Assessment of the Psychometric Properties of the Alcohol-Related Protective Behavioral Strategies Scale (PBSS).
In attempts to better understand and identify factors affecting alcohol consumption and related consequences, research examining cognitive-behavioral strategies titled Protective Behavioural Strategies (PBS) and their associations to various health-related behaviors have become popular (Lewis, Logan & Neighbors, 2009; Palmer, McMahon, Rounsaville & Ball, 2010; Patrick, Lee & Larimer, 2011). PBS are used to limit or reduce the amount of alcohol consumption and in turn, minimize the related negative consequences (Benton et al., 2004; Delva et al., 2004; Glassman, Werch, & Jobli, 2007). They include techniques such as alternating alcoholic drinks with non-alcoholic drinks, spacing drinks and/or using a designated driver (Lewis, Rees, Logan, Kaysen, & Kilmer, 2010). The negative consequences experienced by university students due to alcohol consumption are varied and fall under categories such as health, personal, academic, legal and financial problems (Perkins, 2002; Wechsler et al., 2002). Overall, within the literature, there is support that those who use PBS are less likely to experience alcohol-related negative consequences (Araas & Adams, 2008; Haines, Barker, & Rice, 2006; Ray, Turrisi, Abar, & Peters, 2009).
The small number of previous psychometric assessments of the Protective Behavioral Strategies Scale (PBSS) support a three-factor structure (Martens et al., 2005; Martens, Pedersen et al., 2007; Pearson, Kite, & Henson, 2013). However, these analyses have noted inconsistencies within the outcomes. In their initial investigation, Martens et al. (2005) completed an exploratory factor analysis (EFA) on 437 undergraduate students and provided a three-factor solution, that accounted for 52% of the variance, based on the following subscales: Stopping/Limiting Drinking (SLD; seven items), Manner of Drinking (MOD; five items) and Serious Harm Reduction (SHR; three items). While seeking additional psychometric information on the PBSS, Martens, Pedersen et al. (2007), through confirmatory factor analysis (CFA), again posited a three-factor solution in line with the above subscales.
Reliability checks on the internal consistency of the PBSS' subscales have provided estimates between .81-.85 for SLD, .74-.79 for MOD and .59-.66 for SHR (Martens, Pedersen et al., 2007; Treloar, Martens, & McCarthy, 2014; Walters, Roudsari, Vader, & Harris, 2007). The lower reliability estimates for the SHR have been attributed to the small number of items included in the subscale (three items; Martens et al., 2005; Martens, Pedersen et al., 2007; Walters et al., 2007). Attempts have been made at increasing the reliability estimates of the SHR subscale through revised versions of the PBSS (Madson, Arnau, & Lambert, 2013; Treloar, Martens, & McCarthy, 2015). The revised PBSS utilized by Madson et al. (2013), with added SHR items, was found to best fit a two-factor model for the data collected from White non-Hispanic and African American men and women, through CFA.
To date, studies of PBS have focused on investigating such concepts as outcomes, motives and associated behaviors particularly among university students (Pearson et al., 2013). However, theoretical analysis of PBS and the psychometric structure of scales that seek to measure such, like the PBSS, are still required. This study aims to provide an examination of the psychometric structure of the PBSS within an Irish sample of university students, as the majority of psychometric assessments of the PBSS have been conducted in the United States. It will seek to replicate and build upon the findings of Martens et al. (2005) in their assessment of the original PBSS.
The PBSS data was extracted from two separate studies conducted within the same Irish university collected in Spring 2011 and Spring 2012 for study one and two respectively. Full ethical approval was granted from the University Research Ethics Committee for both studies. Study one utilized a cross-sectional observational design to examine self-regulation and alcohol-related PBS in the context of other key influences on drinking, including motives for drinking and alcohol expectancies. Study two similarly utilized a cross-sectional observational design to examine the relationships among personality, alcohol-related PBS, alcohol consumption and sexual behavior in young women.
For the PBSS data from study one, the age range of the sample was 17-57 years (n = 426; M= 20.82, SD = 4.06) and the sample was predominately female (n = 325, 76.1%). Of the 426 participants, seven were identified as 'non-drinkers' and were therefore removed from the final analysis as theoretically, PBS are used when an individual is consuming alcohol (Martens et. al., 2005). Furthermore, six participants reported being below the age of 18, and due to ethical considerations, they were also removed from the final analysis.
For the alcohol-related PBS data from study two, the age range of the sample was 18-25 years (n = 782; M= 20.27, SD = 1.79). The sample was comprised of all females, of which 95.8% were students. For the final EFA, only data from those identified as full-time students was included.
Measure: Alcohol-related PBSS
The PBSS (Martens, Pedersen et al., 2007; Martens et al., 2005; see Appendix A) was utilized in both studies to access the cognitive-behavioral strategies employed by participants to reduce their alcohol consumption and associated harm. Study one utilized the original PBSS from Martens et al. (2005) study, upon which the revised PBSS (Martens, Pedersen et al., 2007) was developed. For study two, the updated, modified version of the PBSS used by Martens, Pedersen et. al. (2007) was utilized. The questionnaire consisted of 15 items measuring cognitive-behavioral strategies designed to reduce high-risk drinking and resulting negative consequences, e.g., Stopping/Limiting Drinking (SLD; seven items), Manner of Drinking (MOD; five items) and Serious Harm Reduction (SHR; three items). Participants indicated how often they engaged in the behavior when drinking alcohol on a six-point Likert scale (one = Never, six = Always), with higher scores indicating more use of the alcohol-related PBS. One item was reverse scored (for study one this was item 12 and for study two this was item 9).
Descriptive statistics for the PBSS data from both studies are presented in Table 1. As mentioned, the same alcohol-related PBSS was presented to participants in both studies; however, the following slight differences between the two versions of the PBSS should be noted: phrasing of a number of items was slightly different between the two studies and the order in which the 15 items were presented to participants was different due to the updated, modified version being used in study one (differences are highlighted in bold in Appendix A).
Both EFA and CFA were utilized. On the first study, EFA allowed an open examination of the PBSS data (Pett, Lackey & Sullivan, 2003), due to the lack of such previous assessments conducted outside of the United States. By not imposing a predetermined structure (Child, 1990), the EFA aimed to increase the psychometric soundness of the measure, namely in terms of the internal consistency of subscales and external validity of the overall scale. The results of the EFA informed a more focused CFA analysis to be conducted on the PBSS data of the second study to access the structure and relations between the latent factors and items from the initial EFA (Field, 2013). Data were assessed for suitability prior to factor analysis and reliability analyses were also presented for the new factor structures. In terms of accessing missing data, Little's MCAR test for study one indicated that the data was missing at random (i.e., no identifiable pattern exists to the missing data); [chi square](217) = 51.28, p = 1.000. Similarly, for study two, Little's MCAR test indicated that the data was also missing at random; [chi square](274) = 240.34, p = .930.
Study One (EFA)
For the EFA completed study, one PBSS data (n = 305), the scale's dimensionality was investigated using maximum likelihood (ML) with oblique rotation (oblimin, delta set at zero). ML was selected because it is robust when data are not normally distributed (Fabrigar, Wegener, MacCallum, & Strahan, 1999). The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy and Bartlett's test of sphericity indicated that the data were suitable for factor analysis (i.e., Kaiser-Meyer-Olkin Measure of Sampling Adequacy statistic was .88; Bartlett's test was p < .001; Field, 2013). To assist with factor retention, scree plot and traditional parallel analysis (T-PA) were used. Both are considered to be superior to the commonly employed eigenvalue greater than one "rule" (Costello & Osborne, 2005).
Results of the parallel analysis suggested that a two-factor solution was appropriate (i.e., eigenvalues 1 and 2 for the real data exceeded the 95% percentile eigenvalues generated for the random data [5.01 versus 1.40; and 1.45 versus 1.30]). The eigenvalue for the third factor was 1.09, which was lower than the 95% percentile eigenvalue for the random data (1.24). The scree plot also suggested that a two-factor solution should be retained. The proportion of total variance accounted for was 43.07% (Factor 1 = 33.38%, and Factor 2 = 9.69%).
Factors loading at .5 or above were retained. Comparison two-factor models, with all items loading above .4, were tested (the model fit statistics are outlined below in sections 3.2.1. and 3.2.3). The model fit identified in the EFA was optimal. The content of the three items loading on the first factor were all from the original SLD factor; the four items loading on the second factor were all from the original MOD factor (see Table 2).
Reliability and descriptive statistics
The alpha coefficients for both new subscales/factors were acceptable and just below acceptable respectively (MOD = .72, 95% CI [.66-.76]; SLD = .69, 95% CI [.62-.74]), and better than the third original, now deleted factor (SHR = .51). Descriptive statistics for each total sub-scale were: SLD, M = 8.00, SD = 3.39, 95% CI [7.63-8.38]; MOD, M= 13.74, SD = 3.32. 95% CI [13.37-14.10].
Study Two (CFA)
The full sample from study two (n = 768) was randomly divided into two smaller groups (labelled subsamples 1 and 2: ns = 385 and 383, respectively) in order to test the replicability of the two-factor, seven-item (2:7) PBSS model identified using EFA. As they represent nested models, and as the evidence base in relation to an optimal PBSS model is limited to less than five studies, two competing models (original, three-factor model; one-factor model) were also tested.
Using a maximum likelihood parameter estimate, which was used for all CFAs, multiple criteria; comprising a recommended 'minimal set' (Schweizer, 2010); were used to assess model goodness-of-fit: normed chi-square (Chi Square statistic divided by the degrees of freedom: [chi square]/df), Comparative Fit Index (CFI), the Standardized Root Mean Square Residual (SRMR) and the Root Mean Square Error of Approximation (RMSEA). The following thresholds denote good fit: CFI [greater than or equal to] .95, RMSEA [less than or equal to] .05, SRMR <. 08, and [chi square]/df < 2. Acceptable fit is reflected when CFI [greater than or equal to] .90, RMSEA [less than or equal to] .08, SRMR < .10, and [chi square]/df < 3 (Hooper, Coughlan, & Mullen, 2008; Kline, 2011; Tabachnick & Fidell, 2007).
Subsample 1 (ns = 385)
All CFA fit statistics are rounded to two decimal places for text; see Table 2 for full numbers. The fit statistics for the 2:7 PBSS model were good: CFI = .96, RMSEA = .070 (.045-.097), SRMR = .04, and [chi square]/df = 2.9 (see Table 3). As retention of items with factor loadings of .5 and above may be considered overly conservative, we assessed items with loading above .4. This retained five of the eight deleted items, leaving six items in each factor. However, model fit for this solution was sub-optimal (CMIN/DF = 3.79; CFI = .83; RMSEA = >.08). Inspection of the modification indices suggested that deleting two items would improve model fit (which had been below .5 in the optimal model). Their deletion improved model fit, which was 'acceptable' but not good (e.g., CFI = .9 < .95). Alpha coefficients for the two factors in this model were: SLD/SHR combination = .58 (95% CI [.51-.64]), and MOD = .75 (95% CI [.71-.79]). The model fits for both the original three-factor, and one-factor models were not acceptable (e.g., [chi square]/df > 3; CFI < .90, etc.).
Subsample 2 (ns = 383)
The two-factor, seven-item (2:7) model which was confirmed within subsample 1, was tested again with subsample 2. Fit was again good: CFI = .99, RMSEA = .034 (.000-.066), SRMR = [.031.sup.3], and [chi square]/df = [1.4.sup.4] (see Table 3), with the modification indices suggesting that no further alterations in terms of item deletion or shared error covariance were warranted.
Scale score reliability and descriptive statistics for two-factor, seven-item PBSS
The alpha coefficients for the MOD subscales ranged across from acceptable (.76) to poor (.58) between the two sub-samples. Similarly, the SLD subscale changes from questionable (.62) to acceptable (.72) between the two sub-samples. The full sample was re-combined for total analysis of alpha coefficients. Both subscales achieved coefficients at the upper level of questionable, thus approaching acceptable: SLD = .67, 95% CI [.63-.71]; MOD = .67 [.63-70]). Descriptive statistics for the 2:7 PBSS across Study 2 are given in Table 4.
Two-factor model for the PBSS
A two-factor model through EFA was found to best fit the PBSS data from an Irish university sample. This was further supported by CFA. The results go against that of previous research examining the PBSS, where support was found for a three-factor solution (Martens et al., 2005; Martens, Pedersen et al., 2007; Pearson et al., 2013). In comparison to Martens and colleagues' (2005) original model, the two-factor solution from the current analysis accounted for less of the overall variance. However, the amount of variance explained by the current SLD and MOB factors was similar to the original EFA, with fewer items per factor and similar or higher factor loadings. It should be noted that in the original EFA, Martens et al. (2005) found that seven items loaded onto the SLD factor; however, the current analysis found a similar amount of explained variance with just three-item loadings.
Researchers in the past have questioned the psychometric validity of the SHR subscale, in comparison to the SLD and MOD subscales (Madson et al., 2013; Treloar et al., 2015). This is despite other examples highlighting the significant relationships between the uses of PBS, especially the SHR subscale, and fewer negative consequences from drinking (Walters et al., 2007; Martens, Ferrier, & Cimini, 2007). Problems with the SHR subscale have been attributed to its low number of items. As discussed, both Madson et al. (2013) and Treloar et al. (2015) found support for three factor models of the PBSS by the inclusion of additional SHR items. Our analysis adds to this body of literature examining the psychometric properties of the PBSS and raises questions about the structure and consistency of the PBSS across university samples from different nations. When examining the reliability estimates for internal consistency, our estimate for MOD was found to be within the range of those of previous studies (.74-.79 for MOD; Treloar et al., 2014; Walters et al., 2007). However, the alpha coefficient for the SLB subscale was lower (between .81-.85 for SLD). It should be noted that both were above or close to the .7 threshold deemed acceptable for Cronbach's alpha values (Nunnally, 1978).
Given that the scale was developed in the United States, and the vast majority of its psychometric testing has been completed on American samples (Pearson et al., 2013; Madson et al., 2013; Treloar et al., 2015), the current results may suggest gender and cultural differences. Improving reliability estimates, particularly on the SHR subscale, is a noted area for future investigation (Martens et al., 2005; Martens, Pedersen et al., 2007) and attempts are continuing through the development of revised versions of the PBSS (Madson et al., 2013; Treloar et al., 2015). Overall, the results of the current EFA and CFA supporting a two-factor solution for the PBSS, combined with the less than ideal reliability estimates, raise questions about the content validity and overall reliability of the PBSS across different cultures and genders. Given that PBS have been utilized as brief interventions for reducing problematic alcohol consumption and the resulting consequences, with mixed results (Brett, Leffingwell, & Leavens, 2017; Dvorak et al., 2017; Montes, LaBrie, & Froidevaux, 2016), future research may consider a refined structure, not just as a psychometric exercise for the scale, but additionally in consideration of its practical application.
There are a number of limitations to the study that should be noted. Firstly, the PBSS data were collected as part of two separate studies and two marginally different versions of the PBSS were used; however, it was the original version of the PBSS (Martens et al., 2005) used in study one (see Appendix A for differences), and the updated version of the same PBSS (Martens, Pedersen et al., 2007) utilized in study two. Secondly, both studies were conducted within one university, therefore limiting the generalizability of the findings due to sample constraints. Thirdly, a large majority of the sample were female, due to the research design of study two.
Future research and conclusions
Future research may consider examining the factor structure of the PBSS across different cultures and genders, as it may hold relevance. Studies have found that women overall are more likely to utilize PBS and use them more effectively (Benton, Downey, Glider, & Benton, 2008; Benton et al., 2004; Delva et al., 2004; Haines et al., 2006). In their research on a student population, Lewis and colleagues (2010) found that women (50.2% of their population) who used PBS experienced significantly less sex-related alcohol-negative consequences when compared to men. Some researchers have posited that the motives behind the use of PBS are different between men and women, such as women's physiology (LaBrie, Kenney, Lac, & Garcia, 2009), social pressure (DeMartini, Carey, Lao, & Luciano, 2011) and personality traits (Pervin & John, 1999). The findings of the current examination of the PBSS conducted on a predominantly young, female sample raise questions about the psychometric structure and reliability of the PBSS. In order to establish the most effective PBS for specific age and ethnic groups, as well as genders, robust investigations of the theoretical and psychometric constructs (namely, factor structure, reliability and validity) of the PBSS are warranted.
Correspondence concerning this article should be addressed to: Sinead Moylett, Department of Psychiatry, University of Cambridge School of Clinical Medicine, Level E4 Cambridge Biomedical Campus, Cambridge, CB2 0SP, UK.
Acknowledgements: The authors would like to sincerely thank John McCaffrey for his input and advice.
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. doi:10.2190/DE.38.3.b
Benton, S. L., Downey, R. G., Glider, P. J., & Benton, S. A. (2008). College students' norm perception predicts reported use of protective behavioral strategies for alcohol consumption. Journal of Studies on Alcohol and Drugs, 69(6), 859-865. doi:http://dx.doi.org/10.15288/jsad.2008.69.859
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. doi:http://dx.doi.org/10.15288/jsa.2004.65.115
Brett, E. I., Leffingwell, T. R., & Leavens, E. L. (2017). Trait mindfulness and protective strategies for alcohol use: Implications for college student drinking. Addictive Behaviors, 73(), 16-21. doi:10.1016/j.addbeh.2017.04.011.
Child, D. (1990). The Essentials of Factor Analysis (2nd Ed.). London: Cassel Educational Limited.
Costello, A. B., & Osborne, J. W. (2005). Best practices in exploratory factor analysis: Four recommendations forgetting the most from your analysis. Practical Assessment, Research & Evaluation, 10(1), 1-9.
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. doi:10.3200/JACH.53.1.19-27
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. doi:10.1016/j.addbeh.2010.12.009
Dvorak, R.D., Kramer, M.P., Stevenson, B.L., Sargent, E.M., & Kilwein, T.M. (2017). An application of Deviance Regulation Theory to reduce alcohol-related problems among college women during spring break. Psychology of Addictive Behaviors, Epud ahead of print. doi:10.1037/adb0000258
Fabrigar, L. R., Wegener, D. T, MacCallum, R. C, & Strahan, E. J. (1999). Evaluating the use of exploratory factor analysis in psychological research. Psychological Methods, 4(3), 272-299. doi:10.1037/1082-989X.4.3.272
Field, A. (2013). Discovering Statistics Using IBM SPSS Statistics (4th Ed.). London: SAGE Publications
Glassman, T, Werch, C, & Jobli, E. (2007). Alcohol self-control behaviors of adolescents. Addictive Behaviors, 32(3), 590-597. doi:10.1016/j.addbeh.2006.06.003
Haines, M. P., Barker, G., & Rice, R. M. (2006). The personal protective behaviors of college student drinkers: Evidence of indigenous protective norms. Journal of American College Health, 55(2), 69-75. doi:10.3200/JACH.55.2.69-76
Hooper, D., Coughlan, J. & Mullen, M. R. (2008). Structural equation modelling: Guidelines for determining model fit. Electronic Journal of Business Research Methods, 6(1), 53-60.
Kline, R. B. (2011). Principles and Practice of Structural Equation Modeling. New York, NY: Guilford Press.
LaBrie, J. W., Kenney, S. R., Lac, A., & Garcia, J. A. (2009). Mental and social health impacts the use of protective behavioral strategies in reducing risky drinking and alcohol consequences. Journal of College Student Development, 50(1), 35-49. doi:10.1353/csd.0.0050
Lewis, M. A., Logan, D. E., & Neighbors, C. (2009). Examining the role of gender in the relationship between use of condom-related protective behavioral strategies when drinking and alcohol-related sexual behavior. Sex Roles, 61(9), 727-735. doi:10.1007/sl1199-009-9661-1
Lewis, M. A., Rees, M., Logan, D. E., Kaysen, D. L., & Kilmer, J. R. (2010). Use of drinking protective behavioral strategies in association to sex-related alcohol-negative consequences: The mediating role of alcohol consumption. Psychology of Addictive Behaviors, 24(2), 229-238. doi:http://dx.doi.org/10.1037/a0018361
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. doi:10.1037/a0031788
Martens, M. P., Ferrier, A. G., & Cimini, M. D. (2007). Do protective behavioral strategies mediate the relationship between drinking motives and alcohol use in college students? Journal of Studies on Alcohol and Drugs, 68(1), 106-114. doi:http://dx.doi.org/10.15288/jsad.2007.68.106
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 and Drugs, 66(5), 698-705. doi:http://dx.doi.org/10.15288/jsa.2005.66.698
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. doi:10.1037/0893-164X.21.3.307
Montes, K. S., LaBrie, J. W., & Froidevaux, N. M. (2016). Do protective behavioral strategies mediate the effect of preparty motives on event-level preparty alcohol use? Substance Use & Misuse, 57(8), 1047-1055. doi:10.3109/10826084.2016.1152 495
Palmer, R. S., McMahon, T. J., Rounsaville, B. L., & Ball, S. A. (2010). Coercive sexual experiences, protective behavioral strategies, alcohol expectancies and consumption among male and female college students. Journal of Interpersonal Violence, 25(9) 1563-1578. doi:10.1177/0886260509354581
Patrick, M. E., Lee, C. M., & Larimer, M. E. (2011). Drinking motives, protective behavioral strategies, and experienced consequences: Identifying students at risk. Addictive Behaviours, 36(3), 270-273. doi:10.1016/j.addbeh.2010.11.007
Pearson, M. R., Kite, B. A., & Henson, J. M. (2013). The assessment of protective behavioral strategies: Comparing the absolute frequency and contingent frequency response scales. Psychology of Addictive Behaviors, 27(4), 1010-1018. doi:10.1037/a0031366.supp
Pedersen, E. R., & LaBrie, J. (2006). Drinking game participation among college students: Gender and ethnic implications. Addictive Behaviors, 37(11), 2105-2115. doi:10.1016/j.addbeh.2006.02.003
Pedersen, E. R., & LaBrie, J. (2007). Partying before the party: Examining pre-partying behavior among college students. Journal of American College Health, 56(3), 237-245. doi:10.3200/JACH.56.3.237-246
Perkins, H. W. (2002). Social norms and the prevention of alcohol misuse in collegiate contexts. Journal of Studies on Alcohol, Suppl. No. 14, 164-172.
Pervin, L. A., & John, O. P. (1999). Handbook of Personality: Theory and Research (2nd Ed.). New York: Guilford Press.
Pett, M. A., Lackey, N. R., & Sullivan, J. J. (2003). Making Sense of Factor Analysis: The Use of Factor Analysis for Instrument Development in Health Care Research. Thousand Oaks, CA: Sage Publications
Ray, A. E., Turrisi, R., Abar, B, & Peters, K. E. (2009). Social-cognitive correlates of protective drinking behaviors and alcohol-related consequences in college students. Addictive Behaviors, 34(11), 911-917. doi:10.1016/j.addbeh.2009.05.0l6
Schweizer, K. (2010). Some guidelines concerning the modeling of traits and abilities in test construction. European Journal of Psychological Assessment, 26(1), 1-2. doi:10.1027/1015-5759/a000001
Tabachnick, B. G. & Fidell, L. S. (2007). Using Multivariate Statistics. Boston, MA: Pearson Education.
Treloar, H., Martens, M. P., & McCarthy, D. M. (2015). The Protective Behavioral Strategies Scale-20: Improved content validity of the Serious Harm Reduction subscale. Psychological Assessment, 27(1), 340-346. doi:10.1037/pas0000071
Treloar, H. R., Martens, M. P., & McCarthy, D. M. (2014). Testing measurement invariance of the protective behavioral strategies scale in college men and women. Psychological Assessment, 26(1), 307-313. doi: 10.1037/a0034471
Walters, S. T., Roudsari, B. S., Vader, A. M., & Harris, T. H. (2007). Correlates of protective behavior utilization among heavy-drinking college students. Addictive Behaviors, 32(11), 2633-2644. doi: 10.1016/j.addbeh.2007.06.022
Wechsler, H., Lee, J. E., Kuo, M., Seibring, M., Nelson, T. F., & Lee, H. (2002). Trends in college binge drinking during a period of increased prevention efforts. Findings from 4 Harvard School of Public Health College Alcohol Study Surveys: 1993-2001. Journal of American College Health, 50(5), 203-217. doi:10.1080/07448480209595713
Alcohol-related protective behavioral strategies scale for Study 1.
Please indicate the degree to which you engage in the following behaviors when using alcohol.
Never Rarely Occasionally Sometimes Use a designated driver 1 2 3 4 Know where your drink 1 2 3 4 is at all times Make sure you go home 1 2 3 4 with a friend Alternate alcoholic and 1 2 3 4 non-alcoholic drinks Get a friend to let you 1 2 3 4 know when you've ha enough to drink Don't exceed a 1 2 3 4 predetermined number of drinks Leave bar or party at 1 2 3 4 predetermined time Stop drinking at 1 2 3 4 predetermined time Put extra ice in your 1 2 3 4 drink Drink water while 1 2 3 4 drinking alcohol Avoid drinking games 1 2 3 4 Drink shots of liquor 1 2 3 4 Avoid trying to keep 1 2 3 4 up or out-drink others Drink slowly rather 1 2 3 4 than gulping/chugging Avoid mixing different 1 2 3 4 types of drinks Usually Always Use a designated driver 5 6 Know where your drink 5 6 is at all times Make sure you go home 5 6 with a friend Alternate alcoholic and 5 6 non-alcoholic drinks Get a friend to let you 5 6 know when you've ha enough to drink Don't exceed a 5 6 predetermined number of drinks Leave bar or party at 5 6 predetermined time Stop drinking at 5 6 predetermined time Put extra ice in your 5 6 drink Drink water while 5 6 drinking alcohol Avoid drinking games 5 6 Drink shots of liquor 5 6 Avoid trying to keep 5 6 up or out-drink others Drink slowly rather 5 6 than gulping/chugging Avoid mixing different 5 6 types of drinks Note. Item 12 is reverse scored. Differences between the two versions of the PBSS are highlighted in bold.
Alcohol-related protective behavioral strategies scale for Study 2.
Please indicate the degree to which you engage in the following behaviors when using alcohol.
Never Rarely Occasionally Sometimes 1. Use a designated driver 1 2 3 4 2. Determine not to exceed 1 2 3 4 a set number of drinks 3. Alternative alcoholic 1 2 3 4 and non-alcoholic drinks 4. Have a friend let you 1 2 3 4 know when you have had enough to drink 5. Avoid drinking games 1 2 3 4 6. Leave the bar/party at 1 2 3 4 a predetermined time 7. Make sure that you go 1 2 3 4 home with a friend 8. Know where your drink 1 2 3 4 has been at all times 9. Drink shots of liquor 1 2 3 4 10. Stop drinking at a 1 2 3 4 predetermined time 11. Drink water while 1 2 3 4 drinking alcohol 12. Put extra ice in your 1 2 3 4 drink 13. Avoid mixing different 1 2 3 4 types of alcohol 14. Drink slowly, rather 1 2 3 4 than gulp or chug 15. Avoid trying to "keep 1 2 3 4 up" or "out-drink" others Usually Always 1. Use a designated driver 5 6 2. Determine not to exceed 5 6 a set number of drinks 3. Alternative alcoholic 5 6 and non-alcoholic drinks 4. Have a friend let you 5 6 know when you have had enough to drink 5. Avoid drinking games 5 6 6. Leave the bar/party at 5 6 a predetermined time 7. Make sure that you go 5 6 home with a friend 8. Know where your drink 5 6 has been at all times 9. Drink shots of liquor 5 6 10. Stop drinking at a 5 6 predetermined time 11. Drink water while 5 6 drinking alcohol 12. Put extra ice in your 5 6 drink 13. Avoid mixing different 5 6 types of alcohol 14. Drink slowly, rather 5 6 than gulp or chug 15. Avoid trying to "keep 5 6 up" or "out-drink" others Note. Item 9 is reverse scored. Differences between the two versions of the PBSS are highlighted in bold.
University of Cambridge School of Clinical Medicine
University College Dublin
Chloe Walsh, Padraig MacNeela & Brian M. Hughes
National University of Ireland Galway
TABLE 1 Descriptive statistics (n, M and SD) for PBSS data across study one and two Descriptive Statistics N M (SD) PBSS Study 1 Study 2 Study 1 Study 2 PBSLimTotal 302 717 18.91 (7.06) 19.83 (6.94) PBSManTotal 305 721 17.48 (4.31) 17.07 (5.43) PBSConseqTotal 307 726 14.41 (3.25) 14.55 (3.15) PBSTotal 294 698 50.76 (12.29) 51.50 (12.53) Note. PBSLimTotal = Total score for the seven items based on limiting/stopping drinking. PBSManTotal = Total score for five items based on manner of drinking. TotalConseqTotal = Total score of three items based on serious harm reduction or consequence reduction. PBSTotal = PBSLimTotal + PBSManTotal + PBSConseqTotal TABLE 2 Factor loading and alpha coefficients of retained items and remaining factors for study one Factor loadings SLD MOD [alpha] (95% CI) [alpha] (95% CI) = = .69 (.62, .74) .72 (.66, .76) SLD 2: Get a friend to let .602 - you know when you've had enough to drink SLD 4: Leave bar or party .646 - at predetermined time SLD 5: Stop drinking at .694 - predetermined time MOD 1: Avoid drinkin - -.723 games MOD 2: Drink shots of - -.544 liquor (*) MOD 4: Drink slowly - -.618 rather than gulping /chugging MOD 5: Avoid mixing - -.571 different types of drinks Note. SLD = Stopping/Limiting Drinking; MOD = Manner of Drinking. TABLE 3 Confirmatory fit statistics for competing factor models for subsamples from study two Model [chi square] CFI RMSEA SRMR /df Subsample 1 3-factors, 3.1 .89 .074 (.064-.084) .054 15-items 2-factors, 2.9 .96 .070 (.045-.097) .041 7-items 1-factor, 4.8 .79 .099 (.090-.109) .070 15-items Subsample 2 3-factors, 2.5 .91 .062 (.052-.073) .050 15-items 2-factors, 1.4 .99 .034 (.000-.066) .031 7-items 1-factor, 4.4 .78 .095 (.085-.104) .069 15-items TABLE 4 Descriptive statistics (M, SD and 95% CI) for 2:7 PBSS model from study two Descriptive Statistics 2:7 PBSS Model M SD 95% CI Stopping/Limiting Drinking 8.52 3.43 8.27, 8.76 (SLD) SLD 2: Get a friend to let you 2.92 1.62 2.81,3.04 know when you've had enough to drink SLD 4: Leave bar or party at 2.79 1.41 2.69, 2.89 predetermined time SLD 5: Stop drinking at 2.76 1.37 2.67, 2.86 predetermined time Manner of Drinking (MOD) 12.91 4.42 12.60, 13.23 MOD 1: Avoid drinking games 2.89 1.62 2.77,3.00 MOD 2: Drink shots of liquor (*) 3.01 1.33 2.91,3.10 MOD 4: Drink slowly rather 3.85 4.41 3.75, 3.96 than gulping/chugging MOD 5: Avoid mixing different 3.17 1.51 3.06, 3.28 types of drinks
|Printer friendly Cite/link Email Feedback|
|Author:||Moylett, Sinead; Regan, Daniel; Walsh, Chloe; Macneela, Padraig; Hughes, Brian M.|
|Publication:||Journal of Alcohol & Drug Education|
|Date:||Apr 1, 2019|
|Previous Article:||Effects of Social Capital on the Culture of College Drinking.|
|Next Article:||Self-Help Book on Dealing with Alcohol and Drug Addictions.|