. cit. 88 drawn concerning the effect of hospital ownership on turnover. Nurse-Patient Ratio. Two of the studies attempted to use the nurse-patient ratio as a measure of pressure on the RN which might lead to her resignation. Dodge found that a larger RN-patient ratio as well as a larger total staff-patient ratio could reduce voluntary turn over. Hixson also found that the RN-patient ratio was negatively related to turnover. He found, however, that increasing the auxiliary-RN ratio would actually lead to an increase in turnover. He suggests that hospital aides, etc. working with an RN, add to her supervisory responsibilities, whereas RN's would rather do only 110 nursing care. Seasonal Variations in Turnover. Three of the studies consulted found that turnover rates of hospital nurses fluctuated seasonally.^1 Nash found that ^■^Dodge, pp. clt. ^^Jesse Sharp Hixson, "The Demand and Supply of Professional Hospital Nurses: Intro-Hospital Resource Allocation," (unpublished Ph.D. dissertation, Michigan State University, 1969), pp. 99-101 and 118-121. 111Nash, op. clt., p. 29; Hough, op. clt., p. 46; and Null, op. clt., p. 422. 89 resignations were highest in May, August and September, while Hough found July, August and September to be highest. Null found resignations to be highest in June. Other Nurse Characteristics. Several studies found that most turnover occurred in the younger age groups. The median age of RN's terminating in the Hough study was 23.7 years. Lotspeich found the typical age to be 24.8 years, and Wright found most of the turnover 112 occurred between the ages of 20-24 years. Levine also found most turnover was by young RN's. He also found that nurses in supervisory positions had lower Job change rates than other nurses. Smith suggested that married and single nurses be treated separately in studies of turn- 114 over. ■^^Hough, op. clt., p. 44; Lotspeich, op. clt., p. 469; and Wright, op. clt., p. 51. HR •^Levine, op. cit., pp. 51-52. ^^Smith, op. clt. CHAPTER III STATEMENT OF RELATIONSHIPS TO BE INVESTIGATED AND DESIGN OF SURVEY There are two main objectives of this chapter. The first is to outline the relationships to be investi gated in Chapters IV and V, and the second is to describe the design of the present survey. It should be noted that much of the analysis of Chapters IV and V will be of a descriptive nature, due to the fact that there is little available data on nurse mobility patterns. In this re spect it is felt that a significant contribution of the present study is that it has been designed in such a man ner that this data will be forthcoming. This might well serve as basis for future nurse mobility studies. Statement of Relationships to be Investigated in Chapters IV and V This section will be divided into two parts, the first dealing with turnover and related subjects, and the second with the mobility of nurses. The turnover analysis which will be discussed in Chapter IV will be concerned 90 91 with terminating RNs only, while the mobility analysis of Chapter V will consider only hires. Fully describing the mobility pattern of nurses can only be accomplished at the end of the Job movement, i.e., at the time the RN is hired or rehired. This is so because many terminating RNs had no idea concerning where they would be or what they would be doing a short time after leaving a hospital. The mobility pattern for these terminating nurses was simply unknown. Nurse Turnover. One of the first tasks of Chapter IV will be to calculate turnover rates for each of the 18 hospitals in our study. These turnover rates will then be compared with the turnover rates found in the nursing and other studies in Chapter II. We have no reason a priori to expect the turnover rates in our study to be different from those in other nursing studies. We do, how ever, anticipate that the turnover rates in our sample will be higher than those of many other groups of workers. Turnover rates of workers in manufacturing industries are probably exceptions in the above. A chi square test will be run on the turnover rates of the 18 hospitals in our sample to determine whether or not there are significant differences among them. We anticipate that there are differences which may, 92 in part, be attributable to such things as hospital loca tion, payment of shift differential, starting wage level, hospital size and ownership, retirement benefits, etc. If this is in fact the case, a regression will be run with the hospital turnover rates as the dependent variable and the above, hospital characteristics as the independent variables. In four of the hospitals, additional information was obtained which will enable us to calculate turnover rates by various RN and hospital characteristics. In Chapter I these rates were called refined turnover rates. The refined rates will be calculated for age, race, level of education, and shift worked and will then be tested to determine whether or not there are significant differ ences between them. We expect that there are such dif ferences. Two other measures of nurse turnover will be cal culated for the present sample of hospitals. These are the relative frequency of turnover and the instability rates. These rates will be compared with those found in the nursing literature. A chi square test will be run to determine if there are significant differences among the hospitals in our study. In addition to the above turnover measures, an approximation of the expectancy of service concept will also be calculated for the 18 hospitals 93 combined. These percentages will be compared with those of workers in other occupations as well as with those found in other nurse studies. As was mentioned in Chapter I, one of the ques tions about which we hope to obtain information has to do with the motivational forces behind job changes. Consid erable effort was expended to learn the reason for each termination. These reasons will be grouped into five major groups. These five reason groups are as follows: (1) RN not controllable — not job related; (2) RN con trollable — not Job related; (3) RN controllable — wage related; (M) RN controllable — Job related dissatisfac tion; and (5) RN not controllable — involuntary or fired. The percentage of RNs falling into these groups will be found. In addition, these reason groups will be related to several RN characteristics to determine whether or not the reason distribution is affected by them. Chi square tests will be employed wherever possible. The average length of stay will be calculated for each hospital. Chi square tests will be run to determine whether or not there are significant differences in the lengths of stay of the hospitals in our sample. If the differences among the 18 hospitals are significant, an attempt will be made to explain them. Cross tabs 94 relating the length of stay to various RN characteristics will be run using the length of stay as the dependent variable and the RN and hospital characteristics as the independent variables. Having conducted the exit and entrance interviews over the period of one entire year, we are interested to determine whether there are any seasonal differences in turnover. To do this we will construct frequency distri butions of the count of terminations and hires by month. Chi square tests will be employed to determine whether or not there are significant differences among the months of the year. It is expected that there are. Nurse Mobility. Chapter V, which contains the analysis of the mobility flows of registered nurses, might well be considered the heart of the present study. The turnover relationships to be discussed in Chapter IV, although of considerable interest, are really only one aspect of the total movement pattern of nurses. Our major task in Chapter V is to fully describe such patterns. These mobility flows will then be compared to the extent possible with those in the literature. The first task to be accomplished in Chapter V will be to construct a cross-tab of the 28 mobility flows versus the five reason groups discussed in the turnover 95 section above. A count of nurses in each of the 28 flows will be made. If it can be shown that some of the flows are not pertinent to our sample (i.e., there are few or no RNs making those kinds of moves), and that their exclusion will not bias the analysis, they may be dropped from some of the future cross-tabs. There are several studies (non-nursing) reviewed in Chapter II which considered certain aspects of mobility such as geographic, interindustrial and Interoccupational mobility. In order to better facilitate comparison, the 28 mobility flows of our study will be collapsed back Into the eight less detailed mobility types discussed in Chap ter I. We will determine the percentage of nurses in each of these mobility types.'1 ' Both row and column chi square tests will be run to determine whether or not there are differences among the various mobility types with respect to the five reason groups, and vice versa. Any significant differences will be noted. We anticipate that there will be very few frictional moves (i.e., move ment from unemployment back into employment), the main reason being that at the established wage level, most hospitals are unable to fill all of their vacant staff ^It should be noted again, that these eight types are not mutually exclusive groups, where as the 28 mo bility flows are. positions. We also expect few interoccupational moves, i.e., movement from a non-nursing Job back into the nursing profession. The main reason for this expectation is that nurses have a fairly large investment in their training, and therefore, are reluctant to work outside the nursing profession in the first place. Of primary concern to us will be the relationship between the reason group "wage-wage related" and the mobility type "geographic." It will be recalled from Chapter I that if the monopsony-oligopsony argument is to be upheld, wage-induced geographic mobility must be relatively low. This aspect of the mobility type-reason analysis will be given special consideration. Comparisons will be made with all the relevant mobility literature in an attempt to establish some feeling as to the relative "highness" or "lowness" of this important concept. The last task to be completed in Chapter V will be to investigate the effects of race, age, marital status and level of education on the above mobility flows. We expect to find that young nurses are more mobile than older nurses, that Filipinos are more mobile than any other racial group, and that single women are more mobile than married women. Other relationships will also be noted and compared with the relevant literature. Chi square tests will be run wherever possible. 97 Survey Design and Data Description Having briefly outlined the analyses to be per formed in Chapters IV and V, it may be useful to describe some aspects of the present survey's design. The major objective of this study is to design a survey which would enable us to collect data describing the Job movement patterns of nurses and to relate these patterns to the reasons for job change. In undertaking this task one of the first things to be done was to determine the method by which the data was to be obtained. Of the two most com monly used methods, i.e., questionnaire and in-depth per sonal interview, the latter method was chosen. It was felt that questionnaires are frequently too rigid, and therefore often unsatisfactory as a means of obtaining information on the reason for job change and movement between jobs. Further, attempting to make a written questionnaire more flexible, sometimes results in an overly complicated set of questions which only serves to confuse the person being interviewed and leads to non response. In these respects, it was felt that a personal in-depth interview conducted by a neutral person would give superior results. Some writers, however, have objected to the use of interviews as a means of data col lection on the grounds that the "true" underlying reason 98 for a job change will not be uncovered because the respon dents may rationalize in terms of "socially accepted" 2 criteria. While this may be true to some extent, it was nevertheless felt that an in-depth interview by a neutral observer is superior to existing questionnaires or brief exit interviews by hospital staff. The next task was to determine the questions which should be asked in each of the interviews. A list of ques tions used in other studies served as a basis. Beyond this however, additional questions were included in order that a complete description of the mobility flows discussed in Chapter I could be obtained. This resultant group of ques tions was finally included in a questionnaire which was to serve as a guide to the person conducting the interview. A copy of this questionnaire appears as Appendix B. At the same time the interview method and ques tions were being considered, a sample of representative hos- pitals from the local universe was selected. It was felt that the exit and entrance interviews in approximately 2Abraham Bluestone, "Major Studies of Worker's Reasons for Job Choice," Monthly Labor Review. LXXVII (March, 1955), p. 301. ■^As noted in Chapter I, because of the prepon derance of hospital nurses, it was felt that little bias would be introduced by limiting the scope of the study to them. 20 hospitals could be adequately handled. Due to the relatively small size of this sample, it was further decided that the study would be limited to general hospi tals. A further decision to be made was with respect to the size of the area from which to draw the sample. It was finally decided that a reasonable distance to travel to obtain the interviews would be approximately 15 miles. Consequently, with only two exceptions (these were only 17 or 18 miles away), the 20 hospitals selected were located within a 15 mile radius of downtown Los Angeles. The sample was drawn from the Annual Roster and Press Manual, 1968-1969 of the Hospital Council of Southern California. Some 63 hospitals met the criteria above. Each was assigned a number, and 29 were then selected by the use of a table of random numbers. These 29 hospitals were contacted by the Hospital Council of Southern California and were given a brief explanation of the proposed project.1 * Of the 29 hospitals contacted, several felt they could not participate at that time due to construction projects or other research projects already in progress. Only three hospitals ^We are very much indebted to the Hospital Council for the full cooperation which they extended us. They were very influential in gaining us admission into several of the hospitals in our sample. 100 refused to participate. Of the 20 hospitals in the final sample, one was subsequently dropped because of failure to notify us of all hires and terminations. Another very small hospital was later dropped because there was only one staff change during the entire year. The remaining 18 hospitals comprised the working sample. The cooperation of these 18 hospitals was excellent. It is felt that the sample of hospitals was fairly representative of those in the area. Each geographical location was represented. The size of the hospitals ranged from about 39 to over *100 beds. Hospital ownership varied from proprietary to church owned, non-profit. Un fortunately, no government hospitals ended up in the sample although one was among those which could not accommodate us at that time. After the initial contact by the Hospital Council, the actual interviewing arrangements were worked out with the director of nurses, or someone in the nursing office designated by her. In each case a place was provided where the interview could take place privately, generally while the nurse was on duty. It was agreed that someone in the nursing office would call each time a nurse began work at the hospital, and each time a letter of resigna tion or other such notice was received. In those instances where there was no notice given, the nurse had 101 to be interviewed by telephone or written questionnaire. The cooperation of the nurses was excellent. Out of the nearly 1500 nurses in the study, only five or six nurses contacted at the hospital refused to be interviewed. It was felt that the answers given by the nurses were honest. Several girls admitted that the reason they had left their last job was because they were fired. Two single girls acknowledged that their reason for leaving was due to pregnancy. It was decided that the period of data collection would be one year in order that possible seasonal fluctua tions in turnover might be noted. Actual data collection was preceded by a one month presample collection period. During this trial period several improvements were made in the collection procedure and in the interview technique. Only registered nurses (RNs) who were hired, re hired or terminated were included in the sample. In all but two hospitals, both full-time and part-time RNs (various categories) were included. In two hospitals, only full-time RNs were interviewed. Every RN, including the director of nurses, supervisor, etc., down to the general duty staff RN, was a potential interview candi date. An attempt was made to interview RNs who were con sidered to be "helping out only temporarily." This proved to be difficult, and undoubtedly some nurses in this group 102 were missed. This was also the case with a few per diem and on call part-time help. All nurses who worked in Inpatient units were interviewed, while those in the out patient departments were not. Transfers into or out of the inpatient units as well as transfers between the hos pitals of a group of hospitals with common ownership were considered to be Job changes warranting interview. Some groups of nurses not included in the sample were: (1) new graduates and foreign nurses who were not yet licensed;^ (2) any nurse taking an extended leave of absence due to pregnancy, sickness, etc.^ and (3) nurses employed out of the Nurse Registry. The information from the interview sheet was later coded onto IBM cards, one card per RN. The explana tion of the coding format appears as Appendix C. Some of the information collected includes the following: (1) personal characteristics of the RN (i.e., age, race, marital status, year of graduation, etc.); (2) her length of stay at her present or her last Job; (3) place from which she was coming or to which she was going; (4) what "These nurses by law could not perform the duties of RNs. If, and at such time as they became licensed, they were included in the sample. For further explanation see Appendix A. ^If they did not return at the specified time, an attempt was made to interview them. 103 she had been doing or what she intended to do; (5) a description of her last or present job (i.e., position, full-time or part-time, shift worked); (6) reasons for leaving last or present job; etc. An additional IBM card containing characteristics of the hospitals was coded for each hospital. Such information as size, ownership, and age were included. Other characteristics related to hospital policies were also coded. These included such things as starting salary level, overtime payment for weekends and holidays, extra payment for experience and education, retirement plans, the size of the nurse-patient ratio, etc. A fur ther measure of the attractiveness of a hospital was given by the level of Income in the surrounding area and also by the rate of certain crimes in the vicinity of the hos pital. The explanation of the hospital card appears in Appendix D. Of the 1,^53 nurses included in this study, 1,085 were personally interviewed, 132 were interviewed by telephone, 127 were contacted via mail and questionnaire, and the information for 109 RNs had to be furnished by the hospital as we were unable to locate the nurses involved. There was a total (full-time and part-time) of 66M terminations and 789 hires and rehires. Of these 6^2 hires and 529 terminations were full-time employees. 104 Problems Encountered One of the most difficult, and to a great extent unresolved problems encountered during the study, had to do with the fact that quite often an RN was unable to indicate which of several reasons was the primary reason for making a job change. Even rephrasing the questions and putting forth hypothetical cases did not always help. In those instances where the nurse could not decide be tween two or three reasons, the interviewer was forced to make the decision on the basis of the tone of the entire interview. Initially, it was felt that we would encounter few problems taking the nurse off the floor for her 15 to 20 minute interview. In some instances, however, this proved to be erroneous. It seems that many nurses were so busy that they felt they simply could not spare the time. In those cases the nurse became nervous and anxious to return to her floor. This, of course, detracted from the validity of the interview. To alleviate this problem the interviewer offered to return at another time, or left a questionnaire and self-addressed envelope with the nurse. This was only a problem in the case of nurses who were terminating. New hires were generally interviewed during their orientation period and, hence, their services were not as critical to the overall functioning of the 105 floor. In spite of above mentioned problem, this method of interviewing was still deemed better than trying to interview the RN after working hours or trying to find her at home. In this connection, it should be noted that the participating hospitals were most cooperative in allowing the nurses to leave the floor for an inter view. In many cases, a supervisor or someone in the nursing office would cover for the nurse being inter viewed. Another problem encountered had to do with the identification and comparison of various part-time cate gories. The hospitals used different classifications of part-time nurses, and sometimes the same classification was defined differently in different hospitals. For example, per diem in one hospital was the same as on call in another. In one hospital, all nurses who were not full-time were per diem, etc. Because of this problem and also due to the fact that some "part-time" nurses may have been on the records several years and yet only worked a few days a year, it was felt that for many pur poses the analysis should be restricted to full-time nurses only. It was noted that occasionally a newly-hired nurse would stay only a day or two at a hospital and then leave. In those instances neither an entrance nor an exit 106 interview was possible. Quite often the interviewer was unable to contact these nurses. In these few instances, as much information as possible was obtained from the nursing office. The two pieces of information most often lacking in these cases was the origin or destination of the RN and also her reason for coming or leaving. A similar procedure was followed in the case of nurses who left with no notice. No attempt was made to interview nurses who were fired; however, information about the nurse was obtained from the nursing office and the RN's coworkers. One problem which was easily remedied had to do with the fact that there was a tendency for smaller hospi tals to be late in notifying the interviewer of staff changes. This probably happened because changes occurred so seldom that no set procedure was established. It was decided that the interviewer should call each of these hospitals on a weekly basis to check for staff changes. This method worked quite well. It the next chapter of this dissertation, the turnover data will be analyzed and compared. In Chapter V we will do the same for the mobility data, and finally in Chapter VI the major conclusions of this study will be summarized and suggestions for future research will be noted. CHAPTER IV INTERPRETATION OF TURNOVER DATA As was mentioned in Chapter I, although the pri mary concern of this study is with nurse mobility pat terns, some useful and interesting data on nurse turnover was also collected. The task of the present chapter is to analyze and interpret these data. Where possible, com parisons will be made with the turnover studies reviewed in Chapter II. It might again be noted that all of the analyses in this and the following chapter will deal only with full-time RNs. Reasons for this discussion were noted in Chapter III. An additional reason is simply that in two of the hospitals we were able to interview only full-time nurses. Hospital Turnover Rates The turnover rate for each of the 18 hospitals was calculated. These ranged from a low of seven per cent to a high of 69 per cent. The turnover rate for all 18 hospitals combined was 36 per cent. These turnover rates are listed in Table 6. While comparison with other nurse turnover studies cannot be too rigorous, due to the 107 108 differences among the several studies, it is nevertheless interesting to do so. Comparison with the turnover rates in Table 2 in Chapter II seems to indicate that there are more low rates (i.e., in the 20's and low 30's) in the present study. The median rate in Table 6 is about 35 per cent. If the rates in the present study are in fact lower than those in the earlier studies, this might indi cate that the concern over high turnover rates in the 1950's and early 1960's has brought about action which has reduced nurse turnover rates. A further explanation might lie in regional differences, i.e., Southern Califor nia may not be typical of other areas. When the turnover rate for all of the hospitals in this study (36 per cent) is compared with those of all workers in Chapter II (see Table 2) it is seen that only workers in manufacturing industries have higher turnover rates (55.2 per cent for men, and 67.5 per cent for women in 1968). The turnover rate estimated for all workers in 1961 was 17.2 per cent while that for all women workers was lM.5 per cent. Rates for female office workers for 1965 and 1967 were 28 and 32 per cent respectively. Turnover rates for female professional workers (teachers, etc.) were found to be around lM to 18 per cent. These comparisons tend to substantiate the prevailing feeling that relatively high turnover is characteristic of hos- TABLE 6 109 HOSPITAL TURNOVER RATES Hospital Number Hate (%) Hospital Number Rate (%) 1 42 10 7 2 36 11 35 3 54 12 37 4 40 14 23 5 69 15 67 6 25 16 34 7 54 17 30 8 34 18 29 9 36 19 35 All Hospitals 36 pltal nurses. A chi square test was next applied to the turn over rates of the 18 hospitals In Table 6 to determine whether they were significantly different from their overall average. At a confidence level of .005, the rates were found to differ very significantly from the average (the calculated chi square * 100.40, chi square at .005 » 35.7185). 110 In an attempt to explain some of the above dif ferences in turnover rates, several regressions were run with turnover rate as the dependent variable, and various hospital characteristics (i.e., size, ownership, wage level, crime rate in the area, etc.) as the dependent variables. Three of the most significant regressions are described in Appendix E. Experience and wage theory lead us to make certain predictions as to the signs of the variables. The predicted signs of some of the vari ables along with their actual signs are shown in Table 7. Table 7 shows that about 50 per cent of the actual coefficient signs in the regressions coincided with the predicted value. Of these, the wage variable, while con sistently the correct sign, was never significant. The probable explanation for this is that there were few differences in the starting wage levels of the 18 hospi tals. This lends some support to the contention made in Chapter I that there may be little wage competition among hospitals in a given area. These wage levels are in fact suggested by the local hospital council. The next four variables — presence of a retirement plan, overtime pay ment for weekend work, robbery rate, and the RN/patient ratio — consistently had the expected signs and were fairly significant. The latter variable, i.e., the RN/ patient ratio was also found to be significant in two of Ill TABLE 7 PREDICTED AND ACTUAL SIGNS OF THE COEFFICIENTS IN THE HOSPITAL REGRESSIONS Variablea Predicted Actual Level of Significance Wage (-) (-) below .50 Ret (-) (-) about .10 Over (-) (-) about .10 Rob ( + ) ( + ) about .10 Ft/P (-) (-) below .25 Diff 1 (-) ( + ) about .05 Inc ( + ) lower sions) about .05 (much in other regres- Size ( + ) (-)usually below .50 Own (+) (+) & (-) below .50 See Appendix E for variable description. the studies reviewed In Chapter II. Payment for shift differential, although significant, had an unexpected (+) sign. It was suggested that a possible explanation might be that those hospitals with turnover problems were forced to pay higher shift differentials. This, however, was not the case; the size of the differential paid and the 112 turnover rate were not related. The median level of family income in the area was thought to be a proxy for the crime variable (i.e., the higher the income level, the safer the neighborhood, and thus, the lower the turn over rate). While quite significant, the (+) sign was unexpected. Hospital size usually had a negative sign indicating that larger hospitals had lower turnover rates. The T value for this variable, however, was not signifi cant. A possible explanation for the (-) sign might be that large hospitals have better orientation and education programs and that they may also have better organization and more clearly defined policies, etc. They also are more likely to have new and modern equipment. In Chapter II it was found, although the evidence was not conclusive, that the larger size was related to lower turnover. The sign for the hospital ownership variable varied from re gression to regression and was never significant. Similar results were reported in the literature. It should be noted that approximately 16 regres sions were run, and various combinations of variables tried. Effort was made to remove multi-collinearity. Only the three regressions with the highest R/SE were included in Appendix E. The highest R/SE was .3911. While these R-squares are quite low, indicating that the selected variables explain only a relatively small per 113 cent of the total variance among the hospitals, it is nevertheless felt that the inclusion of these regressions is warranted on the grounds that the use of the regression here is simply to describe rather than to predict. Refined Turnover Rates In Chapter I it was suggested that for some pur poses it is instructive to examine turnover rates for various sub-groups of hospital nurses. To do this, how ever, additional information concerning the characteris tics of the nurses remaining on the hospital staff is necessary. This additional information was obtained in four of the hospitals in our sample. As of a given day, each of the four hospitals provided us with a count of all of their RN staff according to (1) race, (2) age, (3) marital status, (4) shift worked and (5) educational level. From this information, together with the infor mation obtained by the interviews of terminating RNs, turn over rates for each of the five groups were calculated. These rates are shown in Table 8. Inspection of the rates calculated in Table 8 shows that turnover in these four hospitals varied in versely with age. The turnover rate for RNs under 26 years old was 82.8 per cent while for RNs 50 and over it was only 8.6 per cent. One would expect the turnover of TABLE 8 REPINED TURNOVER RATES FOR POUR HOSPITALS Age Group Turnover Rate Marital Status Turnover Rate Race Turnover Rate Under 26 82.8jl Single 32.0? Caucasian 30. 45S 26-29 51.756 Married 34. 95S Negro 33.3? 30-39 30.25S Other 67.85S Spanish & Other 36.6? 40-49 33.25S Filipino 40 . 45S 50 & Over 8.6JS Oriental 32.6% Chi Square Tests: Age: 74.76 (significance at .005) Marital Status: 17.59 (significance at .005) Race: 1.74 (not significant at .05) TABLE 8 (Continued) Educational Status Turnover Rate Shift Worked Turnover Rate AA 53.8% Days 32.2? Diploma 33. OSt P.M. 34.3? BS or Higher 35.2? Nights 48.3? Chi Square Tests: Educational Status: 6.41 (significance at .05) Shift Worked: 3.98 (not significant at .05) younger nurses to be greater, because most of them have fewer family ties and are freer to make job changes in connection with travel, social and educational plans. This finding was also substantiated in the literature review. The chi square test on age indicated very signi ficant differences at the .005 level. The chi square test on race differences was not significant at the .05 level. The turnover rates, however, did show that turnover for Filipinos (40.4 per cent) was considerably higher than that for Caucasians (30.4 per cent). The turnover rates by marital status must be considered cautiously. Some problem was encountered in determining from hospital rec ords the actual marital status. This status is likely to change over a period of years (particularly for young single women). These changes were not always noted on the RN’s file. It does seem likely that the turnover rate for separated, divorced and widowed RNs ("other”) would be higher than for that of single or married women. These women often have family or other problems which interrupt their work. The turnover rate for "other" nurses was 67.8 per cent. The rates for married and single nurses were in the 30's. Due to the above-mentioned problem, nothing conclusive can be said about the differences be tween married and single women, although one would expect the turnover of single women to be greater. 117 Significant differences are noted among the three educational groups. Nurses with Associate of Arts (AA) degrees had considerably higher turnover rates (53.8 per cent) than either diploma (33*0 per cent) or BS (35.2 per cent) degree RNs. This might be explained on the grounds that many AA degree holders were older women whose family responsibilities often required them to leave the labor force. From his interviewing experience, this writer felt that AA degree RNs were, for some undefinable reason, less professional than many of their co-workers, and, hence, would be more likely to make job changes. From Table 8 it is also seen that the turnover for BS , degree holders is slightly higher than for diploma grad uates. One possible explanation for this is the fact that a very large percentage of Filipinos (who also had a fairly high turnover rate) had BS degrees. Roughly 42 per cent of all Filipinos had a BS degree, while only 14 per cent of the Caucasions did. The last refined turnover rate was that by shift worked. These rates were not significant at the .05 level. They nevertheless did increase in the expected direction, i.e., higher turn over among night workers. The turnover rates for RNs working days, evenings and nights were 32.2, 34.3 and 48.3 per cent respectively. 118 Other Turnover Measures The relative frequency of turnover and the Insta bility rates were calculated for each hospital in the sample and these rates were then compared with the turn over rates of Table 6. All three of these rates are presented in Table 9 to facilitate comparison. Chi square tests were run, and significant differences among the hos pitals were noted. The chi square value for the relative frequency of turnover rate was less significant than that for turnover rates. This would be expected because the relative frequency measure includes in the denominator the average number of nurses on staff, but also the number of terminations. Those hospitals with higher turnover rates would thus have relatively lower frequency of turn over rates. When relative frequency of turnover rates are used to compare turnover rates among hospitals, there is less likelihood of random sampling fluctuations. The relationship between turnover and frequency of turnover is readily apparent when the respective rates for the 18 hospitals are shown diagramatically. This is done in Figure 2. From Figure 2 and Table 9 it is noted that the range for turnover rates extends from a low of seven per cent to a high of 69 per cent, or a spread of 62 per centage points. The range of the relative frequency of 119 TABLE 9 THREE TURNOVER RATES FOR THE 18 HOSPITALS Hospital Number Turnover Rate Relative Frequency Instability Rate 1 42? 282 17? 2 36? 262 20? 3 54? 352 32? 4 40? 292 25? 5 692 41? 23? 6 252 20? 15? 7 542 352 32? 8 342 252 24? 9 362 272 25? 10 72 62 0? 11 352 26? 18? 12 372 27? 22? 14 232 19? 17? 15 672 40? 42? 16 34? 26? 26? 17 302 23? 20? 18 292 222 20? 19 352 26? 19? Chi Square ■ 100.4 Chi Square*40.77 Chi Square a 32.43 Sig. at .005 Sig. at .005 Sig. at .10 70 65 60 55 50 *♦5 U0 35 30 25 20 15 10 5 Turnover Race Relative Fre quency of Turnover Instability Rate 2 3 < * 5 6 7 8 9 10 11 12 **♦ 15 16 17 18 FIGURE 2 THREE TURNOVER RATES COMPARED 19 Hospital Number ro o 121 turnover rates is six per cent to 41 per cent, or a spread of only 35 points. The relative positions of the hospi tals have not changed, but the percentage spread between them has. The third measure of turnover included in Table 9 and also in Figure 2 is the instability rate. This rate, it will be recalled from Chapter I, is found by dividing the number of terminations of RNs who were on staff at the beginning of the year by the average number of RNs on staff during the year. This rate is of considerable interest because it indicates, to some degree, the overall stability of the nursing staff. High rates indicate that most of the turnover is due to "old staff members" leaving the hospital and, hence, higher instability. Low rates indicate that old staff members are remaining, and that the turnover is the result of newly hired nurses leaving the hospital. This last situation might suggest that there are problems with the orientation procedures of the hospital. The instability rate need not be lower than the other measures of turnover. Reference to Figure 2 verifies this point. The instability rate in hospital 15 was slightly greater than its relative frequency of turn over rate. A fourth measure of turnover which was defined in Chapter I was the expectancy of service measure. The 122 difficulties of obtaining such a rate were mentioned there. A crude proxy for this rate can be obtained, however, with fairly little effort. This proxy is nothing more than the relative frequency distribution of the length of stays of the terminating RNs. This frequency distribution was constructed and it was found that in our sample approx imately 13 per cent of the RNs left their jobs with one month or less service. About 28 per cent left with three or less months, 44 per cent with six or less months, 60 per cent with one year or less, 77 per cent with two years or less, and 86 per cent with three years or less servi-ce. It is of considerable interest to compare these results with those found in the nursing turnover studies reviewed in Chapter II. This comparison is made in Table 10 below. TABLE 10 COMPARISON OF EXPECTANCY OF SERVICE RATES (NURSES) Study 6 Mo. or less 1 Yr. or less 3 Yrs or less Present study 44? 60? 86? From Chapter II Study 1 58? Study 2 15? 38? 51% Study 3 44? Study 4 43? 80? Source: Previously cited studies in Chapter II 123 Table 10 shows that the expected length of service in the present study is somewhat lower than that of the other nursing studies. This finding is interesting in view of the fact that the turnover rates in the present sample appeared to be lower than those in the other nursing studies. The apparent contradiction might be explained by the fact that the studies which calculated turnover rates were, with only one exception, different studies than those providing the expectancy of service estimates. Usually we would expect hospitals with high turnover rates to have relatively low expectancy of service rates. A possible exception might be the termination in a given year of a large number of nurses who have been on staff for a long time. In this case, a high turnover rate might well be associated with a relatively long expectancy of service. Such a situation, however, could not long persist. When expectancy of service rates in our sample are compared with those in non-nursing occupations (see Chapter II), it is immediately apparent that the former are considerably lower. This is consistent with the fact that nursing turnover is much higher than other non-manu facturing industries. Terminations by Reason Groups Of considerable Interest to employers of nurses is Information concerning the reasons behind turnover. In this section the percentage of terminations by various reason groups will be calculated. These reason groups will also be related to various demographic characteristics of the terminating nurses. Table 11 shows the percentages for five major reason groups.'*' Here it is seen that of the 516 nurses who terminated, 8.9 per cent did so involuntarily. Com pared to the workers in other occupations, this figure is very low. In Chapter II it was noted that the percen tages of involuntary separations in manufacturing ranged from 18 to 55 per cent. The percentage for "all females," though hard to determine, appeared to be between 15 and 30 per cent. Professional workers, as would be expected, had fewer involuntary separations than other workers, probably in the neighborhood of 15 to 20 per cent. The percentage of nurses leaving their jobs for involuntary reasons was still lower. In the seven nursing studies summarized in Table 5, the percentages ranged from one to 6.8 per cent. ■^For a description of the reasons included in each of the reason groups, see the coding information sheet column 64-65 in Appendix C. 125 These values are somewhat lower than those found In our sample of hospitals. Explanations of this difference probably lie in the different staff composition among the hospitals dealt with in the various nursing studies. Un fortunately, no information was contained in the nurse turnover studies reviewed which would allow such compari sons . TABLE 11 PERCENTAGE OF TERMINATIONS BY REASON GROUP Percent of Reason Group Number of Terminations Total Term- _____________________________________________1 nations 1. RN not control lable - not job related 131 25.4 2. RN controllable- not job related 186 36.1 3. RN controllable- wage and wage- related 26 5.0 4. RN controllable- job related dis satisfaction 127 24.7 5. RN not control lable - fired- involuntary 46 8.9 TOTAL 516 126 When the Involuntary category is expanded to in clude "RN not controllable (i.e., groups one and five) it is seen from Table 11 that approximately 34.3 per cent of the RNs in our sample left their jobs for reasons over which they had no control. This percentage was somewhat lower than the crude estimates made in the nursing studies of Table 5j but still represents a rather sizable propor tion. Approximately 25.4 per cent of all terminations were due to "RN not controllable — not job related" reasons (Reason Group one). This group includes such reasons as husband moving to take new job, RN needed at home, pregnancy, etc. Of the five reason groups, numbers three, four and five represented "hospital controllable" reasons. In our sample, this was about 38.6 per cent. This figure compared quite well with those of the hospital studies in Chapter II. Hospitals interested in reducing the turnover of their nurses are probably compelled to concentrate their efforts in these reason groups. In our sample, only five per cent of the RNs left for wage or wage-related reasons. It would appear that the prospect of higher wages was not an important factor in motivating nurses to change jobs. The review of the nursing literature in Chapter II supported this finding even though the percentages in Table 5 appear not to 127 2 do so. The above finding seems to support the contention of Chapter I, i.e., that the market for nurses is not com petitive. In that case, we expect all the hospitals in our sample to have about the same wage levels, and conse quently few wage-induced job changes. From Table 11 it is further noted that almost one-fourth of the nurses who terminated in our sample did so because they were dis satisfied with some non-wage factors of their Jobs. This, too, lends support to the contentions of Chapter I. The last category in Table 11 to be discussed is group number two, "RN controllable - not job related." This category is quite large (36.1 per cent) and further substantiates the fact that nurses make many Job changes for reasons not related to the Job. These include such things as "wanted a vacation," "wanted to travel," "move to Join family and friends," "return to school," "mar riage," etc. The nursing literature also indicates that the percentage in this category is quite high (30 to 50 per cent). Considerably more will be said about the p The figures in Table 5 are misleading (high) due to the problems encountered in attempting to combine rea son groups defined according to criteria different from that of the present study. Many of the reasons grouped under number 3» Wage and Wage-Related, should have been in either group two or four (i.e., "RN controllable - non job related" and "RN controllable - dissatisfaction") but we were unable to separate them out. 128 reasons for job change In the mobility chapter which follows. It is of some interest to look at the percentage of nurses who terminated by individual reasons. Several of the most frequently given reasons are enumerated below. Of the 516 terminations, 45 nurses or 8.7 per cent left their jobs because their husbands were accepting other jobs elsewhere. About 5.4 per cent left due to pregnancy, and 6.4 per cent for personal or family illness. In the reason group "RN controllable - not job related," approximately six per cent left In order to get married, 3.9 per cent moved to join family or friends and six per cent simply wanted to move to a different area or location. These six reasons together accounted for over 36 per cent of the responses to the 47 possible reasons used in our study. Reason Group by Race. Having looked briefly at all terminations by reason group, it is Instructive to look at the effect of race on the number of terminations and the reasons for termination. Table 12 shows the number of terminations in each racial group by the five reason groups. Here It Is noted that Caucasians were by far the largest group. The "other" race groups was next followed by Filipinos and then Negros. In Table 13 the percentage of each racial group is shown by reason for TABLE 12 REASON GROUP BY RACE - TERMINATIONS REASON GROUP Cauc Neg Fil Other Total Per cent 1. RN not controllable— not job related 88 9 13 19 129 25.5 2. RN controllable — not job related 120 12 23 25 180 35.6 3. RN controllable — Wage-Wage Related 6 2 3 5 26 5.1 4. RN controllable — job related - dis satisfaction 76 8 22 19 125 24.7 5. RN not controllable— fired-involuntary 28 3 7 8 46 9.1 TOTAL 328 34 68 76 506 PER CENT 64.8 6.7 13.4 15.0 TABLE 13 REASON GROUP BY RACE - TERMINATIONS PERCENTAGE DISTRIBUTION REASON GROUP Cauc Neg Fll Other Chi square = 1. RN not controllable— not job related 26.8 26.5 19.1 25.0 1.721 2. RN controllable — not job related 36.6 35.3 33.8 32.9 .326 3. RN controllable — wage-wage related 1.8 5.9 4.4 6.6 2.798 4. RN controllable — job related - dissatisfaction 23.2 23.5 32.4 25.0 2.553 5. RN not controllable— fired-lnvoluntary 8.5 8.8 10.3 10.5 .423 With 3° Freedom Chi Square at the .25 level = 4.11 131 leaving. Chi square tests revealed that even at the .25 level, there were no significant differences among the percentages leaving for the five reason groups. This finding was somewhat surprising because we expected Filipinos to have a considerably higher percentage of terminations due to "RN controllable — not job related" reasons (travel, etc.) and a much lower percentage of "RN not controllable" reasons. It is noted in Table 13 that while not statistically significant, the percentage of Filipinos leaving for reason group one is lower than the percentage in the other race groups. Again, while not significant, it appears from Table 13 that Caucasians tended to leave less often for wage-related reasons than other groups. One might have expected the percentage of involuntary terminations of foreign-speaking RNs to be considerably higher than that of English-speaking nurses. The percentages in Table 13 indicate such a tendency, but again the differences were not statistically significant. The percentages of involuntary separations for Negroes and whites were almost identical. Reason Group by Age. Table 14 shows the number of nurses who terminated according to the five reason groups by age. Approximately 51 per cent of all the nurses under 30 terminated, while only 12.3 per cent were over 50. TABLE 14 REASON GROUP BY AGE - TERMINATIONS REASON GROUP Under 26 26- 29 30- 39 40- 49 50- over Total Per Cent 1 . RN not controllable— not job related 30 38 34 13 16 131 25.5 2. RN controllable — not Job related 72 45 44 10 14 185 36.0 3. RN controllable — wage-wage related 8 5 8 4 1 26 5.1 4. RN controllable — job related — dissatisfaction 30 23 32 22 19 126 i n • - = T C V J 5. RN not controllable— fired-involuntary 2 10 10 11 13 46 8.9 TOTAL 142 121 128 60 63 514 PER CENT 27.6 23.5 24.9 11.7 12.3 133 The various percentage breakdowns are shown in Table 15. As expected, the percentage of RNs leaving for reason group one was smallest in the "under 26" age group. Un fortunately, however, the chi square on this row was not significant at .25 level. About 50.7 per cent of the RNs in the youngest age group left for reason group two (RN controllable — not job related). This relationship was also expected and was very significant at the .05 level. The above relationships indicate that younger RNs left more often for "RN controllable" reasons than did older ones. The chi square on the wage group (number three) in Table 15 indicated that there were no significant dif ferences attributable to age. We expected younger RNs to be more concerned with wages than older RNs, but, in our sample, this was not the case. The chi square on the percentages in row four (RN controllable - dissatisfac tion), was significant at the .05 level and indicates that older nurses were more likely to leave for reasons of job dissatisfaction than were younger nurses. This finding was not expected. This writer felt that greater job dissatisfaction would prevail in the youngest age group primarily because of the disappointment encountered when the idealistic expectations of the new graduate were not met in the "real" hospital situation. This apparently TABLE 15 REASON GROUP BY AGE - TERMINATIONS PERCENTAGE DISTRIBUTION REASON GROUP Under 26 26- 29 30- 39 40- 49 50- over Chi Square = 1. RN not controllable — not job related 21.1 31.4 26.6 21.7 25.4 2.738 2. RN controllable — not job related 50.7 37.2 34.4 16.7 22.2 21.751 3. RN controllable — wage-wage related 5.6 4.3 6.3 6.7 1.6 3.361 RN controllable — job related — dissatisfaction 21.1 19.0 25.0 36.7 30.2 9.118 5. RN not controllable — fired — involuntary 1.4 8.3 7.8 18.3 20.6 31.805 With 3° Freedom Chi Square at the .05 level=7.8l. 135 is not the case. Inspection of row five (involuntary separations) reveals the expected finding that the percen tage of RNs fired increased with age. The chi square test in this row was very significant at the .05 level. One obvious reason for this is that many older nurses have been away from nursing for some time and may be less com petent in the performance of their assigned duties. Reason Group by Marital Status. Reference to Table 16 reveals that of the 51** terminations, 198 nurses or 38.9 per cent were single, 258, or 50.2 per cent were married, and 58 or 11.3 per cent were "other" (i.e., di vorced, separated, and widow). As was expected, the per centage of single nurses leaving for "RN not controllable" reasons was very small (7.1 per cent), while the corres ponding percentage for married women was high (**1.5 per cent). Almost the reverse was found for reason group two (RN controllable - not job related). These relationships are noted in Table 17. The chi square tests on these two rows were very significant at the .05 level. The percen tage of nurses leaving for wage-related reasons was higher in the single group, but was significant only at the .25 level. The result was nevertheless expected. The table also shows that women in the "other" category were more concerned with higher wages than married women. TABLE 16 REASON GROUP BY MARITAL STATUS - TERMINATIONS REASON GROUP Single Married Other Total Per cent 1. RN not controllable — not job related 14 107 10 131 25.5 2. RN controllable — not Job related 110 61 14 185 36.0 3. RN controllable — wage-wage related 16 7 3 26 5.1 4. RN controllable — job related — dissatisfaction 45 62 19 126 24.5 5. RN not controllable — fired— Involuntary 13 21 12 46 8.9 TOTAL 198 258 58 514 PER CENT 38.5 50.2 11.3 TABLE 17 REASON GROUP BY MARITAL STATUS - TERMINATIONS PERCENTAGE DISTRIBUTION REASON GROUP Single Married Other Chi Square = 1. RN not controllable — not job related 7.1 Ml.5 17.2 26.018 2. RN controllable — not job related 55.6 23.6 2M.1 18.876 3. RN controllable — wage-wage related 8.1 2.7 5.2 2.896 RN controllable — job related — dissatisfaction 22.7 2M.0 32.8 2.111 5. RN not controllable — fired— involuntary 6.6 8.1 20.7 16.311 With 2° of freedom chi square at .05 = 5.99 137 138 Although not significant at the .25 level, it appears that women in the "other" category were more likely to be dissatisfied with the job than either married or single women. Table 17 shows that a greater percentage of divorced, separated and widowed women left for involun tary reasons (group five) than did single or married women. The chi square test indicated that these rela tionships were very significant at the .05 level. Reason Group by Highest Educational Level. In Table 18, it is seen that most of the terminating nurses were diploma graduates (71.1 per cent) while only 9.6 per cent were AA degree holders and 19.3 per cent were BS and MS degree holders. Of the latter group, only ten RNs had the Master's degree. Table 19 reveals that while a larger percentage of the AA degree holders quit for "RN not controllable" reasons (group one) than the other RN groups, the percentages were not significant at .25 level. Type of degree was not significant at .25 level when the percentages of nurses quitting for reason group two were considered. The data from Table 19 indicate that the percentage of nurses who terminated for wage- related reasons was highest among BS degree holders. The chi square test on this row was significant at the .05 level. The literature review also suggested that profes sional workers and those with higher educational levels TABLE 18 REASON GROUP BY HIGHEST EDUCATION LEVEL - TERMINATIONS REASON GROUP AA Diploma BS+ Total Per cent 1. RN not controllable — not job related 16 91 22 129 25.5 2. RN controllable — not job related 20 126 36 182 36.0 3. RN controllable — wage-wage related 1 15 10 26 5.1 4. RN controllable — Job related — dissatisfaction 11 94 21 126 24.5 5. RN not controllable — fired-involuntary 1 35 9 45 8.9 TOTAL 49 361 98 508 PER CENT 9.6 71.1 19.3 TABLE 19 REASON GROUP BY HIGHEST EDUCATIONAL LEVEL - TERMINATIONS PERCENTAGE DISTRIBUTION REASON GROUP AA Diploma BS+ Chi Square = 1. RN not controllable — not job related 32.7 25.2 22.4 2.413 2. RN controllable — not Job related 40.8 34.9 36.7 1.926 3. RN controllable — wage-wage related 2.0 4.2 10.2 7-143 4. RN controllable — job related — dissatisfaction 22.4 26.0 21.4 .664 5. RN not controllable — fired— involuntary 2.0 9.7 9.2 5.431 With 2° freedom chi square at .05 - .25 = .10 = 5.99 2.77 4.61 Ot?T m tended to make more job changes for economic reasons than did other educational groups. The chi square test re vealed that there were no significant differences in the proportions of nurses leaving for reason group number four (RN controllable — dissatisfaction). The last reason group (fired or involuntary termination) was significantly different at the .10 level. Prom this row in Table 19 it is seen that a greater proportion of diploma and BS degree holders were fired than were the AA degree group. On the basis of the turnover results, the opposite rela tionship might have been expected. Problems. One of the problems with the above sec tion on reason versus RN characteristic is the fact that we were unable to take into account any interaction among the RN characteristics due to the size of our sample. For example, it is likely that part of the high involuntary separation percentage of BS degree holders was due to the fact that a large percentage of them were also Filipinos who, as was shown in Table 13, had a some what higher percentage of "fires" than other groups. Interaction of age with race, marital status with age, etc. all could alter to some extent the conclusions drawn above. These interaction relationships will be investi gated to the extent possible in the chapter on mobility. 142 Average Length of Stay The average length of stay (ALOS) of terminating nurses provides another useful indicator as to the condi tions surrounding the nurse staffing patterns of hospitals. While an inverse relationship between the ALOS and nurse turnover cannot be claimed (i.e., it is possible for the turnover rate to be quite high while the ALOS of those terminating is also quite large), the two measures are nevertheless related. The ALOS for each hospital in our sample was calculated and a chi square test was run. These values appear in Table 20. The ALOS for all hospi tals was 634 days. The chi square test indicated that there were very significant differences amont the 18 hospitals. In an attempt to explain these differences, the ALOS for three demographic characteristics was calculated. The values appear in Table 21. As was expected, because of the very significant chi square in Table 20, there were significant differences among the different race, marital, and educational statuses. The ALOS of Filipinos was 265 days, compared with 713 days for Caucasians. The low ALOS for Filipinos was expected. As noted above, their turn over rates were much higher than those of other racial groups. Most Filipino nurses are attracted to the United 143 TABLE 20 AVERAGE LENGTH OF STAY BY HOSPITAL Hospital Number ALOS (Days) Hospital Number ALOS (Days) 1 132 10 8 2 601 11 286 3 696 12 313 4 427 14 606 5 140 15 1,699 6 812 16 739 7 567 17 651 8 778 18 981 9 749 19 594 All Hospitals 634 Chi Square = 1526 was very significant at .005 TABLE 21 AVERAGE LENGTH OF STAY BY RACE, MARITAL STATUS SHIFT AND EDUCATION (18 HOSPITALS) RACE ALOS (Days) MARITAL STATUS ALOS (Days) EDUCATIONAL STATUS ALOS (Days) Cauc 708 SINGLE 473 AA 367 Negro 531 MARRIED 809 Diploma 698 Spanish 745 DIVORCED 382 BS 412 Filipino 281 SEPARATED 329 Chinese 890 WIDOW 482 Korean 516 Japanese 688 Thai 676 Other 401 Chi Sq. - 469 Sig at .005 Chi Sq. = 282 Sig at .005 Chi Sq. = 131 Sig at .005 145 States because of significantly higher wages and more plentiful job opportunities. Many of them come for a three or four year period and find that their training, coupled with the demand for their services, accommodates their personal and travel plans. The other race cate gories are fairly small, and conclusions drawn on the basis of them should be guarded. From these results it appears that Negro nurses had a shorter ALOS than whites. The turnover rate in Table 8 also indicated this might be the case, although the turnover rates there were not significant at the .05 level. As might have been expected, the ALOS for married women was greater than that for single and other women. Married women are more likely to be older, and to have additional family ties which would tend to keep them in one area longer. Table 8 shows that the turnover rate of married RNs exceeded that of single nurses. Assuming that the data from which the turnover figures were derived is reliable (as mentioned above, there is some doubt of this), it is possible, though not likely, that married women had a higher turnover rate, but at the same time had a longer average length of stay. The ALOS data for "other" women complemented the turnover data of Table 8. On the basis of the turnover rates according to educational status in Table 8, the findings 146 with respect to education and ALOS in Table 21 were ex pected; i.e., BS and AA degree holders had a shorter ALOS than did diploma holders. The reasons for this are essen tially the same as those given for the turnover rate dif ferences . Regression on ALOS In an attempt to further verify the demographic- ALOS relationships indicated above, and to determine the affect of additional variables on ALOS, multiple regres sions were run. The length of stay of the RN was the dependent variable, and various nurse and hospital charac teristics were the independent variables. These regres sions are specified in Appendix F. It must be noted at the outset that the regressions had low predictive power (most of them had an R/SE of about .14 to .15). Regres sion analysis was used here not as a predictive tool, but rather as a descriptive one. In this sense, the regres sions performed quite well. Had we been interested in obtaining an equation which could be used to predict the length of stay of RNs, it would have been better to interview RNs who stayed at the hospital rather than those who left. Separate regressions were run on all RNs col lectively (both full-time and part-time), and then on 147 full-time RNs, single RNs, and married RNs individually. In one regression the dependent variable was measured in log terms and in another as the inverse of the length of stay. None of these adaptations had any significant effect on the R/SE which continued at about .15. The equation vfoich was most explanatory was run on all full time nurses. Some of the more significant variables and their expected and actual signs are shown in Table 22. Out of the eighteen independent variables listed in the table, seven were significant at .25 or higher. Only three of the expected signs of the coefficients differed from the actual signs found in the several regressions. We expected the ALOS of nurses working in critical care units (U2) to be lower than in other units. This we felt would be due to the pressure associated with the care of the critically ill. The regressions indicated the con trary. One possible explanation may have to do with the fact that, generally speaking, only RNs work in these units and therefore the staff RN has no "supervisory responsibilities" with respect to auxiliary help. It will be recalled from Chapter II that Hixson suggested that RNs do not want these responsibilities, and that turnover might be reduced by lowering the RN-auxiliary ratio. The level of nurse wages (CR) was another variable whose sign was different than expected. Again, however, 148 TABLE 22 PREDICTED AND ACTUAL SIGNS OF THE COEFFICIENTS IN THE LENGTH OF STAY REGRESSIONS3 VARIABLE PREDICTED ACTUAL LEVEL OF SIGNIFICANCE NAG ( + ) ( + ) at .10 level R2 (-) (-) below .25 R3 (-) (-) at .10 M2 ( + ) ( + ) at .25 M3 (-) (-) at .25 El (-) (-) below .25 E3 (-) (-) below .25 U2 (-) ( + ) below .25 U3 ( + ) ( + ) below .25 U4 ( + ) ( + ) at .05 H2 ( + ) ( + ) below .25 H3 (-) (-) below .25 H4 ( + ) ( + ) at . 10 CSH ( + ) ( + ) at .25 CR ( + ) (-) below .25 CRO (-) (-) below .25 CFT ( + ) (-) below .25 CRET ( + ) ( + ) below .25 aSee Appendix F for equation and explanation of variables. 1M9 it was very insignificant, and, as noted earlier, there could be very little effect of wage differences, because most of them were the same. The variable full-time RN- patient ratio (CFT) also differed from the expected sign. Its level of significance was also very low. The general findings of the length of stay regres sions and cross-tabs are summarized below. In our sample of full-time terminations, age was positively related to ALOS. Negroes had shorter stays than Caucasians; and Filipinos had even shorter stays. Married women had longer ALOS than single women; separated, divorced and widowed women had shorter stays than single nurses. Nurses with AA and BS degrees had shorter ALOS than did RNs with diplomas. RNs who worked in critical care units, in obstetrics and pediatrics, and in supervisory capa cities had longer ALOS than nurses who worked in medical and surgical floors (not specialty units). Women whose husbands were blue collar workers or laborers, and those whose husbands were retired, unemployed or in the military had longer ALOS than nurses whose husbands were self-em ployed or professionals. This latter result might have been expected on the grounds that the need to work for wives of professional and self-employed husbands is prob ably less than for other groups. Wives of students had a shorter ALOS than did those whose husbands were 150 professionals and self-employed. The ALOS in hospitals which paid shift differentials and had retirement plans was longer than in those which did not. And, finally, hospitals located in areas of high robbery rates had a shorter ALOS than those in other areas. It was reassuring to note that, with the two or three exceptions cited above, none of the ALOS regression results contradicted the findings of the turnover and ALOS analysis described above. Seasonal Turnover The review of the nursing literature revealed that there appeared to be a seasonal pattern for nurse turnover. These studies all indicated that turnover was highest in the summer months May through September. August and September were most often mentioned. frequency dis tributions of the number of terminations and hires by month were obtained for the present study. These distri butions are presented in Table 23. The chi square test Indicated that there were significant differences among the twelve months. Inspection of these distributions revealed that the number of terminations in our sample was greatest in August (65) and June (57). The month of fewest terminations was February (31). For hires, it was found that June (73) and September (78) were the two 151 TABLE 23 HIRES AND TERMINATIONS BY MONTH IN THE 18 HOSPITALS3 Month Number of Hires Number of Terminations January 63 43 February 45 31 March 46 42 April 38 33 May 35 41 June 73 57 July 67 43 August 69 65 September 78 52 October 54 50 November 36 33 December 38 39 TOTAL 642 529 a Figures include both Full-time and Part-time RNs. 152 highest months while May (35) and November (36) were the months of fewest number of hires. The large number of terminations in June might be accounted for by vacation plans and children being out of school. The explanation for the high turnover in August is not apparent. Part of it might have been due to the termination of June diploma graduates who remained at their mother hospital until they became licensed. The large number of hires in June is undoubtedly accounted for, in large part, by new graduates. Also, many June graduates who took a vacation began their work in September. Many nurses who were wives of students also resumed work in September. A possible reason for the scarcity of hires in May probably has to do with the anticipated summer vacation, i.e., those currently not working in May would probably want to wait until after the summer to resume work. Similar reasoning might apply to November, i.e., anticipation of Christmas Holidays. It is interesting to note that the months of the highest numbers of terminations and hires did not always coincide. In those months where they did not, the hos pitals simply operated with a reduced staff until more RNs became available. 153 Summary In Chapter IV the data pertaining to the various turnover concepts defined in Chapter I have been described. It will be remembered that these data are a byproduct of the mobility data to be investigated in Chapter V. In the first section of this chapter, turnover rates (according to the definition in Chapter I) were calculated for each of the 18 hospitals separately, along with an overall rate for the 18 hospitals combined. Chi square tests were run, and significant differences among the hospitals were noted. It was found that the turnover rates in the present study were somewhat lower than those in the other nursing studies, but considerably higher than the turnover rates in other occupations with the exception of manufacturing. Multiple regressions were run in an attempt to explain the differences noted above. Several hospital characteristics which had some explanatory power were: (1) the presence of a retirement plan; (2) the overtime payment for weekends; and (3) the robbery rate in the area. Several other variables, such as wages, had the expected signs but were not statistically significant. Four of the hospitals provided us with additional information which allowed us to calculate the refined turnover rates. These rates were calculated by age group, 154 marital status, race, level of education and shift worked. Turnover varied inversely with age. Filipinos had a higher turnover rate than other races. Both married and single women had lower turnover rates than "other" nurses. Associate of Arts degree RNs had higher turnover rates than other educational groups. Relative frequency of turnover and instability rates were next calculated and compared with the turnover rates of the 18 hospitals. Chi square tests indicated (as expected) that there were significant differences among the hospitals. A crude estimate of the expectancy of service measure was next calculated. It was noted that the nurses of our sample had about the same expec tancy of service as nurses in other studies, but this was much shorter than for workers in other occupations. The next section dealt with the reasons for ter- ination. Here it was seen that nurses are "fired" much less frequently than workers in other occupations. It was shown, however, that when the involuntary cate gory was expanded to include "RN not controllable" rea sons, the percentage was a sizable 35 per cent. A significant proportion (36.1 per cent) of the termina tions were for non-Job related reasons. Only five per cent of the terminations were for wage-related reasons. Reason groups were next calculated with race, age 155 marital status and education. The differences among the racial groups were not significant, although some tenden cies were noted. One of the most significant findings with respect to age was that almost 51 per cent of the RNs under 26 left for "RN controllable - not job related" reasons. Also very significant was the fact that the per centage of RNs who were fired increased with age. Only seven per cent of the single nurses left for "RN not controllable" reasons compared to 41.5 per cent of the married women. The per cent of nurses who were fired was greatest among those in the "other" marital group. A greater percentage of AA degree holders left for "RN not controllable" reasons than the other educational groups. Roughly ten per cent of the BS degree terminations left their jobs for higher wages. The next section dealt with the average length of stay of nurses. The ALOS for all hospitals was 634 days. In order to explain the differences among the hospitals, cross-tabs relating ALOS to race, marital status and education were constructed. Chi square tests were run and some significant differences noted. Filipinos and AA degree holders were found to have relatively short lengths of stay. Finally a regression was run to further clarify these and other relationships. The two findings above were verified. It was noted that age at the time of hire was positively related to ALOS. Most of the predicted signs were the same as those found by the regression. Finally, it was found that there were seasonal differences in the number of terminations and hires in our sample, which agreed with findings in the literature. CHAPTER V INTERPRETATION OF MOBILITY DATA As noted in Chapter III, the only feasible way to obtain a thorough description of the mobility patterns of nurses is to interview them as hires, i.e., at the time when their job change or work status change has been completed. The information obtained from RNs who were terminating is incomplete in this "mobility'’ sense. For this reason it was discussed in connection with turnover concepts in Chapter IV. The present chapter will consider only nurses who are full-time hires and rehires (from here on denoted only as hires). Mobility Flows One of the first tasks in the analysis of mobility was to construct a cross-tab of the 26 (actually 28 when new graduates were included) mobility flows versus the five reason categories discussed in Chapter IV. ^ This 1The 28 flows and the 8 mobility types are discus sed in the definition of terms section of Chapter I. The abbreviations used in the tables of this chapter are noted in Table 1 in that chapter. 157 158 cross-tab is depicted in Table 24. Prom the table it is seen that the 26 flows versus the reasons accounted for 556 RNs. There were 79 RNs who, as new graduates, did not have accompanying reasons. Of these new graduates, 75 made no geographic move, while only four had done so. This suggests that most new nurses upon graduation either continue to work in the hospital from which they graduated or accept a position in a nearby hospital. BS and AA degree graduates likely choose a hospital in which they had received some of their training. This seems very reasonable as a new graduate would feel secure in familiar surroundings for her first job. One of the first aspects of Table 24 which warrants discussion is the fact that there were several mobility flows for which there were no, or possibly only one or two observations. There were no observations for the fol lowing mobility flows: (1) INT-O; (2) INT-O-G; (3) INT- O-FTPT; (4) INT-O-FTPT-G; (5) E; and (6) INT-O-IND-FTPT-G. It is noteworthy that five of the six flows included inter- occupation (0) moves. This suggests that interoccupational mobility of nurses might be very low. More will be said about this at a later juncture. Six of the flows contained only one observation in each. These were as follows: (1) INT-FTPT-V-G; (2) INT-IND-FTPT-G; (3) E-G; (4) not certain; (5) INT-O-IND; and (6) INT-O-IND-FTPT. It is TABLE 24 MOBILITY FLOW VERSUS REASON (HIRES) 1 2 3 4 5 6 INT- 7 INT- 8 INT- 9 10 INT- INT- INT- INT- INT- FTPT FTPT FTPT INT- 0- Reason Group I NT G V V-G FTPT -G -V -V-G 0 G 1. RN not controllable- not Job related 14 21 2 6 0 1 0 0 0 0 RN Controllable— not Job related 26 96 12 24 3 l 1 0 0 0 RN controllable— wage-wage related 10 10 6 6 0 0 0 1 0 0 RN controllable— Job related - dissatisfaction 60 4 18 1 0 0 2 0 0 0 RN not controllable involuntary-fired 6 0 1 1 0 0 0 0 0 0 Column Totals 116 131 39 38 3 2 3 1 0 0 Per cent 20.9 23.6 7.0 6.8 .5 .4 .5 .2 - _ VJl VO TABLE 2*1 (Continued) Reason Group 11 INT- 0- FTPT 12 INT-O FTPT- G 13 INT- IND 14 INT- IND- G 15 INT- IND- FTPT 16 I-IND -FTPT -G 17 L 18 L-G 19 E 20 E-G l.RN not controllable- not job related 0 0 1 3 0 0 37 24 0 1 2.RN controllable- not job related 0 0 18 20 0 1 22 31 0 0 3.RN controllable- wage-wage related 0 0 6 4 2 0 1 7 0 0 4.RN controllable- job related- dissatIsfactIon 0 0 12 1 1 0 13 0 0 0 5.RN not controllable' involuntary-fired 0 0 1 0 0 0 0 0 0 0 Column Totals 0 0 38 28 3 1 73 62 0 1 Per cent - - 6.8 5.0 .5 .2 13.1 11.2 - .2 T = r a\ o TABLE 24 (Continued) Reason Group 21 Not Cer tain 22 Extra Job 23 INT- 0- IND 24 INT-O -IND- G 25 INT-O -IND- FTPT 26 INT-O IND- FTPT- G To tals Per cents 27 Grads 28 Grads G l.RN not controllable not Job related 0 0 0 0 0 0 110 19.8 - - 2.RN controllable- not job related 1 10 1 2 1 0 270 48.6 — — 3.RN controllable- wage-wage related 0 2 0 0 0 0 55 9.9 — — 4.RN controllable- job related - dissatisfaction 0 0 0 0 0 0 112 20.1 5.RN not controllable- involuntary-fired 0 0 0 0 0 0 9 1.6 Column Totals(635a) 1 12 1 2 1 0 556 75 4 Per cent .2 2.2 .2 .4 .2 aThere were seven FT hires for,which no mobility information was obtained. The grand total of FT hires is 635 + 7 = 642. No reasons were given for new graduates as h> this was their first job. 162 noted in the 12 flows containing one or no observations, that the intensity type (FTPT) appears six times. Only two flows contained two observations each, and these were (1) INT-FTPT-G and (2) INT-O-IND-G. Three of the flows contained three observations each. These were: (1) INT- FTPT; (2) INT-FTPT-V; and (3) INT-IND-FTPT. When the ob servations contained in the 17 above-mentioned mobility flows are counted, it is seen that they contain only 19 out of the 556 total observations. These 19 observations comprise only 3.4 per cent of the sample. Of the remaining nine mobility flows (see Table 24), the most common was that of an (INT-G) flow. There were 131 nurses (23.6 per cent) which belonged to this flow. The next most common was that of an (INT) move. This flow contained 116 RNs, or 20.9 per cent of all the observations. Labor force flows (L) came next with 13.1 per cent being (L) moves and 11.2 per cent being (L-G) moves. Two vertical flows, (INT-V) and (INT-V-G) con tained 6.8 and 7.0 per cent respectively of all the obser vations. Two interindustry moves were also found to be among the 9 most important flows. The (INT-IND) flow contained 38 RNs or 6.8 per cent of the total sample. The (INT-IND-G) flow contained 28 nurses or 5.0 per cent. The last category or flow to be included in the nine is that of Extra Job. This group accounted for only 2.2 163 per cent of all the RNs in the sample. Mobility Types and Reason Groups In order to facilitate comparison with the litera ture and for analytical purposes, it is useful to collapse the 26 mobility flows into the eight mobility types dis cussed in Chapter I. For sake of completeness we have added one additional mobility type, i.e., non-geographic (NG). These nine mobility types are shown in Table 25. As noted earlier in this dissertation, these mobility types are not mutually exclusive categories. Interoccupatlonal Mobility (0). Table 25 shows that there were only four such moves in our sample. This mobility type accounted for but 0.7 per cent of the total number of moves. The literature review in Chapter II % revealed no nursing studies which contained estimates of the magnitude of this mobility type. There were, however, several non-nursing studies which did consider this cate gory of mobility. From this literature it was noted that women had greater occupational loyalty than did men. The interoccupational mobility rate for all female workers was estimated to be between 30 and 40 per cent. Professional workers had still greater occupational attachment. Their range was anywhere between 15 and 35 per cent. Even TABLE 25 EIGHT MOBILITY TYPES VERSUS REASON GROUPS Reason Group I NT 0 IND L E Ga V FTPT NGa l.RN not controllable— not job related 48 0 4 61 1 56 8 1 54 2.RN controllable— not job related 216 4 43 53 0 175 37 7 84 3.RN controllable— wage related 47 0 12 8 0 28 13 3 25 4.RN controllable— job related— dissatisfaction 99 0 14 13 0 6 21 3 106 5.RN not controllable— involuntary— fired 9 0 1 0 0 1 2 0 8 Total 419 4 74 135 1 266 81 14 277 Per cent of Total Moves (556) 75.4 0.7 13.3 24.3 0.2 47. 8 14.6 2.5 49.8 aThe total of G and one uncertain mover + NG which (266 + 277) fail were excluded. to sum to 556 due to 12 extra job movers a\ -Er 165 allowing for severe definitional differences, the inter- occupational rate for the nurses in our study was dras tically lower than for other professional workers. While the magnitude of the difference between the interoccupa- tional rates of our study and those in the literature was surprising, the general finding was expected. Al though there were only four such moves in our sample, it is noted in Tables 25 and 26 that all four moves were made for RN controllable - not job related reasons. Frictional Mobility (E). There was only one fric tional move found in our sample. We expected that there would be few such moves, but were somewhat surprised that there was only one. This finding could lead one to question the definition which governed this mobility type. While it might be conceded that the 30 day search period was too long, this writer, based on his experience inter viewing the RNs, feels that the results would not have been significantly different had a period of only a week been used. The conclusion to be drawn, then, is that there was virtually no unemployment of nurses in our sample. Unfortunately there was no data found in the literature with which the findings of our study could be compared. TABLE 26 COLUMN PERCENTAGES FOR EIGHT MOBILITY TYPES VERSUS REASON GROUPS Reason Group I NT 0 IND L E G V FTPT NG l.RN not controllable— not Job related 11. 4 5.*» 45.1 100.0 21.1 9.9 7.1 19.5 2.RN controllable — not job related 51.5 100.0 58.1 39.2 — 65.8 45.7 50.0 30.3 3.RN controllable — wage related 11.2 — 16.2 5.9 — 10.5 16.0 21.4 9.0 4.RN controllable — job related — dissatisfaction 23.6 18.9 9.6 2.3 25.9 21.4 38.3 5.RN not controllable— involuntary-fired 2.1 — 1.3 — — .4 2.5 — 2.9 Chi square on Row 1 for Row 2 Row 3 Row 4 INT, IND, L, G, V and NG = 54.1 17.1 7.0 40.7 very significant significant at . significant at . very significant at .005 005 25 at .005 167 Intensity Mobility (FTPT). Prom Table 25 It Is seen that there were lM intensity moves. These constituted only 2.5 per cent of the total. Because we elected to study only full-time nurses, all of these moves were from a part-time to full-time status. As might be expected, a fairly high proportion of them (21.^ per cent, from Table 26) were for wage related reasons. Again, we were unable to find comparable statistics in the literature. Vertical Mobility (V). There were 81 vertical moves noted in our study. This number was 14.6 per cent of the total number of moves. Due to an oversight in programming the cross-tab, we were unable to accurately distinguish the directions of such moves. While the per cent of nurses making this type of job change for wage related reasons was fairly high (16.0 per cent, from Table 26) we are nevertheless led to conclude that a greater proportion of them must have been downward moves. The reasoning for this is that roughly the same propor tion of nurses left for wage related reasons in two of the other mobility types. Also, the chi square test for this reason group was significant only at the .25 level. In addition to this, a relatively large percentage (25.9) of vertical movers left their last Job because they were dissatisfied; hence, the likelihood is greater that they 168 would have to settle for a lower position. Interindustry (IND). It 3s seen In Table 25 that 74 RNs made Interindustry job changes. This was 13.3 per cent of the total number of hires. One of the significant statistics about this mobility type is that only 5.4 per cent of the women made such moves because of RN not con trollable - not job related reasons (from Table 26). This was the lowest percentage of any of the mobility types. The chi square value for this reason group was very signi ficant at the .01 level. The percentages in Table 26 sug gest that, at least in our sample, interindustry moves were most often made for reasons over which the RN had control. The percentage of nurses in this mobility type making the change for wage-related reasons was among the highest (16.2 per cent), but still was not significantly higher than two or three of the other mobility types. Although there were no nursing studies with which to com pare our results, there were several non-nursing ones which allowed comparison. It was found that the inter industry mobility rates for all workers were quite high, i.e., in the 40-60 per cent range. The rate of inter industry mobility for professional workers was somewhat lower — about 30 per cent. As in the case of inter- occupational mobility, there were some definitional 169 problems. Een allowing for these, the rate in our study (13.3 per cent) is very much lower even than that of other professionals. Labor Force Mobility (L). When analyzing labor force moves, two points should be kept in mind. First, the reason associated with this mobility type is the rea son why the RN initially left the labor force. Many years may have elapsed since the nurse withdrew from the labor force. Second, the labor force figures in Table 25 include only the nurses who were re-entering the labor force. New graduates entering the labor force for the first time were excluded. The total (i.e., both entering and re-entering) number of nurses making this type of move can be obtained from Table 24 by adding columns 17, 18, 27, and 28. Table 25 indicates that labor force moves comprised almost 25 per cent of the moves in our study. We would expect this on several grounds, one being that the profes sion is comprised mostly of women, many of whom have family and home responsibilities which sometimes dictate that they withdraw from the labor force. Another reason has to do with the relative ease with which an RN can enter and leave the labor force. As was noted above, there is virtually no unemployment of nurses. A nurse can be on the job two or three days after deciding to 170 return to work. Clearly, the longer the nurse remains out of the labor force the more difficult the re-entry will be. This, however, is mainly because she has forgotten or lost some of her nursing skills. As was expected, Table 26 shows that a relatively high percentage of the nurses in this group initially left their job because of "RN not controllable — not job related" reasons. This was 45.1 per cent. The chi square test for this reason group indicated very significant differences among the mobility types. Approximately 39 per cent of the RNs re-entering the labor force left their last job for "RN controllable — not job related reasons." Many of the nurses in this group took extended vacations abroad, or were simply resting. Relatively few labor force movers left the labor force due to wage-related reasons (5-9 per cent). All of the other mobility types had ten per cent or more of the RNs leaving the last job for this reason. Nothing was found in the literature with which the above results could be compared. One study mentioned in Chapter II, however, did estimate that approximately 17 per cent of the married women left the labor force each year. Geographic Mobility (G). Aside from the inter firm and non-geographic mobility types, the geographic mobility type was the largest group. This group accounted 171 for 266 observations, or 47.8 per cent of all full time hires. Table 26 reveals several Interesting relationships. One of the first things to be noted is that a fairly high percentage (21.1 per cent) of the nurses in this mobility type moved for "RN not controllable — not job related reasons." The only mobility type with a higher percentage for this reason group was that of labor force movement. This finding lends very definite support to the contention that many of the geographic moves made by nurses are for reasons beyond their control, i.e., husband moving to accept another position, etc. Table 26 also shows that 64.8 per cent of all geographic moves were made for "RN controllable — not job related" reasons. This was the highest percentage for any of the mobility types. This, too, suggests that while a large number of geographic moves were made, a significantly large proportion of them were made for non-wage or job related reasons. The most direct support for the idea that wage-induced geographic mobility of nurses was low is seen in the wage reason group in Table 26. Only 10.5 per cent of the geographic moves were made for wage-related reasons. Of the eight (plus the temporarily added NG) mobility types, only RNs making labor force moves and those making NG moves had a lower percentage in this reason group. When the percentages in reason group four (RN 172 controllable — job related - dissatisfaction) are com pared with the respective mobility types, it is noted that nurses making geographic moves had the lowest (2.3 per cent) of any mobility type. This too, while not directly supporting the monopsony argument, surely does it no harm. We would expect nurses making geographic moves to do so less often for reasons of dissatisfaction with the job. In subsequent sections of this chapter, we will examine various nurse characteristics to see whether they alter the tentative conclusion: that the wage-induced geogra phic mobility of nurses is relatively low. A common finding of the mobility studies reviewed in Chapter II was that the rate of geographical mobility of all workers ranged at about five to nine per cent per year. The rate of geographic mobility for all females was slightly less than that for males. Two or three of these studies indicated that the geographic mobility of professional workers was higher than that for all workers, approximately nine to 15 per cent per year. Disregarding for the moment the reason behind geographic moves, it appears that the nurses of our study had a much higher occurrence of geographic moves than did other professional groups. The rate in our study as noted above was 47.8 per cent, while the estimates in the literature ranged up to 15 per cent. Even allowing for the fact that 173 Southern California is probably somewhat atypical as far as labor markets are concerned, the above differences are very significant. Several studies in Chapter II investigated the reasons behind the geographic movement of workers. Ac cording to one study, approximately 49.5 per cent of those men who moved geographically did so for "work related reasons." Only 15.5 per cent of the nurses in our study moved for job-related reasons (groups three and four). Professional workers in the literature were found to make an even higher proportion of geographic moves due to job related factors than that of other workers (i.e., 49.5 per cent). Another study suggested that 33 per cent of all job shifts were to "improve status." For professional women the estimate was about 30 per cent. The term "sta tus improvement" is probably analogous to our "wage-wage related" group (group three). Only 10.5 per cent of the nurses in our sample fell into this group. When this figure is compared with those cited above, it seems rea sonable to suggest that the wage-induced geographical mobility of the nurses in our study was relatively low. Non-geographlc Mobility (NG). The next mobility type to be considered was non-geographic mobility. This type was added to lend completeness to the discussion of mobility types. From Table 25 it is seen that approxi mately 50 per cent of all moves were non-geographic in nature. Table 26 reveals that the greatest single per centage of these moves was for "RN controllable — job related dissatisfaction" reasons. This figure (38.3 per cent) was significantly higher than those of any other mobility type. We would expect this to be the case, par ticularly in a large city such as Los Angeles where there are over 100 hospitals within easy commuting distance. The percentages of the other reason groups reveal that about 19*5 per cent of the non-geographic movers left the job due to "RN not controllable — not job related reasons." Very likely a sizable proportion of these were nurses who were leaving the labor force for pregnancy or other family reasons. Roughly 30 per cent of the nurses who made non-geographic moves did so because of "RN con trollable — not job related reasons." This was the lowest of all the mobility types. The "wage-wage related reason" group was found to account for about nine per cent of the non-geographic moves. This percentage was only slightly lower than that for "wage-induced geographic" moves (10.5 per cent) a fact which further indicates the relative infrequency of the wage-induced geographic mobility of the nurses in our sample. Probably the most significant relationship to be mentioned about the non 175 geographic mobility type is the fact that so many (38.3 per cent) of the nurses in this mobility type made moves because they were dissatisfied with their last job. The last mobility type depicted in Tables 25 and 26 is that of lnterflrm mobility. Because this mobility type includes so many of the others, it does not merit much discussion here. It might, however, be noted that this group includes all of the other mobility types with the exception of the labor force moves and employment or frictional moves. Geographic Mobility by Reason and Demographic Characteristic One of the facts which is probably quite apparent at this point is the almost infinite number of relation ships which could be analyzed. Every relationship which has been discussed so far could be repeated for the four race groups, the five age groups, the three educational groups and the three marital status groups. Clearly such detailed analysis would only tend to obscure many of the conclusions which might otherwise be drawn. For this reason we have decided to limit the scope of the remainder of the analysis to a discussion of geographic mobility. We will consider the reasons associated with this mobility type in connection with the four demographic characteris tics mentioned above. 176 Race. Table 27 shows the number of hires in our sample who made a geographical move by race. Table 28 shows the corresponding percentages. A row chi square test which was run on the geographic mobility type indi cated that there were very significant differences between the various racial groups. Filipinos were found to have the highest percentage (7^.2) of geographical moves of any other racial group. The next highest group, Negroes, made geographic moves ^5 per cent of the time. The finding that Filipinos made a large proportion of geogra phic moves in our study might have been expected. In Chapter IV it was noted that they had higher turnover rates and shorter lengths of stay than other racial groups. This, in addition to the fact that most of them remain in the United States a relatively short time, desiring to see as much of the United States as possible, makes the above finding seem reasonable. It is noteworthy that when the Filipinos were taken out of the sample and the chi square test was again run, the value was not signi ficant at the .25 level. This suggests that Filipinos were responsible for a disproportionately large part of the total geographic mobility. To enable a further investigation of the effects of race on geographic mobility, a cross-tab was construct ed which related mobility flow-reason to race. TABLE 27 NUMBER OF GEOGRAPHIC MOVES BY RACE Kind of Move Cauc Negro Fil Other Total Geographic 100 18 118 29 265 Non-geographic 180 22 41 46 289 Total 280 40 159 75 554 TABLE 28 PERCENTAGE OF GEOGRAPHIC MOVES BY RACE Kind of Move Cauc Negro Fil Other Ave. Chi Square Geographic 35-7 45.0 74.2 38.7 48.4 22.6 Sig at .005 Non-geographic 64.3 55.0 25.8 61.3 51.6 18.1 Sig at .005 Table 29 shows by race and reason group, the number of nurses who made geographical moves. Table 30 contains the corresponding percentages. The cells In Table 30 show the percentage of each of the race groups leaving for a particular kind of reason. Chi square tests were run on each row in order to determine whether or not the observed differences attributable to race were signifi cant. It was found that the differences for reason groups one and three were very significantly different from the average. First, the column for Filipinos demon strates that a very small percentage (7.6) of them left their last job because of "RN not controllable — not job related reasons." This might be expected on the grounds that many of them were single. Also, some of the married Filipinos with families, had left them temporarily in the Philippines. A high percentage (72.0 per cent) left for "RN controllable — not job related reasons" which would also be expected given their situation. It also is seen that a relatively high percentage (16.9 per cent) left their last job for wage or wage related reasons. This value was significant at the 0.25 level. Contrasting the case of the Filipino with the overall findings on geographic mobility described above (Tables 25 and 26) it is observed that while the nurses in general had high rates of geographic mobility (^7.8 TABLE 29 NUMBER OF GEOGRAPHIC MOVERS BY RACE AND REASON GROUP Reason Group Cauc Negro Fil Other Total Per cent 1. RN not controllable — not job related 36 4 9 6 55 20.8 2. RN controllable — not job related 57 13 85 20 175 66.0 3. RN controllable — wage - wage related 5 1 20 2 28 10.6 4. RN controllable — job related — dissatisfaction 2 0 3 1 6 2.3 5. RN not controllable — involuntary-fired 0 0 1 0 1 .4 Total 100 18 118 29 265 TABLE 30 PER CENT OF GEOGRAPHIC MOVERS BY RACE AND REASON GROUP Reason Group Cauc Negro Fil Other Ave. Chi Square 1. RN not controllable — not job related 36.0 22.2 7.6 20.7 21.6 18.7 Sig at .005 2. RN controllable — not job related 57.0 72.2 72.0 70.0 67.8 2.3 Not sig at .25 3. RN controllable — wage-wage related 5.0 5-5 16.9 6.9 8.6 11.0 Sig at .025 4. RN controllable — job related — dissatisfaction 2.0 2.5 3.4 2.6 0.4 Not sig at .25 5. RN not controllable — involuntary— fired — — .8 — — per cent), much of this was attributable to "RN not con trollable" reasons, and relatively little to "wage induced mobility." Filipinos, on the other hand, had a still higher geographic mobility rate (74.2 per cent), relatively more of which was attributable to "RN con trollable" reasons, with "wage-wage related" being par ticularly high. This seems to suggest that if this "unique" RN group were removed from the sample, the con clusions drawn above about geographic mobility in general might be stated even more positively. As a test for this hypothesis, we compared in Table 31 the number of geogra phic moves by reason for the sample, both with and with out Filipinos. Percentages were calculated, and the above possibilities were confirmed. The "RN not control lable — not wage related" percentage rose more than 10 percentage points (from 21.1 to 31.8 per cent) while the "wage-related" percentage dropped from 10.5 per cent to 5.4 per cent. Both of these changes were in a direction favorable to contentions of Chapter I, and the conclusions drawn above. Marital Status. Table 32 shows the count of RNs making geographical moves by marital status. Table 33 shows the associated percentages. From Table 33 it is seen that 57.4 per cent of all single nurses, 43.9 TABLE 31 GEOGRAPHIC MOBILITY TYPE BY REASON GROUP WITH AND WITHOUT FILIPINOS Reason Group Total Number RNs Per Cent Number of RNs— Filipinos Excluded Per Cent 1. RN not controllable— not job related 56 21.1 47 31.8 2. RN controllable — not job related 175 65.8 90 60.8 3. RN controllable — wage-wage related 28 10.5 8 5.4 4. RN controllable — job related - dissatisfaction 6 2.3 3 2.0 5. RN not controllable — involuntary-fired 1 .4 0 — Total 266 148 co r\j TABLE 32 NUMBER OF GEOGRAPHIC MOVERS BY MARITAL STATUS Kind of Move Single Married Other Total Geographic 143 108 15 266 Non-geographic 106 138 46 290 Total 249 246 61 556 TABLE 33 PERCENTAGE OF GEOGRAPHIC MOVERS BY MARITAL STATUS Kind of Move Single Married Other Average Chi Square Geographic Non-geographic 57.4 42.6 43.9 56.1 24.6 75.4 42.0 58.0 12.9 Sig at 9.4 Sig at .005 .05 184 per cent of all married nurses and 24.6 per cent of "other" nurses in our sample made geographic moves. A chi square test indicated that there were significant differences attributable to marital status. The result for married and single nurses was expected. In Table 34 we find the number of single, married and other nurses who made geographic moves by reason group. Table 35 shows the percentage distribution of each marital group by reasons for the geographic Job changes. Chi square tests were made on each of the rows, and these values indicated significant differences attributable to marital status. As was expected, the percentages of single and "other" nurses making geographical job changes for "RN not controllable — not job related" reasons were very low, 4.2 and 6.7 per cent respectively. Married nurses on the other hand, made such moves 45.4 per cent of the time. Again, in the case of the "RN controllable — not job related" reason group, single and "other" nurses made such moves a very large percentage of the time — 81.8 and 73.3 per cent respectively — while the per cent of married RNs was much lower, 43.5 per cent. Although the differences were not highly significant, Table 35 indicates that "other" nurses left more often for "wage and wage related" reasons than did married and single nurses. The percentages for all three were 9.1» 11.1 TABLE 34 NUMBER OF GEOGRAPHIC MOVERS BY MARITAL STATUS AND REASON GROUP Reason Group Single Married Other Total Per cent 1. RN not controllable — not job related 6 49 1 56 21.1 2. RN controllable — job related 117 47 11 175 65.8 3. RN controllable — wage-wage related 13 12 3 28 10.5 4. RN controllable — job related - dissatisfaction 6 0 0 6 2.3 5. RN not controllable — involuntary-fired 1 0 0 1 .4 Total 108 15 266 TABLE 35 PER CENT OF GEOGRAPHIC MOVERS BY MARITAL STATUS AND REASON GROUP Reason Group Single Married Other Average Chi Square 1. RN not controllable — not Job related 4.2 45.4 6.7 18.8 56.8 Sig at .005 2. RN controllable — not Job related 81.8 43.5 73.3 66.2 12.2 Sig at .005 3. RN controllable — wage-wage related 9.1 11.1 20.0 13.4 5.0 Sig at .25 4. RN controllable — Job related - dissatisfaction 4.2 5. RN not controllable — involuntary-fired .7 — — 187 and 20.0 per cent respectively. Of the three marital groups, only single nurses made geographical moves due to "job dissatisfaction," i.e., reason group four. This figure was 4.2 per cent, and supports an earlier finding, i.e., that most job changes for reasons of job dissatis faction are non-geographic in nature. Although nothing was specifically cited in the literature review, one usually thinks of single women as being more geographically mobile than married women. We have just shown in our sample that this was the case. The percentages for single and married nurses respectively were 57.4 and 43.9 per cent. It was noted that almost half of the geographic moves made by married nurses were for "RN not controllable — not job related reasons." It is instructive to calculate the above geographic mobility rates for only those moves which were "RN controllable." When this is done the rates for single and married nurses respectively are 54.6 and 24.0 per cent. Single women are more than twice as mobile when only "RN con trollable" moves are considered. Age. The next demographic characteristic to be considered is that of age. In Table 36 we find the number of nurses in the various age groups who made geo graphical moves. In Table 37 the percentages are TABLE 36 NUMBER OF GEOGRAPHIC MOVERS BY AGE GROUP Type of Move Under 26 26-29 30-39 40-49 50-0ver Total Geographic 93 79 57 24 13 266 Non-geographic 18 56 76 40 30 220 Totals 181 135 133 64 43 556 TABLE 37 PER CENT OF GEOGRAPHIC MOVERS BY AGE GROUP Type of Move Under 26 26-29 30-39 40-49 50-Over Average Chi Square Geographic 51.4 58.5 42.9 37.5 30.2 44.1 11.3 Sig at .025 Non-geographic 48.6 41.5 57.1 62.5 69.8 55.9 10.1 Sig at .05 calculated. From this table it can be noted that 51.4 per cent of the RNs under 26 moved geographically, while 58.5 per cent of those 26-29 years old did so. About 42.9 per cent of the nurses 30-39 made such moves, and the percentages for the 40-49 and 50 and over groups were 37.5 and 30.2 per cent respectively. A chi square test revealed that there were significant differences in the percentages attributable to age differences. As one would expect, with only the exception of the 26-29 year group, the percentage of nurses making geographic moves declined with age. The geographic mobility rate of the 26-29 year old group may have increased over that of the under 26 group due to marriages and moves associated with young married men in search of their life's work. Table 38 shows the number of nurses who made geo graphic moves by reason group while Table 39 shows the percentage of each age group making moves for the five reason categories. Chi square tests were again made to determine whether or not there were significant differences attributable to the age factor. lor the reason group "RN not controllable — not job related" the differences were significant. Here it was noted that the percentage of nurses under 26 who made such a job change was con siderably higher than that for RN in other age groups up to 50 and over. In fact, the percentage declined with age TABLE 38 NUMBER OF GEOGRAPHIC MOVERS BY AGE AND REASON GROUPS Reason Group Under 26 26-29 30-39 40-49 50-Over Total Per Cent l.RN not controllable— not Job related 24 16 10 1 5 56 21.1 2.RN controllable — not Job related 62 53 35 18 7 175 65.8 3.RN controllable — wage-wage related 6 8 9 4 1 28 10.5 4.RN controllable — Job related - dissatisfaction 1 2 3 0 0 6 2.3 5.RN not controllable— involuntary-fired 0 0 0 1 0 1 .4 Totals 93 79 57 24 13 266 TABLE 39 PER CENT OF GEOGRAPHIC MOVERS BY AGE AND REASON GROUP Reason Group Under 26 26-29 30-39 40-49 50-Over Ave. Chi Square l.RN not controllable— not job related 25.8 20.3 17.5 4.2 38.5 21.3 29.3 Sig at .005 2.RN controllable — not Job related 66.7 67.1 61.4 75.0 53.8 64.8 3.8 Not Sig .25 at 3.RN controllable — wage-wage related 6.5 10.1 15.8 16.7 7.7 11.4 7.6 Sig at . 25 4.RN controllable — job related- 1.1 dissatisfaction 2.5 5.3 — — 3.0 3.0 Not Sig .25 at 5.RN not controllable— involuntary-fired — — 4.2 — — 192 up to the last age group. This was contrary to expecta tions. One might have expected that the percentage would increase at least up through age 39, due to the greater probability of an RN being married up to that age. The chi square test on reason group "RN control lable - not job related" was not significant. Most of the percentages for the various age groups were high (65-75 per cent) and at about the same level. There was a slight tendency for younger age groups to have a higher per cent. Another somewhat unexpected finding was that the percentage of RNs changing jobs for wage-related rea sons increased with age up through 49 years old. The chi square test indicated that these differences were signi ficant. No explanation for this was readily apparent, unless the interrelationship between age and marital status (married women tended to be older and had a slightly higher percentage of wage-related moves) was responsible. It should be noted, however, that the number of observa tions in the group was quite small. Overall, the effects of age on geographic mobility were less clear than the other demographic characteristics discussed above. It is certain that other characteristics (i.e., race, marital status, education, etc.) were interacting with the age variable, which is one .of the weaknesses of this kind of descriptive study. 193 Educational Level. The literature review indi cated that workers with greater skills or higher educa tional levels were more likely to make geographic moves than other workers. We expected that RNs with BS degrees were more geographically mobile than other RNs. Table 40 shows the number of nurses who made geographical moves by educational level. Table 41 con tains the respective percentages. In that table it can be seen that the percentage of nurses making geographic moves increased with the number of years in nursing school, i.e., the percentage for AA degree holders was 29.8 for diploma holders 46.5 and for BS and higher degree holders it was 61.7. The chi square test indicated that these differences were very significant. In Table 42 we find the count of RNs in each educational group classified by reason. In Table 43 the corresponding percentages are found. It is seen from Table 43 that the percentage (35.7) of AA degree nurses who left their last job due to "RN not controllable — not wage related reasons" was higher than in the other two groups. The percentage for BS degree RNs was lowest at 15.9 per cent. The chi square test indicated that these differences were significant. Here again we have the interaction problem. Many of the AA degree nurses were older married women, while a high proportion of the BS TABLE 40 NUMBER OF GEOGRAPHIC MOVERS BY EDUCATION Type of Move AA Diploma BS+ Total Geographic 14 173 82 269 Non-geographic 33 199 51 283 Totals 47 372 133 552 TABLE 41 PER CENT OF GEOGRAPHIC MOVERS BY EDUCATION Type of Move AA Diploma BS+ Ave. Chi Square Geographic 29.8 46.5 61.7 46.0 11.1 Sig at .005 Non-geographic 70.2 53.5 38.3 54.0 9.4 Sig at .01 i —1 v o j = - TABLE 42 NUMBER OF GEOGRAPHIC MOVERS BY REASON AND EDUCATION Reason Group AA Diploma BS+ Total Per Cent 1. RN not controllable — not job related 5 36 13 54 20.1 2. RN controllable — not job related 8 115 53 176 65.4 3. RN controllable — wage-wage related 1 19 11 31 11.5 4. RN controllable — job related- dissatlsfaction 0 3 4 7 2.6 5. RN not controllable — involuntary-fired 0 0 1 1 .4 Totals 14 173 82 269 TABLE 43 PER CENT OF GEOGRAPHIC MOVERS BY REASON AND EDUCATION Reason Group AA Diploma BS+ Ave. Chi Square 1. RN not controllable — not job related 35.7 20.8 15.9 24.1 8.8 Sig at .025 2. RN controllable — not job related 57.1 66.5 6b.6 62.7 0.8 Not Sig at .25 3. RN controllable — wage-wage related 7.1 11.0 13. b 4.8 24.5 Sig at .005 4. RN controllable — job related - dissatisfaction w e 1.7 b.9 3.3 1.6 Not Sig at .25 5. RN not controllable — involuntary-fired - — 1.2 197 degree holders were Filipinos who were single or had no immediate family constraints. The differences attributable to education were not significant with respect to "RN controllable — not job related" reason group. These percentages were fairly high, i.e., 57.1 to 66.5 per cent. There were significant differences with respect to the wage related reason group. A greater proportion of BS degree holders made such moves than did either of the other two educational categories. The percentages were 7.1, 11.0 and 13.4 for AA, diploma, and BS degree nurses respectively. Problems. As was discussed in Chapter IV, one of the problems associated with this kind of analysis is that it is very difficult to take into consideration the interaction of two or more demographic (or other) charac teristics. We had initially planned to run several cross tabs, adding an extra demographic dimension to each. Two problems became immediately apparent. The first had to do with our limited sample size (642 full time hires). Each additional dimension greatly increased the possible number of cells. We found that by carrying the cross tab beyond the reason-flow-age, or race, etc., level, there were so many null sets that it was almost meaning less to analyze. The second problem had to do with the 198 sheer magnitude of such a task. An added dimension could increase the potential number of cross-tabs by 12 or 15. It becomes almost impossible to discuss adequately the multitudes of relationships. For these two reasons, then, we have elected to step aside from the interaction anal ysis. Reason Distributions of Chapter IV and V Compared It will be recalled that the reason groups analyzed in Chapter IV contained the reasons nurses left our sample hospitals. Those given in the present chapter refer to the reason the RN left her last job (whenever and wherever that may have been). We now attempt to compare the per centage distributions of the two groups to determine how well the samples from essentially two different popula tions coincide. In Table 44 we show the number and percentages of hires and terminations by reason group. The overall chi square calculated in columns one, three and five showed that there were significant differences between both fre quency distributions. When chi square tests on the five rows were conducted, three of them showed significant dif ferences at the .25 level or better, while two were not significant at this level. from the table it is seen that a greater percentage of RNs (25.4) left our sample for TABLE 44 NUMBER AND PER CENT OF TERMINATIONS AND HIRES BY REASON GROUP 1 2 Terminationsa 3 4 Hiresb 5 Total Hires and 6 Ave. Reason Group Number Per cent Number Per cent Terms. Per cent 1. RN not controllable not job related 131 25.4 110 19.8 241 22.6 2. RN controllable — not job related 186 36.1 270 48.6 456 42.4 3. RN controllable — wage-wage related 26 5.0 55 9.9 81 7.5 RN controllable — dissatisfaction 127 24.7 112 20.1 239 22.4 5. RN not controllable— involuntary-fired 46 8.9 9 1.6 55 5.3 Totals 516 556 1,072 aExcludes 13 for whom no reason was given ^Excludes new graduates (79) and (7) RNs for whom no reason was given. Overall chi square = 52.1 Sig at .005 level Row 1 = .7 Not sig at .25 Row 2 = 1.8 Sig at .25 Row 3-1.6 Sig at .25 Row 4 = .5 Not sig at .25 Row 5 = 5.0 Sig at .05 V O V O 20.0 "RN not controllable — not job related reasons" than those who moved into it (19.8). However, these differ ences were not very significant. A greater proportion of RNs moving into our sample did so for "RN controllable— not job related reasons" than did those who left it. The percentages were 48.6 and 36.1 respectively. This dif ference was significant at only a .25 level. More women moved into the sample for wage related reasons (9*9 per cent) than those who left it (5.0 per cent). Further reference to Table 44 shows that there were no signifi cant differences between the percentage of hires and ter minations who moved for "RN controllable — job related dissatisfaction" reasons. Significant differences were noted in the last reason group, i.e., the involuntary fired group. We feel that the 1.6 per cent for hires is somewhat understated. The reason probably lies in the fact that some new hires may not have told the interviewer the com plete truth about leaving the previous job. The 8.9 per cent found for terminations in our sample was quite high in relation to the nursing literature and that cited above for hires. This might have been the case due to the large number of foreign speaking nurses in our sample. In cluded in this group were Filipinos (by far the most numerous), Koreans, Japanese, Chinese, Thai nurses and several nurses from Europe. Movement to Areas of Net Economic Advantage 201 One of the most common findings in the literature was that geographic moves were from regions of low net advantage to regions of high net advantage. While this finding cannot be fully tested here, some information was obtained from the new hires which sheds light on this sub ject. Most of the new hires were asked whether their new salary would be more than, less than, or the same as the old salary. By cross-classifying these responses with the move type we can at least determine which kinds of moves are most often associated with an improved financial position after the move. In Table 45 we have constructed such a cross classification. The three kinds of moves considered are (1) interstate moves; (2) local moves, i.e., moves of less than 30 miles; and (3) intra-California moves, i.e,, moves within the state of 30 miles or more. The corresponding percentage figures are presented in Table 46. Chi square tests were conducted on each row, and very signi ficant differences were noted for two groups, i.e., RNs making more, and those making less. Of the nurses moving into California, 69.4 per cent earned more money here than at their previous job, while only 11.3 per cent earned less. Nurses making local moves earned more only TABLE 45 COUNT OF NURSES BY FINANCIAL SITUATION BY MOVE TYPEa Financial Situation After Move Interstate Move Local Intra- California Total Makes More 86 33 9 128 Same 16 27 5 48 Less 14 57 4 75 Don't Know 8 8 0 16 Totals 124 125 18 267 It should be noted that nurses making labor force, inter-occupatlonal and inter industry moves were excluded from the Table. Wage comparisons for RNs making these kinds of moves are not meaningful. t\j o ro TABLE 46 PERCENTAGE OF NURSES BY FINANCIAL SITUATION BY MOVE TYPE Financial Situa tion After Move Interstate Move Intra- Local California Ave. Per Cent Chi Square Makes More 69.4 26.4 50.0 48.6 19.1 Sig at .005 Same 15.3 21.6 27.8 21.6 3.6 Sig at . 25 Less 11.3 45.6 22.2 26.4 23.3 Sig at .005 Dorft Know 6.5 6.4 ro o co 20 4 26.4 per cent of the time, while they made less 45.6 per cent of the time. This finding in our sample supports to some degree the finding in the literature. When the intra- California moves are considered, the finding is the same, though somewhat weaker. These findings may appear to contradict the earlier findings — that the wage-induced geographic mo bility of nurses is relatively low. This is not the case. As mentioned above, the reason for making the job change is not a consideration here, but only the financial status of the RN before and after changing jobs. It is entirely possible that an RN could move for "RN not controllable — not job related reasons" and, nevertheless, find that her wage was higher in Los Angeles than at her prior location. It is equally possible that a single nurse, who made her job change for "RN controllable — not job re lated reasons," for example, to travel, or to be with friends, or to live in a warmer climate, could also find that her wage was higher here than at the location of her previous job. In other words, the findings in Table 46 in no way contradict those of earlier sections of this chapter. There is one last point to be made from the per centages in Table 46. The fact that such a large pro portion of local movers (45.6 per cent) eventually earn 205 less at their new job lends support to the monopsony argument. This finding suggests that because the wage levels in the area are roughly the same, an RN making a job change is penalized for moving by being dropped back a pay step. We would expect this in a local market characterized by market imperfections. This lack of wage-induced movement in the local market, and the lack of wage induced geographic mobility between markets tends to interfere with the operation of the market system and prevents the "optimal" allocation of nursing resources. Summary The overall purpose of this chapter has been to analyze the data with respect to the various mobility pat terns or flows. We have attempted to measure the magni tude of these flows, and have also described the effects of various demographic characteristics upon them. A major difference exists between this and the preceding chapter. In the present chapter we were interested in the situation of the RN both before and after her move, while in Chapter IV, we were simply interested in the fact that she had made a Job change. Of necessity, the con centration of the present chapter was upon hires. A cross-tab was constructed which related the 28 mobility flows to the five reason groups. It was noted 206 that there were few or no observations in 17 of the 28 mobility flows. This suggested that for our purpose a less detailed breakdown of flows might be warranted. Several of the most frequented flows were the following: (1) INT-Gj (2) INT; (3) L; (4) L-G; and (5) INT-V-G. Having briefly described the more important flows, it was determined that by collapsing the 28 mobility flows into the eight mobility types, useful comparisons and analyses could be made. It was shown the lnteroccupatlonal moves were made by only 0.7 per cent of the nurses in our sample. This percentage was drastically lower than those found for other professions. Frictional mobility or movement back into employment from unemployment was also rare. Movement from part-time to full-time status was made by only 2.5 per cent of the nurses in our study. A sizable proportion of these were made for wage related reasons. Vertical movement was made by 14.6 per cent of the RNs studied. Many of these left their last job due to dissatisfaction. Interindustry movement was made by 13*3 per cent of our sample. The greatest percentage of these moves were for reasons over which the RN had control. This type of mobility was much lower than that found in the literature. Labor force movement was one of the most frequently made mobility types, almost 25 per cent of the 207 nurses in the study made such a move. As expected, a relatively high percentage (45.1) of these movers origin ally left the labor force for "RN not controllable — not job related reasons." It was noted that almost 50 per cent of moves made in our study involved geographic moves. It was seen, however, that a relatively small percentage of these moves were made for wage related reasons, while a relatively large percentage (21.1) were made for "RN not controllable — not job related reasons." Non-geo- graphlc mobility was also considered. Here it was seen that approximately 38.3 per cent of the moves in this mobility type were made for "RN controllable — job rela ted dissatisfaction" reasons. Due to the fact that our major concern dealt with the geographic mobility of nurses, the remainder of the analysis was concentrated on this mobility type. The affects of four demographic characteristics upon this mobility type and reason groups were investigated. One of the most significant findings of the study concerned the affect which Filipinos had on the overall geographic mobility statistics. They were found to have made geo graphic moves 74.2 per cent of the time. Negroes were somewhat more geographically mobile than whites. All nurses in our study were more geographically mobile than other worker groups. Filipinos made a greater percentage 208 of geographic moves for wage related reasons than other racial groups. Marital status was found to be an important determinant of geographic mobility. Approximately 57.4 per cent of the single nurses and 43.9 per cent of the married ones made such moves. As expected, married women most often (45.4 per cent of the time) moved for "RN not controllable — not job related reasons." The young nurses (under 30) in our study were more geographically mobile than older ones; however, it appeared that older nurses tended to be more interested in wages than young nurses. It was seen that nurses with BS degrees made a greater percentage of geographic moves than AA or diploma nurses. One reason, however, probably had to do with the fact that so many Filipinos were BS degree holders. AA degree nurses made fewer geographic moves, and a rather high percentage (35-7) of these were for "RN not control lable — not job related" reasons. When the distributions of reasons for job change for Chapter IV and the present chapter were compared, it was seen that generally fewer than ten percentage points separated any reason group. It was seen, however, that a higher proportion of hires left their last job for "RN controllable — not job related reasons" than those who terminated from our sample. A much higher percentage of 209 the nurses who left our sample did so because they were fired than did those RNs who were beginning work in our sample. The percentage of nurses leaving a job for wage related reasons was higher among new hires than termina tions . It was finally shown that the percentage of nurses making more money after the Job change than before was very much higher among those who moved geographically. This fact, it was suggested, did not conflict with the monopsony-oligopsony argument. It was also seen that a substantial proportion (^5.6 per cent) of local movers earned lower wages at their new job than they did at their old one. The subsequent and final task of this dissertation is to present the conclusions which can be drawn from an analysis of the preceding chapters. This we undertake in Chapter VI. CHAPTER VI SUMMARY AND CONCLUSIONS One of the major objectives of this study was to design a survey which would obtain information on the mobility patterns of Registered Nurses. As was mentioned in Chapter I, a lack of such data has greatly impeded research into the question of the "shortage" of nurses. Wage theory suggests that an economic shortage cannot persist long in a competitive market. In other words, by responding to their own best interests, nurses would leave low wage areas (causing the wage levels to rise) in favor of areas in which wages were higher, thus lowering the wages. Wage levels would also regulate the entry and exits of nurses into and out of the nursing profession. The concept of mobility is the mechanism by which these adjustments take place. It was suggested that the market for nurses was not competitive, but rather is characterized by monopsony and/or oligopsony; hence, the adjustment mechanism was unable to operate. If this is the case, then the reported shortage of nurses is due only to the imperfections of the 210 211 market. It was suggested that In order for the monop sony hypothesis to be correct, three conditions must be met, i.e., there must be: (1) profit maximizing behavior on the part of the hospital administrators; (2) few em ployers of nurses; and (3) little wage induced geographic mobility of nurses. Evidence supporting the first two conditions was fairly conclusive, but that supporting the third condition was only circumstantial and needed empiri cal verification. By carefully tracing out the movement patterns of nurses, classified by reason for move and certain demographic characteristics, it was hoped that such evidence could be produced. The remainder of Chapter I consisted of an explana tion of the definitions and terms to be used in the body of the study. Most Important of these was the distinc tion between turnover and mobility. Turnover is simply a count of terminations or hires as viewed by the hospi tal. Mobility, on the other hand, is an analysis of the nurse's situation both before and after her job change. A distinction was made among voluntary, involuntary, and "RN not controllable" movement. The various mobility flows and mobility types were discussed and differentiated. Geographic mobility was the most important of those men tioned. The definition section was then followed by a brief outline of the remainder of the dissertation. 212 Chapter II contained a review of some of the more relevant literature on mobility and turnover. One of the facts which became apparent in this chapter was the almost complete absence of nurse mobility data. Although there were several non-nursing mobility studies, comparison with the present study was sometimes very difficult. Chapter III contained an outline of the tasks and relationships to be investigated in Chapter IV and V. It was pointed out that the analysis and discussion of turnover would comprise Chapter IV, while the treatment of the mobility would be in Chapter V. This last mentioned Chapter is the most important because it is primarily on the basis of the analysis carried out here, that the monopsony hypothe sis of Chapter I stands or falls. The last section of Chapter III consisted of a description of the survey design and the data collected. Turnover Findings The findings of the turnover analysis of Chapter IV, while of considerable interest to hospital adminis trators, shed little additional light on the questions raised in Chapter I. Nevertheless, some of the more significant conclusions to be drawn from the turnover data will be mentioned. A comparison of the turnover rates in the 18 sample hospitals with those found in other 213 studies indicated that the overall rate for our sample seemed to be somewhat lower. Compared to the turnover rates of other non-nursing groups of workers, the 36 per cent turnover rate of our study was significantly higher. The only exception was in the case of manufacturing workers. These findings simply substantiated those other studies. When the turnover rates of the 18 individual hospitals were subjected to a chi square test, significant differences were noted among the hospitals. Regressions showed that the presence of a retirement plan and overtime payment for weekends significantly reduced turnover in our sample. The "safeness" of the area as measured by the robbery rate per 10,000 persons was also a factor related to turnover. The fact that few wage differences among the hospitals were noted provided some evidence of market imperfections (i.e., wage rates were suggested by the Hospital Council) and also explained their lack of signi ficance in the regressions. Other institutional factors such as size, ownership, etc, were not significant. These findings indicate that there are some measures available to hospital administrators which could reduce somewhat the high turnover rates. The calculation of refined turnover rates provided some insight into the probable effect of certain 214 demographic characteristics In turnover. As expected, turnover was found to vary Inversely with age of the RN. Due to problems with the data, no definite conclusion could be reached as to the turnover of married versus single RNs. However, it was shown that divorced, separated and widowed nurses had almost twice as great a turnover rate as married and single nurses. The educa tional status was found to be a factor in our sample, with AA degree holders having the highest turnover rates. BS degree holders also had high turnover rates, but possibly due to interaction with race. Filipinos had the highest turnover of any racial group. It was seen that nurses who worked nights had higher turnover rates than those working evenings or days. The above findings suggest that turnover is af fected by demographic factors, and while the employers of nurses are not always able to select or reject nurses on such a basis, they nevertheless should incorporate this information into their staff management and planning programs. Another finding of our study which should be of interest to nurse employers concerns the expectancy of service of RNs. The results In our sample indicated that the expectancy of service of terminating nurses was very short. Approximately 60 per cent of those RNs who 215 terminated did so within one year. This was much shorter than for non-nursing workers. This information could also be of use to hospitals. An examination of the number of terminations by reason revealed that approximately 3^.3 per cent of the terminations in our sample were for reasons over which the RN had no control. It was further found that the reason for termination was affected by the demographic characteristics mentioned above. The average length of stay of all nurses in our sample was 631 * days. When this figure is compared with the turnover rate (36 per cent) for the 18 hospitals, it suggests that most turn over was among nurses with relatively short lengths of stay. This fact was verified by the expectancy of service measure cited above. Once again, the demographic charac teristics affected the magnitude of the dependent variable, i.e., the average length of stay. The findings here sup ported the earlier turnover findings. Regressions which were conducted using the average length of stay as the dependent variable and various RN and hospital charac teristics as the independent variables further substan tiated many of the above findings. The final point to be made regarding the turnover chapter is that there definitely are factors which have an influence on magnitudes of the several turnover measures. 216 Any Information concerning the magnitude of these factors can be of great value to hospitals concerned with solving the problem of high nurse turnover. Mobility Findings In Chapter V we constructed a cross-tab which re lated the 28 mobility flows to the five reason groups. The analysis of these mobility flow (and type) relation ships, it was hope<J, would shed additional light on the nature of the market for registered nurses. Information regarding movement patterns of RNs, classified by reason for Job change, was specifically needed. It became apparent that the 28 mobility flow classifications were more detailed than necessary; several flows contained only one or two observations, if any. Two of the remaining 11 categories pertained to new gradu ates for whom reason for Job change was not applicable. These categories were discussed separately from the other mobility flows. The percentage of all new hires falling into each of the remaining nine flows was then given and their significance discussed. For purposes of comparison, and to facilitate meaningful analysis, the 26 flows (excluding new graduates) were compressed into the eight mobility types. The per centages of RNs making moves in each mobility type were noted. These were then broken down by the five reason groups. Of particular interest to this study was the high rate of geographic mobility, i.e., 47.8 per cent of all moves. Based on this finding alone, one might con clude that wage-induced geographic mobility is relatively high (rather than low), and that the monopsony hypothesis should be rejected. When these geographic moves were broken down by reason, it was noted that only 10.5 per cent of them were made for wage induced reasons. This was only slightly higher than the percentage of wage in duced moves of non-geographic Job changers. This latter figure was 9.0 per cent. The percentage of wage induced geographic mobility was lower than most other mobility types. While the nurses in our study were found to be much more geographically mobile than other worker groups, the rate of wage induced job change was very much lower. The conclusion reached on the basis of the above suggests that the monopsony hypothesis cannot yet be rejected. Further support for the contention that wage- induced geographic mobility is relatively low was found when other reasons behind job changes were explained. In Chapter I it was suggested that many nurses who make geographic job changes do so for reasons beyond their control. Not considering demographic characteristics, 218 It was shown In Chapter V that approximately 21 per cent of the grographlc moves in our study were for RN not controllable — not job related reasons. The percentage for married women, as expected, was much higher, I.e., 45.4 per cent. In addition to the above, about 43.5 per cent of the married women who made geographic moves did so for reasons over which they had control, but which were not job or wage related. When the same analysis was carried out for single women it was noted that a somewhat greater proportion of them made geographic moves than did married women (57.4 compared to 43.9 per cent). The distribution by reason was considerably different, but still showed that rela tively few moves were made for wage-related reasons (9.1 per cent). As expected, few single RNs made geographic moves for "RN not controllable — not job related" rea sons. This was only 4.2 per cent. On the other hand, however, 81.8 per cent of them made moves for "RN con trollable — not job related" reasons. This too offers substantial support for the low wage-induced geographic mobility of nurses. On the basis of these and other find ings, we are led to conclude that the monopsony hypothe sis cannot be rejected. The geographic mobility by reason and other demo graphic characteristics were also investigated. Most of 219 these findings substantiated the findings of the litera ture. It was found that young RNs (29 and under) were more geographically mobile than older RNs. The distribu tion of reasons by age were largely as one would expect. There was a tendency, though not strong, for older RNs to make more wage-induced moves. This was unexpected but might be explained on the grounds that many older nurses were also married, and as shown above, married nurses had a tendency to make more wage induced job changes. It was found that BS degree holders were more geographically mobile than diploma graduates who in turn were more mobile than AA degree graduates. Of the three educational groups, BS degree holders were most likely to make wage induced geographic moves. This, however, was due in part to the fact that so many of them were Fili pinos, who, as was shown, were the most geographically mobile of any racial group. Of the four demographic characteristics consid ered race probably had the most significant effect. This, however, was largely due to the presence of a very highly geographically mobile group of Filipinos. Approximately 7^.2 per cent of them made geographical moves in our study. When this racial group was removed from the sample, it was noted that the percentage of geographic 220 moves made for wage related reasons dropped still lower from 10.5 to 5.4 per cent. Also the per cent of "RN not controllable — not job related" reasons increased by more than 10 per cent. Both of these effects offer still further support for the conclusions reached concerning the low geographic mobility of nurses. A cross-tab relating financial situation after job change to the move type revealed that the 69.4 per cent of those who made geographical moves in our study ended up making more, while only 26.4 per cent of the local movers did so. These figures lend support to the hypo thesis that net geographic movement is from areas of low to areas of high wages. This finding of our study in no way detracted from the above conclusions, because it did not consider reason for move, but rather the financial status after move. In other words, it is very probable that an RN who moved because of her husband's job change could end up financially better off in her new Job. There were several problems which were encountered during the course of this study. One of them was the al most total lack of any nursing mobility statistics with which to compare our findings. It would appear that further studies should be undertaken using similar defini tives and methodology before the conclusions and rela tionships pointed out here can be fully accepted. There 221 is a particular need for such studies in other geogra phical locations. The Los Angeles labor market is probably not typical of those across the country. Partly due to the small size of our sample, we were unable to deal adequately with the interaction of demographic characteristics. Probably larger studies are in order. This, however, raises another problem: the ability of the cross-tab method of analysis to cope with the almost infinite number of possible reason-mobility flow combinations which are possible. Some better means of analysis should be found. BIBLIOGRAPHY Books American Nursing Association, Research and Statistics Department. Facts about Nursing: A Statistical Summary, 1955-195b. Mew York: ANA, l95t>. Bowen, William G. and Aldrich T. Finegan. The Economics of Labor Force Participation. New Jersey: Princeton University Press, 1969. Cain, Glen. Married Women in the Labor Force. Chicago: University of Chicago Press, 19bb. Hicks, J. R. The Theory of Wages. London: Macmillan and Company, Ltd., 1935. Hunter, Laurence C. and Graham L. Reid. Urban Worker Mobility. Paris: Organization for Economic Co operation and Development, 1968. Mincer, Jacob. Labor Force Participation of Married Women, Aspects of Labor Economics. Princeton* Princeton University Press, 1962. Myers, Charles A. and George Shultz. The Dynamics of a Labor Market. Englewood Cliffs, Mew Jersey: Prentice Hall, Inc., 1951. OECD. Wages and Labor Mobility. Paris: OECD, 1965. Palmer, G. L. Labor Mobility in Six Cities, 19^0-19^1. New York! Social Sciences Research Council, 195^. Parnes, Herbert S. Research on Labor Mobility, An Appraisal of Research Findings in the United States. Mew York: Social Sciences Research Coun- cT171954. 222 Reynolds, Lloyd G. The Structure of Labor Markets. New York: Harper and brothers, 1951. Smith, J. H. and B. C. Roberts, eds. The Analysis of Labor Mobility in Manpower Policy and Employment Trends.— London:— B'elT.'T^'b'b-; -------------------- Smith, Phil M. Influence of Wage Rates on Nurse Mo bility. Chicago: Graduate School, March, 1962. Publications of the United States Government, Learned Societies and Other Organizations Heneman, Herbert G., Jr. "Differential Short-run Labor Mobilities: St. Paul, 1941-1942," Minnesota Manpower Mobilities. Bulletin No. 10, University of Minnesota: Industrial Relations Cente?, 1950. Kerr, Clark. "Migration to the Seattle Labor Market Area, 1940-1942." University of Washington Publications in the Social Sciences, XI (August, 194'?).------------------------------ Routh, Guy. "Geographical Mobility of Manpower." Pinal Report of the International Joint Seminar on Geographical and Occupational Mobility of Man power. OECD: Paris, 1964. United States Bureau of Labor Statistics. "Employment and Earnings Statistics for the United States, 1909-1964." Washington, D.C.: Government Printing Office, 1964. Employment and Earnings. XII (February, Employment and Earnings. XV (February, 1966) . T9TO. United States Department of Commerce, Bureau of Census. "Mobility of the Population of the United States, March 1963 to March 1964," Current Population Re ports: Population Characteristics, Series p-20, (September, 1965). 224 United States Department of Health, Education and Welfare, Hospital Personnel: Report of a Personnel Re search Project. Public Health Service Bulletin No. 930-C-9, Washington, D.C.: Government Printing Office, 1964. United States Department of Labor. Manpower Report of the President. Washington, D.C.: Government Printing Office, 1962. United States Department of Labor. "The Operation of Severance Pay Plans and Their Implications of Labor Mobility," Bureau of Labor Statistics Bulletin. No. 1462, Washington, D.C.: Govern ment Printing Office, 1966. Weiss, Jeffrey H. "Nursing Manpower Program Analysis," A Study for the Department of Health, Education and Welfare. Washington, D.C.: Government Printing Office, 1967. Yett, Donald E. "Yes, Virginia, There is a Shortage of Nurses — But It's Not as Simple as all That ... A Paper Presented at the Second Conference on the Economics of Health. Baltimore, Maryland: December 5-7, 196b. I f Periodicals and Journals Administrative Management Society. "Administrative Management Society Reports Results of Turnover Survey," Administrative Management, July, 1966, pp. 43-46. _. "Personnel Turnover: AMS's Findings," Admin istrative Management, July, 1968, pp. 43-4£ _. "Survey: Turnover of Office Personnel," Administrative Management, June, 1965, pp. 55-58. Bain, William. "Turnover of Nursing and Paramedical Staff in 23 Ontario Hospitals," Canadian Hospital, XLVI (June, 1969). Bancroft, Gertrude and Stuart Garfinkle. "Job Mobility in 1961," Monthly Labor Review, August, 1963, ppl 897-90^ Batchelder, Alan B. "Occupational and Geographic Mobility: Two Ohio Area Studies," Industrial and Labor Relations Review, XVIII (July, 19b5) pp. 570-583. Behrend, Hilde. "Normative Factors in the Supply of Labor," The Manchester School of Economics and Social Studies. January, 1955. Bluestone, Abraham. "Major Studies of Worker's Reasons for Job Choice," Monthly Labor Review. LXXXVII (March, 1955), PP~ 301-306. Bramhall, David F. and Herrington J. Bryce. "Interstate Migration of Labor-Force Age Population," Industrial and Labor Relations Review. XXII, No. 4 (July, 1969). Bunting, Robert L. "A Test of the Theory of Geographic Mobility," Industrial and Labor Relations Review, XV (October, 1961), pp. 75-82. _________________ . "Labor Mobility: Sex, Race and Age," Review of Economics and Statistics. XLII (May, I960), pp. 229-231. Bunting, Robert L., Lowell D. Ashby, and Peter A. Prosper. "Labor Mobility in Three Southern States Industrial and Labor Relations Review. XIV (April, 1961), pp. 432-435. Burton, John F., Jr., and John E. Parker. "Interindustry Variations in Voluntary Labor Mobility," Industrial and Labor Relations Review. XXII, No. 2 (January, 1969), pp. 199-216. Catania, James J. "Why do Nurses Change Jobs," Hospital Management. XCVIII (August, 1964). Cohen, M. S. "Participation of Married Women in the Labor Force," Monthly Labor Review. XCII (Octo ber, 1969), pp. 31-35. Dauw, Dean C. "Human Factors Outrank Salary as Turnover Cause in This Hospital," Hospital Topics. XLIV (August, 1966), pp. 65-67. 226 Dernburg, Thomas and Kenneth Strand. "Hidden Unemploy ment 1953-1962: A Quantitative Analysis by Age and Sex," American Economic Review, LVI (March, 1966), pp.“71-75':------------------ Diamond, Lorraine K. and David J. Fox. "Turnover Among Hospital Staff Nurses," Nursing Outlook, VI (July, 1958), pp. 388-391. Dodge, Joan S. "Why Nurses Leave — And What to Do About It," Modern Hospital. XCIV (May, I960). Dunlap, Henry B. "Employee Turnover Costs Millions," Hospital Forum. II (July, 1968), pp. 60-63. Elder, Frances. "Why Nurses Don't Stay Put," Registered Nurse. XXI (May, 1958), pp. 60—63- Eldridge, P. and I. Wolkstein. "Incidence of Employer Change," Industrial and Labor Relations Review. X (October, 1956), pp. 101-107. Foley, Margaret. "Can We Minimize Staff Turnover," Hospital Progress. XXXIII (March, 1952), pp. 66-67. Gallaway, Lowell E. "Age and Labor Mobility Patterns," Southern Economic Journal. XXXVI (October, 1969), pp. 171-180. __________________. "Industry Variations in Geographic Labor Mobility Patterns," Journal of Human Re sources . II (Fall, 1967), pF! ^61-^7^. __________________ . "Interindustry Labor Mobility Among Men, 1957-1960," Social Security Bulletin. (Sep tember, 1966). "Labor Mobility Resource Allocation and Structural Unemployment," The American Economic L, LIII, No. ^ (September, 1963), pp. 69^-716. "The Effect of Geographic Labor Mobility on Income: A Brief Comment," The Journal of Human Resources. IV (Winter, 1969), pp. 103- 109. 227 Gegan, Vincent F. and Samuel H. Thompson. "Worker Mo bility in a Labor Surplus Area," Monthly Labor Review. VIII, pt. 2 (December, 1957), pp. 1451- 1463. Gordon, M.S. "Women in the Labor Force," Industrial Relations. (May, 1968), pp. 187-192. Greenwood, M. J. "An Analysis of the Determinants of Geographic Labor Mobility in the United States," Review of Economics and Statistics. LI, No. 2 (May, 1969), pp. 189-194. Hamel, Harvey R. "Job Tenure of Workers, January, 1966," (Special Labor Force Report) Monthly Labor Review. XC (January, 1967). Holen. A. s. "Effects of Professional Licensing Arrange ments on Interstate Labor Mobility and Resource Allocation." Journal of Political Economics. LXXIII (October, 1965), pp. 492-498. Hough, Louis B. "What Are the Reasons for Nursing Ser vice Turnover?" Hospital Management. LXXIX (January, 1955), pp. 43-46. International Labour Office. "Labour Mobility in the United States," International Labor Review. (March, 1959). Ladinsky, Jack. "Industry Variations in Geographic Labor Mobility Patterns," Journal of Human Resources. II, No. 4 (Fall, 1967), pp. 475-495. ______________ . "Occupational Determinants of Geographic Mobility Among Professional Workers," American Sociological Review. XXXII (April, 1967), pp. 253-264. ______________ . "Sources of Geographic Mobility," Demography. IV, No. 1 (1967), pp. 293-309. Lansing, John B. and James N. Morgan. "The Effect of Geographic Mobility on Income," Journal of Human Resources. II (Fa-1, 1967), pp. ^9-460. Levine, Eugene. "Analyzing Turnover Among Hospital Per sonnel," Part II "Turnover Among Nursing Person nel in General Hospitals," Hospitals. XXXI (Sep tember, 1957), pp. 50-53* 228 Levine, Eugene and Stuart Wright, "New Ways to Measure Personnel Turnover in Hospitals," Journal of the American Hospital Association, XXXI (August, l95/), p p T - 3 F - _IT 2 .-------------------------------------------------- Lotspeich, Ruth L. "Why Do General Duty Nurses Resign," American Journal of Nursing. LI (July, 1951), pp. 68— 69. Lurie, M. and E. Rayack. "Racial Differences in Migration and Job Search: A Case Study," Southern Economic Journal. XXXIII (July, 1966), pp. 81-95. Lurie, M. "The Effect of Non-vested Pensions on Mobility: AvStudy of the Higher Education Industry," Industrial and Labor Relations Review. XVIII (January, 1965), pp. 224-237. Mahoney, T. A. "Factors Determining the Labor-force Participation of Married Women," Industrial and Labor Rexatlons Review. XIV (July"^ 1967), pp. 563-577. Marsh, R. E. "Geographical Labor Mobility in the United States: Recent Findings," Social Security Bulletin. XXX (March, 1967), pp. 14-20. McNally, G. B. "Patterns of Female Labor Force Activity," Industrial Relations. (May, 1968), pp. 204-218. Nash, Mary K. "Turnover of Psychiatric Staff Nurses," Nursing Outlook. XIV (August, 1966). Null, Virginia M. "Facts Tell A Story," Nursing Outlook. Ill (August, 1955). Parrella, Vera C. "Women and the Labor Force," Monthly Labor Review. (February, 1968), pp. 1-12. Raimon, Robert L. "Interstate Migration and Wage Theory," Review of Economics and Statistics, XLIV (Novem- ber, 1962), pp. 428-438. ________ ^ ______ . "Labor Mobility and Wage Inflexibility," Industrial and Labor Relations Review. XVI (December, 1963), pp. 116-129. Rosenfeld, Carl and Vera C. Perrella. "Why Women Start and Stop Working: A Study in Mobility," Monthly Labor Review, XXCVIII, No. 9, (September,“1965), pp. 1077-1082. Rottenberg, Simon. "On Choice in Labor Markets," Industrial and Labor Relations Review, IX (January, 1956), pp. 183-199. Saben, Samuel. "Geographic Mobility and Employment Status March 1962 - March 1963," Monthly Labor Review. XXCVII (August, 1964), pp. 873-881. _____________ . "Occupational Mobility of Employed Workers Monthly Labor Review. XC (June, 1967), pp. 31-38. Saleh, Shoukry D., Robert J. Lee and Erich P. Prien. "Why Nurses Leave Their Jobs — An Analysis of Female Turnover," Personnel Administration. XXVIII (January-February, 1965). Tarver, J. D. "Metropolitan Area Intercounty Migration Rates: Reply," Industrial Labor Relations Review XIX (January,1966), pp. 213-223. Taub, Doris L. "The High Cost of Replacing a Nurse," Hospitals. XL (October, 1966). Telia, Alfred. "Labor Force Sensitivity to Employment by Age, Sex," Industrial Relations. IV, No. 2 (February, 1965), pp. 69-^3• Ulman, L. "Labor Mobility and the Industrial Wage Struc ture in the Postwar United States," Quarterly Journal of Economics. LXXVIIII (February, 1965), pp. 73-97. Wieland, George P. "Studying and Measuring Nursing Turn over," International Journal of Nursing Studies. VI (July, 1969). Wright, Stuart. "Analyzing Turnover Among Hospital Per sonnel," Part III "Turnover and Job Satisfaction," Hospitals, XXXI (October, 1957), pp. 47-52. 230 Unpublished Material Altman, Stuart Harold. "Factors Affecting the Unemploy ment of Married Women: A Study of the Dynamics of the Labor Force Behavior of Secondary Family Workers." Unpublished Ph.D. dissertation, Univer sity of California at Los Angeles, 1964. American Hospital Association, Division of Research. "Survey of Personnel Turnover in Volunteer Hos pitals." An unpublished abstraction from the final report, 1962. American Nursing Association, Division of Research. "Salary Ranges and Other Employment Conditions for RN Staff Nurses in Non-federal Short-term General Hospitals." Unpublished data supplied by ANA. August, 1967. _______________. "Salary Ranges for RN Staff Nurses in Non-federal Short-term General Hospitals." Unpublished data supplied by ANA. April, 1968. _______________ . "Walary Ranges for RN Staff Nurses in fion-federal Short-term General Hospitals." Un published data supplied by ANA. January, 1969. Hixson, Jesse Sharp. "The Demand and Supply of Professional Hospital Nurses: Intra-Hospital Resource Alloca tion." Unpublished Ph.D. dissertation, Michigan State University, 1969. Yett, Donald E. "An Economic Analysis of the Hospital Nursing Shortage." Unpublished Ph.D. dissertation, The University of California, Berkeley, 1968. o APPENDIX A STATE LICENSING REQUIREMENTS AND UNIONS State Licensing Requirements There was no evidence which we could find to indi cate that the state licensing requirements in California inhibited the geographical movement of nurses educated in the United States. RNs who came to California from other states were not required to take the California State Boards. The only requirements which had to be met were: (1) providing proof of graduation and possession of a current license from another state, and (2) filling out the necessary application forms. The waiting period usually took from three to six weeks and generally was completed while the RN was still at her last job. Out of the almost 1500 interviews, in no instance did an RN feel that these requirements served as a serious impedi ment to her movement. The state licensing requirements for RNs educated outside of the United States were somewhat more cumber some. Nurses coming directly to California were required to show proof of an active license in the foreign country 231 232 and also had to provide evidence that they had completed course work In five major nursing areas. Prior to January, 1969, after having met the above requirements, foreign nurses were interviewed by a member of the Board of Nursing Education and Nurse Registration in Sacramento. One of the main purposes of this interview was to deter mine the applicant's ability to communicate well in English. Since January, 1969, all foreign nurses were required to take the State Board Exams. There can be little doubt that this added requirement has discouraged some foreign nurses from working in California. Those foreign nurses who do come are allowed to work only as nurses' aides until such time as they have met all the necessary require ments. In spite of this, as will be noted in the text, Filipino nurses are the most geographically mobile of all nurse groups. Unions None of the nursing departments in the hospitals of our study were represented by unions. To the best of our knowledge there are few if any bona fide nurses' unions. To this point in time, it can be concluded that unions have had neither a positive nor a negative affect on nurse mobility. There does seem to be a trend toward more collective bargaining by nurses (for example the nurses in some San Francisco and New York hospitals in recent years), but most of this will likely be conducted by the nurses' professional associations such as the American Nurses Association or the California Nurses Association. Roughly 80 per cent of the nurses inter viewed in our study were opposed to unions, and most of these were also opposed to strikes. APPENDIX B QUESTIONNAIRE FOR NURSING MOBILITY STUDY Date of Interview: New Hire: Name of Hospital: Termination: Year of Graduation: Unit or Department: Type of Education (Associate 2 years; diploma 3 years, BS A years; Post BS other, etc.) Name of College: When did you last take a refresher course: (For Nurse reentering the Nursing Profession) Age: Race: Sex: Which best describes your marital status: (Married, Single, Divorced, Separated, Widow) Number of children: Ages of children: Are they home: Baby sitting arrangements: Age of husband: Occupation of husband: Geographic Information — Termination to where City: State: Will you be working: New Job if any: Shift: Full or Part-time: If hospital, number of beds: How long were you employed by this hospital: Full or Part- time: Shi ft Worked: Position here: 234 235 Geographic Information — New Hire from where City: State: Were you previously employed: What were you you doing: How long: Till when: If hospital, number of beds: Position then: Now: Did you work full or part-time: then: Now: Shift: Shift: If not working, what enabled you to return to nursing: Reason for Job Change Financial Status after Job Change \ APPENDIX C Column 1-2 3-5 6 7-8 9 10 11 12 13 14-15 CODING INFORMATION — RNs Description Hospital Number RN Number RN Interview case number AGE of RN Race Sex Marital status Dependent Children Child status What RN did prior to being out of Labor Force Code 1-19 1- (total in any facility) 1- (when hired* l,when term*2, etc.) 1-99 l=Cauc,2*Negro 3*Spanish speak ing, 4*Filipino, 5*Chinese, 6* Korean, 7“Japan ese, 8*Thai, 9“ Other l*Female, 2*Male l*Single,2*Mar- ried, 3“Divorced ^■Separated, 5*Widow l*yes, 2*no l*pre school, 2*not pre school See col. 42-43 for description 236 o 237 Column 16-17 18 19 20-21 22-23 24 25-26 Description Code RN's 1st grad year Type of Grad l=AA,2=Diploma, 3=BS Additional Grad type l=AA,2=Dip,3=BS (highest achieved) 4=MS,5=PhD Year of highest Grad State of Country of Grad l=Maine,2=New (1st) Hampshire,3=Ver- mont, 4=Mass., 5® R.I., 6=Conn. , 7= NY., 8=NJ., 9=Pen., 10=0hio., ll=Ind., 12=111., 13=Michigan, l4=Wisc., 15“Minn., l6=Iowa., 17=Missouri, 18-N.D., 19=S.D., 20=Neb., 21=Kansas., 22=Del., 23=Md., 24-D.C., 25=Vlr., 26«W. Vir., 27=N.C. 28=S.C., 29=Ga., 30=Fla.,31=Ken,, 32=Tenn., 33*Ala., 34*Mississippi, 35=Ark., 36=La., 37=Ok. 38=Tx.,39=Montana, 40=Idaho, 4l*Wy., 42=Colo., 43=N Mex., 44=Ari., 45=Utah, 46=Nev., 47= Wash., 48=Ore., 49=Calif., 50=Alaska, 51=Hawaii Northeast Region =1-9 North Central =10-21 South = 22-38 West = 39-51 60=Phillippines, 6l= South and Central America, 62=Mexico, 63=Canada, 64=Thailand, 65=Korea, 66=Taiwan or China, 67=England, 68=Other 69=Europe-Main New England Div.=l-6, Middle Atlantic Div.=7-9 East N. Central Div. =10-14, West N. Central Div. = 15-21, South Atlantic Div.=22-30, East South Central Div.=31-34, West South Central Div=35-38, Mountain Div=39-46,Pacific Div =47-51 Husband's Occupation l=Self employed, 2=professional, 3=Blue Collar, 4=Laborer, 5=student, 6=mili- tary service, 7=unemployed, 8=retired Husband's Age 1-99 238 Column 27 28-31 32-34 35-36 37 38-40 41 42-43 44-48 49 50 Description Termination or Hire When Hired or Terminated For Hire-Time between Jobs To where or from where— Moved within greater LA Code l=Hire,2=ReHire 3*Termination Month and Day 1-999 weeks Same as 22-23 (job site)l=3mi 2*6 ... 9=27 mi Moved Intra-Calif, but by 1-999 miles more than 29 miles Reenter or leaving the labor l=Reent,2*leav., force 3=Continuing within the labor force, 4=not certain (90 days is criterian), 5* Enter-lst time-Grad, 6=Extra Job What have you been doing 2=hosp,3“indus. , or what do you plan to do 4=Drfs office, 5=PH, 6=school nurse, 7°PVT duty, 8 * Nursing home, 9®instructor nurse,10*other nursing, ll*housewife, 12*secretary, 13=study(student) l4=will work, not certain what,15=other non- nursing, l6=vacation length of stay at Job leaving or from job left In days For Hire-position now l*Staff (code For Term-position leaving relief charge as staff and also team leader), 2=Head Nurse and Asst. HN, 3= Supervisor (Asst. Sup and Inservice Instruc tor also as Supervisor), 4=Director of nurses and assistants, 5=charge nurse For Hlre-posit. at last job For Term-posit. in new job Same as above Column Description Code 51 52 53 5^ 55 56 57-58 59-60 61-62 For Hire-Present Job For Term-Job leaving For Hire-last job For Term-new Job if any l=FT,2=PT,3=Per Diem,4=0n Call Same as above For Hire-present shift l=days,2=PMs,3= For Term-Shift on job leaving nights, ^rota tion For Hire-shift at last Job For Term-shift at new job if any Same as above If returning to nursing after 90 days or more Able to return now because: l=Children now grown, 2=can now leave kids with a sitter, 3*want or need more income, 4=want something to do, 5*want to saty in touch with nursing or just get back to work after vacation, 6= 1&3, 7= 3&4, 8= 3&5, 9= 4&5 If out of nursing more than five years, did you have a refresher course Irregular re fresher course, 2=lengthened orientation, 3*no refresher per se Type of mobility flow (30mi) l=Int,2=Int-G, 3=Int-V,4=Int-V -G, 5=Int-FTPT, 6=Int-FTPT-G,7=FTPT-V,8=Int-FTPT-V-G,9-Int-O, 10*Int-0-G,ll*Int-0-FTPT,12-Int-O-FTPT-G, 13=Int-Ind,l4«Int-Ind-G,15*Int-Ind-FTPT, l6=Int-Ind-FTPT-G,17=L,18-L-G,19*E,20=E-G, 21*Not certain, 22=Extra job,23®Int-0-Ind, 24«Int-0-Ind-G,25*Int-0-Ind-FTPT,26=Int-0- Ind-FTPT-G, 27»Grad,28*Grad-G Type of mobility flow Same as above (So. Calif.) Type of mobility flow Same as above (SMSA) 240 Column 63 64-65 Description Code Move within sample l=yes,2=no Primary reason for leaving this or last job #1: l=Husband's or Father's new job, 2=Baby- sitting or other Family responsibility, 3=Hus- band forbids, 4=Pregnancy, 5-Sickness RN or Family, 6=Retirement, 7-Died. #2: 10=Marriage, ll»Family and/or RN decision to move out of area, 12*Move to Join or avoid family or friends, 13*Wanted rest or extended vacation, l4»Return to school, 15=Move to fur ther experience or ed., l6=disliked location of hospital, 17=area offered too few social activities, l8=regular job return after relief or work during vacation from regular job, 19=too far to commute, 20=tlred of particular kind of nursing or work, 21=tired of nursing, 22=in a tur — want change, 23=travel, 24= enter military, 25=misc. §3: 0=Reason not known,26=Found job paying higher wages, 27ssfound job with better working conditions, 28=found job which offers promo tion #4: 32=dissat. with shift, 33=dissat. with hours or unit or job, 34=dissat. with day/W. ends, 35=dissat. with hospital's philos. patient care, 36=dissat. with hospital's or ganization, etc. 37=pressure-supervisors, 38=pressure-understaffing, 39=clash-supervisor, HN 40=clash-co worker, 4l=no advance, oppor. 42=poorly trained, insubordinate, lazy auxiliary, 43=too little p&tient contact, 44=no vacation due,45=no LOA due, 46=unfilled promises, shift, wages, etc., 47®lack of appreciation & consideration shown, 48=not trained to do Job given, work too hard or demanding job not defined, 49=discrimination, not treated as a pro, 50-misc. #5: 52=asked to resign, 53®fired. 241 Column 66-67 68-69 70 71 72-73 74 75-77 Description Code Secondary reason Same as above Tertiary reason Same as above Financial Status after change l=more,2=same 3=less,4=don’t know Did you have this job before l=yes,2=no you left last job (Hires) Worked in which unit or l=Med Surg,2=0r dept. (Terms only) 3=Peds,4=ICU, 5=CCU, 6=EF, 7=Rec Room, 8=0B, 9=Del Rm,10=Nurs Admin,11= Post Pard, 12=Nursery,13=Psyc.,l4=inservice 15=IV,l6=Float,17=0rtho,l8=Urology,19=Ent 20=Gyn,21=Center for Crit Ill,22=Neuro How obtained information l=personal inter view, 2=per telephone, 3=via letter & questionnaire, 4=from hospital, 5=test questionnaire Distance moved (within USA 1=30 mi, 2=60 . & outside California) . .110=3300 mi. 78-80 Blanks APPENDIX D Column 1-2 3-5 6 7-8 9-10 11 12-14 15 16 17-19 20 21-22 CODING INFORMATION — HOSPITAL CHARACTERISTICS Description Hospital number Hospital size Ownership Hospital age Blank Presence of retirement plan Size of shift differential paid Payment of time and a half for weekends and holidays Payment for education Wage paid Blank Turnover rate— all RNs Code 1-19 l=Voluntary non profit, 2=pro- prietary, 3= church related non-profit, 4=non-profit corporation, 5®corporation for profit l=yes,0=no l*=yep,0=no l*yes,0=no 23-24 Turnover rate— FT RNs 242 Column Description 25-27 Average number of RNs on staff during year 28-30 Average number FT RNs on staff during year 31 Type of hospital 32 Dees hospital have a health plan 33 Shift arrangement 34 Payment of time and a half for overtime 35 Vacation time 36 Salary policy-experience 37 Does hospital employ part- time RNs 38 Does hospital grant leave of absence 39-43 Median family income in area 44 Blank 45-46 Homicide rate in area of hospital per 10,000 population 47-49 Rape— same as above 243 Code l=teaching & inservice, 2= inservice only, 3=neither,4= inservice & residency l=yes,2=no l=rotation, 2=permanent, 3=other l=yes,2=no 1=2 wks after 1 year,2= 3 wks after 1 yr l=pay extra for experience,2= dn't pay extra l=yes, 2=no l»yes, 2=no Column Description 50-53 Robbery — same as above 54-57 Assult — same as above 58-61 Burglary — same as above 62-66 Theft — same as above 67-70 Auto theft — same as above 71 Blank 72-74 RN patient ratio — all RNs 75-76 Blank 77-78 RN patient ratio — PT RNs APPENDIX E MULTIPLE REGRESSION EQUATIONS USED TO EXPLAIN HOSPITAL NURSE TURNOVER Definitions Dependent Variable: T-FT - Turnover rates of full-time nurses for each hospital Independent Variables: SIZE* Hospital bed size - measured continuously OWN = Hospital ownership - dummy variable 1 * non profit, 0 = profit AGE = Hospital age - continuous RET = Presence of retirement plan - dummy 1 = yes, 0 = no. DIFF=Payment of shift differential - continuous. OVER=overtime payment for weekends - dummy 1 = yes, 0 = no ED = Extra payment for educational attainment - dummy - 1 = yes, 0 = no FT/P* Full-time nurse-patient ratio - continuous INCb=Median family income in the area of the hospital- continuous WAGE=Starting wage level for new RN - continuous bSource: Los Angeles Times, Marketing Research, Los Angeles Marketing Area, 1970, p. 2. 245 246 RAPEa»Number of rapes per 10,000 population in the area of hospital - continuous R0Ba = Number of robberies per 10,000 population in the area of the hospital - continuous BURGa*Number of burglaries per 10,000 population in the area of hospital - continuous Regression 1 (Note: The T test value appears in paren theses below the coefficient) T-FT = 10.6041 - 5.8962 OWN + 4.2455 INC + 0.5087 ROB (.9024) (-.9223) (2.7955)° (3.1634)° - .1713 FT/P R/SE = .3911 (-.0796) Regression 2 T-FT = 111.8425 - 14.6984 RET + .3645 DIFF - 17.9873 QVER (.8076) (-1.7236) (2.4962)b (- 1.8801)d + 3-7535 ED - 0.1343 WAGE R/SE = .2035 (.5268) (-.6258) aSource: Los Angeles Police Department and State of California, Department of Justice, Division of Law Enforcement, Bureau of Criminal Statistics. Crime and Delinquency in California for 1966, 1967 and 1968. Significant at 1% level Significant at 5% level Significant at 10? level 247 Regression 3 T-FT - 126.9522 - 10.7669 RET - 0.1259 WAGE + 0.3514 ROB (1.9913) (-1.4258)a (-.7158) (1.9554)° - 0.1872 FT/P R/SE » 0.2479 (-.9724) Significant at 25? level Significant at 10? level APPENDIX P MULTIPLE REGRESSION EQUATIONS US TO EXPLAIN LENGTH OF STAY Definitions Dependent Variable: LOS = Length of stay Independent Variables: NAG Nurse's age — continuous - reflects age ; tim'. of hire R1 = Caucasian — base R2 = Negro R3 Filipino R4 = Other Ml = Single M2 = Married M3 = Other El = AA degree E2 s Diploma — base E3 = BS + U1 - Medical-surgical & related units — base U2 s Critical care units 248 U3 a Obstetrics & Pediatrics U4 * Supervisory HI ■ Husband self-employed & professional — base H2 = Blue collar and laborer H3 = Student H4 = Retired, military service and unemployed CSH = Hospital shift differential — continuous CR * Wage range — continuous CRO = Robbery — continuous DFT = FT RN - patient ratio — continuous CRET * Presence of retirement plan — 1 = yes, 0 = no. SH2 = Night, shift CS = Hospital size — continuous CT1 = Time and a half for weekends DIS = Reason for leaving 1 = dissatisfaction, 0 = other 250 Best Regression Equation (Note: The T test value appears in parentheses below the coef ficient .) LOS = -317.67 + 15.69 NAG - 535.66 R3 + 194.86. M2 (3.11)a (3•53)a (1.70)° - 292.91 M3 + 901.81 U4 + 667.03 H4 + 5.66 CSH (1.58)c (3.03)a (2.84)a (2.95) + 123.59 DIS (1.04) R/SE = .11049 Degrees of Freedom * 446 Significant at .005 level. Significant at .100 level. Significant at .250 level.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Forecasting Selected Statewide Recreation Requirements

PDF

Economic Aspects In The Evolution Of The Great Lakes Freighter

PDF

Participation Of Rank-And-File Member In The Decision Making Process And His Access To The Grievance Machinery Of His Union: Including A Survey And Case Study Of Uaw Local 1531, Amarillo, Texas

PDF

The Industrial Relations System In Sweden With Suggestions Concerning Itsrelevance To An African Nation, Ghana

PDF

Arbitration Of Technological Change

PDF

Competition And Concentration In The Brewing Industry

PDF

Economic Analysis Of Factors Underlying Pricing In The Southern California Tuna Canning Industry

PDF

An Economic Analysis Of The Factor Market For Pharmacists

PDF

An Analysis Of Management Functions As Influenced By Organized Labor: Theearly Traditional Approach

PDF

Collective Bargaining, Disputes And Dispute Settlement Procedures In Public Employment

PDF

The reorientation of Japanese trade patterns: Selected prewar and postwar periods

PDF

Effects Of Swedish Fiscal Policy And Planning On Economic Growth And Stability, 1930-1964

PDF

Financing Of Medical Care For The Aged: A Comparative Evaluation Of Compulsory Versus Voluntary Health Insurance

PDF

American Political Economics: An Emerging Consensus

PDF

An Economic Analysis Of The Changes In Retail Store Sales And Location Inthe Central Business Districts Of One Hundred And Nine Central Cities In The United States, 1948-1958

PDF

Planning Institutional Change For Economic Growth: A Study Of The Effect Of Public Institutions On The Rise Of Indigenous Entrepreneurs In East Africa

PDF

Economic and administrative aspects of capital budgeting in municipalities

PDF

Economic Quality Control (Eqc)

PDF

Critical analysis of central banking in Iraq

PDF

Concepts Of Entrepreneurship In Recent Economic Thought

Asset Metadata

Creator Payne, Richard Dee (author)

Core Title An Economic Analysis Of Nurse Mobility Patterns

Degree Doctor of Philosophy

Degree Program Economics

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag economics, general,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Phillips, E. Bryant (committee chair), Barker, Edward H. (committee member), Pollard, Spencer D. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-448991

Unique identifier UC11362035

Identifier 7107732.pdf (filename),usctheses-c18-448991 (legacy record id)

Legacy Identifier 7107732.pdf

Dmrecord 448991

Document Type Dissertation

Rights Payne, Richard Dee

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