Skip to main content
NCBI home page
International Journal of Methods in Psychiatric Research logoLink to International Journal of Methods in Psychiatric Research
. 2019 Mar 7;28(3):e1777. doi: 10.1002/mpr.1777

Protective Behavioral Strategies Scale‐20: Psychometric properties of a French and German version among young males in Switzerland

Véronique S Grazioli 1,✉, Joseph Studer 1, Mary E Larimer 2, Melissa A Lewis 3, Simon Marmet 1, Mélissa Lemoine 1, Jean‐Bernard Daeppen 1, Gerhard Gmel 4,5,6
PMCID: PMC6877153  PMID: 30848002

Abstract

Objective

The Protective Behavioral Strategies Scale (PBSS‐20) is one of the most commonly used measures of engagement in protective behavioral strategies (PBS). This research aimed to examine the psychometric properties of a French and German version of the PBSS‐20 in a large sample of young males in Switzerland.

Method

The sample included 5,017 young males (mean age = 25.44) participating in the Cohort Study on Substance Use Risk Factors in Switzerland. Measures of PBS use, total drinks per week, and alcohol‐related consequences were used from a second follow‐up assessment.

Results

Confirmatory factor analysis testing different models previously documented in the literature provided initial support for a four‐factor model. Fit statistics indicated that this model adequately reflects the structure of data. Further findings also provided support for adequate internal consistency and for convergent validity of this four‐factor model, whereas metric—but not scalar—measurement invariance across linguistic regions was demonstrated.

Conclusion

Although further research testing measurement invariance across linguistic regions and gender is warranted, results of the current study suggest that the French and German PBSS‐20 is reliable and that it may represent a promising research and clinical tool that can be used in both French‐ and German‐speaking countries.

Keywords: alcohol, French, German, protective behavioral strategies scale, young adults

1. INTRODUCTION

Young adults commonly engage in risky drinking behaviors, such as heavy episodic drinking (i.e., HED; reporting ≥60 g of pure alcohol on a single occasion; Wechsler, Davenport, Dowdall, Moeykens, & Castillo, 1994). In this population, prevalence rates of HED reach 32% in the past 2 weeks in the United States and 38% in the past month in Switzerland (Gmel, Kuendig, Notari, & Gmel, 2017; Johnston, O'Malley, Bachman, & Schulenberg, 2010). Importantly, among young adults, males have been consistently identified as being at greater risk of unhealthy alcohol use than females of the same age (Johnston, O'Malley, Bachman, Schulenberg, & Miech, 2014; Schulenberg et al., 2017). For instance, prevalence rates of HED in the past 30 days are twice as high among males aged 15–24 years compared with females of the same age (Delgrande Jordan & Notari, 2011). Of particular concern, these drinking behaviors have been consistently related to negative consequences, ranging in severity from health‐related consequences (e.g., vomiting) to engagement in risky behaviors (e.g., unprotected sexual activity), interpersonal violence, injuries, and death (Abbey, 2002; Hingson, 2010; Hingson, Heeren, Winter, & Wechsler, 2005; Hingson & White, 2012; Perkins, 2002; World Health Organization, 2014).

In response, tremendous research efforts have been dedicated to identifying factors that may alleviate alcohol‐related harm among young adults. This line of research has established that protective behavioral strategies (PBS) represent such a factor (Pearson, 2013). PBS are behavioral strategies that individuals can use before or while drinking to reduce alcohol‐related harm (e.g., avoid drinking games, use a designated driver; Martens et al., 2005; Martens, Pederson, Labrie, Ferrier, & Cimini, 2007). A substantial body of cross‐sectional research has established that the use of PBS is negatively related to alcohol use and alcohol‐related consequences among young adults (i.e., college students; Araas & Adams, 2008; Benton et al., 2004; Delva et al., 2004; Martens et al., 2005; Martens et al., 2007; Pearson, 2013). Although less commonly evaluated in the literature, longitudinal research has reached similar findings, with greater use of PBS being generally related to less alcohol use (Grazioli et al., 2015; Napper, Kenney, Lac, Lewis, & LaBrie, 2014) and fewer alcohol‐related problems over time (Grazioli et al., 2015; Grazioli, Lewis, et al., 2015; Luebbe, Varvel, & Dude, 2009; Napper et al., 2014). Further evidence for the association between the use of PBS and alcohol outcomes stems from clinical trials that found increase of PBS use following brief interventions (e.g., personalized mailed feedback) as a significant mechanism accounting for the intervention's effect on alcohol use (Barnett, Murphy, Colby, & Monti, 2007; Larimer et al., 2007), although results have not always been consistent across studies (Kulesza, Apperson, Larimer, & Copeland, 2010; Neighbors, Lee, Lewis, Fossos, & Walter, 2009; Reid & Carey, 2015).

PBS use may be especially important in young males because of their high rates of risky drinking behaviors. Unfortunately, past findings among young adults has revealed that males endorse less positive attitudes toward PBS (Demartini, Carey, Lao, & Luciano, 2011) and engage in PBS less often (Jongenelis et al., 2016; Walters, Roudsari, Vader, & Harris, 2007) than females. Further studying PBS use in males is therefore warranted.

Given the promising findings regarding PBS use effectiveness and considering that brief motivational interventions often comprise information on PBS (Cronce & Larimer, 2011), adequate measurement of engagement in PBS is critical. Most research on PBS has used the Protective Behavioral Strategies Scale (PBSS; Martens et al., 2005; Martens et al., 2007). PBSS psychometric properties have been evaluated in several studies conducted among college students in the United States. Table 1 presents a summary of different structures previously found in the literature. The PBSS was initially developed more than a decade ago in drinker college students in the United States. Participants are asked to indicate how often they engage in a series of behaviors when using alcohol or partying, using a 6‐point scale, where 1 = never and 6 = always. The authors conducted an exploratory factor analysis (EFA, N = 437; Martens et al., 2005) and a confirmatory factor analysis (CFA, N = 505; Martens et al., 2007). Findings revealed a three‐factor model to best fit the data, including, (a) Limiting/Stopping Drinking (LSD; seven items), (b) Manner of Drinking (MoD; five items), and (c) Serious Harm Reduction (SHR; three items). Fit indices indicated acceptable fit. Furthermore, the authors found evidence for convergent validity and good internal consistency except for the SHR subscale.

Table 1.

Summary of different structures found in past research evaluating the PBSS among college students

References Population Internal structure Fit indices
Subscale Items (n) Cronbach's alpha CFI RMSEA
PBSS, Martens et al., 2005 437 U.S. college student drinkers (≥1 drink in the past 30 months) Three‐factor model:
LSD 7 0.81 — —
MOD 5 0.73
SHR 3 0.63
PBSS, Martens et al., 2007 505 U.S. college student drinkers (≥1 drink in the past 30 months) Three‐factor modela: 0.91 0.07
LSD 7 0.82
MOD 5 0.74
SHR 3 0.59
Walters et al., 2007 288 U.S. college student heavy drinkers (≥1 HED in the past 2 weeks) Three‐factor model in females (n = 187):
LSD 7 — — —
MOD 5
SHR 3
Four‐factor model in males (n = 101)
LSD‐mixing 3
LSD‐Planned 4
MOD 5
SHR 3
PBSS‐R, Madson, Arnau, & Lambert, 2013 747 U.S. college student drinkers (≥1 drink in the past month) Two‐factor modelb: 0.98 0.06
Controlled consumption 12 0.90
6 0.79
SHR
PBSS‐20, Treloar, Martens, & McCarthy, 2015 603 college student drinkers (≥1 drink in the past year) Three‐factor model: 0.93 0.099
LSD 7 0.87
MOD 5 0.83
SHR 8 0.86
Open in a new tab

Note. CFI: comparative fit index; HED: heavy episodic drinking; LSD: Limiting/Stopping Drinking; MOD: Manner of Drinking; PBSS: Protective Behavioral Strategies Scale; RMSEA: root mean square error of approximation; SHR: Serious Harm Reduction.

a

The error terms for two items were freely estimated (leave the party at a predetermined time and stop drinking at a predetermined time).

b

The error terms of two pairs of items were freely estimated (drink water while drinking alcoholic drinks and alternate alcoholic and nonalcohol drinks; Drink slowly rather than gulp or chug and avoid trying to keep up our out‐drink others).

The PBSS's internal structure has been further evaluated in three other studies conducted in college students. In the first study (N = 288), the authors explored the factor structure of the PBSS using EFA and found that a three‐factor model best fit the data for females, whereas a four‐factor model best fit data for males (Walters et al., 2007). Specifically, the LSD factor was split into two factors, (a) LSD‐Mixing nonalcoholic drinks with alcohol, and (b) LSD‐Planned limits on drinking. Notably, Walters et al. (2007) did not conduct CFAs, limiting the ability to determine which model had the best fit. A more recent study (N = 747) tested the fit of different factor structures using CFA (Madson et al., 2013). Findings showed that a two‐factor model best fit the data (referred by the authors as to PBSS‐R; Madson et al., 2013). Fit indices for this two‐factor model indicated a good fit. The authors also showed that the two‐factor structure was invariant across White non‐Hispanic and African American males and females (Madson et al., 2013). Finally, Treloar et al. (2015) recently conducted a study aiming at improving the content validity of the SHR subscale, using both EFA and CFA (N = 603). Analysis resulted in the PBSS‐20, which comprises five additional items in the SHR subscale (e.g., avoid combining alcohol with marijuana) and a substitution in the MoD subscale (i.e., drink shots of liquor replaced by avoid pregaming previously included in the SHR subscale). Internal consistency was improved after this revision. That said, fit indices were not optimal, and findings showed a lack of scalar measurement invariance across gender (Treloar et al., 2015).

Taken together, these findings indicate that research further evaluating PBSS psychometric properties is warranted. Furthermore, to the best of the authors' knowledge, the PBSS has not been evaluated in other countries than the United States and in other languages than English. Therefore, research expanding PBS measurement to different cultures and diverse populations is needed (Pearson, 2013). Doing so is important to enable international studies on PBS. Finally, given that brief intervention often provides information on PBS (Cronce & Larimer, 2011), developing a PBSS in other countries and languages may represent an important first step prior to implementing PBS‐brief intervention outside the United States. In response, the current study aimed to test the psychometric properties of a French and German version of the PBSS among young males in Switzerland.

Specifically, this study aimed to test the fit of the previously reported factor structures of the PBSS (i.e., two‐, three‐ and four‐factor models) and to explore measurement invariance across linguistic regions (i.e., French‐ and German‐speaking) using CFA in a representative sample of young males in Switzerland. Further aims include the evaluation of the abovementioned factor structures' internal consistency and convergent validity across linguistic regions. To stay consistent with most past research using the PBSS, we decided to translate items included in the most recent version of the PBSS (i.e., PBSS‐20; Treloar et al., 2015).

2. METHODS

2.1. Study design

This study used data from the Cohort Study on Substance Use Risk Factors (C‐SURF) designed to examine substance‐use trajectories among young males in Switzerland. Recruitment took place between August 2010 and July 2011 in three of a total of six recruitment centers covering 21 of the 26 Swiss cantons (including French‐ and German‐speaking cantons). In Switzerland, all Swiss males aged around 19 years must undergo a recruitment procedure to determine their eligibility for military service. Thus, virtually all males aged 19–20 years in the 21 covered cantons were eligible for inclusion in the study and were asked consent for participation. Inclusion in the study was independent of whether they were deemed eligible or not to serve in the army. Participants completed questionnaires outside of the army environment. At baseline, conscripts who gave consent to participate received a mail or an email within 2 weeks to complete a paper pencil questionnaire or an online questionnaire. Next, participants were invited by email (or by mail if they preferred) to complete the follow‐up assessments using an online questionnaire (or a paper pencil questionnaire if they asked so; 45–60 min duration on average). Participants received a voucher for completion of each survey (i.e., CHF 30—around $30 for the baseline and for the follow‐up 1 and CHF 50—around $50 for the follow‐up 2 assessment). More details regarding the parent study procedures are provided elsewhere (Gmel et al., 2015). All procedures were approved by the institutional ethical review board at the home institution (research protocol number 15/07).

2.2. Participants

A total of 7,556 conscripts provided written informed consent to participate in the study. Of those, 5,987 (79.2%) completed the baseline assessment between September 2010 and March 2012, and 5,435 completed the second follow‐up assessment between April 2016 and March 2018. Nonresponse analysis at baseline showed that nonrespondents were more likely to be at‐risk drinkers than respondents. These differences were commonly small, and significance was due to the large sample size, indicating a small nonresponse bias (Studer et al., 2013). For instance, nonrespondents (9.5%) were more often abstainers than respondents (9.2%), but the difference was not significant (OR = 1.04, 95% CI [0.91, 1.17]), whereas nonrespondents reported HED at least monthly significantly more often than respondents (49.4% vs. 45.1%, OR = 1.19, 95% CI [1.10, 1.28]).

In line with Martens et al. (2007), abstainers (reporting no drink in the past year; n = 361, 6.6%) were not included in the current study. The sample included 138 partial completers (i.e., missing values on key variables, 2.5%). Attrition tests indicated that there was no significant difference between completers (participants with complete key variables, N = 5,017) and partial completers (participants removed from the sample, n = 138) on total drinks per week, age, and on linguistic region (all ps > 0.1), suggesting that data were missing completely at random. Further tests indicated that, among partial completers on the PBSS, there was no significant difference in percentages of missing data by PBS item across linguistic regions (all ps > 0.2). Given the results of the attrition tests and the low percentage of missing data (2.5%, < 5%; Bennett, 2001), missing data were listwise deleted, resulting in a final sample of 5,017 participants. The mean age of participants was 25.44 (SD = 1.24). More than half of the sample was French‐speaking (57%). Primary school (i.e., obligatory school, 8–10 years; 2.90%) was the least commonly reported highest level of education completed, followed by secondary school (obligatory school plus basic apprenticeship or vocational school, 12 years; 39.20%) and tertiary school (vocational school diploma, high school diploma or bachelor; 57.9%). The PBS measure was only assessed in the second follow‐up assessment. Therefore, in the current study, we used data from the second follow‐up exclusively.

2.3. Measures

Measures in the current study were part of a larger assessment battery and are described below. The PBSS was at the end of the fourth section of the questionnaire (out of 10 sections) focusing on alcohol use. The complete questionnaire can be downloaded on C‐SURF's website—http://www.c-surf.ch/en/30.html.

2.3.1. Sociodemographic variables

Age, linguistic region (i.e., German‐ or French‐speaking), and highest achieved education served to describe the sample and as covariates in the analyses.

2.3.2. Protective Behavioral Strategies Scale

The most recent version of the PBSS (PBSS‐20; Treloar et al., 2015) was translated into French and German by an expert committee (G. G., V. G., J. S., and the research coordinator of the study). For both versions (i.e., French and German), initial translation was completed independently by two translators fluent in both the original and the target languages and familiar with both cultures. One of the translator had expertise in the domain being assessed (i.e., the content of the scale and the intention of each item), whereas the other translator was not an expert in the field, ensuring a more “naïve” translation. Next, independent translations were discussed in meetings wherein any discrepancies were resolved until reaching a consensus. Both translations are provided as supplementary material (Table S1).

2.3.3. Alcohol outcomes

Alcohol use was assessed with a quantity–frequency measure. Specifically, the average number of drinking days and the number of standard drinks consumed per drinking day over the past 12 months were measured. The average number of drinks per week (referred to as total drinks per week hereafter) over the past 12 months was computed by multiplying the number of drinking days by the number of standard drinks (i.e., a standard drink = 10 g of ethanol) per drinking days (Gmel et al., 2014). The number of alcohol‐related consequences was assessed with nine items adapted from Knight and colleagues (Knight et al., 2002; Wechsler et al., 1994). Items included, for example, getting in trouble with the police, having unplanned sex, and having an argument or a fight. Participants were asked to indicate whether they had experienced these alcohol‐related consequences over the past 12 months (coded 0 = never, 1 = at least once). Both total drinks per week and alcohol‐related consequences served as dependent variables in the analyses. Furthermore, total drinks per week served as a covariate in the convergent analysis testing the association between PBS use and alcohol‐related consequences.

2.4. Statistical analyses

2.4.1. Internal structure analyses

CFAs were conducted to examine the fit of the factorial structures of the PBSS. Building on past research, we tested three model structures: (a) a three‐factor model as proposed by Treloar et al. (2015) and Martens et al. (2007); (b) a two‐factor model with a factor including items of SHR and another factor comprising items of MoD and LSD (Madson et al., 2013); and (c) a four‐factor model inspired from Walters et al. (2007), wherein items of LSD were split into two factors—LSD: Planned limits on drinking (LSD‐Pl) and LSD: Mixing nonalcoholic drinks with alcohol (LSD‐M). To stay consistent with most past research using the PBSS, evaluating PBSS psychometric properties using CFA, we kept the PBSS‐20 overall structure in this third model as well (i.e., MoD, SHR, and LSD items' belonging). Thus, in clear, the only difference between model (a) and model (c) was that LSD subscale was split into two factors following Walters et al.'s (2007) findings (see Table 2). In all models tested, factors were allowed to correlate.

Table 2.

Item means, standard deviations, corrected item–subscale correlations, and standardized factor loadings for PBS items by model and by linguistic regions

PBS items Model 1 (three‐factor) Model 2 (two factors) Model 3 (four factors)
Total French German French German French German French German
M SD M SD M SD r t b r t b r t b r t b r t b r t b
1. Use a designated driver 3.73 1.94 3.86 1.89 3.56 1.99 0.44 0.51 0.41 0.50 0.44 0.51 0.41 0.50 0.44 0.51 0.41 0.49
2. Determine not to exceed a set number of drinks 2.80 1.66 2.94 1.69 2.61 1.61 0.55 0.69 0.55 0.68 0.54 0.65 0.54 0.64 0.54 0.72 0.54 0.74
3. Alternate alcoholic and nonalcoholic drinks 3.48 1.59 3.55 1.58 3.38 1.60 0.56 0.59 0.53 0.58 0.54 0.55 0.52 0.54 0.57 0.59 0.54 0.57
4. Have a friend let you know when you have had enough to drink 1.59 1.13 1.62 1.16 1.56 1.09 0.45 0.61 0.44 0.60 0.42 0.58 0.40 0.57 0.43 0.64 0.48 0.65
5. Avoid drinking games 2.99 1.87 3.18 1.89 2.72 1.81 0.53 0.70 0.52 0.70 0.54 0.67 0.54 0.68 0.53 0.70 0.52 0.70
6. Only go out with people you know and trust 3.78 1.85 3.90 1.80 3.62 1.90 0.59 0.69 0.58 0.70 0.59 0.69 0.58 0.70 0.59 0.69 0.58 0.70
7. Leave the bar/party at predetermined time 2.46 1.47 2.59 1.49 2.30 1.42 0.54 0.70 0.55 0.69 0.51 0.66 0.52 0.65 0.57 0.73 0.59 0.75
8. Make sure that you go home with a friend 2.68 1.70 2.50 1.64 2.93 1.76 0.50 0.68 0.54 0.66 0.50 0.68 0.54 0.66 0.50 0.68 0.54 0.66
9. Know where your drink has been at all times 3.79 1.98 3.52 1.97 4.15 1.92 0.52 0.61 0.63 0.70 0.52 0.61 0.63 0.70 0.52 0.61 0.63 0.70
10. Avoid combining alcohol with marijuana 4.29 2.09 4.45 2.02 4.09 2.15 0.47 0.62 0.58 0.72 0.47 0.62 0.58 0.72 0.47 0.62 0.58 0.72
11. Avoid “pre‐gaming” 2.75 1.71 2.66 1.68 2.87 1.74 0.54 0.72 0.53 0.68 0.53 0.68 0.51 0.66 0.54 0.72 0.53 0.68
12. Refuse to ride in a car with someone who has been drinking 4.12 1.93 4.12 1.84 4.13 2.04 0.49 0.61 0.50 0.63 0.49 0.61 0.50 0.63 0.49 0.60 0.50 0.63
13. Stop drinking at predetermined time 2.34 1.51 2.54 1.56 2.07 1.38 0.61 0.78 0.57 0.74 0.60 0.73 0.56 0.69 0.64 0.82 0.60 0.80
14. Make sure you drink with people who can take care of you if you drink too much 3.25 1.86 3.20 1.81 3.32 1.92 0.59 0.74 0.57 0.72 0.59 0.74 0.57 0.72 0.59 0.74 0.57 0.72
15. Drink water while drinking alcohol 4.00 1.63 3.94 1.62 4.08 1.63 0.48 0.55 0.45 0.61 0.49 0.51 0.49 0.57 0.56 0.55 0.55 0.60
16. Put extra ice in your drink 2.82 1.64 2.79 1.64 2.86 1.64 0.34 0.51 0.33 0.52 0.35 0.48 0.38 0.49 0.27 0.52 0.30 0.51
17. Eat before or during drinking 4.47 1.34 4.48 1.29 4.44 1.40 0.42 0.56 0.49 0.60 0.42 0.56 0.49 0.60 0.42 0.56 0.49 0.60
18. Avoid mixing different types of alcohol 3.35 1.64 3.46 1.64 3.19 1.63 0.55 0.69 0.60 0.68 0.55 0.65 0.55 0.66 0.55 0.69 0.60 0.68
19. Drink slowly, rather than gulp or chug 3.76 1.59 3.93 1.58 3.53 1.57 0.61 0.69 0.66 0.76 0.56 0.65 0.64 0.72 0.61 0.70 0.66 0.76
20. Avoid trying to keep up or out‐drink others 3.50 1.76 3.90 1.75 2.98 1.64 0.60 0.68 0.66 0.76 0.55 0.64 0.65 0.72 0.60 0.68 0.66 0.76
Open in a new tab

Note. French: French‐speaking participants. German: German‐speaking participants. r t = corrected item–subscale correlation. Model 1: three‐factor model following Treloar et al. (2015). Model 2: two‐factor model following Madson et al. (2013). (i.e., Factor 1 = Serious Harm Reduction; Factor 2 = Manner of drinking and Stopping/Limiting). Model 3: four‐factor model inspired by Walters et al. (2007) (Factor 1 = Serious Harm Reduction; Factor 2 = Manner of drinking; Factor 3 = Stopping/Limiting‐Planned limits on drinking; Factor 4 = Limiting/Stopping Drinking‐Mixing nonalcoholic drinks with alcohol).

graphic file with name MPR-28-e1777-g001.jpg

Model fit was examined with the root mean square error of approximation (RMSEA) and the comparative fit index (CFI). RMSEA values close to 0.06 or lower are generally considered as showing a good fit, although according to some authors, values <0.08 indicate a fair fit (Hoyle, 2012). For CFI, values >0.90 indicate a fair fit. CFAs were also conducted to test configural, metric, and scalar model invariance across linguistic regions (i.e., French‐ and German‐speaking participants). We conducted a hierarchical approach by constraining model parameters and comparing models' fit across models (Vandenberg & Lance, 2000). Three models were successively conducted: first, a baseline model where all parameters were freely estimated (unconstrained) across linguistic regions (i.e., configural model); a second model where factor loadings were constrained to be equal (i.e., metric model); and a third model where thresholds of each item and factor loadings were constrained to be equal (i.e., scalar model). Multigroup models' fit was examined with CFI and RMSEA. The fit's differences between the three models (i.e., configural with metric; metric with scalar) were used to examine invariance across linguistic regions. Although it is common to use differences in the χ 2 values to examine invariance, differences in χ 2 statistics are sensitive to large sample size (i.e., N > 300; Chen, 2007). Accordingly, we used a combination of changes in CFI and RMSEA (|∆CFI| and |∆RMSEA|) between the unconstrained and the constrained models to evaluate invariance. The null hypothesis of invariance should not be rejected when |∆CFI| and |∆RMSEA| equal or do not exceed 0.010 and 0.015, respectively (Chen, 2007).

2.4.2. Internal consistency and convergent validity

We tested internal consistency with Cronbach alphas and finally convergent validity as follows for each model and by linguistic regions (i.e., total sample and French‐ and German‐speaking participants). We first computed correlations between PBSS scores and measures of alcohol use and related consequences. Then, multiple regression analyses were conducted to test for the association between PBSS scores and alcohol use and related consequences, after controlling for demographic factors, and for alcohol use in model testing alcohol‐related consequences. Inspection of total drinks per week showed the presence of outliers. Therefore, total drinks per week was winsorized using a tail of 0.01. Specifically, values under the first or over the 99th percentile were replaced with the first or 99th percentile, respectively. Winsorized total drinks per week was used in the convergent validity analyses to limit the influence of outliers (Tukey, 1962). Next, total drinks per week (skewness = 2.91, kurtosis = 14.07) and alcohol‐related consequences (skewness = 1.62, kurtosis = 2.62) showed positively skewed distributions approximating a negative binomial distribution with the exception of a disproportionately large number of zero values for alcohol‐related consequences (50.4%). A zero‐inflated binomial regression (ZINB) was therefore selected to examine the association between PBS and alcohol‐related consequences, whereas the association between PBS and total drinks per week was examined with a negative binomial regression (Atkins & Gallop, 2007; Hilbe, 2007). The zero‐inflated count model allows modeling simultaneously two dimensions of a variable. The logistic portion examines the likelihood of the observation being an excess of zero value (predicting the excess zeros that would be expected in a negative binomial distribution). The second portion examines the count portion of the model. In the ZINBs, the magnitude of the association between the independent variable and the outcomes were examined with odds ratios (for the ZINB logistic submodel) and incident rate‐ratios. Both logistic and count portions of the ZINB models are reported in the results; however, only count portions were considered in the discussion, given that use of PBS has been commonly related to fewer alcohol outcomes but not to no alcohol outcome at all. Incident rate‐ratios describe the proportion increase (>1) or decrease (<1) in outcomes for each unit increase in the covariate (Atkins, Baldwin, Zheng, Gallop, & Neighbors, 2013; Atkins & Gallop, 2007). Mplus 8 was used for CFAs, SPSS 23 for descriptive statistics, and STATA 14 for convergent validity analyses.

3. RESULTS

Participants reported a mean of 7.72 drinks per week (SD = 9.40, median = 4.5) and a mean of 1.10 alcohol‐related consequences (SD = 1.51, median = 0.00). Table 3 presents means and standard deviations of PBS scores for the whole sample and by linguistic regions. Paired sample t tests indicated that there were significant differences between PBS MoD (mean = 3.27) and LSD (mean = 2.78) scores, t = 33.35, p < 0.001; similarly, PBS SHR score (mean = 3.77) was significantly higher than PBS LSD score, t = −69.72, p < 0.001. Finally, PBS SHR score (mean = 3.77) was higher than PBS MoD score, t = −33.02, p < 0.001.

Table 3.

Means and standard deviations of PBS scores and Spearman rank‐order correlations between PBS total scale, PBS subscales, and alcohol use and alcohol‐related consequences by linguistic regions

Total SHR (M1,2,3) MoD (M1,3) MoD2 (M2) LSD (M1) LSD: Pl (M3) LSD: M (M3)
Mean (SD) t a
Total 3.30 (0.98) 4.86*** 3.77 (1.20) −0.75 3.27 (1.27) 10.21*** 2.99 (0.99) 8.74*** 2.78 (0.99) 5.66*** 3.00 (1.10) 9.41*** 3.43 (1.24) −0.47
French 3.36 (0.97) 3.75 (1.15) 3.43 (1.25) 3.09 (0.99) 2.85 (1.00) 3.43 (1.23) 3.43 (1.23)
German 3.22 (1.00) 3.78 (1.26) 3.06 (1.25) 2.85 (0.97) 2.70 (0.95) 2.13 (1.05) 3.44 (1.24)
Total Spearman rank‐order correlations
Drinks/week −0.15*** −0.11*** −0.16*** −0.14*** −0.08*** −0.08*** −0.05**
Consequences −0.13*** −0.10*** −0.18*** −0.12*** −0.03 −0.02 −0.03
French
Drinks/week −0.19*** −0.15*** −0.21*** −0.19*** −0.12*** −0.05** −0.14***
Consequences −0.17*** −0.16*** −0.21*** −0.16*** −0.06** −0.05** −0.06 *
German
Drinks/week −0.09*** −0.05 * −0.11*** −0.09*** −0.03 −0.01 −0.03
Consequences −0.06** −0.04 * −0.14*** −0.07** 0.03 0.05 * −0.01
Open in a new tab

Note. LSD: Limiting/Stopping Drinking; LSD‐Pl: LSD‐Planned limits on drinking; LSD‐M: LSD‐Mixing nonalcoholic drinks with alcohol; MoD: Manner of Drinking; SHR: Serious Harm Reduction.

a

t tests comparing PBS means across linguistic regions.

*

p < 0.05.

**

p < 0.01.

***

p < 0.001.

3.1. Internal structure analyses

For CFAs, the weighted least square mean and variance adjusted estimator was used in order to take into account the ordinal nature of the PBS indicators (Muthén & Muthén, 1998). Table 4 presents fit statistics for the different models tested. Consistent with Madson et al. (2013), inspection of modification indices indicated that two pairs of items had high correlated error terms, item 19 (i.e., drink slowly, rather than gulp or chug) with item 20 (i.e., avoid trying to keep up or out‐drink others), and item 3 (i.e., alternate alcoholic and nonalcoholic drinks) with item 15 (i.e., drink water while drinking alcohol). On the basis of the content of both pairs of items, correlating the error terms was considered as defensible given that in each pair, items measure very similar behaviors. As shown in Table 4, after freeing the error term covariance, the fit was still poor for the two‐factor and the three‐factor models, whereas the four‐factor model showed a better fit (CFI = 0.922, RMSEA = 0.079). Means, SD, and standardized factor loadings by models and linguistic regions are presented in Table 2.

Table 4.

Fit statistics: Confirmatory factor models of the French and German PBSS by models

Models Fit indices
CFI RMSEA
Model 1a 0.870 0.100
Model 2a 0.831 0.114
Model 3a 0.905 0.087
Model 1b 0.908 0.085
Model 2b 0.890 0.093
Model 3b 0.922 0.079
Open in a new tab

Note. CFI: comparative fit index; RMSEA: root‐mean‐square error of approximation. Model 1a: three‐factor model following Treloar et al. (2015). Model 2a: two‐factor model following Madson et al. ( 2013) (i.e., Factor 1 = Serious Harm Reduction; Factor 2 = Manner of drinking and Stopping/Limiting). Model 3a: four‐factor model inspired by Walters et al. (2007) (Factor 1 = Serious Harm Reduction; Factor 2 = Manner of drinking; Factor 3 = Stopping/Limiting‐Planned limits on drinking; Factor 4 = Limiting/Stopping Drinking‐Mixing nonalcoholic drinks with alcohol). Models 1b, 2b, and 3b: models with freed error covariance for item pairs 3–15 and 19–20.

3.2. Multigroup analyses testing invariance by linguistic region

Table 5 presents multigroup analyses (French‐ and German‐speaking participants) results by model. Results indicated that |∆CFI| between the configural models and the metric models constraining equal factor loadings between the two groups did not exceed 0.010 (|∆CFI| ranging from 0.002 to 0.004). Similarly, |∆RMSEA| between the configural model and the metric model was below 0.015 (|∆RMSEA| ranging from 0.000 to 0001). Thus, in line with Chen (2007), these findings indicate that the null hypothesis of metric invariance should not be rejected and that the four‐factor PBSS assesses similar underlying factors across German‐ and French‐speaking participants.

Table 5.

Multigroup analysis results by models

Models Configural models Metric models ∆ Scalar models ∆
CFI RMSEA CFI RMSEA ∆CFI ∆RMSEA CFI RMSEA ∆CFI ∆RMSEA
Model 1 0.911 0.085 0.907 0.085 0.004 0.000 0.889 0.084 0.018 0.001
Model 2 0.893 0.093 0.891 0.092 0.002 0.001 0.877 0.088 0.004 0.004
Model 3 0.924 0.080 0.921 0.079 0.003 0.001 0.900 0.080 0.021 0.001
Open in a new tab

Note. CFI: comparative fit index; RMSEA: root‐mean‐square error of approximation. Model 1: three‐factor model following Treloar et al. (2015). Model 2: two‐factor model following Madson et al. (2013) (i.e., Factor 1 = Serious Harm Reduction; Factor 2 = Manner of drinking and Stopping/Limiting). Model 3: four‐factor model inspired by Walters et al. (2007) (Factor 1 = Serious Harm Reduction; Factor 2 = Manner of drinking; Factor 3 = Stopping/Limiting‐Planned limits on drinking; Factor 4 = Limiting/Stopping Drinking‐Mixing nonalcoholic drinks with alcohol).

Next, whereas |∆RMSEA| between the metric and scalar model was below 0.015 (|∆RMSEA| ranging from 0.001 to 0.004), |∆CFI| between the metric and the scalar model exceeded 0.010 for the first and the third models (|∆CFI| = 0.018 and 0.021), indicating that the null hypothesis of scalar invariance should be rejected for these two models. In contrast, for Model 2, |∆CFI| between the metric and the scalar model did not exceed 0.010 (|∆CFI| = 0.004).

3.3. Reliability and convergent validity

Item–subscales correlations are presented in Table 2. Cronbach's alphas were computed for the total sample and by linguistic regions. For total scores, Cronbach's alpha coefficients ranged from 0.89 to 0.90; similarly, they ranged from 0.79 to 0.85 for the SHR and the two MoD subscales. Regarding LSD subscales, Cronbach's alphas ranged from 0.74 to 0.75 for both LSD subscale and LSD‐Pl. For LSD‐M, they ranged from 0.64 to 0.65. Alpha coefficients indicate an adequate internal consistency for all subscales except for LSD‐M subscale, which was slightly below the most cited standard in the literature (i.e., 0.70; Nunnally & Bernstein, 1994). The low consistency may be attributed to the small number of items included in this subscale. Mean inter‐item correlation (=0.38) indicated however a good internal consistency (Briggs & Cheek, 1986).

Next, correlation and regression analyses were conducted to test convergent validity. Table 3 presents correlations between the French and German PBSS‐20 scores by models and linguistic regions with total drinks per week and alcohol‐related consequences. For total scores, SHR and MoD subscales correlations were statistically significant and in the expected direction in both French‐ and German‐speaking participants. Next, whereas correlations between LSD subscales and alcohol outcomes were significant and in the expected direction in French‐speaking participants, they did not reach significance in German‐speaking participants (except for the correlation between LSD‐Pl subscale and alcohol‐related consequences, which was positive).

Table 6 presents regression analysis results after controlling for demographic factors, and for alcohol use in models testing alcohol‐related consequences. All subscales from each model were related to total drinks per week in the expected direction in both French‐ and German‐speaking participants. Regarding alcohol‐related consequences, total score, SHR, and MoD subscales were negatively related to consequences in both samples, except for MoD from Model 2, which did not reach significance in German‐speaking participants. Next, the association between LSD subscale (Model 1) and consequences was not significant in both sample, whereas LSD‐M was significantly related to fewer consequences in both samples. Finally, consistent with correlations findings, LSD‐Pl was positively related to consequences in German‐speaking participants.

Table 6.

Negative binomial and zero‐inflated regression models examining the associations between alcohol outcomes and the PBSS scores by linguistic regions

PBS scores Total sample French‐speaking participants German‐speaking participants
b SE b 95% CI IRR/OR b SE b 95% CI IRR/OR b SE b 95% CI IRR/OR
Total drinks/week models Negative binomial regression models
PBS total score −0.20 *** 0.02 [−0.23, −0.16] 0.82 −0.26 *** 0.02 [−0.30, −0.21] 0.77 −0.13 *** 0.03 [−0.18, −0.07] 0.88
Model 11
Serious Harm Reduction −0.10 *** 0.01 [−0.13, −0.07] 0.90 −0.15 *** 0.02 [−0.19, −0.11] 0.86 −0.05 ** 0.02 [−0.12, −0.03] 0.95
Manner of Drinking −0.19 *** 0.01 [−0.21, −0.16] 0.83 −0.22 *** 0.02 [−0.26, −0.19] 0.80 −0.15 *** 0.02 [−0.19, −0.11] 0.87
Limiting–Stopping (LS) −0.14 *** 0.02 [−0.17, −0.11] 0.87 −0.18 *** 0.02 [−0.22, −0.13] 0.84 −0.09 ** 0.03 [−0.14, −0.04] 0.91
Model 22
Serious Harm Reduction −0.10 *** 0.01 [−0.13, −0.07] 0.90 −0.15 *** 0.02 [−0.19, −0.11] 0.86 −0.05 ** 0.02 [−0.20, −0.03] 0.95
Manner of Drinking −0.22 *** 0.02 [−0.25, −0.18] 0.80 −0.27 *** 0.02 [−0.32, −0.22] 0.77 −0.16 *** 0.03 [−0.21, −0.12] 0.85
Model 33
Serious Harm Reduction −0.10 *** 0.01 [−0.13, −0.07] 0.90 −0.15 *** 0.02 [−0.19, −0.11] 0.86 −0.05 ** 0.02 [−0.20, −0.32] 0.95
Manner of Drinking −0.19 *** 0.01 [−0.21, −0.16] 0.83 −0.22 *** 0.02 [−0.26, −0.19] 0.80 −0.15 *** 0.02 [−0.19, −0.11] 0.87
LS: Planned limits −0.14 *** 0.02 [−0.17, 0.11] 0.87 −0.19 *** 0.02 [−0.23, −0.15] 0.83 −0.08 *** 0.02 [−0.13, −0.04] 0.92
LS: Mixing −0.05 *** 0.01 [−0.08, −0.02] 0.95 −0.06 ** 0.02 [−0.09, −0.02] 0.95 −0.04 ** 0.02 [−0.08, −0.01] 0.96
Consequences models
PBS total score Zero‐inflated negative binomial regression models: logistic portion of the models
−0.12 0.06 [−0.25, 0.00] 0.89 −0.07 0.09 [−0.25, 0.10] 0.93 −0.17 0.09 [−0.34, 0.01] 0.84
Zero‐inflated negative binomial regression models: count portion of the models
−0.14 *** 0.02 [−0.19, −0.09] 0.87 −0.18 *** 0.03 [−0.25, −0.11] 0.83 −0.08 * 0.03 [−0.15, −0.02] 0.92
Model 11 Zero‐inflated negative binomial regression models: logistic portion of the models
Serious Harm Reduction −0.09 0.05 [−0.19, 0.01] 0.91 −0.01 0.07 [−0.16, 0.13] 0.99 −0.15 * 0.07 [−0.29, −0.01] 0.86
Manner of Drinking −0.03 0.05 [−0.14, 0.08] 0.97 −0.02 0.08 [−0.17, 0.13] 0.98 −0.04 0.07 [−0.19, 0.11] 0.96
Limiting–Stopping (LS) −0.11 0.06 [−0.23, 0.00] 0.90 −0.09 0.08 [−0.24, 0.07] 0.91 −0.16 0.09 [−0.34, 0.02] 0.85
Model11 Zero‐inflated negative binomial regression models: count portion of the models
Serious Harm Reduction −0.10 *** 0.02 [−0.14, −0.06] 0.90 −0.13 *** 0.03 [−0.19, −0.08] 0.88 −0.06 * 0.02 [−0.11, −0.01] 0.94
Manner of Drinking −0.15 *** 0.02 [−0.19, −0.11] 0.86 −0.18 *** 0.03 [−0.23, −0.12] 0.84 −0.12 *** 0.03 [−0.17, −0.07] 0.89
Limiting–Stopping (LS) −0.02 0.02 [−0.06, 0.03] 0.98 −0.05 0.03 [−0.11, 0.01] 0.95 0.02 0.03 [−0.04, 0.09] 1.02
Model 22 Zero‐inflated negative binomial regression models: logistic portion of the models
Serious Harm Reduction −0.09 0.05 [−0.19, 0.01] 0.91 −0.01 0.07 [−0.16, 0.13] 0.99 −0.15 * 0.07 [−0.29, −0.01] 0.86
Manner of Drinking −0.09 0.06 [−0.21, 0.04] 0.91 −0.07 0.09 [−0.24, 0.10] 0.93 −0.11 0.09 [−0.30, 0.07] 0.90
Zero‐inflated negative binomial regression models: count portion of the models
Serious Harm Reduction −0.10 *** 0.02 [−0.14, −0.06] 0.90 −0.13 *** 0.03 [−0.19, −0.08] 0.88 −0.06 * 0.02 [−0.11, −0.01] 0.94
Manner of Drinking −0.11 *** 0.02 [−0.16, −0.07] 0.89 −0.15 *** 0.03 [−0.22, −0.09] 0.86 −0.06 0.03 [−0.13, 0.00] 0.94
Model 33 Zero‐inflated negative binomial regression models: Logistic portion of the models
Serious Harm Reduction −0.09 0.05 [−0.19, 0.01] 0.91 −0.01 0.07 [−0.16, 0.13] 0.99 −0.15 * 0.07 [−0.29, −0.01] 0.86
Manner of Drinking −0.03 0.05 [−0.14, 0.08] 0.97 −0.02 0.08 [−0.17, 0.13] 0.98 −0.04 0.07 [−0.19, 0.11] 0.96
LS: Planned limitsa −0.01 0.05 [−0.12, 0.09] 0.99 −0.03 0.07 [−0.17, 0.11] 0.97 −0.01 0.08 [−0.16, 0.15] 0.99
LS: Mixingb −0.15 ** 0.05 [−0.24, −0.06] 0.86 −0.10 0.06 [−0.23, 0.03] 0.90 −0.22 ** 0.07 [−0.36, −0.09] 0.80
Zero‐inflated negative binomial regression models: count portion of the models
Serious Harm Reduction −0.10 *** 0.02 [−0.14, −0.06] 0.90 −0.13 *** 0.03 [−0.19, −0.08] 0.88 −0.06 * 0.02 [−0.11, −0.01] 0.94
Manner of Drinking −0.15 *** 0.02 [−0.19, −0.11] 0.86 −0.18 *** 0.03 [−0.23, −0.12] 0.84 −0.12 *** 0.03 [−0.17, −0.07] 0.89
LS: Planned limits 0.03 0.02 [−0.01, 0.07] 0.98 −0.02 0.03 [−0.08, 0.03] 0.98 0.09 ** 0.03 [0.03, 0.14] 0.90
LS: Mixing −0.05 ** 0.02 [−0.08, −0.02] 0.95 −0.05 ** 0.02 [−0.10, −0.01] 0.95 −0.05 * 0.02 [−0.10, −0.01] 0.95
Open in a new tab

Note. OR = odd ratios are presented for the logistic portion and IRR = incident rate ratios are presented for the count portion of the ZINB models and for the negative binomial regression models. All models were adjusted for age, linguistic region and education and models testing alcohol‐related consequences were also adjusted for total drinks per week. Bold coefficients are statistically significant (p > .05).

a

Limiting/Stopping Drinking: Planned limits on drinking.

b

Limiting/Stopping Drinking: Mixing nonalcoholic drinks with alcohol.

1

Model 1 = 3‐Factor model following Treloar al., 2015.

2

Model 2 = 2‐Factor model following Madson et al., 2013. (i.e., factor 1 = Serious Harm Reduction, factor 2 = Manner of drinking and Stopping/Limiting).

3

Model 3 = 4‐Factor model inspired by Walters et al; et al., 2007, (1 = Serious Harm Reduction, factor 2 = Manner of drinking, factor 3 Stopping/Limiting‐Planned limits on drinking, factor 4 = Limiting/Stopping Drinking‐Mixing nonalcoholic drinks with alcohol).

*

p < .05.

**

p < .01.

***

p < .001.

4. DISCUSSION

Previous research on the PBSS has found support for several factor structures and identified the need to further improve its psychometric properties and to broaden its use to other cultures and populations (e.g., Martens et al., 2007; Pearson, 2013; Treloar et al., 2015). The current research aimed to contribute to this line of research by examining the psychometric properties of PBBS' previous‐reported structures in French and German in a large sample of young males in Switzerland. In testing factor models that were previously documented in the literature (Madson et al., 2013; Martens et al., 2007; Treloar et al., 2015; Walters et al., 2007), we found that neither the two‐ nor the three‐factor PBBS adequately fit our data. Results provided, however, support for a four‐factor model.

Our results that a four‐factor model best fits our data are consistent with Walters et al. (2007) who conducted an EFA and found support for a four‐factor structure among males, whereas a three‐factor structure best fit data among females. Notably, fit statistics yielded in the CFA in the current study provide further support for a four‐factor model in which LSD was split into two factors (i.e., LSD: Planned limits on drinking and LSD: Mixing nonalcoholic drinks with alcohol). Importantly, the two other factors comprising our four‐factor model (MoD and SHR) are consistent with most past research evaluating the PBSS psychometric properties (Treloar et al., 2015), suggesting that both MoD and SHR factors may be stable across cultures. It is important to note that our sample included males exclusively, preventing us from testing gender invariance. This may be critical given that previous research testing PBSS measurement invariance across gender has yielded mixed findings. Whereas Madson et al. (2013) found their PBSS‐R model to be invariant across gender, at least two other recent studies have shown opposite findings with the PBSS and the PBSS‐20 (Treloar et al., 2015; Treloar, Martens, & McCarthy, 2014). Specifically, in their recent study conducted among college students, Treloar et al. (2014) found that measurement parameters of the PBSS differed for males and females. Hence, future research testing the French and German PBSS‐20 structure with both males and females and evaluating gender measurement invariance is warranted.

Next, our findings provided support for metric, but not scalar, invariance across linguistic regions (i.e., French and German speaking). Evidence of metric invariance indicates that the factors have the same meaning across French‐ and German‐speaking participants and that the scale measures the same constructs in both groups, which allows group comparison of factor variances and covariance. That said, lack of scalar invariance implies that some influences, such as cultural norms or the questionnaire translation, may affect the way participants answer to items across linguistic regions. Echoing these results, we found that, except for PBS‐SHR subscale and LSD‐M, PBS scores were significantly lower in German‐speaking participants than in French‐speaking participants (see Table 3). Taken together, these findings suggest that the French and German PBSS‐20 may be used in international studies conducted in French‐ and German‐speaking samples, although caution is needed when comparing mean scores across French‐ and German‐speaking samples.

Next, although most correlations reached significance in the expected direction (PBS factors related to fewer drinks and consequences), they were in general weaker than those reported in previous research evaluating the PBSS in college students (e.g., ranging from −0.15 to −0.34; Martens et al., 2007), which indicates small meaningful effect sizes. Further findings showed that correlation coefficients were weaker in the German‐speaking sample than in the French‐speaking sample. Similar findings were yielded in the negative binomial and zero‐inflated regression models. Taken together with the lack of scalar invariance, these findings suggest that further evaluation of the PBSS content among German‐speaking young adults may be warranted. That said, although effect sizes were overall small, findings showed that most PBS scores were significantly related to fewer drinks and consequences in both German‐ and French‐speaking participants, which shows initial evidence for convergent validity of the French and German four‐factor PBSS.

Consistent with past research among college students (Araas & Adams, 2008; Martens et al., 2004) and chronically homeless individuals with alcohol dependence (Grazioli, Hicks, Kaese, Lenert, & Collins, 2015), findings documented that PBS aiming to limit amount of drinking (i.e., LSD‐Pl) were the least endorsed strategies. These findings suggest that PBSs aiming to reduce alcohol‐related harm and to change the manner of drinking may better fit individuals' needs and expectations across different cultures (i.e., the United States and French‐ and German‐speaking Switzerland) and diverse populations (i.e., young adults, college students, and chronically homeless individuals with alcohol dependence).

Although future research is needed to improve the content of the PBBS in German, these findings provide initial support for the French and German PBSS‐20 use for prevention purpose. Brief motivational alcohol‐related interventions targeting young adults (i.e., college students) in the United States typically include a PBS component, consisting of providing participants with a list of PBS they might use when drinking or partying (Larimer et al., 2007). As previously mentioned, this approach is promising as increased use of PBS has been found to explain motivational intervention's efficacy in decreasing alcohol use among college students (Larimer et al., 2007), although findings have not always been consistent (Reid & Carey, 2015). Therefore, if future research further confirms the negative association between French and German PBSS‐20 and alcohol outcomes, items included in the French and German PBSS‐20 may be used in brief alcohol‐related interventions tailored to French‐ and German‐speaking young adults.

Although major strengths of the current study include its large sample size, it is not without limitations. First, the sample was limited to young males. Given the PBSS measurement invariance across gender previously described in the literature, the French and German PBSS‐20 should not be used among females until future research further validates the tool for both males and females. Similarly, given our results of lack of scalar measurement invariance, comparison across French and German speaking participants requires caution. Finally, consistent with past research on PBS (Martens et al., 2007), the study relied on response to self‐reported questionnaires, which may raise validity concerns.

Despite these limitations, we believe that the current study contributes to the PBS literature by providing a French and German version of the PBSS‐20. Although study and psychometric limitations require caution in comparing French‐ and German‐speaking males' scores, the French and German PBSS‐20 appears to have four reliable factors as well as a good convergent validity. Therefore, the French and German PBSS‐20 represents a promising research and clinical tool that can be used among young males in both French‐ and German‐speaking countries.

CONFLICT OF INTEREST

The authors declare that they have no conflict of interest.

Supporting information

Table S1. French Version of the PBSS‐20

Table S2. German Version of the PBSS‐20

Click here for additional data file. (19.5KB, docx)

ACKNOWLEDGEMENTS

This study was funded by the Swiss National Science Foundation (FN 33CSC0‐122679, FN 33CS30‐139467, and FN 33CS30‐148493).

We are grateful to Céline Gachoud and Charlotte Eidenbenz for their extensive efforts in the coordination of this study.

Grazioli VS, Studer J, Larimer ME, et al. Protective Behavioral Strategies Scale‐20: Psychometric properties of a French and German version among young males in Switzerland. Int J Methods Psychiatr Res. 2019;28:e1777 10.1002/mpr.1777

REFERENCES

  1. Abbey, A. (2002). Alcohol‐related sexual assault: A common problem among college students. Journal of Studies on Alcohol. Supplement, 14, 118–128. 10.15288/jsas.2002.s14.118 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Araas, T. E. , & Adams, T. B. (2008). Protective behavioral strategies and negative alcohol‐related consequences in college students. Journal of Drug Education, 38(3), 211–224. 10.2190/DE.38.3.b [DOI] [PubMed] [Google Scholar]
  3. Atkins, D. C. , Baldwin, S. A. , Zheng, C. , Gallop, R. J. , & Neighbors, C. (2013). A tutorial on count regression and zero‐altered count models for longitudinal substance use data. Psychology of Addictive Behaviors, 27(1), 166–177. 10.1037/a0029508 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Atkins, D. C. , & Gallop, R. J. (2007). Rethinking how family researchers model infrequent outcomes: A tutorial on count regression and zero‐inflated models. Journal of Family Psychology, 21(4), 726–735. 10.1037/0893-3200.21.4.726 [DOI] [PubMed] [Google Scholar]
  5. Barnett, N. P. , Murphy, J. G. , Colby, S. M. , & Monti, P. M. (2007). Efficacy of counselor vs. computer‐delivered intervention with mandated college students. Addictive Behaviors, 32(11), 2529–2548. 10.1016/j.addbeh.2007.06.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bennett, D. A. (2001). How can I deal with missing data in my study? Australian and New Zealand Journal of Public Health, 25(5), 464–469. 10.1111/j.1467-842X.2001.tb00294.x [DOI] [PubMed] [Google Scholar]
  7. Benton, S. L. , Schmidt, J. L. , Newton, F. B. , Shin, K. , Benton, S. A. , & Newton, D. W. (2004). College student protective strategies and drinking consequences. Journal of Studies on Alcohol, 65(1), 115–121. 10.15288/jsa.2004.65.115 [DOI] [PubMed] [Google Scholar]
  8. Briggs, S. R. , & Cheek, J. M. (1986). The role of factor analysis in the development and evaluaton of pesonality scales. Journal of Personality, 54, 106–148. 10.1111/j.1467-6494.1986.tb00391.x [DOI] [Google Scholar]
  9. Chen, F. F. (2007). Sensitivity of goodness of fit indexes to lack of measurement invariance. Structural Equation Modeling, 14, 464–504. 10.1080/10705510701301834 [DOI] [Google Scholar]
  10. Cronce, J. M. , & Larimer, M. E. (2011). Individual‐focused approaches to the prevention of college student drinking. Alcohol Research & Health, 34(2), 210–221. [PMC free article] [PubMed] [Google Scholar]
  11. Delgrande Jordan, M. , & Notari, L. (2011). Consommation d'alcool en Suisse. une analyse des données de l'Enquête suisse sur la santé 2007. Lausanne: Addiction Info Suisse. [Google Scholar]
  12. Delva, J. , Smith, M. P. , Howell, R. L. , Harrison, D. F. , Wilke, D. , & Jackson, D. L. (2004). A study of the relationship between protective behaviors and drinking consequences among undergraduate college students. Journal of American College Health, 53(1), 19–26. 10.3200/JACH.53.1.19-27 [DOI] [PubMed] [Google Scholar]
  13. Demartini, K. S. , Carey, K. B. , Lao, K. , & Luciano, M. (2011). Injunctive norms for alcohol‐related consequences and protective behavioral strategies: Effects of gender and year in school. Addictive Behaviors, 36(4), 347–353. 10.1016/j.addbeh.2010.12.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Gmel, G. , Akre, C. , Astudillo, M. , Bähler, C. , Baggio, S. , Bertholet, N. , … Wang, J. (2015). The Swiss cohort study on substance use risk factors—Findings of two waves. Sucht, 61(4), 251–262. 10.1024/0939-5911.a00380 [DOI] [Google Scholar]
  15. Gmel, G. , Kuendig, H. , Notari, L. , & Gmel, C. (2017). Monitorage suisse des addictions: consommation d'alcool, tabac et drogues illégales en Suisse en 2016. Lausanne.
  16. Gmel, G. , Studer, J. , Deline, S. , Baggio, S. , N'Goran, A. , Mohler‐Kuo, M. , & Daeppen, J. B. (2014). More is not always better—Comparison of three instruments measuring volume of drinking in a sample of young men and their association with consequences. Journal of Studies on Alcohol and Drugs, 75(5), 880–888. 10.15288/jsad.2014.75.880 [DOI] [PubMed] [Google Scholar]
  17. Grazioli, V. S. , Dillworth, T. , Witkiewitz, K. , Andersson, C. , Kilmer, J. R. , Pace, T. , … Larimer, M. E. (2015). Protective behavioral strategies and future drinking behaviors: Effect of drinking intentions. Psychology of Addictive Behaviors 10.1037/adb0000041, 29, 355–364. [DOI] [PubMed] [Google Scholar]
  18. Grazioli, V. S. , Hicks, J. , Kaese, G. , Lenert, J. , & Collins, S. E. (2015). Safer‐drinking strategies used by chronically homeless individuals with alcohol dependence. Journal of Substance Abuse Treatment, 54, 63–68. 10.1016/j.jsat.2015.01.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Grazioli, V. S. , Lewis, M. A. , Garberson, L. A. , Fossos‐Wong, N. , Lee, C. M. , & Larimer, M. E. (2015). Alcohol expectancies and alcohol outcomes: Effects of the use of protective behavioral strategies. Journal of Studies on Alcohol and Drugs, 76(3), 452–458. 10.15288/jsad.2015.76.452 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hilbe, J. M. (2007). Negative binomial regression. Cambridge, U.K.: Cambridge University Press; 10.1017/CBO9780511811852 [DOI] [Google Scholar]
  21. Hingson, R. W. (2010). Focus on: College drinking and related problems: Magnitude and prevention of college drinking and related problems. Alcohol Research & Health: The Journal of the National Institute on Alcohol Abuse and Alcoholism, 33(1), 45–54. [PMC free article] [PubMed] [Google Scholar]
  22. Hingson, R. W. , Heeren, T. , Winter, M. , & Wechsler, H. (2005). Magnitude of alcohol‐related mortality and morbidity among U.S. college students ages 18‐24: Changes from 1998 to 2001. Annual Review of Public Health, 26, 259–279. 10.1146/annurev.publhealth.26.021304.144652 [DOI] [PubMed] [Google Scholar]
  23. Hingson, R. W. , & White, A. M. (2012). Prevalence and consequences of college student alcohol use In Correia C. J., Murphy J. G., & Barnett N. P. (Eds.), College student alcohol abuse: A guide to assessment, intervention, and prevention (Wiley ed.) (pp. 3–24). New Jersey: John Wiley & Sons, Inc. [Google Scholar]
  24. Hoyle, R. H. (2012). Handbook of structural equation modeling. New York: The Guildford Press. [Google Scholar]
  25. Johnston, L. D. , O'Malley, P. M. , Bachman, J. G. , & Schulenberg, J. E. (2010). Monitoring the future. National survey results on drug use, 1975–2009: Volume I, Secondary school students. Bethesda, MD: National Institute on Drug Abuse. [Google Scholar]
  26. Johnston, L. D. , O'Malley, P. M. , Bachman, J. G. , Schulenberg, J. E. , & Miech, R. A. (2014). Monitoring the future. National survey results on drug use, 1975–2013: Volume 2, College students and adults ages 19–55. Ann Arbor: Institute for Social Research, the University of Michigan. [Google Scholar]
  27. Jongenelis, M. I. , Pettigrew, S. , Pratt, I. S. , Chikritzhs, T. , Slevin, T. , & Liang, W. (2016). Predictors and outcomes of drinkers' use of protective behavioral strategies. Psychology of Addictive Behaviors, 30(6), 639–647. 10.1037/adb0000194 [DOI] [PubMed] [Google Scholar]
  28. Knight, J. R. , Wechsler, H. , Kuo, M. , Seibring, M. , Weitzman, E. R. , & Schuckit, M. A. (2002). Alcohol abuse and dependence among U.S. college students. Journal of Studies on Alcohol, 63(3), 263–270. 10.15288/jsa.2002.63.263 [DOI] [PubMed] [Google Scholar]
  29. Kulesza, M. , Apperson, M. , Larimer, M. E. , & Copeland, A. L. (2010). Brief alcohol intervention for college drinkers: How brief is? Addictive Behaviors, 35(7), 730–733. 10.1016/j.addbeh.2010.03.011 [DOI] [PubMed] [Google Scholar]
  30. Larimer, M. E. , Lee, C. M. , Kilmer, J. R. , Fabiano, P. M. , Stark, C. B. , Geisner, I. M. , … Neighbors, C. (2007). Personalized mailed feedback for college drinking prevention: A randomized clinical trial. Journal of Consulting and Clinical Psychology, 75(2), 285–293. 10.1037/0022-006X.75.2.285 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Luebbe, A. M. , Varvel, S. , & Dude, K. (2009). Reciprocal relations of protective behavioral strategies and risk‐amplifying behaviors with alcohol‐related consequences: Targets for intervention with female college students. Journal of Alcohol and Drug Education, 53, 7–29. [Google Scholar]
  32. Madson, M. B. , Arnau, R. C. , & Lambert, S. J. (2013). Development and psychometric evaluation of the Revised Protective Behavioral Strategies Scale. Psychological Assessment, 25(2), 556–567. 10.1037/a0031788 [DOI] [PubMed] [Google Scholar]
  33. Martens, M. P. , Ferrier, A. G. , Sheehy, M. J. , Corbett, K. , Anderson, D. A. , & Simmons, A. (2005). Development of the Protective Behavioral Strategies Survey. Journal of Studies on Alcohol, 66(5), 698–705. 10.15288/jsa.2005.66.698 [DOI] [PubMed] [Google Scholar]
  34. Martens, M. P. , Pederson, E. R. , Labrie, J. W. , Ferrier, A. G. , & Cimini, M. D. (2007). Measuring alcohol‐related protective behavioral strategies among college students: Further examination of the Protective Behavioral Strategies Scale. Psychology of Addictive Behaviors, 21(3), 307–315. 10.1037/0893-164X.21.3.307 [DOI] [PubMed] [Google Scholar]
  35. Martens, M. P. , Taylor, K. K. , Damann, K. M. , Page, J. C. , Mowry, E. S. , & Cimini, M. D. (2004). Protective behavioral strategies when drinking alcohol and their relationship to negative alcohol‐related consequences in college students. Psychology of Addictive Behaviors, 18(4), 390–393. 10.1037/0893-164X.18.4.390 [DOI] [PubMed] [Google Scholar]
  36. Muthén, L. , & Muthén, B. O. (1998). Mplus: The comprehensive modeling program for applied researchers: User's guide. Los Angeles: Author. [Google Scholar]
  37. Napper, L. E. , Kenney, S. R. , Lac, A. , Lewis, L. J. , & LaBrie, J. W. (2014). A cross‐lagged panel model examining protective behavioral strategies: Are types of strategies differentially related to alcohol use and consequences? Addictive Behaviors, 39(2), 480–486. 10.1016/j.addbeh.2013.10.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Neighbors, C. , Lee, C. M. , Lewis, M. A. , Fossos, N. , & Walter, T. (2009). Internet‐based personalized feedback to reduce 21st‐birthday drinking: A randomized controlled trial of an event‐specific prevention intervention. Journal of Consulting and Clinical Psychology, 77(1), 51–63. 10.1037/a0014386 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Nunnally, J. C. , & Bernstein, I. H. (1994). Psychometric theory (3rd ed.). New York: McGraw‐Hill. [Google Scholar]
  40. Pearson, M. R. (2013). Use of alcohol protective behavioral strategies among college students: A critical review. Clinical Psychology Review, 33(8), 1025–1040. 10.1016/j.cpr.2013.08.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Perkins, H. W. (2002). Social norms and the prevention of alcohol misuse in collegiate contexts. Journal of Studies on Alcohol. Supplement, 14, 164–172. 10.15288/jsas.2002.s14.164 [DOI] [PubMed] [Google Scholar]
  42. Reid, A. E. , & Carey, K. B. (2015). Interventions to reduce college student drinking: State of the evidence for mechanisms of behavior change. Clinical Psycholy Review, 40, 213–224. 10.1016/j.cpr.2015.06.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Schulenberg, J. E. , Johnston, L. D. , O'Malley, P. M. , Bachman, J. G. , Miech, R. A. , & Patrick, M. E. (2017). Monitoring the future. National survey results on drug use, 1975–2016: Volume II, College students and adults ages 19–55. Ann Arbor: Institute for Social Research, The University of Michigan; Retrieved from http://monitoringthefuture.org/pubs.html#monographs [Google Scholar]
  44. Studer, J. , Baggio, S. , Mohler‐Kuo, M. , Dermota, P. , Gaume, J. , Bertholet, N. , … Gmel, G. (2013). Examining non‐response bias in substance use research—Are late respondents proxies for non‐respondents? Drug and Alcohol Dependence, 132(1–2), 316–323. 10.1016/j.drugalcdep.2013.02.029 [DOI] [PubMed] [Google Scholar]
  45. Treloar, H. , Martens, M. P. , & McCarthy, D. M. (2014). Testing measurement invariance of the protective behavioral strategies scale in college men and women. Psychological Assessessment, 26(1), 307–313. 10.1037/a0034471 [DOI] [PubMed] [Google Scholar]
  46. Treloar, H. , Martens, M. P. , & McCarthy, D. M. (2015). The protective behavioral strategies scale‐20: Improved content validity on the serious harm reduction subscale. Psychological Assessessment, 27(1), 340–346. 10.1037/pas0000071 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Tukey, J. (1962). The future of data analysis. The Annals of Mathematical Statistics, 33(1), 1–67. 10.1214/aoms/1177704711 [DOI] [Google Scholar]
  48. Vandenberg, R. J. , & Lance, C. E. (2000). A review and synthesis of the measurement invariance literature: Suggesions, practices, and recommendations for organizational research. Organizational Research Methods, 3(1), 4–70. 10.1177/109442810031002 [DOI] [Google Scholar]
  49. Walters, S. T. , Roudsari, B. S. , Vader, A. M. , & Harris, T. R. (2007). Correlates of protective behavior utilization among heavy‐drinking college students. Addictive Behaviors, 32(11), 2633–2644. 10.1016/j.addbeh.2007.06.022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. 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]
  51. World Health Organization . (2014). Global status report on alcohol and health. Geneva: WHO. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Table S1. French Version of the PBSS‐20

Table S2. German Version of the PBSS‐20

Click here for additional data file. (19.5KB, docx)

Articles from International Journal of Methods in Psychiatric Research are provided here courtesy of Wiley

RESOURCES