Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Jun 8.
Published in final edited form as: Exp Clin Psychopharmacol. 2022 Nov 10;31(3):643–651. doi: 10.1037/pha0000615

Validation of the Brief Young Adult Alcohol Consequences Questionnaire Among Student and Nonstudent Young Adults

Amy L Stamates 1, Manshu Yang 1, Cathy Lau-Barraco 2,3
PMCID: PMC10249775  NIHMSID: NIHMS1901152  PMID: 36355679

Abstract

The Brief Young Adult Alcohol Consequences Questionnaire (BYAACQ) is a widely used measure designed to assess alcohol-related consequences experienced by young adults, but little psychometric work has been done outside of college student populations. The primary goal of this study was to investigate if there were group differences between nonstudents (i.e., those without any postsecondary education) and college students (i.e., currently enrolled in a 4-year institution) on the BYAACQ in terms of (a) the “difficulty” level of a consequence based on endorsement and (b) the association between each consequence and the underlying overall level of consequences. Participants were 724 young adult drinkers (Mage= 20.40, SD = 2.19; 58.8% female) who was either a student currently enrolled in a 4-year college (n = 560; 77.3%) or a nonstudent if they had no current or prior college attendance (n = 164; 22.7%) that completed a survey in-person. A confirmatory factor analysis supported the unidimensional structure of the BYAACQ for both students and nonstudents. Differential item functioning (DIF) analyses indicated that six items showed significant DIF in the item difficulty parameters, with college students more likely to experience social- and short-term health-related alcohol consequences, while nonstudents more likely to experience consequences related to impaired control and dependence. Thus, using the latent or sum scores of the entire scale could be problematic, as this overall score is unable to capture specific differences in the types of alcohol consequences experienced between college students and nonstudents. Implications for interventions are discussed.

Keywords: alcohol use, alcohol consequences, education attainment, young adults

Young adulthood (i.e., between the ages of 18 and 25) is a high-risk period for alcohol use (Substance Abuse and Mental Health Services Administration, 2021) and consequences (e.g., Lee et al., 2018; White & Ray, 2014). Most alcohol research has focused on young adults in 4-year colleges, despite that 53.3% of young adults are without any postsecondary educational attainment (United States Census Bureau, 2022). There are well-documented health-related disparities based on educational attainment (Lawrence, 2017; Zajacova et al., 2012), including alcohol-related harms. For example, although college students, as compared to nonstudents, report greater rates of alcohol use in the past month (56% vs. 49%, respectively), binge rates in the past month are the same between groups (24%; Schulenberg et al., 2021). Previous research has suggested that nonstudent drinkers are less likely to mature out of heavy drinking patterns (Evans-Polce et al., 2017; Lanza & Collins, 2006; White et al., 2005), are more likely to develop an alcohol use disorder in adulthood (Harford et al., 2006), and may benefit less from existing alcohol treatments (Davis et al., 2017). Most of this prior research, however, has focused on single indicators of alcohol use and consequences (i.e., met criteria for alcohol use disorder vs. not) or mean levels of alcohol-related consequences, rather than examining specific types of consequences. Although students and nonstudents may experience similar types of alcohol-related consequences, there may be differences given that some alcohol-related consequences may be contextually or socially driven (Stone et al., 2012).

One of the most widely used validated measures of alcohol-related consequences is the Young Adult Alcohol Consequences Questionnaire (YAACQ; Read et al., 2006). The YAACQ was designed to capture a broad range of alcohol-related consequences experienced by college students. It includes 48 items across eight different subscales (Social Interpersonal Problems, Impaired Control, Self-Perception, Self-Care, Risky Behaviors, Academic/Occupational Consequences, Physiological Dependence, and Blackout Drinking) that have strong internal consistency reliabilities and load onto one higher order alcohol consequence scale. The YAACQ has been validated in various college populations (Bravo et al., 2019). A brief version was created, known as the Brief Young Adult Alcohol Consequences Questionnaire (BYAACQ), which includes a 24-item unidimensional scale of consequences that has adequate reliability and construct validity among college students (Kahler et al., 2004). Shorter versions of scales may be particularly beneficial in community-based research as they have been associated with increased completion rates (Kost & de Rosa, 2018).

Although both the full and brief scales are widely used among young adults, there has been little psychometric validation research outside of the college student population. One recent study examined measurement invariance of the YAACQ across students (i.e., those currently enrolled or completed a 4-year college degree or higher) and nonstudents (i.e., those that completed an associate’s degree or less and are not currently enrolled in a 4-year institution; Campbell et al., 2021). Confirmatory factor analysis (CFA) revealed support for measurement equivalence of the eight-factor version of YAACQ across students and nonstudents, with nonstudents reporting greater scores in the impaired control, physical dependence, and risky behavior domains. Thus, while there was some similarity between students and nonstudents on the types of consequences endorsed, there also were some notable differences. Comparisons between student and nonstudent groups on the BYAACQ have not been explored.

In order to enhance the comparability of the BYAACQ among different groups of young adults, it is critical to assume that students and nonstudents with the same overall level of alcohol-related consequences would respond in the same way to a particular question. If this assumption of measurement equivalence does not hold, group comparisons based on the BYAACQ scores could be problematic. Differential item functioning (DIF) analysis is a procedure, based on the item response theory (IRT), to investigate whether items measure the same latent trait (i.e., overall level of alcohol consequences) in the same way across two or more groups of individuals. A key feature of IRT is that a person’s likelihood of endorsing an item is determined not only by this person’s underlying trait (e.g., overall level of alcohol consequences) but also by the item’s properties that are characterized by one or more item parameters, which provides additional information compared to a linear CFA model. For example, in a 2-PL (2-parameter logistic) model (Birnbaum et al., 1968), two parameters per item (i.e., the discrimination parameter and the difficulty parameter) are used to capture the item properties. The discrimination parameter is analogous to the factor loading in a CFA model; it estimates how well an item differentiates among individuals who have higher versus lower levels of the measured trait (i.e., overall level of alcohol consequences), and it also indicates how strong the item is related to the trait being measured. The difficulty parameter provides information about the relative location of an item; specifically, it represents the points along the latent trait continuum at which a respondent has a 50% chance of endorsing the item. The higher the difficulty parameter for an item, the less likely a person would endorse this item. For instance, an item with a difficulty parameter of 1.5 has fewer people endorsing it, as compared to an item with a difficulty parameter of 0.2. If the item parameter estimates differ significantly between groups (e.g., students vs. nonstudents), then there is evidence that DIF occurs, and measurement equivalence does not hold—in other words, the item is not measuring the latent trait in the same way across groups.

The primary goal of this study was to investigate the potential DIF of the BYAACQ between young adults currently enrolled in college and young adults with no postsecondary education. Specifically, the present study aimed to determine if there were group differences between nonstudents and students on the BYAACQ in terms of (a) the “difficulty” level of a consequence based on endorsement and (b) the association between each consequence and the underlying overall level of consequences.

Method

Participants and Procedure

The sample consisted of 724 young adult drinkers (Mage = 20.40, SD = 2.19; 58.8% female) recruited from the community (n = 164) and campus (n = 560) of a mid-size, urban southeastern city in the United States. Community members were recruited via study flyers and Craigslist postings as part of a larger study (Lau-Barraco et al., 2018). Only the baseline data from the larger study were used in the present study. A comparison group of students using the same inclusion criteria was recruited from the participating university via flyers posted on campus and the psychology participant pool (i.e., Sona-systems). Students could access the survey via an online portal and volunteer to participate. For study inclusion criteria, participants must have been between 18 and 25 years old and reported engaging in at least two heavy drinking episodes (i.e., consumed at least 4+/5+ standard drinks for women/men on one occasion) in the past month. We chose to include drinkers that would be considered “high risk” and engaging in at least two heavy drinking episodes over the past 30 days is consistent with criteria established by Wechsler et al. (1994). Exclusion criteria included consumption above 40 drinks weekly, as these individuals may represent a more severe pattern of drinking and may not be suitable for the parent brief intervention study (e.g., Darkes & Goldman, 1993, 1998; Lau-Barraco & Dunn, 2008).

Respondents were considered a college student if they were enrolled in a 4-year college/university at the time of data collection. Respondents were considered a nonstudent if they had no current or prior college attendance and were not currently enrolled in high school. Individuals who were currently enrolled or had a history of training in technical/vocational programs or associate degree programs were excluded. Individuals with previous but not current enrollment in a 4-year college were not eligible for the study. Descriptive statistics by student status are provided in Table 1.

Table 1.

Descriptive Statistics and Comparisons Between College Students and Nonstudents

Variable College students (n = 560)
 Nonstudents (n = 164)
M (SD) M (SD) df t p
Age 19.96 (2.03) 21.98 (2.02) 713 −10.99 <.001
Typical alcohol use 14.43 24.16 722 −7.99 <.001
Variable n (%) n (%) df χ2 p
Gender 1 53.79 <.001
 Men 189 (33.8) 108 (65.9)
 Women 370 (66.2) 56 (34.1)
Race/ethnicity 2 13.87
 Caucasian/White 291 (52.3) 67 (42.1) <.001
 Native American/Indian 0 2 (1.3)
 African American/Black 169 (30.4) 79 (49.7)
 Hispanic/Latino 31 (5.6) 11 (6.9)
 Asian 20 (3.6) 0
 More than once race 31 (5.6) 0
 Other 14 (2.5) 0
 Missing 4 5
Relationship status 1 42.32
 Single/never married 531 (95.2) 117 (71.3)
 Living with partner 18 (3.2) 26 (15.9) <.001
 Married 8 (1.4) 11 (6.7)
 Divorced 1 (0.2) 10 (6.1)
Employment 3 67.04
 Yes, part-time only 206 (36.8) 42 (25.8)
 Yes, full and part-time 19 (3.4) 15 (9.2)
 Yes, full-time only 16 (2.9) 31 (19.0) <.001
 Unemployed 319 (57.0) 75 (46.0)
 Missing
Year in college — —
 Freshman 211 (37.7) —
 Sophomore 155 (27.7) —
 Junior 104 (18.6) —
 Senior 85 (15.2) — —
 Other 5 (0.9)
Individual income 2 74.99
 Under $10,000 453 (81.8) 81 (50.9)
 $10,001-$20,000 71 (12.8) 42 (26.4)
 $20,001-$40,000 17 (3.1) 29 (18.2)
 $40,001-$60,000 4 (0.7) 6 (3.8) <.001
 $60,001-$80,000 3 (0.5) 1 (0.1)
 $80,001-$100,000 4 (0.7) 0
 $100,000 or more 2 (0.3) 0
Open in a new tab

Note. Categories with fewer than 5 responses were removed from overall chi-square tests.

All participants were informed that the purpose of the study was to gain more knowledge about alcohol use among young adults to develop more efficacious strategies to help young adults who drink. Those who met the study criteria were scheduled to attend an in-person meeting where they obtained informed consent and completed a battery of survey assessments. The surveys took approximately 45 min to 1 hr to complete. Nonstudents were compensated financially (between $40 and $60 depending on their time commitment), and college students were compensated course credit. All data collection procedures and questionnaires were the same for each group except for compensation. The present study was approved by the university’s institutional review board and followed all APA ethical guidelines (American Psychological Association, 2017). These data have not been previously published or presented, but data from the nonstudent sample has been published on from the parent study (Lau-Barraco et al., 2018). The study was not preregistered. Study materials are available upon request from the first author.

Measures

Alcohol Use

The Daily Drinking Questionnaire (Collins et al., 1985) was used to assess typical drinking behavior in the past 30 days. Participants were instructed on the definition of a standard drink and asked to report the number of standard drinks that they consumed on each day of a typical week in the past 3 months. Alcohol quantity was derived by summing the total number of drinks and used as descriptive data of the sample.

Alcohol-Related Consequences

The 24-item BYAACQ (Kahler et al., 2005) measured alcohol-related consequences in the past 12 months using a “yes” (1) or “no” (0) response format. In the present study, Cronbach’s α was .88 and .87 for nonstudents and students, respectively.

Statistical Analyses

Item-level descriptive statistics were computed as a check on data entry validity and to verify that there were no empty (zero frequency) response categories for any BYAACQ item. We then conducted analyses to examine the measurement equivalence of the BYAACQ items by college status (i.e., college students vs. nonstudents of similar ages). Specifically, we examined (a) the unidimensionality and the overall model fit of the BYAACQ among college students and nonstudents separately, (b) the presence of DIF by college status, and (c) the latent or sum score differences between college students and nonstudents.

Unidimensionality and Overall Model Fit.

The BYAACQ items were found to measure a single underlying dimension of alcohol-related consequences among college students (Kahler et al., 2004). The scale unidimensionality is also a key assumption for fitting the IRT models used in this study. To investigate whether the single-factor structure can be extended from college students to nonstudents and to verify that the unidimensionality assumption has been met before IRT analyses, we evaluated scale dimensionality by examining the fit of a one-factor categorical CFA model to the college students’ data and the nonstudents’ data, respectively. The CFA also evaluated the extent to which the empirically obtained survey data matched the specified IRT model. CFA was conducted using Mplus (Muthén & Muthén, 1998–2017) with the WLSMV estimator (the weighted least squares with adjustments for the mean and variance). Because there is not a single statistic universally accepted for all tests of model fit, we examined multiple indicators to evaluate dimensionality. Model fit indices included the Comparative Fit Index (CFI), the Tucker–Lewis Index (TLI), and the root-mean-square error of approximation (RMSEA), with CFI > 0.95, TLI > 0.95, and RMSEA< 0.06 considered as good fit (Hu&Bentler, 1999).

Presence of DIF.

IRT-based DIF analyses were carried out using the software IRTPRO (Cai et al., 2011). We evaluated two types of DIF. The uniform DIF indicates a difference only in the difficulty parameter between groups; that is, one group is consistently more likely than the other group to endorse the item, regardless of the level of the underlying trait. The nonuniform DIF suggests a difference between groups in the discrimination parameter (alone or coupled with difference in the difficulty parameter); it implies that, at certain levels of the underlying trait, one group is more likely to endorse the item, while at other levels, the other group is more likely to endorse the item. The BYAACQ was originally validated using the 1-parameter logistic model (Rasch, 1960) that estimates a single difficulty parameter per item to assess the level of alcohol consequences reflected in each item (Kahler et al., 2004). Given that the 1-PL model can only detect uniform DIF, we extended the IRT analyses to a 2-PL model to evaluate both uniform and nonuniform DIF. We examined the possibility of DIF by college status for each BYAACQ item using the Wald test (Langer, 2008; Lord, 1980). For the Wald test, a nonsignificant χ2 value indicates no detectable DIF for an item, or that the item parameter estimates are not significantly different between college students and nonstudents. We used the Benjamin–Hochberg procedure (Benjamini & Hochberg, 1995; Thissen et al., 2002) to adjust α levels and control for the multiple comparisons involved in checking DIF for all the 24 items. It is possible that an item shows statistically significant DIF yet has a small effect size, making its practical significance trivial. Therefore, for items exhibiting statistically significant DIF, we further examined the effect size of DIF, with a difference of 0.5 or greater between groups in the difficulty parameter considered as a large effect and practically meaningful (Steinberg & Thissen, 2006).

DIF Sensitivity Analysis.

Preliminary analyses showed that the study samples of college students and nonstudents differed in sociodemographic characteristics other than the college status. Therefore, sensitivity analyses were conducted using the moderated nonlinear factor analysis (MNLFA) approach (Bauer, 2017; Bauer & Hussong, 2009; Curran et al., 2014), to examine whether DIF by college status remained after controlling for other confounding factors. The MNLFA removes the limitation in the conventional DIF analysis methods that evaluate DIF only for a single grouping variable (e.g., college students vs. nonstudents) at a time. MNLFA addresses the sample heterogeneity issue in the present study and allows researchers to evaluate DIF between groups while controlling for confounding background variables that differ between the students’ and nonstudents’ samples. The MNLFA models were fit in Mplus Version 8.6 (Muthén & Muthén, 1998–2017).

Score Comparisons by College Status.

Upon detection of DIF, we further compared (a) the latent IRT-based scores for the entire BYAACQ scale, (b) the raw sum scores for the entire scale, and (c) the raw sum scores for items with significant DIF, between college students and nonstudents, to demonstrate the impact of DIF on group comparison results. Group comparisons were conducted using the independent-sample t-test in SAS (SAS Institute Inc., 2002–2014).

Results

Descriptive Statistics

The BYAACQ was completed by a total of 724 respondents, including 560 college students and 164 nonstudents. Demographic information for the respondents by college status is shown in Table 1. College students and nonstudents differed on age, gender, race/ethnicity, relationship status, employment status, and level of individual income, p < .001.

Item nonresponses were minimal, with no more than 1.2% of missing responses for each of the 24 items. As shown in Table 2, the percentage of respondents endorsing each item ranged from 4.1% to 80.2%, indicating that these items cover a wide range of alcohol-related consequences. The three most commonly endorsed items were the same for college students and nonstudents, including Item 1 (said/done embarrassing things; 80.2% and 64.6% endorsed by college students and nonstudents, respectively), Item 5 (had a hangover; 77.9% and 79.3% endorsed by college students and nonstudents), and Item 8 (sick to stomach or vomiting; 72.1% and 64.6% endorsed by college students and nonstudents). However, the three least endorsed items differed by college status. College students least endorsed Item 12 (need drink upon awakening; 4.1%), Item 23 (overweight; 8.4%), and Item 16 (neglected obligations; 12.3%), whereas nonstudents least endorsed Item 2 (suffered quality of work/schoolwork; 11.6%), Item 22 (harmed physical appearance; 11.6%), and Item 16 (neglected obligations; 14%). No item had empty (zero frequency) response categories.

Table 2.

Item Endorsement, Item Parameters Estimates, and DIF Testing by College Status

Item no. Item content College students
 Nonstudents
 DIF in item slope
 DIF in item severity
% endorsed Item slope Item severity % endorsed Item slope Item severity χ 2 χ 2 ES
1 Said/done embarrassing things 80.2 1.42 −1.34 64.6 1.63 −0.55 0.3 11.3 0.79
5 Had a hangover 77.9 1.24 −1.30 79.3 1.45 −1.32 0.3 0.4
8 Sick to stomach or vomiting 72.1 1.28 −0.97 64.6 0.58 −1.15 8.4 4.1
7 Taken foolish risks when drinking 48.6 2.29 0.06 50.6 1.59 0.00 3.2 0
17 Unplanned drinking on nights 47.0 1.47 0.12 53.7 1.23 −0.14 0.7 1.6
22 Harmed physical appearance 48.6 0.36 0.14 11.6 1.24 2.10 5 35.4 1.96
21 Memory loss when drinking 42.3 1.83 0.26 28.7 1.23 1.00 2.9 15.5 0.74
18 Later regretted impulsive things 40.7 2.12 0.31 39.0 2.27 0.42 0.1 0.4
6 Passed out from drinking 41.3 1.23 0.37 32.3 1.18 0.84 0 4.6
24 Less energy or felt tired 28.2 1.39 0.89 50.6 1.37 −0.06 0 23.9 0.96
20 Later regretted sexual situations 27.9 1.38 0.92 30.5 0.90 1.10 2.9 0.1
11 Missed work or classes 24.8 1.78 0.93 17.7 1.30 1.59 1.5 6.3
4 Too much drink to drive safely 25.2 1.64 0.95 22.6 1.02 1.50 3.7 1.2
15 Tolerance to alcohol 29.8 1.10 0.95 35.4 1.00 0.76 0.1 1.1
3 Felt badly about self 27.5 1.25 0.99 24.4 1.95 1.00 2.3 0.4
2 Suffered quality of work/schoolwork 20.5 2.01 1.07 11.6 2.40 1.56 0.3 7.2 0.49
13 Rude/obnoxious/insulting 20.7 1.55 1.20 30.5 1.37 0.86 0.3 4.1
14 Woken up in an unexpected place 19.1 1.47 1.33 23.8 1.31 1.19 0.2 1
9 Problems with partners/relatives 20.0 1.37 1.33 18.9 1.43 1.43 0 0.3
19 Difficulty limiting drinking 15.2 1.81 1.41 29.9 1.62 0.83 0.2 12.4 0.58
16 Neglected obligations 12.3 2.25 1.44 14.0 3.07 1.32 0.9 0
10 Spent too much time drinking 12.5 1.85 1.56 25.6 2.04 0.93 0.1 10.2 0.63
23 Overweight 8.4 1.16 2.49 14.6 1.42 1.72 0.3 4.6
12 Need drink upon awakening 4.1 1.35 2.90 20.1 0.97 1.73 1.1 33.7 1.17
Open in a new tab

Note. ES = Effect size; DIF = differential item functioning; items are organized in ascending order of item severity for college students; boldface values/text indicate items exhibiting significant DIF with overall α = 0.05 and with higher item severity for nonstudents than for college students; boldface and italic values/text indicate items exhibiting significant DIF with overall α = 0.05 and with lower item severity for nonstudents than for college students.

Unidimensionality and Overall Model Fit

Overall model fit results from the single-factor categorical CFA analyses suggested that the unidimensional structure and the 2-PL IRT model fit the empirical data well for college students and nonstudents, allowing us to apply the IRT model and test for DIF across the 24 items of the BYAACQ. For the college student sample, χ2(252) = 678.656 with p < .001, CFI = 0.916, TLI = 0.908, RMSEA = 0.055 with 95% confidence interval (CI) [0.05, 0.06]. For the nonstudent sample, χ2(252) = 319.233 with p = .003, CFI = 0.956, TLI = 0.952, RMSEA = 0.04 with 95% CI [0.025, 0.053]. Although the CFI and TLI for college students were slightly below the conventional thresholds of 0.95 (Hu & Bentler, 1999), they were within an acceptable range. In addition, Cook et al. (2009) and Marsh et al. (2004) recommended an investigative approach rather than overly relying on traditional cutoffs of fit indices when examining unidimensionality, when a relatively large number of items (e.g., 24 items in the present study) were loaded on a single factor.

DIF

Results of DIF testing by college status (college students vs. nonstudents) are shown in Table 2. None of the items exhibited significant DIF in terms of the item discrimination parameters (i.e., the association between a consequence and the underlying overall level of consequences). However, a total of eight items showed significant DIF in the item difficulty parameters, with four items less likely to be endorsed by nonstudents than by college students, and the other four items less likely to be endorsed by college students than by nonstudents. These items showed uniform DIF, indicating that one group was consistently more likely than the other group to endorse these items, at each level of the overall alcohol consequences score. All the items showing significant DIF had large effect sizes greater or close to 0.5 (i.e., at least a half standard unit difference in item difficulty between college students and nonstudents), indicating that the DIF was practically meaningful. Specifically, as compared to college students with the same overall level (i.e., latent score) of consequences, nonstudents were less likely to endorse Item 1 (said/done embarrassing things), Item 2 (suffered quality of work/schoolwork), Item 21 (memory loss when drinking), and Item 22 (harmed physical appearance). In contrast, Item 10 (spent too much time drinking), Item 12 (need drink upon awakening), Item 19 (difficulty limiting drinking), and Item 24 (less energy or felt tired) were more likely to be endorsed by nonstudents, as compared to their college student counterparts with the same overall consequence score.

Sensitivity analyses were conducted using MNLFA to examine DIF by college status, while controlling for participants’ age, gender, race/ethnicity, relationship status, employment status, and typical level of alcohol use, which differed significantly between the study samples of college students and nonstudents. Although the level of individual income also differed between college students and nonstudents, it was not included as a covariate for DIF, given its high correlation with employment status. After controlling for participant characteristics, DIF remained significant in six items (Items 1, 12, 19, 21, 22, 24), while two items (Items 2 and 10) no longer exhibited DIF by college status. Therefore, we considered Items 2 and 10 as not showing DIF.

Group Mean Comparisons

As a by-product of the DIF analyses, latent IRT-based scores were obtained for all participants to measure their overall levels of alcohol-related consequences. In addition, three sets of sum scores were computed by summing up raw item responses (1 = endorsed, 0 = not endorsed) across items, including (a) sum score for the entire BYAACQ scale across 24 items, (b) sum score across items showing significant DIF and that were less likely to be endorsed by nonstudents (Items 1, 21, 22), and (c) sum score across items showing significant DIF and that were more likely to be endorsed by nonstudents (Items 12, 19, 24). Mean score comparisons between college students and nonstudents are presented in Table 3. Based on both latent scores and sum scores of the entire scale, no significant differences were found on average between college students and nonstudents in terms of the overall level of alcohol consequences. However, compared to college students, nonstudents had significantly lower sum scores of the three consequences that they were less likely to endorse (said/done embarrassing things, harmed physical appearance, memory loss when drinking), and nonstudents had significantly higher sum scores of the three consequences that they were more likely to endorse (less energy or felt tired, difficulty limiting drinking, need drink upon awakening). Since DIF occurred in opposite directions across the six items, their impact on the overall scores of the entire scale cancelled out.

Table 3.

Comparisons of Latent IRT Scores and Sum Scores by College Status

BYAACQ score College students
 Nonstudents
 Mean difference
M SD M SD Estimate 95% CI
Latent IRT score (entire scale) 0.00 0.93 0.01 1.01 −0.01 [−0.17, 0.16]
Sum score (entire scale) 8.06 5.06 7.88 5.29 0.18 [−0.74, 1.11]
Sum score (3 DIF items) 1.72 0.91 1.04 0.87 0.68* [0.52, 0.84]
 Said/done embarrassing things
 Harmed physical appearance
 Memory loss when drinking
Sum score (3 DIF items) 0.48 0.72 1.03 0.94 −0.55* [−0.68, −0.41]
 Less energy or felt tired
 Difficulty limiting drinking
 Need drink upon awakening
Open in a new tab

Note. IRT = Item response theory; BYAACQ = Brief Young Adult Alcohol Consequences Questionnaire; Cl = confidence interval; DIF = differential item functioning.

*

Mean difference was statistically significant at p < .0001.

Discussion

The present study aimed to psychometrically validate the widely used BYAACQ and to provide a comparison in the types of consequences experienced by groups of young adults. In particular, we compared nonstudents and college students on the “difficulty” level of a BYAACQ consequence and the association between each consequence and the underlying overall level of consequences. All consequences on the BYAACQ were endorsed in the present study, and the unidimensional structure of the BYAACQ fits well for both students and nonstudents. The overall level of negative alcohol-related consequences did not differ between students and nonstudents, but six items out of 24 were endorsed differently between the two groups.

Examination of the BYAACQ items revealed some overlap and differences between students and nonstudents. Regarding similarities, the most commonly endorsed consequences (e.g., said/done embarrassing things, had a hangover, and sick to stomach or vomiting) were the same across groups. These types of consequences are short-term health-related and social consequences associated with heavy drinking and are experienced by many college-attending young adults (Lau-Barraco et al., 2017; Perkins, 2002). Regarding differences, there were three items less likely to be endorsed by nonstudents than college students and these items reflected domains from the Self-care, Blackout, and Social subscales of the original YAACQ. There were three items less likely to be endorsed by college students, as compared to nonstudents, which were from the impaired control, self-care, and dependence domains of the YAACQ. In other words, when comparing students and nonstudents with the same overall level of alcohol-related consequences, college students were more likely to experience social- and short-term health-related alcohol consequences, while nonstudents were more likely to experience consequences related to impaired control and dependence.

The differences in which types of consequences endorsed by students and nonstudents have implications. For some young adults, individuals may progress from more common to more extreme types of consequences (Vik et al., 2000). Among college students, some consequences (e.g., blackouts, hangovers, social embarrassment) are more acceptable, and even desirable (Mallett et al., 2008). Social consequences have been linked to increase in alcohol use in later semesters among college students (Read et al., 2013), suggesting that social-related consequences may reinforce drinking behavior. Thus, our study findings are in line with previous research that college students would be more likely to endorse these types of consequences. For nonstudents, their greater likelihood of endorsement related to impaired control and dependence is consistent with previous research (Campbell et al., 2021) and may suggest why they are at greater risk for developing alcohol use disorder in the long term (Harford et al., 2006). Research suggests that impaired control is one of the earliest developing signs of problem drinking (Leeman et al., 2014), and physiological dependence symptom in young adulthood is a risk factor for problematic drinking later in life (O’Neill & Sher, 2000). While our findings suggest that the latent or sum BYAACQ score maintains a valid, unidimensional structure for both students and nonstudents, there are specific differences in some items between the two groups that are not captured by the overall BYAACQ latent or sum score, which could be a barrier to fully understanding the harms experienced by health-disparate groups.

Our findings have implications for interventions. Assessing specific consequences with the BYAACQ items may have clinical utility for interventionists regarding which individuals may be at greatest long-term risk for developing alcohol-related problems. Personalized feedback interventions (PFIs) are an effective secondary prevention approach for alcohol use and problems, most commonly used among college students (Miller et al., 2013), but also have been found to be effective in short-term drinking reductions among nonstudents (Lau-Barraco et al., 2018). PFIs could incorporate information about specific types of consequences endorsed by individuals, as those may be more personally relevant. Moreover, because of the greater endorsement of specific BYAACQ items for nonstudents and the potential indicator of these items for later problem drinking, changes in the endorsement of those specific consequences over time in response to intervention efforts should be examined in future research.

There are limitations in the present study. First, the data were self-reported and may be subject to recall bias; however, research supports the consistency of negative consequences reported between retrospective and prospective assessments (Merrill et al., 2020). Relatedly, the BYAACQ was specifically created to measure consequences among college-attending young adults. It is possible that the BYAACQ does not capture the full range of consequences experienced by young adults who are not in college and could be an area for future research. Second, there may be potential bias in the recruitment or payment method approach given that student participants could earn course credit and nonstudents were paid. Research is mixed on whether financial versus course credit incentives influence behavior (Bowen & Kensinger, 2017; Luccasen & Thomas, 2014). If recruitment or payment strategies only influenced students and nonstudents to differ in their overall levels but did not influence their behaviors when answering a specific BYAACQ item, then the DIF analytic approach would produce valid results, because it controlled for the underlying latent trait (Holland & Wainer, 1993). Third, we defined “nonstudent” as an individual who has never attended any postsecondary education because this population may reflect a vulnerable population who has not received any secondary prevention programming. However, this may limit generalizability to other definitions of nonstudent (e.g., individuals who have not completed a 4-year degree and are currently not enrolled) as there is a lack of consistency in how nonstudent status is operationalized in the literature. Fourth, previous research has recommended 200–400 respondents per group for detecting DIF in a 2-PL IRT model (Belzak, 2020; Meade & Bauer, 2007; Woods et al., 2013); although the sample size of college students (n = 560) met this criterion, the smaller sample of nonstudents (n = 164) may have reduced the power of detecting DIF. Given that two items (Items 2 and 10) showed DIF in the primary analyses but not in the sensitivity analysis, future studies with a larger nonstudent sample are needed to evaluate the replicability of the DIF findings. Last, our data were collected at one timepoint, and thus, we are limited in making inferences about how alcohol-related consequences may develop between students and nonstudents over time. Future work is needed to investigate the long-term development of consequences in nonstudents, specifically.

Overall, the present study advanced knowledge regarding the utility of the BYAACQ in assessing alcohol-related consequences among college students and nonstudents. We found that specific items were differentially endorsed by each group. Thus, using the latent or sum scores of the entire scale could be problematic, as this overall score is unable to capture specific differences in the types of alcohol consequences experienced between college students and nonstudents.

Public Health Significance.

This study suggests that although the unidimensional Brief Young Adult Alcohol Consequences Questionnaire (BYAACQ) was psychometrically supported among college students and nonstudents, items were differentially endorsed between the two groups. Specifically, college students were more likely to endorse social- and short-term health-related alcohol consequences from the BYAACQ, while nonstudents were more likely to experience consequences related to impaired control and dependence. Assessing specific BYAACQ items may have clinical utility for young adults.

Acknowledgments

This research was supported by Grant K01-AA018383 (PI: Cathy Lau-Barraco) from the National Institute on Alcohol Abuse and Alcoholism. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. These data have not been previously published or presented, but data from the nonstudent sample has been published on from the parent study (Lau-Barraco et al., 2018). The study was not preregistered. Study materials are available upon request from Amy L. Stamates.

Footnotes

The authors declare that they have no conflicts of interest to disclose.

References

  1. American Psychological Association. (2017). Ethical principles of psychologists and code of conduct (2002, amended effective June 1, 2010, and January 1, 2017). https://www.apa.org/ethics/code/
  2. Bauer DJ (2017). A more general model for testing measurement invariance and differential item functioning. Psychological Methods, 22(3), 507–526. 10.1037/met0000077 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bauer DJ, & Hussong AM (2009). Psychometric approaches for developing commensurate measures across independent studies: Traditional and new models. Psychological Methods, 14(2), 101–125. 10.1037/a0015583 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Belzak WCM (2020). Testing differential item functioning in small samples. Multivariate Behavioral Research, 55(5), 722–747. 10.1080/00273171.2019.1671162 [DOI] [PubMed] [Google Scholar]
  5. Benjamini Y, & Hochberg Y (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society. Series B. Methodological, 57(1), 289–300. 10.1111/j.2517-6161.1995.tb02031.x [DOI] [Google Scholar]
  6. Birnbaum A, Lord FM, & Novick MR (1968). Some latent trait models and their use in inferring an examinee’s ability. In Lord FM & Novick MR (Eds.), Statistical theories of mental test scores (pp. 397–479). Addison-Wesley. [Google Scholar]
  7. Bowen HJ, & Kensinger EA (2017). Cash or credit? Compensation in psychology studies: Motivation matters. Collabra: Psychology, 3(1), 12. 10.1525/collabra.77 [DOI] [Google Scholar]
  8. Bravo AJ, Pilatti A, Pearson MR, Read JP, Mezquita L, Ibáñez MI, & Ortet G (2019). Cross-Cultural examination of negative alcohol-related consequences: Measurement invariance of the Young Adult Alcohol Consequences Questionnaire in Spain, Argentina, and USA. Psychological Assessment, 31(5), 631. 10.1037/pas0000689 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Cai L, Thissen D, & du Toit SHC (2011). IRTPRO: Flexible, multidimensional, multiple categorical IRT modeling [Computer software]. Scientific Software International. [Google Scholar]
  10. Campbell KW, Pebley K, MacKillop J, & Murphy JG (2021). Measurement invariance of the Young Adult Alcohol Consequences Questionnaire across college status, race, and childhood SES in a diverse community sample. Psychology of Addictive Behaviors. Advance online publication. 10.1037/adb0000789 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Collins RL, Parks GA, & Marlatt GA (1985). Social determinants of alcohol consumption: The effects of social interaction and model status on the self-administration of alcohol. Journal of Consulting and Clinical Psychology, 53(2), 189–200. 10.1037//0022-006x.53.2.189 [DOI] [PubMed] [Google Scholar]
  12. Cook KF, Kallen MA, & Amtmann D (2009). Having a fit: Impact of number of items and distribution of data on traditional criteria for assessing IRT’s unidimensionality assumption. Quality of Life Research, 18(4), 447–460. 10.1007/s11136-009-9464-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Curran PJ, McGinley JS, Bauer DJ, Hussong AM, Burns A, Chassin L, Sher K, & Zucker R (2014). A moderated nonlinear factor model for the development of commensurate measures in integrative data analysis. Multivariate Behavioral Research, 49(3), 214–231. 10.1080/00273171.2014.889594 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Darkes J, & Goldman MS (1993). Expectancy challenge and drinking reduction: Experimental evidence for a mediational process. Journal of Consulting and Clinical Psychology, 61(2), 344–353. 10.1037/0022-006X.61.2.344 [DOI] [PubMed] [Google Scholar]
  15. Darkes J, & Goldman MS (1998). Expectancy challenge and drinking reduction: Process and structure in the alcohol expectancy network. Experimental and Clinical Psychopharmacology, 6(1), 64–76. 10.1037/1064-1297.6.1.64 [DOI] [PubMed] [Google Scholar]
  16. Davis JP, Smith DC, & Briley DA (2017). Substance abuse prevention and treatment outcomes for emerging adults in non-college settings: A meta-analysis. Psychology of Addictive Behaviors, 31(3), 242–254. 10.1037/adb0000267 [DOI] [PubMed] [Google Scholar]
  17. Evans-Polce RJ, Maggs JL, Staff J, & Lanza ST (2017). The age-varying association of student status with excessive alcohol use: Ages 18 to 30 years. Alcoholism: Clinical & Experimental Research, 41(2), 407–413. 10.1111/acer.13294 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Harford TC, Yi HY, & Hilton ME (2006). Alcohol abuse and dependence in college and noncollege samples: A ten-year prospective follow-up in a national survey. Journal of Studies on Alcohol, 67, 803–809. 10.15288/jsa.2006.67.803 [DOI] [PubMed] [Google Scholar]
  19. Holland PW, & Wainer H (1993). Differential item functioning. Erlbaum. [Google Scholar]
  20. Hu L, & Bentler PM (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1–55. 10.1080/10705519909540118 [DOI] [Google Scholar]
  21. Kahler CW, Strong DR, & Read JP (2005). Toward efficient and comprehensive measurement of the alcohol problems continuum in college students: The Brief Young Adult Alcohol Consequences Questionnaire. Alcoholism: Clinical and Experimental Research, 29, 1180–1189. 10.1097/01.ALC.0000171940.95813.A5 [DOI] [PubMed] [Google Scholar]
  22. Kahler CW, Strong DR, Read JP, Palfai TP, & Wood MD (2004). Mapping the continuum of alcohol problems in college students: A Rasch model analysis. Psychology of Addictive Behaviors, 18(4), 322–333. 10.1037/0893-164X.18.4.322 [DOI] [PubMed] [Google Scholar]
  23. Kost RG, & de Rosa JC (2018). Impact of survey length and compensation on validity, reliability, and sample characteristics for Ultrashort-, Short-, and Long-Research Participant Perception Surveys. Journal of Clinical and Translational Science, 2(1), 31–37. 10.1017/cts.2018.18 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Langer MM (2008). A reexamination of Lord’s Wald test for differential item functioning using item response theory and modern error estimation. The University of North Carolina at Chapel Hill. [Google Scholar]
  25. Lanza ST, & Collins LM (2006). A mixture model of discontinuous development in heavy drinking from ages 18 to 30: The role of college enrollment. Journal of Studies on Alcohol, 67, 552–561. 10.15288/jsa.2006.67.552 [DOI] [PubMed] [Google Scholar]
  26. Lau-Barraco C, Braitman AL, & Stamates AL (2018). A randomized trial of a personalized feedback intervention for nonstudent emerging adult at-risk drinkers. Alcoholism: Clinical and Experimental Research, 42(4), 781–794. 10.1111/acer.13606 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Lau-Barraco C, & Dunn ME (2008). Evaluation of a single-session expectancy challenge intervention to reduce alcohol use among college students. Psychology of Addictive Behaviors, 22(2), 168–175. 10.1037/0893-164X.22.2.168 [DOI] [PubMed] [Google Scholar]
  28. Lau-Barraco C, Linden-Carmichael AN, Hequembourg A, & Pribesh S (2017). Motivations and consequences of alcohol use among heavy drinking nonstudent emerging adults. Journal of Adolescent Research, 32(6), 667–695. 10.1177/0743558416630812 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Lawrence EM (2017). Why do college graduates behave more healthfully than those who are less educated?. Journal of Health and Social Behavior, 58(3), 291–306. 10.1177/0022146517715671 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Lee CM, Cadigan JM, Fairlie AM, & Lewis MA (2018). Transitions into young adulthood: Extent to which alcohol use, perceived drinking norms, and consequences vary by education and work statuses among 18–20 year olds. Addictive Behaviors, 79, 107–112. 10.1016/j.addbeh.2017.12.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Leeman RF, Beseler CL, Helms CM, Patock-Peckham JA, Wakeling VA, & Kahler CW (2014). A brief, critical review of research on impaired control over alcohol use and suggestions for future studies. Alcoholism: Clinical and Experimental Research, 38(2), 301–308. 10.1111/acer.12269 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Lord FM (1980). Applications of item response theory to practical testing problems. Lawrence Erlbaum. [Google Scholar]
  33. Luccasen RA III, & Thomas MK (2014). Monetary incentives versus class credit: Evidence from a large classroom trust experiment. Economics Letters, 123(2), 232–235. [Google Scholar]
  34. Mallett KA, Bachrach RL, & Turrisi R (2008). Are all negative consequences truly negative? Assessing variations among college students’ perceptions of alcohol related consequences. Addictive Behaviors, 33(10), 1375–1381. 10.1016/j.addbeh.2008.06.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Marsh HW, Hau K-T, & Wen Z (2004). In search of golden rules: Comment on hypothesis-testing approaches to setting cutoff values for fit indexes and dangers in overgeneralizing Hu and Bentler’s (1999) findings. Structural Equation Modeling, 11(3), 320–341. 10.1207/s15328007sem1103_2 [DOI] [Google Scholar]
  36. Meade AW, & Bauer DJ (2007). Power and precision in confirmatory factor analytic tests of measurement invariance. Structural Equation Modeling, 14(4), 611–635. 10.1080/10705510701575461 [DOI] [Google Scholar]
  37. Merrill JE, Fan P, Wray TB, & Miranda R Jr. (2020). Assessment of alcohol use and consequences: Comparison of data collected via timeline followback interview and daily reports. Journal of Studies on Alcohol and Drugs, 81(2), 212–219. 10.15288/jsad.2020.81.212 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Miller MB, Leffingwell T, Claborn K, Meier E, Walters S, & Neighbors C (2013). Personalized feedback interventions for college alcohol misuse: An update of Walters & Neighbors (2005). Psychology of Addictive Behaviors, 27(4), 909–920. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Muthén LK, & Muthén BO (1998–2017). Mplus user’s guide (8th ed.). [Google Scholar]
  40. O’Neill SE, & Sher KJ (2000). Physiological alcohol dependence symptoms in early adulthood: a longitudinal perspective. Experimental and Clinical Psychopharmacology, 8(4), 493–499. 10.1037/1064-1297.8.4.493 [DOI] [PubMed] [Google Scholar]
  41. Perkins HW (2002). Surveying the damage: A review of research on consequences of alcohol misuse in college populations. Journal of Studies on Alcohol, Supplement, (14), 91–100. 10.15288/jsas.2002.s14.91 [DOI] [PubMed] [Google Scholar]
  42. Rasch G (1960). Probabilistic models for some intelligence and attainment tests. Danmarks Paedagogiske Institut. [Google Scholar]
  43. Read JP, Kahler CW, Strong DR, & Colder CR (2006). Development and preliminary validation of the young adult alcohol consequences questionnaire. Journal of Studies on Alcohol, 67(1), 169–177. 10.15288/jsa.2006.67.169 [DOI] [PubMed] [Google Scholar]
  44. Read JP, Wardell JD, & Bachrach RL (2013). Drinking consequence types in the first college semester differentially predict drinking the following year. Addictive Behaviors, 38(1), 1464–1471. 10.15288/jsa.2006.67.169 [DOI] [PubMed] [Google Scholar]
  45. SAS Institute Inc. (2002–2014). SAS® 9.4 help and documentation. [Google Scholar]
  46. Schulenberg JE, Patrick ME, Johnston LD, O’Malley PM, Bachman JG, & Miech RA (2021). Monitoring the future national survey results on drug use, 1975–2020: Volume II, College students and adults ages 19–60. Institute for Social Research, The University of Michigan. [Google Scholar]
  47. Steinberg L, & Thissen D (2006). Using effect sizes for research reporting: Examples using item response theory to analyze differential item functioning. Psychological Methods, 11(4), 402–415. 10.1037/1082-989X.11.4.402 [DOI] [PubMed] [Google Scholar]
  48. Stone AL, Becker LG, Huber AM, & Catalano RF (2012). Review of risk and protective factors of substance use and problem use in emerging adulthood. Addictive behaviors, 37(7), 747–775. 10.1016/j.addbeh.2012.02.014 [DOI] [PubMed] [Google Scholar]
  49. Substance Abuse and Mental Health Services Administration. (2021). Key substance use and mental health indicators in the United States: Results from the 2020 National Survey on Drug Use and Health. Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration. https://www.samhsa.gov/data/ [Google Scholar]
  50. Thissen D, Steinberg L, & Kuang D (2002). Quick and easy implementation of the Benjamini–Hochberg procedure for controlling the false positive rate in multiple comparisons. Journal of Educational and Behavioral Statistics, 27(1), 77–83. 10.3102/10769986027001077 [DOI] [Google Scholar]
  51. United States Census Bureau. (2022). Education attainment in the United States: 2021. https://www.census.gov/data/tables/2021/demo/educational-attainment/cps-detailed-tables.html
  52. Vik PW, Carrello P, Tate SR, & Field C (2000). Progression of consequences among heavy-drinking college students. Psychology of Addictive Behaviors, 14(2), 91–101. 10.1037/0893-164X.14.2.91 [DOI] [PubMed] [Google Scholar]
  53. Wechsler H, Davenport A, Dowdall G, Moeykens B, & Castillo S (1994). Health and behavioral consequences of binge drinking in college. A national survey of students at 140 campuses. JAMA, 272(21), 1672–1677. 10.1001/jama.1994.03520210056032 [DOI] [PubMed] [Google Scholar]
  54. White HR, Labouvie EW, & Papadaratsakis V (2005). Changes in substance use during the transition to adulthood: A comparison of college students and their noncollege age peers. Journal of Drug Issues, 35(2), 281–306. 10.1177/002204260503500204 [DOI] [Google Scholar]
  55. White HR, & Ray AE (2014). Differential evaluations of alcohol-related consequences among emerging adults. Prevention Science, 15(1), 115–124. 10.1007/s11121-012-0360-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Woods CM, Cai L, & Wang M (2013). The Langer-improved Wald test for DIF testing with multiple groups: Evaluation and comparison to two-group IRT. Educational and Psychological Measurement, 73(3), 532–547. 10.1177/0013164412464875 [DOI] [Google Scholar]
  57. Zajacova A, Hummer RA, & Rogers RG (2012). Education and health among U.S. working-age adults: A detailed portrait across the full educational attainment spectrum. Biodemography and Social Biology, 58(1), 40–61. 10.1080/19485565.2012.666122 [DOI] [PubMed] [Google Scholar]

RESOURCES