Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 Dec 1.
Published in final edited form as: Psychol Addict Behav. 2012 Dec 31;27(4):10.1037/a0030821. doi: 10.1037/a0030821

Relationship-Specific Alcohol Expectancies and Relationship-Drinking Contexts: Reciprocal Influence and Gender-Specific Effects over the First 9 Years of Marriage

Ash Levitt 1, Kenneth E Leonard 2
PMCID: PMC3855643  NIHMSID: NIHMS500684  PMID: 23276314

Abstract

Previous research has suggested that individuals in romantic relationships hold beliefs about the effects of relationship-drinking contexts (e.g., drinking with one’s partner) on later relationship functioning, as well as the effects of relationship functioning on later relationship-drinking contexts, and that these reciprocal effects may be stronger among women compared to men. However, little research has directly examined relationship-specific alcohol expectancies, and no studies have tested the reciprocal associations between these expectancies and the extent of drinking in specific relationship contexts. The current study examined these reciprocal effects by testing time-lagged multilevel models using data from a community sample of 470 married couples across the first 9 years of marriage. As expected, results showed reciprocal effects between intimacy and social expectancies, respectively, and the extent of drinking-with-partner when both partners were drinking, which appeared to be adaptive mechanisms for relationships. Also as expected, some of these effects differed or were stronger for wives compared to husbands. Unexpectedly, reciprocal effects were also found between power expectancies and the extent of drinking apart from one’s partner, which appeared to be maladaptive mechanisms for relationships. The current study supports and extends previous research on alcohol use in romantic relationships by showing how relationship-specific alcohol expectancies influence and are influenced from relationship-drinking contexts, and which processes might be adaptive vs. maladaptive for romantic relationships. Implications for future research and theory are discussed.

Keywords: alcohol, alcohol expectancies, marriage, romantic relationships

Drinking alcohol with one’s partner is an integral component of romantic relationships for many couples. Research suggests that the associations between relationship alcohol use and relationship functioning operate as an interrelated, two-part process. In the first process, relationship alcohol use influences subsequent relationship functioning. Research shows that couple members who drink with their partner (compared to drinking apart from their partner) subsequently report greater intimacy and relationship satisfaction, and less relationship problems (Homish & Leonard, 2005; Levitt & Cooper, 2010). In the second process, relationship functioning reciprocally influences subsequent relationship alcohol use. Levitt and Cooper (2010) demonstrated that drinking together, particularly among women, occurred more following poor relationship functioning, suggesting not only that these processes may differ between male and female couple members, but also that these processes are psychologically reinforced. Taken together, research is building that suggests an underlying motivational process for relationship drinking.

Building on the observations of Covington and Surrey (1997), Leonard and Mudar (2004) proposed the idea of a relationship-motivation model of alcohol use, which includes cognitive, emotional, partner social influence, and behavioral factors associated with alcohol use in romantic relationships. However, beyond that study, little research has tested specific elements of a relationship-motivation model of alcohol use. Based on motivational theories of alcohol use (e.g., Cooper, Kuntsche, Levitt, Barber, & Wolfe, in press; Cox & Klinger, 1988), alcohol expectancies serve as a relatively distal predictor of alcohol use behaviors, an association that is mediated through relatively more proximal drinking motives. As an initial step in testing a relationship-motivation model of alcohol use, the current study focuses on the mechanism of relationship-specific alcohol expectancies. We posit that relationship-specific alcohol expectancies should predict subsequent drinking with one’s partner, which should in turn reinforce subsequent expectancies, thus representing a positive motivational feedback loop. To date, little research has examined relationship-specific alcohol expectancies, and no studies have examined the reciprocal associations between relationship-specific alcohol expectancies and the extent of drinking in specific relationship-drinking contexts. Thus, the current study examined these reciprocal associations, and whether they differed between husbands and wives, in a community sample of married couples over the first nine years of marriage to gain a better understanding of relationship-drinking processes and to extend research within a framework of relationship-motivated alcohol use.

Relationship-Drinking Contexts

Although research on the role of alcohol use in romantic relationships often focuses on negative effects of alcohol, which often stem from high levels of consumption (see Marshal, 2003; Roberts & Linney, 2000, for reviews), research is building that shows that alcohol can play a positive role in relationships in certain relationship-drinking contexts. For instance, Levitt and Cooper (2010) showed in a daily diary study of mostly dating couples that drinking low to moderate (vs. heavy) amounts of alcohol and drinking with one’s partner (vs. apart) predicted increases in next-day intimacy and decreases in next-day relationship problems for both men and women. Similar results have been shown in a longitudinal study of married couples. Homish and Leonard (2005) demonstrated that the extent of drinking with one’s partner (compared to the extent of drinking apart from one’s partner) for both husbands and wives was associated with greater reports of relationship satisfaction at the time of marriage and at each of the first two anniversaries.

Research also suggests that some of the potential benefits or protective effects of drinking with one’s partner in relationships may be stronger for women than men. Levitt and Cooper (2010) found that women (but not men) were more likely to drink with their partner later the same day or the next day if they experienced low intimacy or increased perceptions of relationship problems, respectively. Additionally, in the reverse direction, they found that only women were protected against decreases in next-day intimacy as a function of increased alcohol consumption when they drank with their partner the previous day. Homish and Leonard (2005) also found different effects for wives longitudinally. Although wives who drank with their partner reported similar levels of marital satisfaction at the time of marriage compared to nondrinking wives, satisfaction levels decreased significantly more over the first two years of marriage for nondrinking wives compared to those who drank with their partner. These results support and extend previous research showing that frequent (but not heavy) drinking in the relationship is associated with increased intimacy (Roberts, Leonard, & Senchak, 1994) and increased marital satisfaction (Roberts & Linney, 1998) for nonproblem-drinking wives compared to husbands (see Roberts & Linney, 2000, for reviews of these studies).

Taken together, this research demonstrates that certain relationship-drinking contexts, particularly drinking with one’s partner, are associated with positive relationship functioning for both dating and married couples. Additionally, some effects of drinking together, particularly as they pertain to feelings of intimacy, appear to be stronger for women than men. Collectively, this research strongly suggests that the extent of drinking with one’s partner should be associated with specific expectancies that alcohol will have beneficial effects on relationship functioning in this context.

Relationship-Specific Alcohol Expectancies

Alcohol expectancies reflect beliefs that alcohol will have certain effects in general situations (e.g., feeling good, becoming more social; Brown, Christiansen, & Goldman, 1987; Fromme, Stroot, & Kaplan, 1993). An abundant amount of research stemming from alcohol expectancy theory (see Goldman, Brown, & Christiansen, 1987; Goldman, Del Boca, & Darkes, 1999; Goldman, Reich, & Darkes, 2006, for reviews) has demonstrated that expectancies predict alcohol use in a wide range of contexts, including romantic relationships (Leonard & Homish, 2008). However, research also demonstrates that alcohol expectancies that are specific to certain contextual domains (e.g., sexual effects, Dermen & Cooper, 1994a; 1994b; tension reduction effects, Read & Curtin, 2007) are better predictors of reported drinking behavior in those domains than are other expectancies. Thus, some recent research has focused on alcohol expectancies specific to the realm of romantic relationships, in which individuals should theoretically hold unique and more proscribed expected effects of alcohol when drinking with their partner compared to general, nonromantic contexts (Leonard & Mudar, 2004).

To date, however, only three studies have examined relationship-specific alcohol expectancies. First, in a longitudinal study of married couples, and using a measure of relationship-specific alcohol expectancies created for that study, Leonard and Mudar (2004) found that positive expectancies (i.e., increased intimacy/openness, social pleasure/fun, and sexual enhancement effects) moderated husbands’ influence on wives’ drinking over the first year of marriage, such that husbands’ influence was greater among wives who had stronger positive expectancies. These results suggest that women are more likely to engage in relationship drinking when they expect that alcohol will have positive effects on their relationship.

Derrick et al. (2010), using the same measure of relationship-specific alcohol expectancies as Leonard and Mudar (2004), reported evidence that further illustrates gender differences in these expectancies. In a cross-sectional examination of married couples, Derrick et al. (2010) found that wives’ drinking status (heavy vs. light drinker) predicted greater expectancies in all domains (i.e., intimacy/openness, social pleasure/fun, sexual enhancement, and power/assertion) when drinking with their partner, whereas husbands’ drinking status only predicted greater intimacy and power expectancies. Patterns of intimacy expectancies also differed between husbands and wives as a function of whether couples were congruent (i.e., both heavy drinkers or both light drinkers) vs. discrepant (i.e., one heavy-drinking member and one light-drinking member) in their drinking statuses. Wives in congruent heavy-drinking couples reported stronger intimacy expectancies compared to heavy drinking husbands and to couples of other drinking configurations. These results suggest that wives have greater expectations about the effects of alcohol on their relationship than husbands, and also that intimacy expectancies for wives are particularly strong when both couple members are similar drinkers and therefore presumably more likely to drink together. This study supports and extends the work of Leonard and Mudar (2004) and elucidates previous findings of gender differences in drinking with one’s partner following low intimacy (Levitt & Cooper, 2010) and concordant vs. discrepant drinking patterns (Homish & Leonard, 2007; Levitt & Cooper, 2010; Mudar, Leonard, & Soltysinski, 2001) by suggesting that women may have stronger associations between intimacy and other positive expectancies and the extent of drinking with one’s partner than men.

Additionally, Kelly, Halford, & Young (2002) examined mean differences in expectancies between groups of women categorized by 1) both problem drinking and relationship distress, 2) problem drinking only, 3) relationship distress only, or 4) neither problem drinking or relationship distress. Using a measure of relationship-specific alcohol expectancies that the authors created for the study, they found that women who were problem drinkers and/or in a distressed relationship reported greater expectancies, particularly for increased intimacy and relationship efficacy effects, compared to women without problem drinking or relationship distress. Although these results generally support the notion of drinking in response to poor relationship functioning mentioned above, the cross-sectional nature of this study makes temporal interpretations impossible. Additionally, because men were not included in this study, it is unclear what role women’s partners had in these associations and whether these effects exist for men as well.

The Current Study

As a whole, the limited literature on relationship-specific alcohol expectancies is consistent with the literature on relationship-drinking contexts (e.g., the extent of drinking with one’s partner), relationship functioning domains (e.g., intimacy), and gender differences in these effects. The literature on relationship-specific alcohol expectancies suggests that couple members drink together because of expected positive effects of drinking with their partner. The current study directly tests this notion by comparatively estimating the extent of drinking in three different relationship-drinking contexts: drinking with one’s partner when both partners are drinking, drinking with one’s partner when the partner is not drinking, and drinking apart from one’s partner. Additionally, assuming these positive effects are realized following drinking together, the expectations of these effects should be reinforced. That is, relationship-specific alcohol expectancies and the extent of drinking together should reciprocally influence each other. The current study sought to test this hypothesis using temporally-ordered longitudinal data in a community sample of married couples over the first nine years of marriage.

Based on the literature and theory reviewed above, we hypothesized that reports of relationship-specific alcohol expectancies domains (i.e., intimacy/openness, social pleasure/fun, sexual enhancement, power/assertion) would be associated with greater subsequent reports of the extent of drinking in contexts with one’s partner, but not of drinking in contexts apart from one’s partner. In the reverse direction, we hypothesized that the frequency of drinking in contexts with one’s partner (but not in contexts apart from one’s partner) would be reciprocally associated with stronger subsequent relationship-specific alcohol expectancies. Additionally, based on previous findings of gender-specific effects among women (e.g., Derrick et al., 2010; Leonard & Mudar, 2004; Levitt & Cooper, 2010), we hypothesized that the reciprocal associations between intimacy expectancies, specifically, and drinking-with-partner contexts would be stronger for wives compared to husbands. Although similar gender differences might be seen for other positive expectancies domains as well (Leonard & Mudar, 2004), the research evidence for these domains is less clear. Therefore, we did not make gender-specific hypotheses concerning other associations, although tests of gender differences were examined for all relevant associations (see below). To our knowledge, the current study is the first to examine the temporally-ordered, bidirectional associations between relationship-specific alcohol expectancies and the frequency of drinking in different relationship-drinking contexts over time in romantic relationships. Additionally, the current study extends previous analyses of the current dataset (i.e., Homish & Leonard, 2005; Leonard & Mudar, 2004) by 1) independently examining the factors of relationship-specific alcohol expectancies, 2) simultaneously examining the extents of multiple relationship-drinking contexts, and 3) extending the analyses across the full 9-year study period compared to previous 3- and 4-year examinations, respectively.

Method

Participants and Procedure

Couples (N = 634) were drawn from the Adult Development Study, which assessed alcohol involvement and marital functioning in married couples over 6 waves across the first 9 years of marriage. Couples were initially sampled from the Buffalo, NY, community from1996–1999 as they were applying for their marriage licenses. Additional information on the original sample, sampling procedure, and differences between eligible and ineligible couples is presented elsewhere (Leonard & Mudar, 2004). Couple members independently completed questionnaires on alcohol involvement and relationship processes at each wave. The current study analyzes data from a subsample of 470 couples in which both members had consumed alcohol in the past year at the time of marriage. Of this subsample, couple retention rates were adequate over time: wave 2, N = 430 couples (91.5% of wave 1 subsample); wave 3, N = 406 (86.4%); wave 4, N = 382 (81.3%); wave 5, N = 369 (78.5%); wave 6, N = 311 (66.2%). Additionally, no mean differences were found in key study variables (i.e., relationship-specific alcohol expectancies and extents of relationship-drinking contexts) between couples who remained in the study throughout vs. couples who dropped out of the study by wave 6. Relevant demographic data for husbands and wives in the current subsample can be seen in Table 1.

Table 1.

Descriptive Statistics on Study Variables

Variable Husbands Wives
Mean (SD) Mean (SD)
Age 28.44 (5.78) 26.62 (5.25)
Education 5.74 (1.70) 5.92 (1.61)
Race (% White) 0.68 (0.47) 0.71 (0.45)
Children at baseline (% with) 0.56 (0.50) --
Pre-Marriage Cohabitation (%) 0.71 (0.45) --
DWP, Both Partners Drinking 2.30 (1.73) 2.25 (1.77)
DWP, Partner Not Drinking 1.97 (1.82) 0.86 (1.31)
Drink Apart from Partner 2.27 (1.77) 1.25 (1.43)
Average Consumption 3.92 (3.03) 2.79 (2.43)
Intimacy RAE 2.70 (1.25) 2.63 (1.31)
Social RAE 3.25 (1.37) 3.32 (1.42)
Sex RAE 3.09 (1.41) 3.17 (1.48)
Power RAE 2.53 (1.23) 2.48 (1.28)
Open in a new tab

Note. Age, Education, Race, and Pre-Marriage Cohabitation variables are time-invariant, and reflect values at baseline. All other variables are time-varying, and reflect average values collapsed across all waves. Dashes indicate that the measure reflects the couple’s value, and is therefore the same between husbands and wives. DWP = Drink-with-Partner. RAE = Relationship-specific Alcohol Expectancies.

Measures

All measures were assessed in separate reports from both husbands and wives. Descriptive statistics for all measures in the current study are shown in Table 1.

Demographic covariates

A host of demographic variables were assessed as covariates. Age, race, and highest level of education obtained (1 = less than 8th grade; 2 = some high school; 3 = GED or high school equivalency; 4 = high school graduate; 5 = trade school; 6 = some college; 7 = college graduate; 8 = 5 years or more college; 9 = Master’s, doctoral, or other postgraduate degree) were assessed for both husbands and wives. Whether couples cohabited before marriage (1 = Yes; 0 = No) and whether either partner had children at baseline (1 = Yes; 0 = No) were assessed as couple variables. Preliminary analyses also assessed whether the birth to a child in the past year at each assessment served as a significant covariate above and beyond having children at baseline. This variable was not a significant predictor of any outcome of interest, and was therefore not included in models.

Alcohol consumption

Typical quantity of couple members’ most frequently consumed drink type per drinking occasion over the past year was used as a covariate. Couples reported their frequency and typical quantity of consumption of beer, wine, and liquor, respectively. Frequency of consumption for each type of drink was assessed using a 9-point scale ranging from 0 = “not at all” to 9 = “every day.” The typical quantity consumed, assessed using a 10-point scale ranging from 1 drink to 18 or more drinks, of the most frequently consumed drink type was used as the covariate. This measure was included to control for the effect that those who drink heavily might be more likely to drink in all relationship contexts as opposed to a specific context, so that effects found for the frequency of drinking in specific relationship contexts would reflect contextual effects and not effects of heavy consumption.

Relationship-specific alcohol expectancies

Four domains of relationship-specific alcohol expectancies were assessed using the 21-item Relationship-Specific Alcohol Expectancy questionnaire (RSAEQ), which was created for this study (Leonard & Mudar, 2004). The four domains assessed were Intimacy/Openness (6 items), Social/Fun (5 items), Sexual Enhancement (5 items), and Power/Assertiveness (5 items). Participants were instructed to respond about effects of alcohol when drinking “in your partner’s presence.” All items were preceded with the stem, “How likely is it that alcohol will affect you so that you…” Example items are: “…feel closer to your partner?” (Intimacy/Openness); “…become talkative and happy with your partner?” (Social/Fun); “…become a better lover with your partner?” (Sexual Enhancement); and “…feel confident, powerful, or in control with your partner?” (Power/Assertiveness). Items were scored on a 6-point scale with 1 = “Very unlikely,” and 6 = “Very likely.” Composites computed for these domains were highly reliable in the current sample (αs = .92 – .93 for husbands, and .90 – .93 for wives).

Developed in the mid 1990s, some domains of the RSAEQ were newly created while others were partially adapted from the Alcohol Expectancy Questionnaire (AEQ; Brown et al., 1987), which was the most established expectancy measure at the time. RSAEQ items and domains either supplemented the AEQ if certain items or domains did not previously exist, or were adapted from the AEQ as romantic relationship-specific analogs. The newly created domains of the RSAEQ focused on Intimacy/Openness and Power/Assertiveness expectancies, respectively. Concerning positive interpersonal interactions, the Intimacy/Openness domain of the RSAEQ was created to tap expected intimacy and self-disclosure effects of alcohol in romantic relationships (e.g., Miller, Ingham, Plant, & Miller, 1977; Roberts et al., 1994; Wilsnack, Wilsnack, & Klassen, 1987), which the AEQ did not tap. In contrast, concerning negative interpersonal interactions, the Power/Assertiveness domain of the RSAEQ was created in part based on theoretical considerations (e.g., McClelland, Davis, Kalin, & Wanner, 1972), and in part as a relationship-specific adaptation of the AEQ’s Social Assertiveness domain. Aggressiveness was excluded from this domain so as not to overlap with other aggression measures, which were key to other hypotheses of the larger study. Social/Fun and Sexual Enhancement RSAEQ domains were adapted more closely from the similar AEQ domains Social and Physical Pleasure, and Sexual Enhancement, respectively, only the emphasis of the item wording, as described above, was changed to reflect drinking in the presence of one’s partner. Overall, these domains were selected to reflect other-dependent, dyadic processes in romantic relationships. As such, relatively more individualistic, self-directed expectancies domains (e.g., tension reduction, global positive changes) were not included.

Relationship-drinking contexts

Single-item measures were used to assess the past year frequency of three different relationship-drinking contexts: drinking-with-partner when both partners were drinking, drinking-with-partner when the partner was not drinking, and drinking when the partner was not present. Each item was scored on an 8-point Likert scale where 0 = “Not at all during the past year,” 1 = “1–4 times in the year,” 2 = “5–10 times in the year,” 3 = “About once a month,” 4 = “2–3 times a month,” 5 = “1–2 times a week,” 6 = “3–4 times a week,” and 7 = “Every day or nearly every day.”

Data Analyses

Multilevel time-lagged analyses were conducted using the Mixed procedure in SPSS (Version 19, 2010). Repeated measures (Level 1) were nested within persons (Level 2), and matched by couple (Kashy & Donnellan, 2012; Laurenceau & Bolger, 2005). This method allows for missing data at Level 1.

To test the hypothesized reciprocal associations between relationship-specific alcohol expectancies and relationship drinking contexts, models were estimated in two directions. First, reports of the frequency of drinking in each relationship-drinking context outcome domain were individually predicted from temporally prior reports (i.e., reports from the previous study wave) of relationship-specific alcohol expectancies, controlling for prior reports of the outcome, concurrent reports of the other drinking contexts, and covariates. This strategy produced three models predicting the extent of each relationship-drinking context outcome (see Table 2). Second, models were estimated in the reverse direction such that reports of each domain of relationship-specific alcohol expectancies were individually predicted from temporally prior reports of relationship drinking contexts, controlling for prior reports of the outcome, concurrent reports of other expectancies domains, and covariates. This reverse strategy produced four models predicting relationship-specific alcohol expectancies domains (see Table 3). All terms were entered into models grand-mean centered with the exception of gender (coded Husbands = 1; Wives = 0), which was entered uncentered.

Table 2.

Tests of Covariates, Prior Relationship-Specific Alcohol Expectancies, and Gender Moderation Predicting Relationships Drinking Contexts

Predictors Drinking-with-Partner
Both Partners Drinking 		Drinking-with-Partner
Partner Not Drinking 		Drinking Away from Partner
b (SE) p d b (SE) p d b (SE) p d
Intercept 2.344 (.037) .000 3.159 1.258 (.031) .000 2.003 1.658 (.033) .000 2.475
L – Prior Outcome
Control .434 (.017) .000 1.008 .371 (.016) .000 .890 .283 (.017) .000 .662
Gender −.182 (.045) .000 .021 .203 (.053) .000 .193 .239 (.050) .000 .237
Actor Age -- -- -- -- -- -- -- -- -- -- --
Partner Age .012 (.005) .014 .105 −.016 (.004) .000 .187 .014 (.004) .002 .141
Actor Education .040 (.017) .020 .089 -- -- -- -- -- -- -- --
Partner Education -- -- -- -- -- -- -- -- -- -- -- --
Actor Race .168 (.067) .012 .108 -- -- -- -- −.166 (.058) .004 .133
Partner Race -- -- -- -- -- -- -- -- -- -- -- --
Children (Baseline) −.188 (.062) .003 .163 -- -- -- -- .151 (.051) .003 .160
Pre-Marriage Cohabited -- -- -- -- -- -- -- -- .149 (.054) .006 .151
DWP, Both Drinking n/a .091 (.015) .000 .277 .148 (.015) .000 .405
DWP, Partner Not Drinking .071 (.017) .000 .177 n/a .380 (.016) .000 .954
Drink Apart from Partner .160 (.018) .000 .348 .365 (.016) .000 .854 n/a
Actor Avg. Consumption .009 (.010) .392 .033 −.015 (.010) .121 .061 .058 (.010) .000 .233
Partner Avg. Consumption .080 (.014) .000 .151 −.044 (.009) .000 .204 .025 (.009) .009 .108
L - Actor Intimacy RAE .089 (.053) .094 .071 −.000 (.045) .999 .000 −.044 (.046) .338 .037
L - Partner Intimacy RAE −.042 (.049) .397 .034 n/a n/a
L - Actor Social RAE .044 (.038) .249 .045 .039 (.035) .255 .046 −.042 (.035) .236 .046
L - Partner Social RAE .097 (.038) .010 .101 n/a n/a
L - Actor Sex RAE .010 (.036) .785 .011 −.026 (.033) .422 .034 .040 (.033) .228 .045
L - Partner Sex RAE −.023 (.036) .521 .024 n/a n/a
L - Actor Power RAE −.034 (.045) .442 .031 −.047 (.041) .256 .044 .087 (.041) .035 .082
L - Partner Power RAE −.012 (.049) .804 .011 n/a n/a
L - Actor Intimacy RAE X Actor Gender −.109 (.043) .012 .100 -- -- -- -- -- -- -- --
L - Partner Power RAE X Actor Gender .092 (.044) .035 .085 -- -- -- -- -- -- -- --
Open in a new tab

Note. Coefficients are unstandardized. L = Lagged Prior Wave Report. DWP = Drink-with-Partner. RAE = Relationship-specific Alcohol Expectancies. Gender was coded Male = 1; Female = 0. Race was coded White = 1; nonWhite = 0. Dashes indicate the term was dropped from the model due to nonsignificance. n/a indicates the term was not tested in the model.

Table 3.

Tests of Covariates, Prior Relationship Drinking Contexts, and Gender Moderation Predicting Relationship-Specific Alcohol Expectancies

Relationship-Specific Alcohol Expectancies
Intimacy Social Sex Power
Predictors b (SE) p d b (SE) p d b (SE) p d_ b (SE) p d
Intercept 2.665 (.014) .000 9.681 3.273 (.018) .000 8.978 3.148 (.019) .000 7.915 2.508 (.010) .000 12.663
L – Prior Outcome
Control .096 (.008) .000 .423 .108 (.010) .000 .394 .167 (.010) .000 .606 .111 (.010) .000 .430
Gender .012 (.019) .524 .031 .203 (.053) .000 .071 −.039 (.026) .144 .069 n/a
Actor Age -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
Partner Age -- -- -- -- -- -- -- -- −.006 (.002) .012 .109 -- -- -- --
Actor Educ. -- -- -- -- .039 (.008) .000 .192 -- -- -- -- −.023 (.007) .001 .135
Partner Educ. -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
Actor Race -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
Partner Race -- -- -- -- .100 (.028) .000 .152 -- -- -- -- -- -- -- --
Children (Baseline) -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
Pre-Marriage Cohabited .052 (.020) .009 .135 -- -- -- -- -- -- -- -- −.066 (.022) .004 .151
Actor Int. RAE n/a .453 (.022) .000 .755 .103 (.025) .000 .152 .736 (.015) .000 1.780
Actor Soc. RAE .274 (.013) .000 .804 n/a .565 (.016) .000 1.326 −.044 (.015) .005 .104
Actor Sex RAE .053 (.013) .000 .156 .494 (.014) .000 1.344 n/a .141 (.014) .000 .370
Actor Power RAE .575 (.013) .000 1.809 −.064 (.021) .002 .114 .232 (.022) .000 .403 n/a
Actor Avg. Consumption .008 (.003) .021 .091 −.008 (.005) .106 .062 .007 (.005) .129 .060 −.006 (.004) .118 .058
Partner Avg. Consumption −.014 (.003) .000 .163 .006 (.004) .173 .053 −.004 (.005) .375 .034 .012 (.004) .001 .124
L-Actor DWP, Both Drinking .014 (.007) .042 .086 .016 (.009) .077 .073 −.025 (.010) .009 .109 −.018 (.008) .021 .098
L-Partner DWP, Both Drinking −.003 (.007) .700 .017 .011 (.009) .212 .052 −.014 (.009) .130 .065 −.001 (.008) .929 .004
L-Actor DWP, Partner Not Drinking .032 (.010) .001 .157 −.012 (.009) .159 .055 .003 (.009) .770 .011 −.006 (.007) .391 .032
L-Actor Drink Apart from Partner −.008 (.007) .247 .044 −.022 (.013) .080 .081 −.011 (.009) .244 .045 .016 (.008) .033 .079
L-Actor DWP, Partner Not Drinking X Actor Gender −.025 (.011) .027 .085 -- -- -- -- -- -- -- -- n/a
L-Actor Drink Apart from Partner X Actor Gender -- -- -- -- .037 (.014) .011 .102 -- -- -- -- n/a
Open in a new tab

Note. Coefficients are unstandardized. L = Lagged Prior Wave Report. DWP = Drink-with-Partner. Educ. = Education. RAE = Relationship-specific Alcohol Expectancies. Int. = Intimacy. Soc. = Social. Gender was coded Male = 1; Female = 0. Race was coded White = 1; nonWhite = 0. Dashes indicate the term was dropped from the model due to nonsignificance. n/a indicates the term was not tested in the model.

The same iterative procedure was used in the testing of all models. First, all control and demographic covariate terms for each model were estimated simultaneously. Nonsignificant demographic covariate terms were then trimmed from models to increase model stability (Raudenbush & Bryk, 2002), and the models were re-estimated. Second, main effects of prior predictors of interest and their interaction terms with gender were entered simultaneously. Nonsignificant interaction terms were then trimmed, and models were re-estimated. Main effects of prior predictors of interest remained regardless of significance to control for related constructs (e.g., estimating all four domains of relationship-specific alcohol expectancies when predicting the frequency of drinking in relationship contexts). All terms were estimated as fixed effects with random intercept and error components. The final models for each outcome are shown in Tables 2 and 3.

Two special considerations were given to model building concerning the inclusion of tests of partner effects (via the Actor Partner Interdependence Model [APIM]; Kashy & Donnellan, 2012; see also Ackerman, Donnellan, & Kashy, 2011; Kenny, Kashy, & Cook, 2006). First, we viewed drinking-with-partner when both partners were drinking as the only truly dyadic drinking variable (as opposed to drinking-with-partner when the partner was not drinking or drinking apart from one’s partner), because only in this context are both couple members drinking together as a shared relationship activity. Second, the relationship-specific alcohol expectancies items were worded in a manner that reflects a within-person effect as opposed to a between-person or cross-partner effect (i.e., how likely alcohol will have a certain effect on you when drinking in the presence of your partner, not how likely you think alcohol will have a certain effect on your partner). Therefore, taken together, it is logical to include tests of partner effects when estimating the reciprocal associations between relationship-specific alcohol expectancies and drinking-with-partner when both partners were drinking. Doing so controls for interdependencies between husbands’ and wives’ reports as they relate to this dyadic drinking context by estimating actor and partner effects simultaneously (e.g., Kashy & Donnellan, 2012). In contrast, we tested only actor effects when estimating reciprocal associations between expectancies and the other two drinking contexts in which only the actor (not the partner) was drinking. Therefore, our models blend elements of the APIM and multilevel models with individuals.1

Finally, preliminary analyses were conducted to determine whether data for husbands and wives were empirically distinguishable by gender, as well as conceptually, following procedures recommended by Kashy & Donnellan (2012; see also Ackerman et al., 2011; Kenny et al., 2006). Distinguishability in this case is tested by comparing a constrained model (in which the means, variances, and covariances are assumed to be equal between men and women) with an unconstrained model (in which the means, variances, and covariances between men and women are allowed to vary). Deviance statistics from these models (i.e., −2*log likelihood values) are then used to calculate a chi square difference test. Significant chi square differences indicate that the data are empirically distinguishable. For models predicting the frequency of drinking in relationship contexts, husbands and wives were distinguishable in all three outcome models, χ2s(6) ≥ 28, ps < .001. For models predicting relationship-specific alcohol expectancies, husbands and wives were distinguishable in models predicting social and sex expectancies outcomes, χ2s(5) ≥ 12, ps < .05. Husbands and wives were marginally distinguishable in the model predicting intimacy expectancies, χ2(5) = 10, p = .075, and were considered distinguishable for analyses. However, husbands and wives were indistinguishable in the model predicting power expectancies, χ2(5) = 4, p = .549. Thus, models predicting power expectancies did not include tests of gender differences.

Results

Do relationship-specific alcohol expectancies predict the frequency of drinking in subsequent relationship contexts?

Results from models predicting the frequency of drinking in relationship contexts are shown in Table 2. As shown in the top panel of Table 2, wives, and individuals with older (compared to younger) partners, more education, of Caucasian race, and couples with no children reported greater drinking-with-partner when both partners were drinking at each wave. In contrast, husbands, and individuals with less education reported greater drinking-with-partner when the partner was not drinking at each wave. Also, husbands, and individuals with older partners, of nonCaucasian race, and couples with children and who cohabited prior to marriage reported greater drinking apart from their partner at each wave. The frequency of drinking in a specific relationship context also positively predicted drinking in the other contexts, suggesting that individuals who drink frequently in one context also frequently drink in other contexts. As would be expected, average partner consumption was positively associated with actor reports of the extent of drinking-with-partner when both partners were drinking, but negatively associated with actor reports of the extent of drinking-with-partner when the partner was not drinking. Both actor and partner average consumption were positively associated with actor reports of the extent of drinking apart from one’s partner suggesting that this is the only relationship-drinking context associated with heavier consumption for both partners.

As shown in the second panel of Table 2, two main effects of relationship-specific alcohol expectancies were found. First, in line with expectation, greater partner reports of social expectancies at one wave predicted greater actor reports of drinking-with-partner when both partners were drinking for both husbands and wives at the following wave. Additionally, and contrary to expectation, greater reports of power expectancies at one wave predicted greater reports of drinking apart from one’s partner for both husbands and wives at the following wave. No effects of relationship-specific alcohol expectancies were found predicting drinking-with-partner when the partner was not drinking.

Do effects of relationship-specific alcohol expectancies differ between husbands and wives?

Two relationship-specific alcohol expectancies X gender interactions were found predicting drinking-with-partner when both partners were drinking (see bottom panel of Table 2). First, an actor intimacy expectancies X actor gender interaction was found. As shown in the top panel of Figure 1, and as hypothesized, the effect of intimacy expectancies on subsequent drinking-with-partner when both partners were drinking was stronger for wives compared to husbands. Post hoc probing of this interaction revealed that greater reports of intimacy expectancies among wives at one wave predicted marginally greater reports of drinking-with-partner when both partners were drinking at the following wave (b = .089, p = .094), whereas this pattern was not found for husbands (b = −.019, p = .721).

Figure 1.

Figure 1

Open in a new tab

Top panel shows lagged actor intimacy relationship-specific alcohol expectancies (A-Int) X actor gender interaction predicting current actor reports of drinking-with-partner when both partners are drinking. Bottom panel shows lagged partner power relationship-specific alcohol expectancies (P-Pow) X actor gender interaction predicting current actor reports of drinking-with-partner when both partners are drinking. Low (Lo) and high (Hi) values of predictors reflect the 15th and 85th percentiles of their distribution, respectively.

A partner power expectancies X actor gender interaction was also found predicting drinking-with-partner when both partners were drinking. As shown in the bottom panel of Figure 1, husbands’ and wives’ reported opposite patterns of drinking together as a function of their partner’s increased power expectancies at the prior wave. However, further probing of this interaction revealed that neither effect of partner power expectancies was significant within husbands (b = .080, p = .104) or wives (b = −.034, p = .442).

Does the frequency of drinking in relationship contexts reciprocally influence relationship-specific alcohol expectancies?

Results from reverse models (testing reciprocal effects) predicting relationship-specific alcohol expectancies are shown in Table 3. As shown in top panel of Table 3, couples who cohabited prior to marriage reported greater intimacy expectancies at each wave. Men, and individuals with greater education and partners of Caucasian race reported greater social expectancies at each wave. Individuals with younger partners reported greater sex expectancies at each wave. And individuals less education and couples who did not cohabit prior to marriage reported greater power expectancies at each wave. Consistent with theory, the positive relationship-specific alcohol expectancies domains (i.e., intimacy, social, and sex) were positively associated with each other. Associations with power expectancies and other domains were mixed. Power and social expectancies domains were negatively associated, which is consistent with theory. However, power expectancies were also positively associated with intimacy and sex expectancies. The positive association between power and intimacy expectancies could reflect that both domains include items reflecting self-disclosure despite the fact that intimacy-related self-disclosure is thought to be interpersonally positive, whereas power-related self-disclosure is thought to be domineering and at the cost of one’s partner. The positive association between power and sex expectancies could also reflect this domineering quality of our operationalization of power. Additionally, average actor consumption was positively associated with actor reports of intimacy expectancies, whereas an opposite effect was found for average partner consumption, suggesting that heavier partner consumption on average adversely affects actors’ expectations of intimacy-enhancing effects of alcohol when they drink with their partner. Average partner consumption was also positively associated with actor reports of power expectancies suggesting that heavier partner consumption on average increases actors’ expectations of power effects when they drink with their partner.

As shown in the second panel of Table 3, and as expected, actor reports of drinking-with-partner when both partners were drinking predicted all four relationship-specific alcohol expectancies outcomes. This drinking context positively predicted greater intimacy expectancies and marginally greater (p = .077) social expectancies for both husbands and wives at the following wave, whereas it negatively predicted sex and power expectancies for both husbands and wives at the following wave. Partner reports of drinking-with-partner when both partners were drinking did not predict actor reports of relationship-specific alcohol expectancies. One effect was found for drinking-with-partner when the partner was not drinking predicting subsequent intimacy expectancies; however, this effect was moderated by gender (see below). Finally, two effects were found for drinking apart from one’s partner predicting relationship-specific alcohol expectancies. The first effect, predicting social expectancies, was moderated by gender (see below). The second effect showed that drinking apart from one’s partner at one wave predicted greater power expectancies for both husbands and wives at the following wave.

Do the effects of the frequency of drinking in specific relationship contexts differ between husbands and wives?

Two relationship-drinking context X gender interactions were found predicting relationship-specific alcohol expectancies outcomes (see bottom panel of Table 3). First, an actor report of drinking-with-partner when the partner was not drinking X actor gender interaction was found predicting intimacy expectancies. As shown in the top panel of Figure 2, and as expected, the effect of drinking-with-partner when the partner was not drinking on subsequent intimacy expectancies was stronger for wives compared to husbands. Probing this interaction revealed that wives who reported greater drinking-with-partner when the partner was not drinking reported significantly greater intimacy expectancies at the following wave (b = .032, p = .001), whereas such an effect was not found for husbands (b = .007, p = .369).

Figure 2.

Figure 2

Open in a new tab

Top panel shows lagged actor reports of drinking-with-partner when the partner is not drinking (DWP-PND) X actor gender interaction predicting current actor reports of intimacy relationship-specific alcohol expectancies (RAE). Bottom panel shows lagged actor reports of drinking apart from the partner (No DWP) X actor gender interaction predicting current actor reports of social RAE. Low (Lo) and high (Hi) values of predictors reflect the 15th and 85th percentiles of their distribution, respectively.

An actor report of drinking apart from one’s partner X actor gender interaction was also found predicting social expectancies. As shown in the bottom panel of Figure 2, husbands and wives demonstrated different patterns of social expectancies as a function of drinking apart from their partner. Probing this interaction revealed that wives who reported less drinking apart from their partner at one wave reported marginally greater social expectancies at the following wave (b = −.022, p = .080). An opposite, albeit nonsignificant, effect was found for husbands (b = .015, p = .164).

Discussion

The current study examined time-lagged, reciprocal effects between relationship-specific alcohol expectancies and the frequency of drinking in specific relationship contexts in a community sample of married couples across the first nine years of marriage. A schematic summary of results is shown in Figure 3.

Figure 3.

Figure 3

Open in a new tab

Summary schematic of results showing effects (p < .10) that were reciprocal for at least one partner (solid lines) compared to nonreciprocal effects (dotted lines), the valence of effects (+/−), whether the effect differed by gender (W = for Wives only), and whether the effect was a partner effect (P). RAE = Relationship-specific alcohol expectancies. DWP = Drinking with partner.

As expected, effects were found in both directions, indicating reciprocal influence between these two constructs over time in marriage. The greatest number of associations was found between relationship-specific alcohol expectancies and the frequency of drinking with one’s partner when both partners are drinking together. Two of the four expectancies domains predicted subsequent reports of the frequency of drinking in this context, which in turn predicted subsequent reports of all four expectancies domains. Interestingly, none of the expectancies domains similarly predicted the extent of subsequent drinking-with-partner when the partner was not drinking, and the frequency of drinking in this context in turn only predicted an increase in subsequent intimacy expectancies among wives. Taken together, these results suggest that couple members psychologically separate situations of joint, dyadic drinking as opposed to drinking in the presence of a nondrinking partner concerning their expected effects of alcohol. In other words, it is not simply drinking in the presence of one’s partner that seems to be important for couple members, but rather drinking together as a couple.

In the case of positive expectancies (i.e., intimacy and social), it appears as though couple members only expect these effects from alcohol when their partner is also drinking with them. Assuming that these expected effects (e.g., feeling closer to, opening up to, and talking more with one’s partner) actually occur when couple members drink together also explains the reciprocal findings of increased subsequent intimacy and social expectancies for both husbands and wives, suggesting that drinking together reinforces these expectancies. These results support and extend previous research showing bidirectional effects between daily reports of drinking-with-partner and intimacy in dating couples (Levitt & Cooper, 2010), and unidirectional effects of drinking-with-partner predicting greater relationship satisfaction in married couples (Homish & Leonard, 2005).

Insofar as couple members expect positive effects from drinking together, and perceive drinking together as being beneficial for their relationship functioning, the finding that this drinking context also predicted decreased subsequent power expectancies for both husbands and wives would not be unexpected. This finding makes sense considering that intimacy-related and power-related processes have long been thought to be theoretically distinct from one another (McAdams, 1988), and that power is associated with coercive, maladaptive relationship functioning (see Roberts, 2006; Simpson & Tran, 2006, for reviews). Thus, it is no surprise that the extent of drinking together (i.e., a positive experience associated with intimacy and social effects for couple members) fails to positively reinforce subsequent power expectancies.

Interestingly, sex expectancies did not predict drinking-with-partner contexts, and drinking together actually predicted decreased subsequent sex expectancies. This is in line with previous findings by Levitt and Cooper (2010) showing that having sex or the quality of sex was not associated with the link between drinking-with-partner and later intimacy. One explanation for this effect could be that the characteristics of typical drinking environments and situations when couples drink together are not conducive to alcohol’s expected sexual effects. For example, drinking with one’s partner in public at a bar or a friend’s house is not logically consequent from or reinforcing of the expected sexual effects of alcohol (assuming that the couple could not readily act on these effects), whereas it logically precedes and follows for expected intimacy and social effects (on which the couple could immediately act). Examining such varying characteristics of drinking situations is impossible given the current data, and should be examined in future research using an event-level methodology to better understand the mechanisms and processes of relationship-drinking situations.

Additionally, and unexpectedly, power expectancies were positively reciprocally associated with the frequency of drinking apart from one’s partner for both husbands and wives. One explanation for the effect of power expectancies predicting later drinking apart from the partner is that individuals who hold these expectancies understand, perhaps from previous experience, that achieving the expected power effects is adverse to their relationship (Roberts, 2006; Simpson & Tran, 2006), and that they therefore drink apart from their partner when they do drink. In the reverse direction, it is also possible that, despite drinking apart from the partner, the expected effects of power for these individuals are psychologically robust and last beyond the drinking session to be manifested when back in the presence of the partner, thus reinforcing subsequent power expectancies. Although speculative, this delayed-effect interpretation is logically supported by the fact that only the drinking-apart-from-partner context was significantly associated with increased alcohol consumption in the current study (see Table 2), suggesting the possibility of prolonged, adverse effects from increased consumption in this drinking context. This interpretation is also consistent with the notion that drinking apart from one’s partner is associated with poor relationship functioning (Homish & Leonard, 2005; Levitt & Cooper, 2010). Furthermore, considering that power is associated with increased conflict and aggression in relationships (Roberts, 2006; Simpson & Tran, 2006), and that alcohol aggression expectancies predict alcohol-related aggression in marriage (Kachadourian, Homish, Quigley, & Leonard, 2011), these findings together indicate a combination of drinking contexts and expectancies that are potentially highly adverse and problematic for romantic couples. Future research is needed to better illuminate the processes of drinking apart from one’s partner on relationship functioning among individuals with strong power expectancies.

As expected, the associations between intimacy expectancies and the frequency of drinking in contexts with one’s partner were stronger for wives compared to husbands. Wives’ (but not husbands’) intimacy expectancies were associated with marginally greater reports of subsequent drinking together (see top panel of Figure 1), and only wives reported greater intimacy expectancies following drinking with their partner when their partner was not drinking (see top panel of Figure 2). These results suggest that wives value alcohol’s effects on increased intimacy and drink with their partner to a greater extent to achieve these effects. Although the broad pattern of effects, as discussed above, indicates that the context of drinking together is more integrated with relationship-specific alcohol expectancies for couples than drinking-with-partner when the partner is not drinking, the extent of drinking in this latter context nevertheless reinforces intimacy expectancies for wives. However, this effect is not reciprocated in predicting the extent of subsequent drinking-with-partner when the partner is not drinking. Wives also reported marginally decreased social expectancies following a greater extent of drinking apart from their partner (see bottom panel of Figure 2), and exhibited an opposite trend compared to husbands in drinking together as a function of their partner’s power expectancies (see bottom panel of Figure 1). As a whole, these results support previous research showing differential effects in relationship alcohol use and expectancies for women compared to men (Derrick et al., 2010; Leonard & Mudar, 2004; Levitt & Cooper, 2010), and extend it by showing which domains of relationships women expect to be affected by alcohol use.

Most effects found in the current study (see Tables 2 and 3) would be considered small effects (Cohen, 1988), which could be viewed as a limitation of the current study. However, as Cohen (1988) pointed out, it is important to interpret effects in their relative context. Small effects can be additive and can amount to large effects over time (Abelson, 1985). This is particularly meaningful in the context of romantic relationships, where small effects should not be viewed as trivial because they can become large and quite meaningful over the course of a relationship.

The current study has implications for a relationship motivation model of alcohol use, which can guide future research and theory on relationship drinking. The current results suggest that couple members drink together more because of certain positive expected effects of alcohol and that doing so positively reinforces these expectancies. This is in line with previous research (Homish & Leonard, 2005; Leonard & Mudar, 2004; Levitt & Cooper, 2010), and suggests an underlying motivational system for relationship alcohol use. Drinking motives are theoretically believed to be the final common pathway between alcohol expectancies and drinking behavior (Cooper et al., in press; Cox & Klinger, 1988). Thus, it is likely that relationship-specific drinking motives mediate the associations between positive relationship-specific alcohol expectancies and drinking together in relationships. However, just because an individual expects alcohol to have a certain effect does not mean that he or she will drink to achieve that effect (Ajzen & Fishbein, 2005). Whether couple members are motivated to drink to achieve certain relationship effects of alcohol is likely dependent on the situation, context, and other individual difference factors. Future research is needed to fully conceptualize and test a comprehensive relationship motivation model of alcohol use, although the current results offer preliminary support for such a model.

Despite its many strengths, however, the current study had some limitations in the measurement of relationship-specific alcohol expectancies. First, as mentioned above, the RSAEQ assesses individuals’ expectancies of alcohol’s effects only when they are drinking in the presence of their partner (i.e., a within-person effect). That is, they do not assess individuals’ expectancies that alcohol will have certain effects on their partner when they are drinking together (i.e., a cross-partner effect). Another limitation of the RSAEQ is that it does not assess to what extent couple members value expected effects of alcohol in relationship-drinking situations, which has been shown to be an important factor in expectancy research (Fromme et al., 1993). Understanding how much couple members value alcohol’s effects could further enlighten the findings of the current study, and contribute to the development of a relationship-motivation model of alcohol use. Additionally, although the RSAEQ was restricted to only certain domains, it is possible that couple members hold relationship-specific alcohol expectancies in other domains beyond the ones assessed. Future research should revise the RSAEQ to more sensitively and more completely assess relationship-specific alcohol expectancies.

In conclusion, the current study was the first to demonstrate that relationship-specific alcohol expectancies predict the extents of relationship-drinking contexts, which in turn reinforce expectancies. Additionally, the context of drinking together appeared to be more integrated with positive expected effects of alcohol compared to other drinking contexts, and that some of these effects were stronger for wives compared to husbands. The current study supports and extends previous research on the bidirectional effects between alcohol use and relationship functioning, and represents a promising preliminary step towards a comprehensive relationship motivation model of alcohol use.

Acknowledgments

This research was supported by National Institute on Alcohol Abuse and Alcoholism Grant R37-AA09922 awarded to Kenneth E. Leonard. Preparation of this manuscript was partially supported by National Institute on Alcohol Abuse and Alcoholism Grant T32-AA007583 awarded to the Research Institute on Addictions, University at Buffalo, SUNY. The authors would like to acknowledge Jaye L. Derrick, Maria Testa, Gregory G. Homish, Brian M. Quigley, Rebecca J. Houston, and Audrey Kubiak for their comments during the preparation of this manuscript.

Footnotes

1

For interested readers, the authors are willing to provide copies of the syntax used for all analyses. Please contact the first author by email if interested.

Contributor Information

Ash Levitt, Research Institute on Addictions, University at Buffalo, SUNY.

Kenneth E. Leonard, Leonard, Research Institute on Addictions, University at Buffalo, SUNY

References

  1. Abelson RP. A variance explanation paradox: When a little is a lot. Psychological Bulletin. 1985;97:129–133. [Google Scholar]
  2. Ackerman RA, Donnellan MB, Kashy DA. Working with dyadic data in studies of emerging adulthood: Specific recommendations, general advice, and practical tips. In: Fincham FD, Cui M, editors. Romantic relationships in emerging adulthood. New York: Cambridge University Press; 2011. pp. 67–97. [Google Scholar]
  3. Ajzen I, Fishbein M. The influence of attitudes on behavior. In: Albarracin D, Johnson BT, Zanna MP, editors. The handbook of attitudes. Mahwah, NJ: Erlbaum; 2005. pp. 173–221. [Google Scholar]
  4. Brown SA, Christiansen BA, Goldman MS. The Alcohol Expectancy Questionnaire: An instrument for the assessment of adolescent and adult alcohol expectancies. Journal of Studies on Alcohol. 1987;48:483–491. doi: 10.15288/jsa.1987.48.483. [DOI] [PubMed] [Google Scholar]
  5. Cohen J. Statistical power analysis for the behavioral sciences. 2. New York: Lawrence Erlbaum Associates; 1988. [Google Scholar]
  6. Cohen J. The Earth is round (p < .05) American Psychologist. 1994;49:997–1003. [Google Scholar]
  7. Cooper ML, Kuntsche E, Levitt A, Barber LL, Wolf S. Motivational models of substance use: A review of theory and research on motives for using alcohol, marijuana, and tobacco. To appear. In: Sher KJ, editor. The Oxford handbook of substance use disorders. New York: Oxford University Press; (in press) [Google Scholar]
  8. Covington SS, Surrey JL. The relational model of women’s psychological development: Implications for substance use. In: Wilsnack S, Wilsnack R, editors. Gender and alcohol: Individual and social perspectives. New Brunswick, NJ: Rutgers Center of Alcohol Studies; 1997. pp. 335–351. [Google Scholar]
  9. Cox M, Klinger E. A motivational model of alcohol use. Journal of Abnormal Psychology. 1988;97:168–180. doi: 10.1037/0021-843X.97.2.168. [DOI] [PubMed] [Google Scholar]
  10. Dermen KH, Cooper ML. Sex-related alcohol expectancies among adolescents: I. Scale development. Psychology of Addictive Behaviors. 1994;8:152–160. doi: 10.1037/0893-164X.8.3.152. [DOI] [Google Scholar]
  11. Dermen KH, Cooper ML. Sex-related alcohol expectancies among adolescents: II. Prediction of drinking in social and sexual situations. Psychology of Addictive Behaviors. 1994;8:161–168. doi: 10.1037/0893-164X.8.3.161. [DOI] [Google Scholar]
  12. Derrick JL, Leonard KE, Quigley BM, Houston RJ, Testa M, Kubiak A. Relationship-specific alcohol expectancies in couples with concordant and discrepant drinking patterns. Journal of Studies on Alcohol and Drugs. 2010;71:761–768. doi: 10.15288/jsad.2010.71.761. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Fromme K, Stroot E, Kaplan D. Comprehensive effects of alcohol: Development and psychometric assessment of a new expectancy questionnaire. Psychological Assessment. 1993;5:19–26. doi: 10.1037/1040-3590.5.1.19. [DOI] [Google Scholar]
  14. Goldman MS, Brown SA, Christiansen BA. Expectancy theory: Thinking about drinking. In: Blane HT, Leonard KE, editors. Psychological Theories of Drinking and Alcoholism. New York: Guilford Press; 1987. pp. 181–226. [Google Scholar]
  15. Goldman MS, Del Boca FK, Darkes J. Alcohol expectancy theory: The application of cognitive neuroscience. In: Leonard KE, Blane HT, editors. Psychological Theories of Drinking and Alcoholism. 2. New York: Guilford Press; 1999. pp. 203–246. [Google Scholar]
  16. Goldman MS, Reich RR, Darkes J. Expectancy as a unifying construct in alcohol-related cognition. In: Wiers RW, Stacy AW, editors. Handbook of Implicit Cognition and Addiction. Thousand Oaks, CA: Sage Publications; 2006. pp. 105–119. [Google Scholar]
  17. Homish GG, Leonard KE. The drinking partnership and marital satisfaction: The longitudinal influence of discrepant drinking. Journal of Consulting and Clinical Psychology. 2007;75:43–51. doi: 10.1037/0022-006X.75.1.43. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Homish GG, Leonard KE. Marital quality and congruent drinking. Journal of Studies on Alcohol. 2005;66:488–496. doi: 10.15288/jsa.2005.66.488. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kachadourian LK, Homish GG, Quigley BM, Leonard KE. Alcohol expectancies, alcohol use, and hostility as longitudinal predictors of alcohol-related aggression. Psychology of Addictive Behaviors. 2011 Oct 17; doi: 10.1037/a0025842. Advance online publication. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kashy DA, Donnellan MB. Conceptual and methodological issues in the analysis of data from dyads and groups. In: Deaux K, Snyder M, editors. The Oxford Handbook of Personality and Social Psychology. New York: Oxford University Press; 2012. pp. 209–238. [Google Scholar]
  21. Kelly AB, Halford WK, Young RM. Expectations of the effects of drinking on couple relationship functioning: An assessment of women in distressed relationships who consume alcohol at harmful levels. Addictive Behaviors. 2002;27:451–464. doi: 10.1016/S0306-4603(01)00185-X. [DOI] [PubMed] [Google Scholar]
  22. Kenny DA, Kashy DA, Cook WL. Dyadic data analysis. New York: Guilford Press; 2006. [Google Scholar]
  23. Laurenceau JP, Bolger N. Using diary methods to study marital and family processes. Journal of Family Psychology. 2005;19:86–97. doi: 10.1037/0893-3200.19.1.86. [DOI] [PubMed] [Google Scholar]
  24. Leonard KE, Homish GG. Predictors of heavy drinking and drinking problems over the first 4 years of marriage. Psychology of Addictive Behaviors. 2008;22:25–35. doi: 10.1037/0893-164X.22.1.25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Leonard KE, Mudar P. Husbands’ influence on wives’ drinking: Testing a relationship motivation model in the early years of marriage. Psychology of Addictive Behaviors. 2004;18:340–349. doi: 10.1037/0893-164X.18.4.340. [DOI] [PubMed] [Google Scholar]
  26. Levitt A, Cooper ML. Daily alcohol use and romantic relationship functioning: Evidence of bidirectional, gender-, and context-specific effects. Personality and Social Psychology Bulletin. 2010;36:1706–1722. doi: 10.1177/0146167210388420. [DOI] [PubMed] [Google Scholar]
  27. Marshal MP. For better or for worse? The effects of alcohol use on marital functioning. Clinical Psychology Review. 2003;23:959–997. doi: 10.1016/j.cpr.2003.09.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. McAdams DP. Power, intimacy, and the life story: Personological inquiries into identity. New York: Guilford Press; 1988. [Google Scholar]
  29. McClelland DC, Davis WN, Kalin R, Wanner E. The drinking man: Alcohol and human motivation. New York: Free Press; 1972. [Google Scholar]
  30. Miller PM, Ingham JG, Plant MA, Miller T. Alcohol consumption and self-disclosure. British Journal of Addiction. 1977;72:296–300. doi: 10.1111/j.1360-0443.1977.tb00694.x. [DOI] [PubMed] [Google Scholar]
  31. Mudar P, Leonard KE, Soltysinski K. Discrepant substance use and marital functioning in newlywed couples. Journal of Consulting and Clinical Psychology. 2001;69:130–134. doi: 10.1037/0022-006X.69.1.130. [DOI] [PubMed] [Google Scholar]
  32. Raudenbush SW, Bryk AS. Hierarchical linear models. 2. Thousand Oaks, CA: Sage Publications; 2002. [Google Scholar]
  33. Read JP, Curtin JJ. Contextual influences on alcohol expectancy processes. Journal of Studies on Alcohol and Drugs. 2007;68:759–770. doi: 10.15288/jsad.2007.68.759. [DOI] [PubMed] [Google Scholar]
  34. Roberts LJ. From bickering to battering: Destructive conflict processes in intimate relationships. In: Noller P, Feeney JA, editors. Close Relationships: Functions, forms and processes. New York: Psychology Press; 2006. pp. 325–352. [Google Scholar]
  35. Roberts LJ, Leonard KE, Senchak M. Alcohol use, intimacy, and marital adjustment in early marriage. Paper presented at the Annual Meeting of the Research Society on Alcoholism; Maui, HI. 1994. Jun, [Google Scholar]
  36. Roberts LJ, Linney KD. Alcohol problems and couples: Drinking in an intimate relational context. In: Schmaling KB, Sher TG, editors. The psychology of couples and illness: Theory, research, & practice. Washington, DC: American Psychological Association; 2000. pp. 269–310. [Google Scholar]
  37. Roberts LJ, Linney KD. Alcohol use, marital functioning and the family life cycle. Paper presented at the Annual Meeting of the National Council on Family Relations; Milwaukee, WI. 1998. Nov, [Google Scholar]
  38. Simpson JA, Tran S. The needs, benefits, and perils of close relationships. In: Noller P, Feeney JA, editors. Close Relationships: Functions, forms and processes. New York: Psychology Press; 2006. pp. 3–24. [Google Scholar]
  39. SPSS, Inc. SPSS Statistics. Vol. 19. Chicago, IL: 2010. [Google Scholar]
  40. Wilsnack RW, Wilsnack SC, Klassen AD. Antecedents and consequences of drinking and drinking problems in women: Patterns from a U.S. national survey. In: Rivers PC, editor. Nebraska Symposium on Motivation: Vol. 34. Alcohol and addictive behavior. Lincoln: University of Nebraska Press; 1987. pp. 85–158. [PubMed] [Google Scholar]

RESOURCES