University of Southern California Dissertations and Theses

Alcohol expectancies and consumption: Age and sex differences

(USC Thesis Other)

Alcohol expectancies and consumption: Age and sex differences

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS
This manuscript has been reproduced from the m icrofilm master. U M I
films the text directly from the original or copy submitted. Thus, some
thesis and dissertation copies are in typew riter face, while others may be
from any type o f computer printer.
The q u a lity o f this reproduction is dependent upon the quality o f the
copy subm itted. Broken or indistinct print, colored or poor quality
illustrations and photographs, print bleedthrough, substandard margins,
and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send U M I a complete
manuscript and there are missing pages, these w ill be noted. Also, i f
unauthorized copyright material had to be removed, a note w ill indicate
the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and
continuing from left to right in equal sections w ith small overlaps. Each
original is also photographed in one exposure and is included in reduced
form at the back o f the book.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6” x 9” black and white
photographic prints are available fo r any photographs or illustrations
appearing in this copy for an additional charge. Contact U M I directly to
order.
UMI
A Bell & Howell Information Company
300 North Zeeb Road, Ann Arbor MI 48106-1346 USA
313/761-4700 800/521-0600
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ALCOHOL EXPECTANCIES AND CONSUMPTION:
AGE AND SE X DIFFERENCES
by
D erek Davies Satre
A Thesis Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In P a rtia l F u lfillm e n t o f the
Requirem ents fo r the Degree
M aster o f A rts
(P sychology)
May 1999
C opyright 1999 Derek Davies Satre
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 1395137
UMI Microform 1395137
Copyright 1999, by UMI Company. All rights reserved.
This microform edition is protected against unauthorized
copying under Title 17, United States Code.
UMI
300 North Zeeb Road
Ann Arbor, MI 48103
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UNIVERSITY OF SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES. CALIFORNIA 80007
This thesis, w ritten by
Derek Davies Satr e ________
A
under the direction o f &JL.S. Thesis Committee,
and approved by a ll its members, has been p re
sented to and accepted by the D ea n of T h e
Graduate School, in p a rtia l fu lfillm en t of the
requirements fo r the degree of
Master of Arts
D a te F e b ru a ry 1 6 , 1 9 9 9
THESIS COMMITTEE
/
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table o f Contents
tf
List o f Tables \ ii
A b stra ct far
B ackground and S ignificance 1
M ethods
13
P rocedure 14
Results
19
D iscussion
31
R eferences
45
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
List o f Tables
Table 1: R e lia b ility C oefficients fo r
Expectancy Subscales (C ronbach's Alpha)
Table 2: Sample Dem ographic C haracteristics
Table 3: Q uantity: Mean Num ber o f D rinks a t a
Tim e, showing T rim m ed Means and Trim m ed
Standard Errors (A ll Subjects)
Table 4: Q uantity: Mean Num ber o f D rinks a t
a Tim e, showing T rim m ed Means and Trim m ed
Standard Errors (D rin kers O nly)
Table 5: Frequency: Num ber o f D rinkin g
Occasions per M onth, showing T rim m ed
Means and T rim m ed Standard E rrors (A ll Subjects)
Table 6: Frequency: Num ber o f D rin kin g
Occasions per M onth, showing T rim m ed
Means and T rim m ed Standard Errors (D rinkers O nly)
Table 7: Q uantity: Mean Num ber o f D rinks a t a
Tim e, showing T rim m ed Means and Trim m ed
S tandard Errors (A ll W hite Subjects)
Table 8: Frequency: Number o f D rinkin g
Occasions per M onth, showing T rim m ed Means
and Trim m ed Standard Errors (A ll W hite Subjects)
Table 9: Expectancy Score Means by Sex and Age
G roup, showing T rim m ed Means and Trim m ed
S tandard Errors
Table 10: Percentage-Bend C orrelations o f
C onsum ption Measures w ith Expectancies
18
21
22
23
24
25
26
27
28
30
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
iU
A b stra ct
Previous studies have dem onstrated a co rre la tio n between expectancies
th a t adolescents and young a d ults hold ab out the effects o f alcohol and the
am ount o f alcohol they consume. This stu d y exam ined age and sex differences
in expectancies and alcohol consum ption levels in a sam ple o f 92 o ld e r and 83
you nge r adults, using s e lf-re p o rt questionnaires. Results fo u n d low er
q u a n titie s o f alcohol consum ption in o ld e r adults, b u t m ore fre que nt d rin k in g .
O lder adults re ported low er levels o f both positive and negative expectancies.
Am ong o ld e r wom en, there w ere sig n ific a n t co rre la tio n s between negative
expectancies and low er alcohol consum ption. For yo u n g e r wom en, p o sitive
expectancies were correlated w ith increased consum ption. C orrelations
between expectancies and consum ption fo r b o th o ld e r and younger m en were
n o t sig n ifica n t. These results suggest th a t age and sex m ay d iffe re n tia lly
in flu ence how expectancies a ffe ct d rin k in g . For old er a d ults, this w o u ld
suggest th a t w om en’s d rin k in g m ig h t be re ducib le b y increasing negative
beliefs about the effects o f alcohol.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1
Background and S ignificance
Prevalence and M easurem ent o f Alcohol Use in Late Life
A lcohol abuse and alcohol dependence (alcoholism ), the two alcohol use
disorders described in the DSM-IV, are sig n ifica n t problem s among o ld e r
adults in the U nited States. Excessive alcohol consum ption has been associated
w ith num erous physical and psychological problem s in o ld e r people,
in clu d in g liv e r damage, heart disease, stroke, m a ln u tritio n , accidental in ju ry,
dem entia, depression, and suicide (see Bucholz, Sheline, and Helzer, 1995, for a
review ).
These consequences p ro vid e a com pelling im petus to seek a greater
understanding o f the phenom enon o f la te -life alcohol use, and to explore how
the problem s th a t re su lt from excessive d rin k in g m ight be reduced. Although
alcohol abuse and dependence have been stud ied w id e ly in younger
populations, re la tiv e ly little a tte n tio n has been paid to excessive d rin k in g in
late life , n o r have m any studies com pared younger d rin ke rs to older ones
(Bucholz e t a l, 1995). However, one difference between the two populations
has been observed repeatedly— o ld e r adults d rin k less th a n younger ones.
A lthough prevalence varies between studies, it seems cle a r th a t rates o f
alcohol abuse and dependence are also low er in samples o f o ld e r adults than in
younger ones. The E pidem iological Catchm ent Area stu d y fou nd life tim e
prevalence o f alcohol dependence fo r men 18 to 27 was 27%, w ith prevalence
o f 14% fo r men 65 and older. Prevalence rates were 7% fo r younger women
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2
and 1.5% fo r o ld e r women (M ille r, Belkin & Gold, 1991). In a study o f male and
fem ale general hospital patients, Beresford, Blow and Brower (1990) found
cu rre n t prevalence rates fo r alcohol dependence o f 23.3% fo r ages 18 to 59,
10.4% fo r ages 60 to 69, 6.7% fo r ages 70 to 79, and 0% fo r ages 80 and over.
Several cross-sectional com m unity surveys have found low er rates o f
heavy d rin k in g and h ig h e r rates o f abstention among o ld e r adults (see Osgood,
1995 fo r a review ). Cross-sectional studies o f alcohol consum ption th a t include
non-problem d rin ke rs also show a decrease over the life span fo r both men
and women, w ith decreased d rin kin g e vid e n t in W hites, A frican Am ericans
and Hispanics (M olgaard, Nakamura, Stanford, Peddecord, & Morton, 1990). One
cross-sectional study w hich examined prevalence o f alcohol consum ption
versus abstention found an increase in d rin k in g be havio r among both sexes
u n til around age 35, and the n a steady decline w ith age; women o f a ll ages
were m ore lik e ly to re p o rt no d rin kin g a t a ll (M u lfo rd & M iller, 1959). In th is
study, 65% o f males and 57% o f females aged 21 to 25 were drinkers; 84% o f
males and 63% o f females aged 31 to 35 were drinkers; and 49% o f males and
23% o f females over 61 years o f age were drinkers. Prevalence o f d a ily
d rin k in g has been estim ated to be from 10 to 20% in o ld e r adults (Barnes,
1979), although lig h t d rin k in g rather th a n heavy d rin k in g is more com m on.
Studies have consistently dem onstrated th a t rates o f alcohol consum ption,
abuse and dependence are low er fo r women than fo r men, across a ll age
groups. Reasons fo r th is diffe ren ce are n o t w ell established, but may include
greater social stigm a associated w ith fem ale d rin k in g and lower physiological
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3
tolerance fo r alcohol (W ilsnack, Vogeltanz, Kiers & W ilsnack, 1995).
In discussing alcohol use disorders in late life , it is im p o rta n t to
d istin g u ish between e a rly- and Iate-onset. There is general agreem ent in the
lite ra tu re th a t ap pro xim ate ly o n e -th ird o f a ll o ld e r alcoholics are late-onset
alcoholics, and tw o -th ird s are early-onset, w ith certain differences noted
between the two groups. In a study o f 170 o ld e r adults in a day tre atm e nt
program fo r e ld e rly alcoholics, Schonfeld and Dupree (1991) m atched 23 ea rly-
onset abusers to 23 late-onset abusers by age and sex. They found th a t early-
onset alcoholics re p o rte d being intoxicated tw ice as o fte n as late-onset
alcoholics, expected less success from the program , and w ere less lik e ly to
com plete treatm ent.
M any d iffe re n t explanations have been offered fo r the apparent
decline in alcohol consum ption among o ld e r people. The m ost p ro m in e n t o f
these are: the m o rta lity hypothesis, w hich states th a t heavy d rin kers die
earlier, and as a re s u lt do n o t appear in surveys o f la te -life d rin kin g ; the
m o rb id ity hypothesis, w hich posits that o ld e r adults lim it th e ir inta ke o f
alcohol due to increased m edical problem s th a t make d rin k in g uncom fortable
o r unsafe; the bio lo g ica l hypothesis, w hich proposes th a t norm al bio lo gical
changes reduce the am ount o f alcohol th a t o ld e r adults can com fortably
consume; the co h o rt hypothesis, w hich suggests th a t h is to ric a l factors such as
the Tem perance m ovem ent and P ro h ib itio n may make the cu rre n t e ld e rly
cohorts less in clin e d to d rin k than la te r generations; the m atura tion
hypothesis, w hich claim s th a t alcoholism is a s e lf-lim itin g disease and th a t
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4
w ith age heavy d rin k in g goes in to spontaneous rem ission; and the
m easurem ent hypothesis, w h ich claim s th a t prevalence rates in e ld e rly
samples m ay be a rtific ia lly low due to problem s in accurately m easuring the
d rin k in g behavior o f o ld e r adults (see Stall, 1987, fo r a review).
Am ong these theories, the issues o f co h o rt and m easurem ent effects
h ig h lig h t the d iffic u ltie s in h e re n t in o b ta in in g a reasonably accurate
prevalence estim ate o f la te -life problem d rin k in g . Addressing the
m easurem ent issue, G raham (1986) points o u t tha t se lf-re p o rt measures
require accurate m em ory fo r past consum ption, the a b ility to do m ental
averaging, and a w illingness to disclose in fo rm a tio n a b o u t personal behavior
o f w hich others may n o t approve. Also, o ld e r adults m ay need to d rin k less
than younger adults to experience a com parable effect, due to physiological
changes w ith age (Beresford & Lucey, 1995). Thus, even an accurate
m easurem ent o f consum ption m ig h t n o t provide an accurate in d ic a to r o f
in to xica tio n level.
S upport fo r the co h o rt hypothesis has been fo u n d in the results o f three
lo n g itu d in a l studies in d ic a tin g th a t d rin k in g patterns are fa irly stable over
tim e (C hris topherson, Escher, & Bain ton, 1984; G lynn, Bouchard, LoCastro, &
Laird, 1985; Fillm ore & M idanik, 1983). G lynn and associates even fo u n d a
slig h t increase in d rin k in g problem s over tim e, in a nine-year stu d y w ith
2,100 com m unity-dw elling m en bom between 1892 and 1945. O ther studies
co n tra d ict these find ings: Adams, G arry, Rhyne, H unt and Goodwin (1990)
conducted an eight-year lo n g itu d in a l study o f 165 e ld e rly p a rticip a n ts and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
found th a t lig h t o r m oderate drinkers d id n o t d rin k less over tim e. Heavy
d rin ke rs con tinue d to consume alcohol, b u t q u a n titie s declined over tim e.
The m o rb id ity hypothesis, w hich states th a t older adults are lim ite d in
the am ount o f alcohol they can consume due to adverse reactions w ith
m edications as w ell as chronic illnesses, may help to explain reductions in
d rin k in g over the life span, if in fact such reductions occur. Among illnesses
com m on in o ld e r adults, heavy alcohol intake m ay exacerbate g a stroin testin al
problem s and hypertension, as w ell as co n trib u te to the onset o f a v a rie ty o f
illnesses (see Gam bert and Katsoyannis, 1995, fo r a review). If o ld e r a d u lts are
in stru cte d about the effects o f alcohol on these problem s by th e ir p h ysicia n in
the course o f treatm ent, as seems like ly, it co u ld lead to vo lu n ta ry re d u ctio n o f
alco hol consum ption.
In itia tio n and M aintenance o f D rin kin g B ehavior
Social le a rn in g th e o ry proposes th a t in d iv id u a ls who abuse alco h o l
d iffe r fro m m oderate d rin ke rs in th e ir a b ility to cope w ith stressors an d in
th e ir beliefs a b out alcohol. According to this perspective, deficiencies in m ore
adaptive coping skills and positive expectancies concerning the effects o f
alcohol operate in concert to prom ote the use o f alcohol as a coping
mechanism (Cooper, Russell, & George, 1988). Reliance on alcohol to cope,
therefore, should lead to heavier d rin k in g and increase the ris k o f alco h o l
dependence over tim e.
D rin kin g to cope has been defined as the tendency to use alcohol to
escape, avoid, o r otherw ise regulate unpleasant em otions (Cooper et al., 1988).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6
M any studies have shown th a t reliance on alcohol as a coping mechanism is
associated, w ith, heavy o r abusive d rin k in g (F arberr Khavari, & Douglassr 1980;
M u lfo rd , 1983). The desire to reduce negative feelings by d rin k in g has been
re fe rre d to as a coping motive,, w h ile the desire to increase po sitive affect has
been described as an “ enhancem ent” m otive (Cooper, Frone, Russell, & M udar,
1995). .Although em otion re gula tion is a central com ponent o f each m otive,
some researchers consider the two to be separate phenomena, in th a t the
coping m otive may be a reactive process in itia te d by the experience o f
negative em otions, w hile the enhancem ent m otive may be conceptualized as
an a p p e titive process (Cooper e t aL, 1995). D rin kin g to cope m ay also be
conceptualized as negative reinforcem ent, w ith enhancem ent d rin k in g
considered positive reinforcem ent (Farber et aL, 1980).
A lcohol Expectancy
In research conducted w ith younger a d u lts, alcohol expectancies appear
to play a role in explaining how stressors lead to the adoption o f d rin kin g as a
coping strategy. It has been hypothesized th a t expectancies m ediate the
developm ent o f d rin k in g patterns in b o th adolescents and a d u lts (Brown,
G oldm an &. C hristiansen, 1985). Expectancies have been characterized as the
cognitive com ponent o f attitudes— th e beliefs th a t in d ivid u a ls have
concerning how alcohol affects them (Leigh, 1989). In the in itia tio n o f
d rin k in g behavior, an in d iv id u a l has p a rtic u la r expectancies regarding th e
effects o f alcohoL com bined w ith an evaluation o f how d e sirab le th e expected
state is, in m aking a decision to consum e alcohoL Once d rin k in g behavior has
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
been in itia te d , o r once a p a tte rn o f d rin k in g has been established, this
co g n itive com ponent m ay recede fro m awareness, an d the response o f
d rin k in g m ay follo w antecedent s tim u li in a m ore o r less au tom a tic fashion.In
established d rin k in g behavior, which- has also been, described as a
m aintenance phase o f d rin k in g (Oei & Baldwin, 1994), d rin k in g m ay not
d ire c tly in v o lv e cog nitive e va luatio n o f expected results o f consum ing alco hol.
Using fa cto r a n a lytic techniques, Fromme an d her colleagues (From m e,
S troot, & Kaplan, 1993) have id e n tifie d the fo llo w in g dom ains in w hich a lco h o l
m ay fu n c tio n as e ith e r a po sitive o r negative re in fo rce r: s o c ia b ility , tension
re d u ctio n , courage, se xu a lity, c o g n itive im p a irm e n t, ris k and aggression, an d
self-perception. These are s im ila r to the dom ains id e n tifie d by o th e r
researchers (Brown, C hristiansen, & Goldm an, 1987; Young, Oei, & Crook, 1991).
Expectancy scales are se lf-re p o rt measures th a t ty p ic a lly cover m ost o f these
areas, and have been designed in a varie ty' o f form ats (see Leigh, 1989, fo r a
review ). F u rth e r a tte n tio n to expectancy scales is given under M ethods, below .
Research on the a b ility o f expectancy scales to p re d ict a lco hol
consum ption has found th a t expectancies are co rre la te d w ith d rin k in g
behavior in samples o f collegians, adolescents, a n d alcoholics undergoing
treatm ent, w ith positive expectancies (th e enjoyable effects) m ore h ig h ly
correlated th a n negative ones (the undesirable o r unpleasant effects). So
w h ile th in k in g th a t a lco hol is enjoyable is co rre la te d w ith h ig h e r levels o f
usage, th in k in g th a t i t is unpleasant o r has negative effects is less con sistently
correlated w ith low er levels o f usage (Leigh, 1989; Leigh & Stacy, 1993). These
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8
studies have exam ined the p re d ictive capabilities o f expectancy measures
using a v a rie ty o f d e p e n d e n t variables, in c lu d in g frequency o f d rin kin g ,
am ount, type o f a lco h o lic beverage, preferred settin g o f d rin k in g , and
d rin k in g "style s” such as problem d rin kin g s and fam ily d rin k in g (Leigh,
1989). For example, H ittn e r (1995) and Brown, Goldman and Anderson (1980)
found th a t “ h ig h ” versus “ low ” frequency d rin ke rs endorsed s ig n ifica n tly
greater te n sio n -re d u ctio n expectancies.
Measures o f q u a n tity and frequency are predicted d iffe re n tly by
expectancy scales. Expectancies are b e tte r able to p redict q u a n tity o f alcohol
consumed than frequency o f d rin k in g occasions (M ooney, Fromm e, Kivlahan,
& M arlatt, 1987). Leigh (1989) argues th a t this fin d in g makes sense because
many o f the effects o f alcohol are fe lt o n ly a fte r a certain num ber o f drinks.
In a d d itio n , she points o u t th a t college students, on whom m ost expectancy
research has been conducted, m ay be constrained by external factors in
d rin k in g frequency, such as accessibility o f alcohol fo r underage drin kers.
Such factors may n o t lim it q u a n tity consumed once d rin k in g has begun. When
q u a n tity and frequency measures are com bined, expectancy scales have been
found to explain between 10 and 19% o f variance in d rin k in g (Leigh, 1989).
M ost o f the studies o f the re la tion ship between alcohol consum ption and
expectancies are cross-sectional ra th e r than lo n g itu d in a l. As a result, it is
im possible to in fe r a d ire c tio n o f causality between the two (Leigh, 1989).
However, some lo n g itu d in a l studies have been conducted th a t suggest that
expectancies may be p re d ictive o f fu tu re d rin k in g behavior. Sm ith, Roehling,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Goldman and C hristiansen (1987) in a three-year lo n g itu d in a l study, found
tha t expectancies measured a t Year 1 predicted d rin kin g behaviors a t Years 2
and 3. Bauman, Fisher, Bryan and Chenoweth (1985) reported data fro m a
lo n g itu d in a l study sup porting a reciprocal m odel in w hich expectations o f
positive consequences predicte d consum ption, w hich in tu rn predicted
m o dificatio n o f expectancies. As th is study shows, expectancies not o n ly
precede d rin kin g behavior, b u t are a t least p a rtia lly form ed by past
experiences w ith alcohol. This im pact o f p rio r experiences w ith alcohol on the
form atio n o f expectancies seems p a rticu la rly im p o rta n t fo r old er adults. Given
th e ir lon ger life h isto ry, and greater o p p o rtu n ity fo r lea rning how alcohol
affects them past d rin k in g behavior may show a greater influence on
expectancies fo r o ld e r adults, ra th e r than expectancies in flu e n cin g behavior.
Studies exam ining the co rre la tio n o f expectancies to consum ption
measures in samples o f college students have often found sex differences.
However, results are som ewhat d iffic u lt to com pare because o f the d iffe re n t
expectancy and consum ption measures em ployed in each. Mooney and her
colleagues (Mooney e t al., 1987) found that fo r female u n ive rsity students,
frequent consum ption o f alcohol was best predicted by greater expectations
that alcohol w ould reduce tension. For male students, greater frequency was
correlated w ith expectations o f social and physical pleasure, global positive
feelings, and sexual enhancem ent. Q uantity o f alcohol consumed a t one
occasion was associated w ith social and physical pleasure and social assertion,
fo r bo th males and fem ales.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10
Some studies have found sex differences in mean expectancy scores as
w e ll as differences across sex in how expectancies p re d ic t consum ption. Brow n
and her colleagues (Brow n et al., 1980) found th a t wom en were m ore like ly
tha n m en to expect global positive effects o f alcohol. A nother stud y found th a t
wom en expected less pleasure and tension re d u ctio n and m ore cognitive and
m o tor im pairm ent, b u t d id not d iffe r from m en in expected enhancem ent o f
sexual pleasure, aggression o r expressiveness. (Rohsenow, D., 1983). In this
study, alcohol consum ption was sta tistica lly co n tro lle d , so th a t sex differences
were n o t confounded by consum ption.
Because expectancies are correlated w ith d rin k in g , the m o d ifia b ility o f
expectancies and the im pact o f a lte re d expectancies on re d u ctio n o f d rin k in g
is p o te n tia lly o f in te re s t to co g n itive ly o rie nte d therapists. A lcohol
expectancies m ay p ro vid e valuable in fo rm a tio n to clin icia n s regarding w hat
b e n e fit a person seeks in consum ing alcohol. From an in te rv e n tio n
stan dp oint, e ith e r a lte rin g expectancies o r fin d in g alternate ways to satisfy
the need expressed by expectancies may be used in preven tion o r treatm ent.
Thus far, however, little research has been conducted in attem p ting to a lte r
expectancies, and the few re ported results show m ixed success (Fromme,
M ooney, K ivlahan, & M a rlatt, 1986).
Expectancies o f O lder A dults
Some studies have found th a t expectancies va ry by age (M ooney et al,
1987; Lundahl, Davis, Adesso, & Lukas, 1997). However, these studies have been
lim ite d to samples o f adolescents and young adults. For exam ple, the study by
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
11
Lundahl found low er expectancies in college students over 20 as compared to
those under 20. P rio r to the c u rre n t study, no research to date has tested the
u tility o f expectancy scales w ith an old er sample.
In one study w hich exam ined self-reported em otional state o f older
adults ju s t p rio r to d rin kin g , C hristopherson, Escher and B ainton (1984) found
th a t lig h t d rin kers were more lik e ly to d rin k fo r social reasons such as special
occasions, to reduce social anxiety, and to have m ore confidence. Heavy
d rin ke rs, in contrast, were m ore lik e ly to d rin k to feel happier, to forget
problem s, to feel less lonely, and because the y were bored. In th is study, heavy
d rin ke rs were d e fined as those who drank a t least once a week and at least
th re e d rin ks per occasion, w ith 17% o f th e sam ple fa llin g in to this category.
A study by K lein and P ittm an (1993) suggests th a t as people get older,
em otion al state becomes less im p o rta n t in accounting fo r the am ount of
alco hol they consume. This stu d y investigated the re la tio n sh ip between
em otional state ju s t p rio r to d rin k in g and the am ount o f alcohol consumed, in a
na tiona l sample o f 1,069 adults. They found th a t fo r adults fro m 45 to 64 years
o ld , and aged 65 and older, n e ith e r positive n o r negative a ffe ct was predictive
o f greater use o f a n y form o f alcoholic beverage. As this stu d y o n ly considered
a ffect-relate d reasons fo r d rin k in g , no o th e r p re d ictive factors were id e n tifie d
in o ld e r people. Younger respondents, especially those aged 21 to 29, were
m ore lik e ly to o ffe r affect-related reasons fo r d rin k in g . T his study also fo u n d
s ig n ific a n t sex differences: w om en’s consum ption across age ranges was
greatest d u rin g the experience o f positive a ffective states, w h ile men tended to
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
12
d rin k he avily in response to both negative and positive affective states.
K lein and P ittm an’s inferences concerning the reduced ro le o f em otion
in m o tiva tin g d rin k in g behavior in o ld e r people has some support in research
concerned w ith changes in em otion over the lifespan (e.g., Labouvie-Vief,
Hakim -Larson, DeVoe, & Schoeberlein, 1989). A ccording to these theorists, the
em otions o f o ld e r adults are experienced as less intense and less extrem e in
valence. Given the fin d in g s th a t o ld e r adults consume less alcohol and have
low er rates o f alcohol-related problem s than younger adults, and the th e o ry
th a t e ffo rts to regulate em otions p la y a ro le in in itia tin g d rin kin g behavior,
the question then arises i f em otional factors co n trib u te to differences in
d rin k in g behavior. If o ld e r adults experience less extrem e o r less intense
em otions than younger adults, this m ig h t help to explain w hy o ld e r adults
d rin k less: they may have a less pressing need than younger adults to regulate
th e ir em otions. If so, it m igh t be expected th a t expectancies regarding a ffe ct-
related benefits o f alcohol w ould be low er in o ld e r adults than in younger
adults.
Aside fro m em otional factors, health concerns may also be re flected in
expectancy measures o f o ld e r adults. Due to the m any adverse effects on the
health o f o ld e r people as a result o f excessive d rin kin g , including increased
risk o f confusion and falls, it seems probable th a t o ld e r people may have
h ig her expectancies regarding adverse behavioral and cognitive effects. As
noted above, o ld e r adults are more lik e ly to abstain com pletely fro m alcohol
than younger adults. P rio r research has fou nd abstainers more lik e ly than
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
13
d rin ke rs to re p o rt negative expectancies regarding the effects o f alcohol,
although abstainers expect ju s t as m uch positive effect (Leigh, 1987). I f the
o ld e r sample contains hig her num bers o f abstainers, as expected, the o ld e r
sam ple should re p o rt hig her levels o f negative expectancies.
Hypotheses
In consideration o f this p rio r research, the fo llo w in g hypotheses w ere tested:
1. O lder adults w ould re p o rt less alcohol consum ption tha n younger adults.
2. Males w ould re p o rt less alcohol consum ption than females.
3. O lder adults w ou ld have low er scores on the subscales created to test
em otion-related expectancies: increase in positive affect and re d u ctio n o f
negative affect.
4. O lder adults w ould show hig her levels o f negative expectancies
regarding cognitive and behavioral effects o f alcohol than younger adults.
5. Positive expectancy levels w ou ld be m ore h ig h ly correlated w ith
consum ption o f alcohol fo r younger adults than fo r older adults.
Methods
P a rticip a n ts
Two groups o f pa rticipa nts were em ployed in the study: younger adults
and o ld e r adults. The younger a d u lt sample was draw n fro m the Psychology
D epartm ent undergraduate subject pool at the U niversity o f Southern
C alifornia, and ranged in age fro m 17 to 32, w ith a mean age o f 20.7 years. The
o ld e r a d u lt sample was composed o f volunteers a t the Andrus G erontology
Center (32 pa rticip a n ts), the R etired and Senior V olunteer Program o f Los
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
14
Angeles (40 p a rticip a n ts), and the Elks Club o f Anaheim (20 pa rticip a n ts). The
age o f the o ld e r sam ple ranged from 55 to 89 years old, w ith a mean age o f 71.2
years. An e th n ic a lly diverse sample was sought in subject re cru itm e n t by
targetin g senior organizations in a v a rie ty o f geographic areas o f Los Angeles
C ounty. The undergraduate population a t the U n iversity o f Southern C a lifornia
is also e th n ic a lly diverse. Dem ographic characteristics o f th e sample are listed
in Table 2 (below ). A p p ro va l fo r use o f the psychology subject pool was
obtained fro m the Psychology D epartm ent Hum an Subjects Com m ittee.
Procedure
P articipants w ere given two questionnaires to com plete, each fiv e pages
in length: one re garding a lco hol expectancy an d the o th e r regarding alcohol
consum ption and dem ographic in fo rm a tio n . Together these questionnaires
re q u ire d about fifte e n to tw e n ty m inutes to com plete. P articipants were to ld
th a t th e ir p a rtic ip a tio n in the study wr as v o lu n ta ry and th a t they could cease
com pleting the questionnaires a t any tim e and fo r any reason. A ll younger
p a rticipa nts, and m ost o ld e r participants, fille d in the questionnaires w h ile
seated in groups o f 5 to 15 people. About h a lf o f the o ld e r participants, those
who were re c ru ite d th ro u g h th e ir in vo lve m e n t in off-cam pus senior
vo lu n te e r groups, to o k the questionnaires hom e and then re tu rn e d them by
way o f volunteers w ho assisted in data co lle ctio n . For these participants,
envelopes were p ro vid e d in w hich to seal the com pleted questionnaires, to
ensure c o n fid e n tia lity in tra n s it.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
15
D em ographics
Dem ographic variables o f interest, aside from age, include race, sex and
education level. As discussed above, women generally d rin k less than m en at
a ll ages. In d ivid u a ls w ith less education tend to d rin k less than those w ith
m ore, a difference observed in o ld e r as w ell as younger ad ults (Bucholtz,
Sheline & Helzer, 1995). Variations by race in alcohol consum ption have also
been observed. For exam ple, w hile younger A frica n A m erican men re p o rt less
heavy alcohol use th a n w hite men, rates o f heavy d rin k in g are hig her fo r
o ld e r A frica n Am erican men than o ld e r w hite m en (see Gomberg and Nelson,
1995 fo r a review ). As p a rt o f the questionnaire, older participants were asked
to ind ica te th e ir age, sex, race, years o f education, m a rital status, occupation,
and re tire m e n t status. The questions fo r the younger sam ple was the same,
except th a t p a rticip a n ts were asked th e ir m a jo r instead o f occupation, parents’
incom e instead o f th e ir own personal income, and hours w orked per week
instead o f re tire m e n t status. (Not a ll o f these variables are analyzed in the
c u rre n t re p o rt).
A lcohol C onsum ption
To measure frequency, participants were asked how many tim es per
week and how' m any tim es per m onth they ty p ic a lly consume liquo r, w ine and
beer. The la rg e r o f the w e e kly/m o n th ly Figures was accepted as an in d ic a to r
o f frequency in the study, due to the tendency o f in d ivid u a ls to un der-re port
d rin k in g (W erch, 1989). A total fig u re fo r frequency was calculated b y adding
num ber o f occasions fo r liq u o r, w ine and beer together. (P articipants were
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
16
also asked to indicate d rin kin g d u rin g the week as opposed to the weekend.
Due to inconsistencies w ith other frequency measures, these responses a re n o t
included in the results.) Q uantity was obtained by asking how many d rin k s
(liq u o r, w ine o r beer) th a t pa rticipa nts consumed a t one tim e, and taking th e
largest fig u re o f the three types o f drin ks.
Participants were in itia lly asked w hether o r n o t they consume a lco h o l,
before th e y proceeded to answer questions regarding q u a n tity and frequency.
In a d d itio n , participants were asked to indicate th e ir height and weight, in
order to get a more accurate estim ate o f in to xica tio n level.
As discussed above, the re lia b ility o f se lf-re p o rt measures o f a lco h o l
consum ption w ith o ld e r adults has been questioned (Graham, 1986). This stu d y
addresses Graham’s concerns re garding m ental averaging by asking separate
questions regarding q u a n tity and frequency. The mem ory concern may be
somewhat alleviated by asking p a rticip a n ts to separately re p o rt consum ption
fo r d iffe re n t types o f alcohol. The social d e s ira b ility aspect o f this
m easurem ent problem , w hich may be o f greater concern in clin ica l settings
than in a research setting, is reduced by the anonym ous nature o f the
questionnaire.
Alcohol Expectancy Scale
The expectancy scale used in the study is a m odified version o f the
Comprehensive Effects o f Alcohol (CEOA) questionnaire (Fromme, Stroot, &
Kaplan, 1993). This instrum e nt consists o f 45 item s designed to measure
expectancies in fo u r po sitive areas— socia bility, tension re duction, courage,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
17
sexuality; and three negative areas— co g n itive a n d behavioral im pa irm e nt,
ris k and aggression, an d self-perception. This scale was created through
e xp lo ra to ry and c o n firm a to ry factor analyses, based on 139 o rig in a l items
taken from pre-existing expectancy scales. Responses to the item s are checked
on a 4 -p o in t L ike rt scale (l=disagree, 4=agree). Scoring o f the subscales is
conducted by averaging scores o f the co n stitu e n t item s. Scoring o f total
negative and to ta l po sitive expectancies m ay also be obtained b y averaging
subscale scores.
Fromme and h e r colleagues have re ported th a t the scale dem onstrates
adequate in te rn a l consistency, tem poral s ta b ility , and construct v a lid ity
(1 993 ). Regression analyses conducted w ith measures o f alcohol consum ption
showed tha t to ta l po sitive expectancy scores were s ig n ific a n tly correlated w ith
alco hol consum ption. U nfo rtun ate ly, Fromm e does n o t report m ean subscale
scores obtained in her stu d y o r co rre la tio n s o f in d iv id u a l subscales w ith
a lco hol measures.
The CEOA is designed to include a question asking the respondent to rate
the d e sira b ility o f each item , since people may place d iffe re n t values on
p a rtic u la r effects o f alco hol. For exam ple, some people may con sider risk-
ta kin g a negative e ffe ct and some a po sitive e ffe ct o f d rin kin g . The re su ltin g
p o sitive and negative value scales were n o t found to be sig n ific a n tly related to
any o f the alcohol use measures (From m e et al., 1993). Leigh (1989) has also
noted th a t p re d ictio n o f d rin k in g b e h a vio r is n o t im proved by adding
measures o f subjective value to expectancies, tho ug h the value measures are
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
18
useful in d e te rm in in g subject’s a ttitude s tow ard p a rtic u la r effects o f alcohol.
As a result, the present study included o n ly the expectancy measures from the
CEOA.
Two a d d itio n a l subscales have been added to th e CEOA fo r the current
study— a scale fo r expected decrease in negative affect (4 item s) and a scale fo r
increase in po sitive affect (3 item s). A lth o u g h other expectancy scales have
used item s such as these to d ire c tly in q u ire about expectancies concerning
em otion, the o rig in a l Fromme scale does n o t. These are a d d itio n a l item s that I
have created fo r the scale. R e lia b ility analysis (C ronbach’s alpha) was
perform ed fo r each subscale o n the questionnaire, and appears as Table 1.
Table 1:
R e lia b ility C oefficients fo r Expectancy Subscales (C ronbach’s A lpha)
Subscale
Older
P a rtic ip a n ts
Y o u n g e r
P a rtic ip a n ts
S o c ia b ility .92 .89
Tension R eduction .79 .76
Courage .89 .86
S exua lity .85 .81
Increased P ositive A ffe c t .89 .81
Reduced Negative A ffe ct .81 .77
C ognitive and Behavioral
Im p a irm e n t
.94 .87
Risk Taking .90 .85
Negative Self C oncept .78 .61
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19
A nalyses
The analyses used in testin g mean differences in alcohol consum ption
and expectancy scores, as w ell as the correlations between expectancy and
consum ption, are robust statistica l m ethods (W ilcox, 1996). Robust methods are
designed to increase the pow er o f sta tistica l analyses, p a rtic u la rly when the
assumptions o f m ore tra d itio n a l m ethods are n o t met. For MANOVA, these
assum ptions in clu d e n o rm a lity o f d is trib u tio n and e q u a lity o f variance
between groups. The v a lid ity o f these assum ptions is ve ry d iffic u lt to establish
in any study (W ilcox, 1996). The T2way test, w hich compares trim m ed group
means, was used in this study to test significance o f q u a n tity and frequency by
age group and sex. This test is analogous to MANOVA, b u t is more resistant to
o u tlie rs (W ilcox, 1996). T rim m ing 20% o f the cases fro m the analyses helps to
elim in ate o u tlie rs in the data. The sta tistica l m ethod o f trim m ing does not
re su lt in the “ discarding” o f data from the h ig h and low ends o f sam ple in a
way th a t w ould reduce power, as is sometimes thought. The tables in this study
show trim m ed means and trim m ed standard deviations- In calculating the
correlations, the percentage-bend co rre la tio n was em ployed. This test has
advantages o ve r the Pearson co rre la tio n coe fficient, such as greater
resistance to o u tlie rs (W ilcox, 1996).
Results
D em ographics
Women made up a m a jo rity o f bo th the older and younger groups, w ith
the younger group having an even greater percentage o f women (72%) than
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
the older group (57%). A Chi-Square test o f com position o f the o ld e r and
younger groups shows th a t there are s ig n ific a n tly m ore women in the
younger group, X2 (1) = 4.11, p = .043). In the c u rre n t study, numbers o f
m in o rity p a rticipa nts w ere n o t considered s u ffic ie n t to enable detailed
com parisons o f alcohol consum ption and expectancy to be made by race.
However, consum ption fo r w hites o n ly was analyzed by age group and race
(see below). A Chi-Square test o f the racial com position o f the o ld e r and
younger groups found s ig n ific a n t o ve ra ll differences, X2 (7) = 18.15, p = .011).
The younger sample had more Asian Am ericans than the older sample, w h ile
the old er sam ple had m ore A frican Am ericans. No sig n ifica n t differences in
education le ve l were re p o rte d — education level fo r the two groups was alm ost
exactly the same, w ith a mean o f 14.4 years fo r the younger group and 14.5 fo r
the old er group. This education level is som ewhat hig her than usual fo r an
o ld e r sample. The younger sample, o f course, has n o t ye t com pleted th e ir
education. Table 2 sum m arizes the dem ographic characteristics o f the sample:
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
21
Table 2:
Sample D em ographic C haracteristics
Younger A d u lts: O lder A dults:
(n=82) (n=92)
Mean age 20.7 (sd = 3,1) 71.2. (sd = 8.3)
Years education 14.4 14.5
E th n ic ity
W hite 50.0% 60.9%
A fric a n Am. 4.9 10.9
A sian 32.9 13.0
H ispanic 9.8 8.7
Native Am. 0.0 2.2
O th e r/ m ixed 2.4 0.0
Sex
M ale 28% 43%
Female 72 57
A lcohol C onsum ption
O f the o ld e r adults, 34 o f the 92 pa rticipa nts (34%) reported not
d rin k in g alcohol. O f the younger adults, 16 o f 82 (19%) do n o t d rin k alcohol. A
Chi-Square test o f num ber o f d rin ke rs versus abstainers in the two samples
was n o t q u ite sig n ifica n t, X2 (1) = 3.82, p = .051. W hile n o t significant, this
difference is too close to dismiss, given pow er considerations.
The T2way test, w hich com pares trim m e d group means, was used to test
significance o f q u a n tity and frequency by age group and sex. When a ll
pa rticipa nts were in clu d e d in these tests, T2w ay results o f differences in
q u a n tity o f alcohol consum ed showed a s ig n ific a n t m ain effect fo r age group,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
22
w ith younger adults d rin k in g m ore than o ld e r ones,critic a l value = 20.36, p<
.001, sex, w ith men d rin k in g m ore than women, c ritic a l value = 7.5, p < .01, and
a s ig n ific a n t in te ra ctio n between age group and sex, critica l value = 5.15, p<
.05. The difference in q u a n tity consumed is greater between the younger men
and wom en than it is in o ld e r men and women. Table 3 describes q u a n tity o f
alcohol consum ption b y age group and sex, in clu d in g a ll participants.
Table 3:
show ing Trim m ed Means and
T rim m ed Standard E rrors (A ll Subjects)
Sex
M ale Female
Y o u n g e r 3.40 1.7
(.67) (.24)
Age n=23 n=59
Older 1.06 0.78
(.22) (.18)
n=38 n=50
W hen o n ly d rin ke rs are included, T2way results o f differences in
q u a n tity o f alcohol consum ed showed a sign ifica nt m ain effect fo r age group,
w ith younger adults d rin k in g m ore than old er ones,c ritic a l value = 69.19, p<
.001, sex, w ith men d rin k in g m ore than women,critic a l value = 35.98, p <.001,
and a s ig n ific a n t in te ra ctio n between age group and sex, critical value =20.97,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
23
p < .05. The difference in q u antity consumed is greater between the younger
men and women than it is in old er men and women. See Table 4:
Table 4:
Q uantity : Mean Num ber o f Drinks at a Tim e,
showing Trim m ed Means and
Trim m ed Standard Errors (D rinkers O nlv)
Sex
Male Female
Y o u n g e r 4.17 2.03
(.33) (.20)
Age n=18 n=48
O lder 1.56 1.24
(.16) (.14)
n=26 n=33
T2w ay analysis fo r frequency o f alcohol consum ption by age group and
sex found no m ain effect fo r sex, no m ain effect fo r age group, an d no
sig n ifica n t in te ra ctio n between age group and sex w hen a ll subjects were
included. See Table 5 and fo r means and, standard deviations:
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
24
Table 5:
Frequency: Number of Drinking Occasions per Month-
showing Trimmed Means and Trimmed Standard
Errors fAll Subjects)
Sex
M ale Female
Y o u n g e r 5.1 2.12
(1.64) (.47)
Age n=23 n=59
O lder 6.43 3.79
(3.27) (1.48)
n=38 n=50
W hen o n ly d rin ke rs were included, th e re was a sig n ific a n t m ain e ffe ct fo r age
group, w ith o ld e r adults d rin kin g m ore fre q u e n tly th a n younger adults,
critic a l value = 6.57, p < .02. See Table 6 below fo r trim m ed means and trim m ed
standard errors. So, old er drinkers d rin k m ore often th a n younger ones, an
e ffe ct w hich disappears if abstainers (m ore num erous in the o ld e r group) are
includ ed .
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
25
Table 6:
Frequency: Number o f Drinking Occasions per Month,
showing Trimmed Means and Trimmed Standard
Errors (Drinkers Only)
Sex
M ale Female
T o u n g e r /.00 2.91
(1.62) (.47)
Age n=18 n=48
Older 13.67 8.48
(3.90) (2.44)
n=26 n=33
T2way analyses fo r q u a n tity and frequency were also conducted using
o n ly w hite subjects, in o rd e r to c o n tro l fo r the e ffe ct o f th e d iffe rin g e th nic
m in o rity com position o f the o ld e r and younger sample. C ross-tabulation
revealed sig n ifica n t e th nic co m p o sitio n differences in the tw o groups. In
p a rticu la r, the younger sample in clu d e d greater num bers o f Asian Am ericans,
w h ile the o ld e r sample had la rg e r num bers o f A frica n Am ericans. This
diffe ren ce m ay affect the c o m p a ra b ility o f the tw o samples in both alcohol
consum ption and expectancies. P rio r studies have found ra cia l differences in
consum ption levels in college sam ples (e.g., Hare, 1990) in w hich Asian
Am ericans and A frican Am ericans d rin k less th a n whites. Com paring
q u a n tity o f alcohol consumed, re sults found a sig n ifica n t m ain effect fo r age
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
26
group, w ith younger adults d rin k in g m ore than o ld e r one,critica l value =
16.00, p < .001, sex, w ith men d rin k in g m ore than women,critic a l value = 10.00,
p < .01, and a sign ifica nt in te ra ctio n between age group and sex, c ritic a l value
=7.00, p < .05. The difference in q u a n tity consum ed is greater between the
younger men and women than it is in o ld e r men and women. See Table 7.
Table 7:
Q uantity : Mean Num ber o f D rinks at a Tim e,
showing Trim m ed Means and
Trim m ed Standard Errors (A ll W hite Subjects)
Sex
M ale Female
Y o u n g e r 3.00 2.00
(.60) (.30)
Age n=12 n=27
Older 1.00 1.00
(-20) (.20)
n=26 n=28
T2way analysis fo r frequency o f alcohol consum ption by age group an d sex
fou nd no m ain effect fo r sex o r age group, and no sig n ifica n t in te ra ctio n
between age group and sex when o n ly w hite subjects were included. See Table
8 below:
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
27
Table 8:
Frequency: Number o f Drinking Occasions per Month,
showing Trimmed Means and Trimmed Standard
Errors (All White Subjects)
Sex
M ale Female
Y o u n g e r 7.00 2.49
(2.14) (.60)
Age n=12 n=27
Older 8.72 9.11
(3.90) (3.03)
n=26 n=28
Expectancy D ifferences
T2way analysis was used to com pare expectancy subscale scores by age
group and sex. In a ll sig n ifica n t age com parisons, younger ad ults scored
hig her than o ld e r ones. S ig n ifica n t m ain effects fo r age g ro u p were observed
fo r so cia b ility, critical value =22.87, p < .001; courage, critic a l value = 19.71, p <
.001; increased positive em otion, c ritic a l value = 6.27, p < .014; reduced negative
em otions, c ritic a l value = 19.15, p < .001; cognitive and behavioral im p a irm e n t
critical value =6.55, p < .012, and ris k taking c ritic a l value = 14.23, p < .001).
S ignificant differences w ere observed by sex fo r increase p o sitive em otion,
critica l value = 4.99, p <. 028, w ith m en scoring higher. Means and standard
deviations by sex and age group are shown in Table 9. For sex and age group,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28
no sig n ifica n t interactions w ere observed on any o f th e nine subscales.
Table 9:
Expectancy Score Means bv Sex and Age G roup, show ing T rim m ed Means and
t ---- ----- o --------- ■ ---■ ■-------------
Trim m ed S tandard Errors
Group
Subscale Younger Y ounger Older Older
M en W om en M en Women
(n=23) (n=59> (n=38) (n=50)
P ositive
S o c ia b ility * * * 3.16 2.98 2.57 2.30
(.13) (-10) (•13> (.17)
Tension 2.57 2.50 2.51 2.09
R eduction (.19) (.11)
(.14) (.15)
C o u ra g e *** 2.64 2.45 2.01 1.8
(-14) (.13) (.15) (.17)
S exuality 2.23 2.18 1.98 1.99
(.15) (-13) (.13) (.18)
H igher Pos.* 2.8 2.6 2.4 2.05
Em otion (.13) (-14) (.18) (.19)
Low er Neg. 2.68 2.68 2.18 1.96
Em otion
(.11)
(-14) (.13) (.18)
N esative:
Cog./Behav. 2.72 2.84 2.24 2.57
Im p a irm e n t* (.13) (.11)
017) (.18)
R isk T a k in g *** 2.79 2.50 1.90 2.21
(-14) (.12) (.15) 022)
N egative 1.97 2.12 1.67 1.95
S elf concept (.20) (.07) (.14) 015)
Age group d iffe re n ce s;
*S ig n ifica n t a t p<.05
* * * S ig n ifica n t a t p<.001
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29
C orrelations Between Expectancy Subscales and C onsum ption Measures
C orrelations between consum ption and expectancy measures were
calculated fo r d rin ke rs only, given the in te re st o f th is stu d y in reasons fo r
d rin k in g o f subjects who s till d rin k and m ay p o te n tia lly have alcohol
problem s. C orrelations were calculated separately by sex and age group, due to
the sig n ifica n t differences in alcohol consum ption observed. Due to the large
num ber o f correlations calculated (18 x 4 = 72), there is an increased chance
th a t some o f the correlations w ill be due to chance. For th is reason, o n ly those
corre la tio n s w ith p values o f less than .01 were considered sign ifica nt. For
younger m ale d rin ke rs, no sig n ific a n t correlatio ns betw een alcohol
consum ption and expectancy w ere found, probably due to low pow er resulting
from sm all (n=18) sample size: younger males showed sta tistica lly non
sig n ific a n t co rre la tio n s o f q u a n tity w ith courage (p b c o rr = .44, p = .070) and
s o c ia b ility (p b c o rr =.46, p = .052). For younger female d rin ke rs, sig n ifica n t
positive correlatio ns w ith q u a n tity o f alcohol consum ption were observed for
the expectancy subscales o f re d u ctio n o f negative em otion (p b c o rr = .38, p <
.01). For o ld e r male drinkers, the re were no sig n ifica n t correlations a t the .01
level. For o ld e r fem ale d rin ke rs, there was a sig n ifica n t negative co rre la tio n
between the expectancy o f co g n itive and behavioral im p a irm e n t w ith
q u a n tity (p b c o rr = -.49, p < .01). S ignificant correlations are listed on Table 10.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30
Table 10:
Percentage-Bend C orrelations o f C onsum ption Measures w ith Expectancies
Group V ariables C o rre la tio n
O lder Females im p a irm e n t w ith q u a n tity rpb = -.49**
im p a irm e n t w ith fre que ncy rpb = -.38*
poor se lf image w ith q u a n tity rpb = -.43*
poor s e lf image w ith frequency rpb = -.33*
O lder Males courage w ith frequency rpb = .49*
s o c ia b ility w ith frequency rpb = .41*
increased positive em otion
w ith frequency
rpb - .45*
im p a irm e n t w ith q u a n tity rpb = -.3 9 *
Younger Females tension reduction w ith q u a n tity rpb = .31*
low er negative em otion
w ith q u a n tity
rpb = .38**
courage w ith q u a n tity rpb = .34*
positive em otion w ith q u a n tity rpb - .30*
risk takin g w ith q u a n tity rpb = .32*
Younger Males no s ig n ific a n t co rre la tio n s
*sig n iflca n t at p < .05
**S ig n ific a n t a t p<.01
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
31
Discussion
A lcohol C onsum ption
The consum ption measures in this com parison o f tw o convenience
samples revealed d iffe re n t d rin k in g patterns in the o ld e r and younger groups.
It was hypothesized th a t o ld e r adults in this sample w o u ld d rin k less than
younger adults on measures o f both q u a n tity and frequency. This hypothesis
was correct fo r q u a n tity b u t n o t fo r frequency: when o n ly d rin ke rs are
includ ed in th e analysis, o ld e r adults d rin k m ore fre q u e n tly than younger
adults; the yo u n g e r sam ple d rin ks sig n ifica n tly more p e r occasion, w hether
abstainers are includ ed in the analysis o r not.
In the you nge r sam ple, the percentage o f abstainers (19%) was very
s im ila r to th a t observed in o th e r studies using larger samples o f college
populations (e.g., O’Hare, 1990; Haworth-Hoeppner, G lobetti, Stem, & Morasco,
1989). H aw orth and colleagues found 16% abstainers, w h ile O’Hare fo u n d 19%.
Q u a n tity and frequency were also sim ilar to findings fro m o th e r studies.
Mooney, Fromme, Kivlahan and M a ria tt (1987) found a m ean q u a n tity o f 3.2
d rin ks consum ed a t a tim e and 6.4 d rin kin g episodes pe r m onth. So d rin k in g
patterns in the younger sam ple were ro ugh ly com parable to w hat o th e r
investigators have found.
The percentage o f p a rticip a n ts who reported no d rin k in g a t a ll
(abstainers) in the older sam ple was 34%. This result is consistent w ith rates o f
abstention in o ld e r adults observed in other studies. This abstention rate is on
the low end o f the range o f abstainers found in other studies, w hich ranges
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
32
fro m 31% to 65% (Douglass, Schuster, & M cC lelland, 1988). However, o ld e r
d rin ke rs were oversam pled in the cu rre n t stu d y in order to have enough
subjects to use in the co rre la tio n analyses, w h ich exclude abstainers. In the
c u rre n t study, q u a n tity and frequency results w ere reported b o th in clu d in g
abstainers (Tables 3 and 6) and excluding them (Tables 4 and 7). Inclu ding the
abstainers gives a m ore accurate picture o f d rin k in g behavior in the sam ple
as a whole; excluding them is m ore accurate in describing the q u a n tity and
frequency o f the p o rtio n o f the sample th a t does d rin k .
Few studies th a t measure d rin k in g o f o ld e r adults report q u a n tity and
frequency, o p tin g instead to re p o rt c lin ic a lly s ig n ific a n t consum ption levels
o r categories o f d rin k in g such as “ lig h t,” “ m oderate” and “heavy” that
confound q u a n tity and frequency (e.g., Barnes, 1979; Christopherson et al.,
1984; M olgaard e t al., 1990). Findings fro m such studies consistently re p o rt
th a t o ld e r adults are m ore lik e ly to be lig h t d rin k e rs — either d rin k in g ra re ly
in m oderate q u antitie s o r re g u la rly in sm all q u a n titie s. A lthough the
q u a n tity-fre q u e n cy m ethodology used in the c u rre n t study is com m on am ong
expectancy studies in college samples, it is d iffic u lt to compare these raw
alcohol consum ption figures w ith the categories re p o rte d in m ore c lin ic a lly
orie nte d studies o f o ld e r samples. One study w h ich does report q u a n tity and
frequency fou nd th a t o ld e r male drin kers consum ed 1.3 drinks per day, w h ile
fem ale d rin ke rs consumed 1 d rin k per day (Adams e t al, 1990). This q u a n tity o f
d rin k in g is in agreem ent w ith the cu rre n t study, although frequency is n o t
clear fro m Adam ’s re port.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
33
For o ld e r adults who d rin k it is possible th a t frequency m ay become
m ore im p o rta n t than q u a n tity pe r occasion in the id e n tific a tio n o f d rin k in g
problem s. Some authors have suggested th a t thresholds o f d rin k in g used to
define heavy d rin k in g (5 to 7 d rin ks at once) m ay n o t be a p p ro p ria te fo r old er
adults (Beresford, 1995). Physiological studies have shown th a t due to
decreased lean body mass in o ld e r adults, less alcohol is re q u ire d o f a typica l
o ld e r person to achieve the same in to xica tio n level th a t a you nge r person
experiences w ith greater intake (Beresford & Lucey, 1995). C onsum ption levels
reported in the c u rre n t study suggest a pattern o f frequent, lo w -q u a n tity
d rin k in g w hich may be com patible w ith alcohol dependence, especially in
e ld e rly wom en who m ay d rin k sm all amounts and y e t s till develop dependence
w ith adverse consequences (W ilsnack e t al., 1995). However, the consum ption
measures includ ed in the cu rre n t stu d y were designed p rim a rily fo r th e ir
a b ility to be correlated w ith alcohol expectancy measures, and are not
intended to give an estim ation o f alcohol abuse o r dependence in e ith e r the
younger o r the o ld e r sample.
Sex differences fou nd in the younger sam ple are also consistent w ith
p rio r research th a t has found low er d rin k in g rates in fem ales than males in
undergraduate populations (Haworth-Hoeppner, G lobetti, Stem, & Morasco,
1989) as w e ll as across a ll age ranges (Bucholtz, Sheline & Helzer, 1995; M u lford
& M iller, 1959). Results found an in te ra ctio n o f effects o f age group and sex on
am ount o f alcohol consumed; q u a n tity o f d rin k in g seems to converge in the
o ld e r sample, due to low er o f am ounts o f alcohol consumed by o ld e r males as
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
34
com pared to younger males. For the male sample, these results are consistent
w ith studies showing declines in heavy d rin k in g w ith age (Adams et al., 1990).
Post-hoc analyses o f q u a n tity and frequency o f alcohol consum ption
measures were perform ed using o n ly w hite subjects, in o rd e r to c o n tro l fo r
the effect o f d iffe rin g e th n ic m in o rity com position o f the o ld e r and you n g e r
sample. G roup means were com pared by age and sex. Results o f these analyses
were the same as those fou nd using the en tire sam ple (in clu d in g abstainers),
suggesting th a t va ria tio n s in the e th n ic com position o f the tw o samples d id not
change the results o f these group com parisons.
In sum m ary, the alcohol consum ption results suggest th a t p o te n tia l
com plexities exist in attem pts to com pare the d rin k in g habits o f o ld e r and
younger adults: o ld e r adults are m ore like ly to abstain from alcohol an d to
consume less alcohol a t a tim e. They are less lik e ly to be binge d rin ke rs, b u t
m ay d rin k m ore fre q u e n tly than younger adults. In m easuring consum ption
in older samples, researchers and clinician s may need to take in to account th a t
averaging q u a n tity and frequency m ay obscure differences in consum ption
across age ranges, especially if abstainers are in clu d e d in the calculations.
Expectancy Scores
Age Group D ifferences. This study is the firs t to examine alcohol expectancies
in older adults, so the hypotheses regarding expectancy differences w ere
somewhat tentative. It was hypothesized tha t o ld e r adults w ould show h ig h e r
levels o f expectancies on the negative subscale o f cognitive and be havio ral
im pairm ent, due to the increasing im pact o f alco hol on physical ailm ents and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
35
fa lls in o ld e r adults. However, th is hypothesis was n o t supported: the younger
sam ple scored s ig n ific a n tly h ig h e r on th is subscale. One possible explanation
fo r th is difference is th a t the item s on th is subscale do n o t adequately
represent alco hol-relate d concerns o f o ld e r adults. Item s such as “ I w ould feel
clum sy,” o r “ m y senses w ould be d u lle d ” m ay n o t capture w orries about
effects o f alcohol on health. Wh ile th is subscale measures sh o rt-te rm negative
aspects o f in to xica tio n , these m ig h t n o t be tro u b lin g in an o ld e r sample that
d rin ks sm all q u a n titie s o f alcohol.
O lder adults were expected to score low er on the subscales o f decreased
negative a ffe ct and increased po sitive affect, due to a possible greater reliance
o f younger adults on alcohol to regulate em otion as com pared to o ld e r adults.
This hypothesis appeared ju s tifie d on the basis o f the w o rk o f K lein and
P ittm an (1993) w hich fo u n d th a t o ld e r adults offere d less affect-related
reasons fo r d rin k in g than younger adults, and concluded th a t em otional state
was less im p o rta n t in accounting fo r the am ount o f alcohol th a t adults
consume as they get old er. The hypothesis appeared co rre ct fo r the subscale
m easuring increased p o sitive a ffect. In lig h t o f the generally h ig her subscale
scores o f the younger adults, th is re s u lt should be in te rp re te d cautiously: one
h ig h e r subscale m ean score m ig h t n o t be ve ry strong evidence fo r this
the ory. On the o th e r hand, younger wom en w ho made up m ost o f the younger
sam ple showed a s ig n ific a n t c o rre la tio n between q u a n tity o f alcohol consumed
and the subscale o f low er negative affect, and o ld e r wom en d id not. This
fin d in g m ig h t also in d ica te greater reliance o f younger adults on alcohol to
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
36
regulate em otion. It is possible th a t the reduced need o f o ld e r adults to regulate
em otion co u ld explain the lower expectancy scores on the positive subscales o f
courage and socia b ility as well. In o th e r w ords, increased courage and
so cia b ility m ay also be aspects o f em otion re g u la tio n in greater demand by the
younger sam ple than the older one.
There are oth er possible explanations to the low er alcohol expectancy
scores re ported by o ld e r adults. One o f these is th a t alcohol expectancies play a
d iffe re n t ro le in d rin k in g behavior in older ad ults than th e y do in younger
people. As described e a rlie r, expectancies are cognitive com ponents o f a
decision-m aking process th a t may be divide d in to two parts— in itia tio n and
m aintenance (Oei & Baldwin, 1994). In the in itia tio n phase, an in d ivid u a l has
p a rtic u la r expectancies regarding the effect o f alcohol, com bined w ith an
evaluation o f how desirable the e ffe ct is, in m aking a decision to d rin k. Once
d rin k in g behavior has been in itia te d , o r once a pattern o f d rin k in g has been
established, th is cognitive com ponent may recede from consciousness, and the
response o f d rin kin g m ay follow antecedent s tim u li in a m ore o r less autom atic
fashion. This is the m aintenance phase o f d rin k in g , in w h ich e xp licit
evaluation o f consequences may have ceased. Because older adults tend to have
been d rin k in g fo r much longer tha n younger adults, perhaps 50 or m ore
years as com pared to 2 o r 3, au tom a ticity m ay be greater and expectancies less
in flu e n tia l. Oei and Baldw in (1994) have argued th a t expectancies make th e ir
c o n trib u tio n to encouraging o r discouraging d rin k in g b e havio r in the
a cq u isitio n phase, and th a t after d rin k in g habits have been acquired the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
37
in d iv id u a l experiences an “ autom atic process” in w hich cues, d rin k in g
behavior and feelings are in e xtrica b ly bound together. Oei and Baldw in
describe the in itia tio n o f d rin k in g behavior in terms o f c o n tro lle d processing
and instrum ental lea rning , w hich em phasize conscious co g n itive processing.
They argue that the m aintenance phase o f d rin kin g represents autom atic
processing o r classical co n d itio n in g , in w hich a sequence o f b e h a vio r occurs
w ith o u t conscious c o n tro l once begun.
In the cu rre n t study, we asked participants to re p o rt co g n itive
com ponents o f d rin k in g behavior regarding positive and negative effects o f
alcohol. If the model described above is correct, the role o f c o g n itio n in the
d rin k in g behavior o f the two groups in the study may be d iffe re n t Reduced
salience o f alcohol-related cognitions in the older sample, due to a s h ift fro m
co n tro lle d to autom atic processing, may have resulted in lower endorsem ents
o f a range o f effects o f alcohol.
Sex D ifferences. No p a rticu la r sex differences were hypothesized fo r this
study, although previous studies have fou nd differences in expectancy levels
by sex. Brown and h e r colleagues (1980) found that women were m ore lik e ly
than men to expect global positive effects o f alcohol. Rohsenow (1983) found
th a t women expected less pleasure and tension reduction and m ore cognitive
and m o tor im pairm ent, b u t d id n o t d iffe r fro m men in expected enhancem ent
o f sexual pleasure, aggression o r expressiveness. In Rohsenow’s stud y, alcohol
consum ption was sta tistica lly co n tro lle d , so th a t sex differences w ere not
confounded by consum ption. In the c u rre n t study, men scored s ig n ific a n tly
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
38
hig her on the subscale o f increased positive em otion, a re s u lt th a t appears
consistent w ith Rohsenow’s. One possible reason fo r n o t fin d in g more
sig n ifica n t sex differences in the c u rre n t study was the lo w num ber o f males,
unequal sample sizes fo r males and females, and re su ltin g low power o f the
com parisons.
C orrelation Between A lcohol C onsum ption and Expectancy
In the cu rre n t study, few correlations were s ig n ific a n t a t the .01 level.
However, trends in the data, in c lu d in g several co rre la tio n s sig n ifica n t a t the
.05 level, suggest sex and age group differences w hich fu rth e r studies may
seek to replicate. (Table 6). These results suggest th a t in you nge r women,
positive expectancies p re d ic t q u a n tity o f consum ption; in o ld e r men positive
expectancies p re d ict frequency; in o ld e r women negative expectancies p re d ict
bo th q u a n tity and frequency. In the younger sample, fin d in g s regarding sex
difference in the a b ility o f alcohol expectancies to p re d ic t consum ption were
sim ila r to those observed by o th e r researchers. In a s tu d y o f college
undergraduates, Mooney and h e r colleagues (1987) fo u n d th a t fo r both males
and females, q u a n tity was s ig n ific a n tly correlated w ith expectations o f social
and physical pleasure, and social assertiveness. For m ales, frequency was
correlated w ith sexual enhancem ent and social and p h ysica l pleasure, w hile
fo r females frequency was co rre la te d w ith tension re d u ctio n . A lthough the
scales used in M ooney’s stud y are d iffe re n t from those used in the cu rre n t
study, the sex differences she observed are consistent w ith the cu rre n t study.
If it is the case th a t expectancies play a role in in itia tin g d rin kin g
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
39
behavior, age is im p o rta n t in determ ining w hich expectancies are operating.
In these two samples, th is difference was m ost cle a r between younger and
o ld e r wom en. Even though o ld e r adults d id n o t score higher on the subscale o f
co g nitive and be havio ral im pairm ent, fo r exam ple, concerns a b o u t these
effects o f alcohol, a t least fo r o ld e r women, are correlated w ith reduced
d rin k in g be havio r in a w ay th a t is not present in younger wom en. On the
o th e r hand, p o sitive beliefs about the effects o f alcohol appear to play a
greater ro le in the d rin k in g behavior o f younger women than o ld e r women.
Results o f the present stu d y may have c lin ic a l im plications fo r the
p reven tion and tre a tm e n t o f excessive alcohol use. One possible d ire ctio n in
tre a tm e n t is to a tte m p t to a lte r expectancies regarding the effects o f alcohol.
Oei and Baldw in (1994) have argued fo r this approach, stating th a t effective
co g n itive -b e h a vio ra l in te rve n tio n s fo r substance abusers m ust invo lve the
d is ru p tio n o f autom a tic processes o f d rin kin g . Such treatm ent involves
educating in d iv id u a ls re garding in te rn a l states th a t have cued d rin k in g
b e havio r in the past, increasing behavioral alte rn a tive s to d rin k in g behavior,
understanding th e ir ow n expectations regarding alco hol consum ption as w e ll
as expectations regarding behaviors th a t m ay serve as alternatives to d rin k in g
behavior. E xp licit discussion o f expectancies regarding alcohol consum ption
involves d ire c t attem pts to increase negative expectancies and decrease
positive ones. A ttem pts to a lte r expectancies as a com ponent o f prevention o r
tre atm e nt have had m ixed results. Fromme, M ooney, Kivlahan and M a rla tt
(1986) fo u n d th a t an eight-w eek in te rve n tio n program attem p ting to change
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
expectancies was successful in low ering levels o f d rin k in g , b u t d id not re sult
in actual changed expectancies. A lthou gh some authors have recommended
fu rth e r e xp lora tion o f this approach (Oei & Baldwin, 1994), it m ay be that
changing beliefs about the effects o f alcohol are n o t necessary' to reduce
d rin k in g .
The c u rre n t stud y suggests th a t w hich expectancies p la y a significant
ro le in d rin k in g behavior depends to a certain exte nt on an in d iv id u a l’s age
and sex. For o ld e r adults, this w ould suggest th a t wom en’s d rin k in g m ight be
re ducib le by increasing negative beliefs about the effects o f alcohol; d rin k in g
o f o ld e r m en and younger women m ig h t be reducible if positive beliefs are
decreased, o r i f the beneficial effect expected from alcohol can be provided in
some o th e r fashion. In this la tte r approach, alte ra tio n o f expectancies m ight
n o t be the goal o f inte rven tion. Rather, expectancy scales m ig h t be useful as a
diagnostic to o l in the id e n tifica tio n o f coping o r inte rpe rson al skills deficits
w hich can the n be addressed d ire c tly . Prevention and tre a tm e n t programs
designed to address bo th coping and social sk ill d e ficits are w e ll established,
and have been e m p irica lly supported b y a num ber o f outcom e studies (see
M ille r e t al., 1995, fo r a review).
I f low er mean level expectancy levels in o ld e r adults are explainable in
term s o f reduced salience o f cognitive com ponents o f d rin k in g behavior in
fa vo r o f greater autom atic processing a fte r num erous years o f d rin kin g , the
a lte ra tio n o f expectancies o f o ld e r adults m ight be m ore d iffic u lt than for
younger adults. I f d rin k in g levels o f o ld e r adults are re la tiv e ly stable over
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
41
tim e (Adams e t al., 1990), m any years w ill have passed since the acq uisition o f
alcohol consum ption behavior patterns. This could m ake autom atic processing
m ore d iffic u lt to disru pt.
Am ong older alcoholics, age o f onset could p la y a role in de term ining
how am enable expectancies are to change. In late-onset alcoholics,
dependence on alcohol o fte n emerges in co n ju n ctio n w ith , and possibly as a
response to stressors associated w ith age such as bereavem ent and retirem ent.
It seems lik e ly tha t m aladaptive behavior o f this s o rt w h ich has re ce n tly
em erged is m ore easily d isru p te d through cognitive an d behavioral
in te rve n tio n s. Treatm ent studies com paring early- an d late-onset alcoholics
are consistent w ith this hypothesis, in th a t late-onset alcoholics appear to
respond b e tte r to cognitive-behavioral in te rve n tio n s (Schonfeld & Dupree,
1991).
Lim ita tion s o f the study
In re cru itin g p a rticip a n ts, an e ffo rt was made to include enough older
d rin ke rs to obtain a va rie ty o f consum ption levels in th e o ld e r sample, so th a t
correlatio ns o f consum ption w ith expectancies could be measured w ith o u t
having the sample skewed tow ards in clud ing o n ly abstainers o r lig h t
drin ke rs. T his was achieved th ro u g h re c ru itin g o ld e r pa rticipa nts a t social
organizations where d rin k in g was known to take place (e.g., The Elks Club). As
a result, it is probable th a t th e alcohol consum ption levels are not
representative o f the o ld e r p o p u la tio n at large. Because a ll older subjects were
re cru ite d a t com m unity vo lu n te e r organizations, the re are other aspects in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42
w hich the sam ple m ay not be repre senta tive o f the general e ld e rly population,
possibly in c lu d in g hig her levels o f e d ucatio n and be tter physical health.
As noted above, re la tive ly sm all num bers o f men in the younger sample
may have decreased power in the analyses o f sex differences. M ore broadly,
the d iffe re n t com positio n o f the o ld e r a n d younger groups by sex m ust q u a lify
com parisons m ade between the o ld e r a n d younger groups— apparent age
differences m ay have been influ ence d b y sex. It is lik e ly th a t the re la tive ly
large num bers o f wom en and m in o ritie s in the younger sample m ay have
resulted in u n derestim a tion o f d rin k in g q u a n tity', and w ould make
consum ption fin d in g s in the younger sam ple d iffic u lt to generalize to other
samples th a t w ere n o t sim ila rly com posed.
Conclusion and D irectio ns fo r Future Research
To m y know ledge, this stu d y is th e firs t to explore expectancy scores o f
o ld e r adults and th e ir relationship to d rin k in g behavior. It was fo u n d tha t
o ld e r adults re p o rt low er levels o f bo th p o sitive and negative expectancies in
com parison to a sam ple o f college students, a re sult th a t rem ained even after
the difference in q u a n tity o f alcohol consum ed by the two samples was
sta tistica lly c o n tro lle d . The reasons fo r th is broad difference rem ain to be
explored. Expectancies are generally conceived as cognitive evaluations o f the
benefits and costs o f alcohol consum ption. It m ay be the case th a t these
cognitions play a d im in ished ro le in the d rin k in g behavior o f o ld e r adults,
whose d rin k in g b e h a vio r is w e ll established. A lte rn a tive ly, o ld e r adults may
re ly on alcohol to regulate th e ir em o tio n a l states less than younger adults.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
43
Future investigations m ay elaborate these d iffe re n t possibilities. C orrelatio n
analyses in the cu rre n t study indicate th a t the relationship o f expectancies to
alcohol consum ption varies by sex and age group, o ffe rin g im p o rta n t data fo r
those p la n n in g interventions. For o ld e r adults, these results suggest th a t
tre atm e nt and preven tion efforts m ay be designed somewhat d iffe re n tly fo r
men than fo r women.
As discussed above, sam pling issues and the generalizability o f the
samples to the broader population are also o f concern. Future studies m ay wish
to re plicate these results using robust sta tistica l methods and a larger sample
to increase power. The cu rre n t study com pares two samples approxim ately 50
years a p a rt in age. If the re are changes w ith age, this m ethod says little about
when the changes occur. It would be in te re stin g to replicate this stu d y w ith a
broader selection o f age groups, w hich co u ld examine the role o f co g n itio n in
d rin k in g across age ranges, and id e a lly to use life span developm ental designs
to look a t age, cohort and time effects. In a d d itio n , it w ould be useful to extend
expectancy m easurem ent to clin ica l populations o f older a d u lt d rin ke rs and
samples in c lu d in g la rg e r numbers o f e th n ic m inorities, and to analyze
differences in expectancies between e th n ic groups.
Such studies have the p o te n tia l to make valuable co n trib u tio n s to the
un d e rsta n d in g o f d rin k in g behavior in c u rre n tly under-researched
populations. Given dem ographic trends in the population and projected
increases in num bers o f older d rin kers, m ore research in to the d rin k in g
behavior o f o ld e r adults is needed. This fie ld o f study has the po ten tial to make
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
44
a great im pact on one o f the m ajor m ental health and pu blic health issues that
we face, and to s ig n ifica n tly increase q u a lity o f life fo r o ld e r adults.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
45
R eferences
Adams, W. L., Garry, P. j., Rhyne, R., H unt, W. C., & Goodwin, J. S. (1990).
A lcohol intake in the healthy e ld e rly: Changes w ith age in a cross-sectional
and lo n g itu d in a l study, lo u m a l o f the A m erican G eriatrics Society, 38. 211-16.
Barnes, G. M. (1979). A lcohol use am ong old er persons: Findings from a
W estern New York state general p o p u la tio n survey, lo u m a l o f the Am erican
G eriatrics Society. 6. 244-250.
Bauman, K. E., Fisher, L. A., Bryan, E. S., & Chenoweth, R. L. (1985).
R elationship between subjective expected u tility and behavior: A lo n g itu d in a l
stu d v o f adolescent d rin k in g behavior, lo u m a l o f Studies on A lcohol. 41. 952-
958.
Beresford, T. P. (1995). A lcoholic e ld e rly : Prevalence, screening,
diagnosis and prognosis. In T. Beresford an d E. Gomberg (Eds.), A lcohol and
a g in g (pp. 3-18). New York: O xford U n ive rsity Press.
Beresford, T. P., Blow, F. C., & Brower, K. J. (1990). Alcoholism in the
e ld e rly. Com prehensive Therapeutics. 16. 38-43.
Beresford, T. P., & Lucey, M. R. (1995). Ethanol m etabolism and
in to xica tio n in the elderly. In T . Beresford and E. Gomberg (Eds.), A lcohol and
a g in g (pp. 117-127). New York: O xford U nive rsity Press.
Brown, S. A., Christiansen, B. A., & Goldm an, M. S. (1987). The alcohol
expectancy questionnaire: An in stru m e n t fo r the assessment o f adolescent and
a d u lt alcohol expectancies. Toumal o f Studies on A lcohol. 48. 483-491.
Brown, S. A., Goldman, M. S., & C hristiansen, B. A. (1985). Do alcohol
expectancies m ediate d rin k in g patterns o f adults? lo u m a l o f C onsulting and
C lin ica l Psychology. 53. 123-130.
Brown, S. A., Goldman, M . S., Inn, A., & Anderson, L. (1980). Expectations
o f re in forcem e nt fro m alcohol: T h e ir do m ain and re la tio n to d rin k in g
patterns, lo u m a l o f C onsulting and C lin ica l Psychology. 48. 419-426.
Bucholz, K. K., Sheline, Y. I., & Helzer, J. E. (1995). The epidem iology o f
A lcohol use, problem s, and dependence in elders: A review . In T. Beresford and
E. Gomberg (Eds.), A lcohol and aging (pp. 19-41). New York: O xford U niversity
Press.
C hristopherson, V. A., Escher, M. C., & Bainton, B. R. (1984). Alcohol
patterns am ong ru ra l aged in Arizona, lo u m a l o f Studies on A lcohol. 45. 417-
423.
Cooper, M. L., Frone, M. R., Russell, M., & M udar, P. (1995). D rin kin g to
regulate positive and negative em otions: A m o tiva tio n a l m odel o f alcohol use.
lo u m a l o f P ersonality and Social Psychology. 69. 990-1005.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
46
Cooper, M. L., Russell, M., & George, W. H. (1988). Coping, expectancies,
and alcohol abuse: A te st o f social le a rn in g form ulations, lo u m a l o f A bnorm al
Psychology. 97. 218-230.
Cox, W. M. & K linger, E. (1988). A m o tivation al model o f alcohol use.
lo u m a l o f A bnorm al Psychology. 97. 168-180.
Douglass, R. L., Schuster, E. O., & McClelland, S. C. (1988). D rinking
patterns a n d abstinence among the e ld e rly. The In te rn a tio n a l lo u m a l o f the
Addictions. 23. 399-415.
Farber, P. D., K havari, K. A., & Douglass IV, F. M. (1980). A factor a n a lytic
study o f reasons fo r d rin k in g : E m pirical va lid a tio n o f positive and negative
re in fo rce m e n t dim ensions. lo u m a l o f C onsulting and C linical Psychology. 48.
780-781.
Fillm ore, K. M., & M idanik, L. (1983). C h ro n icity o f d rin k in g problem s
among m en: A lo n g itu d in a l study. lo u m a l o f Studies on Alcohol. 45. 228-236.
Fromme, K, Mooney, D. K., K ivlahan, D. R., & M arlatt, G. A. (1986).
Expectancies, risk id e n tific a tio n , an d secondary prevention w ith problem
d rin ke rs. Advances in B ehavior Research and Therapy. 8. 237-251.
Fromme, K., S troot, E., & Kaplan, D. (1993). Com prehensive effects o f
alcohol: D evelopm ent an d psychom etric assessment o f a new expectancy
questionnaire. Psychological Assessment. 5. 19-26.
Gam bert, S. R., & Katsoyannis, K. K. (1995). A lcohol-related medical
disorders o f o ld e r heavy drin kers. In E. Gomberg and T. Beresford (Eds.),
A lcohol and aging (pp. 70-81). New York: O xford U niversity Press.
G lynn, R. J., Bouchard, G. R., LoCastro, J. S., & Laird, N. M. (1985). Aging
and generational effects o f d rin k in g behavior in m en: Results fro m the
norm ative aging study. Am erican lo u rn a l o f Public Health. 75. 1413-1419.
Gomberg, E., & Nelson, B. W. (1 995 ). Black and w hite o ld e r men: A lcohol
use and abuse. In E. Gom berg and T. Beresford (Eds.), Alcohol and aging (pp.
307-323). New York: O xfo rd U n ive rsity Press.
Graham, K. (1 986 ). Id e n tify in g and m easuring alcohol abuse among the
e ld erly: Serious problem s w ith e xistin g instrum ents, lournal o f Studies on
Alcohol. 47. 322-326.
H aw orth-Hoeppner, S., G lobetti, G, Stem, J., & Morasco, F. (1989). The
q u a n tity and frequency o f d rin k in g am ong undergraduates a t a southern
un ive rsity. The In te rn a tio n a l lo u m a l o f the A ddictions. 24. 829-857.
H ittn e r, J. B. (1 995 ). T ension-reduction expectancies and alcoholic
beverage preferences re visite d : A ssociations to d rin k in g frequency and
gender. The In te rn a tio n a l jo u rn a l o f the A ddictions. 30. 323-336.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
47
K lein, H., & Pittm an, D. J. (1993). The relationship between em otional
state and alcohol consum ption. The In te rn a tio n a l lo u m a l o f the A ddictions. 28.
47-61.
Labouvie-Vief, G, Hakim -Larson, J., DeVoe, M., & Schoeberlein, S. (1989).
Emotions and self-reg ulatio n: A life span. view. Human Developm ent. 32. 279-
299.
Leigh, B. C. (1987) Beliefs about the effects o f alco hol on the se lf and
others, lo u m a l o f Studies on A lcohol. 48. 467-475.
Leigh, B. C. (1989). In search o f the seven dwarves: Issues o f
m easurem ent and m eaning in alcohol expectancv research, Psychological
B ulletin. 105. 361-373.
Leigh, B. C., & Stacy, A. W. (1993). A lcohol outcom e expectancies: Scale
co n stru ctio n and p re d ictive u tility in h ig her o rd e r co n firm a to ry models.
Psychological Assessment. 5. 216-227.
Lundahl, L. H., Davis, T. M., Adesso, V. J., & Lukas, S. E. (1997). Alcohol
Expectancies: Effects o f gender, age, and fa m ily h isto ry o f alcoholism ..
A ddictive Behaviors. 22. 115-125).
M idanik, L. (1982). The v a lid ity o f self-reported a lco h o l consum ption
and alcohol problem s: A lite ra tu re review . B ritish lo u m a l o f A ddiction. 77.
357-382.
M ille r, N. S., Belkin, B. M., & Gold, M. S. (1991). A lcohol and drug
dependence am ong the e ld e rly: Epidem iology, diagnosis, an d treatm ent.
C om prehensive Psvchiatrv. 32. 153-165.
M ille r, W.R., Brown, J. M., Simpson, T. L., Handm aker, N. S., Bien, T. H.,
Luckie, L. F., M ontgom ery, H. A., Hester, R. K., & Tonigan, J. S. (1995). What
works? A m ethodological analysis o f the alcohol tre a tm e n t outcome lite ra tu re .
In R. K. Hester & W. M. M ille r (Eds.), Handbook o f alcoholism treatm ent
approaches (pp. 12-44). Boston: A lly n and Bacon.
M olgaard, C. A., Nakamura, C. M., Stanfor, E. P., Peddecord, K., M., &
M orton, D. J. (1990). Prevalence o f alcohol consum ption am ong o ld e r persons.
lo u m a l o f C om m unity Health. 15. 239-251.
Mooney, D. K, Fromme, K., Kivlahan, D. R., & M arlatt, A. (1987).
Correlates o f alcohol consum ption: Sex, age and expectancies relate
d iffe re n tia lly to q u a n tity and frequency. A ddictive Behaviors. 12. 235-240.
M u lford , H. (1983). Stress, alcohol intake and problem d rin kin g in Iowa.
In L. A. Pohorecky & J. B rick (Eds.), Stress and Alcohol Use. Proceedings o f the
F irst In te rn a tio n a l Sym posium on Stress and A lcohol Use (pp. 321-332). New
York: Elsevier.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
48
M ulford, H. & M ille r, D. (1959). D rin kin g in Iowa: S ociocultural
d is trib u tio n o f D rinke rs. Q uarterly foum al o f Studies on A lcohol. 20. 704-726.
Oei, T. P. S., & Baldwin, A. R . (1994). Expectancy th e o ry: A two-process
m odel o f alcohol use and abuse, loum al o f Studies on A lcohol. 55. 525-534.
O’Hare, T. M. (1990). D rinking in college: C onsum ption patterns,
problem s, sex differences and legal d rin k in g age. lou m al o f Studies on
A lcohol. 51. 536-541.
Osgood, N. J., Wood, H. E., & Parham, I. A. (1995). A lcoholism and aging:
An annotated b ib lio g ra p h y and review. W estport. CT: Greenwood Press.
Rohsenow, D. (1983). D rinking habits an d expectancies about a lco h o l’s
effects fo r self versus others, loum al o f C onsulting and C lin ica l Psychology.
51* 752-756.
Schonfeld, L., & Dupree, L. W. (1991). Antecedents o f d rin k in g fo r early-
and late-onset e ld e rly alcohol abusers, lo u m a l o f Studies on Alcohol. 52. 587-
592.
Schuckit, M. A., & Pastor, P . A. (1979). A lcoho l-rela ted psychopathology
in the aged. In O. J. Kaplan (Ed.), Psychopathology o f aging (p p .2 11-227). New
Y ork: Academ ic Press.
Sm ith, G. T., Roehling, P. V., Goldman, M. S., & C hristiansen, B. A. (1987,
August). A lcohol expectancies versus dem ographics to p re d ic t adolescent
d rin k in g lo n g itu d in a lly . Paper presented a t the 9 5 th annual Meeting o f the
A m erican Psychological Association, New Y ork, NY.
Stall, R. (1 9 8 7 ). Research issues concerning alcohol consum ption among
aging populations. D rug and Alcohol Dependence. 19(3). 195-213.
Werch, C. E. (1989). Q uantity-frequency and d ia ry measures o f alcohol
consum ption fo r e ld e rlv d rin kers. The In te rn a tio n a l lo u m a l o f the A ddictions.
24, 859-865.
W ilcox, R. (1996). S tatistics fo r the social sciences. New York: Academ ic
Press.
W ilcox, R. (1998). How many discoveries have been lo s t by ign oring
m odem sta tistica l methods? Am erican Psychologist. 53. 300-314.
W ilsnack, S. C, Vogeltanz, N. D., Diers, L. E., & W ilsnack, R. W. (1995).
D rin kin g and Problem D rin kin g in O lder W omen. In T. Beresford and E.
Gomberg (Eds.), A lcohol and aging fpp. 263-292). New York: O xford U niversity
Press.
Young, R. McD., Oei, T., and Crook, G. (1 991 ). D evelopm ent o f a d rin k in g
se lf-e ffica cy q u e stio n n a ire , loum al o f Psychopathology and Behavior
Assessment. 13. 1-15.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
IMAGE EVALUATION
TEST TARGET (Q A -3 )
.0
l.l
1.2 5
1£
1 2 .8
m
E a
i i i
i l i a
U i
LL
S U S
I l i a
1 . 4
1.8
1 .6
150mm
A P P L IE D A IIV14G E . In c
— 1653 East Main Street
Rochester. NY 14609 USA
- — Phone: 716/482-0300
- = Fax: 716/288-5989
O 1993. Applied Image. Inc.. All Rights Reserved
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Alcohol treatment entry and refusal in a sample of older veterans

PDF

History of depression, antidepressant treatment, and other psychiatric illness as risk factors for Alzheimer's disease in a twin sample

PDF

Expectancies for alternative behaviors predict drinking problems: A test of a cognitive reformulation of the matching law

PDF

Applying the theory of reasoned action to condom use: The effect of immediate consequences on intention-behavior consistency

PDF

Depression as a risk factor for dementia

PDF

Adolescents' social attitudes: Genes and culture?

PDF

Cognitive functioning and dementia following cancer: A Swedish twin study

PDF

Head injury and dementia: A co-twin control study of Swedish twins

PDF

Alcohol's impact on triggered displaced aggression

PDF

Hispanic culture, acculturation, and distress among caregivers of dementia patients

PDF

Aging and the use of context and frequency information in ambiguity resolution

PDF

Depression in offspring following severe prenatal stress

PDF

Frequency and risk factors of poststroke dementia

PDF

Dementia caregiving and ethnicity: African American caregivers and the sociocultural stress and coping model

PDF

Associations and mechanisms among attention deficit hyperactivity disorder symptoms, cognitive functioning, and drinking habits

PDF

Affection and conflict in family relationships

PDF

Articulated thoughts regarding cognitions toward older adults

PDF

Depressed children and the social and behavioral attributes of their best friend

PDF

Determinants of life satisfaction among the high-functioning elderly: Implications of effects of psychological factors and activity participation

PDF

Alcoholism among the elderly: Fact or fallacy

Asset Metadata

Creator Satre, Derek Davies (author)

Core Title Alcohol expectancies and consumption: Age and sex differences

School Graduate School

Degree Master of Arts

Degree Program Psychology

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

Tag Gerontology,health sciences, public health,OAI-PMH Harvest,psychology, behavioral,Psychology, clinical

Language English

Contributor Digitized by ProQuest (provenance)

Advisor [illegible] (committee chair), [illegible] (committee member), Gatz, Margaret (committee member), Weg, Ruth B. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-30703

Unique identifier UC11341996

Identifier 1395137.pdf (filename),usctheses-c16-30703 (legacy record id)

Legacy Identifier 1395137.pdf

Dmrecord 30703

Document Type Thesis

Rights Satre, Derek Davies

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