An Ecological Momentary Assessment of College Women’s Decisions to Use Protective Behavioral Strategies

Abstract

Protective behavioral strategies (PBS) have been associated with reduced risk for sexual assault victimization in college women. Sexual assault risk reduction programs have had limited success increasing PBS use, particularly among heavy drinkers, suggesting a need for additional research on the malleable predictors of PBS use. Whereas longitudinal studies show women’s decisions to use PBS can be both planned and reactive, little is known about the decision-making processes that affect PBS use on drinking days, when sexual assault risk may be elevated. The current study used ecological momentary assessment to examine variability in the associations between decision-making and PBS use within and across drinking days in first-semester college women. Participants (56 female drinkers) completed a 14-day protocol with three daily measures of intentions and willingness to use PBS, and once-daily diaries of PBS use. Multilevel models examined between-day and within-day effects of intentions and willingness to use PBS with regards to sexual assault PBS (e.g., communicating sexual boundaries) and drinking PBS (e.g., limiting alcohol consumption), respectively. On days when sexual assault PBS willingness increased throughout the day, women tended to use more sexual assault PBS. This association was strongest among women who were typically less willing to use these PBS. Among women who were the least willing to use drinking PBS, their drinking PBS use decreased on days when they reported increased willingness to use them. Decisions to use sexual assault and drinking PBS on drinking days were qualified by women’s typical levels of willingness to use the different PBS. This suggests the need for a multi-faceted intervention strategy that targets both typical and event-level risk. Individual-level alcohol and sexual assault risk reduction approaches could be enhanced with event-level PBS messaging and evaluation.

Keywords

ecological momentary assessment protective behavioral strategies sexual assault risk alcohol consumption college women

Nearly one quarter of college women experience sexual assault in the United States (Fisher et al., 2000; Krebs et al., 2016). Risk for victimization is particularly high among younger students (i.e., those under the age of 21) who engage in heavy episodic drinking (HED; i.e., consuming four or more drinks on a single occasion; Abbey et al., 1996; Testa & Livingston, 2009). Even women who typically consume less are at increased risk on days when they consume more than their typical amounts (Parks & Fals-Stewart, 2004; Scaglione et al., 2014). The underlying mechanisms that explain the association between alcohol and sexual assault risk are well-established and multifaceted (see Abbey et al., 2004). Studies show that men perceive intoxicated women to be more sexually available or interested (George & Stoner, 2000); women who are drinking heavily also have a delayed recognition of and response to sexual assault risk. As such, perpetrators may be more inclined to prey on women who are drinking. Although sexual assault risk only exists in the presence of a perpetrator, alcohol is involved in over 50% of sexual assaults that occur on college campuses (Abbey et al., 2004; Krebs et al., 2016), warranting prevention strategies that address alcohol-related risk (Testa & Livingston, 2009).

Sexual assault risk reduction programs are an integral component of comprehensive prevention efforts and address multiple risk and protective factors (Gidycz, 2018; Ullman, 2007). Risk reduction programs assert sexual assault is always the perpetrator’s fault and also empower women to minimize the likelihood they will be targeted as a potential victim (Gidycz, 2018; Orchowski et al., 2020). A core mechanism of this prevention approach has been to increase use of sexual assault protective behavioral strategies (PBS; Gilmore et al., 2015; Orchowski et al., 2008). Sexual assault PBS include strategies for setting and communicating personal boundaries (e.g., determining in advance what one is willing to do in a sexual situation) and deterring harm in situations where sexual assault risk may be present (e.g., walking home with trusted friends, being prepared to use forceful verbal or physical resistance strategies if necessary; Orchowski et al., 2020). Several attempts have been made to tailor this risk reduction approach for heavy drinking women (Gilmore et al., 2015). These programs have typically expanded the discussion of PBS to include drinking PBS, which aim to limit alcohol consumption (e.g., consuming no more than one drink per hour; planning not to exceed three drinks in a night) or modify the manner of drinking (e.g., alternating alcohol and non-alcoholic beverages; avoiding shots) to minimize the potential negative effects of intoxication (Martens et al., 2005; Pearson, 2013). Whereas sexual assault PBS have been directly associated with decreased sexual assault victimization in college women (Gilmore et al., 2018), drinking PBS have been indirectly associated with fewer incidents of forced and unwanted sex by first reducing alcohol consumption (Mallett et al., 2015; Moorer et al., 2013).

Several rigorous evaluation studies of risk reduction programs have reported significant effects on sexual assault victimization (e.g., Gidycz et al., 2015; Senn et al., 2015) and alcohol-involved sexual assault, specifically (Gilmore et al., 2015). Program impacts on PBS use have been less consistent, especially in samples of heavy drinkers (Gilmore et al., 2015). This gap could be a function of measurement specificity. Prospective evaluation studies measure typical PBS use averaged over several months (Gilmore et al., 2015). This level of measurement does not account for whether or how women tailor their PBS use in response to interpersonal or contextual risk on a given day. Existing programs may also be insufficiently addressing the upstream factors that motivate an individual to use PBS (Nielsen et al., 2018). Relatively little is known about the processes that inform women’s decisions to use PBS, and no studies have examined these processes on days when women drink and may be at increased risk for victimization. A better understanding of the malleable constructs that inform PBS use on drinking days is needed to enhance prevention efforts.

Decisions to Use PBS

Research shows women’s decisions to use PBS can be complex, operating through core processes that are both planned (i.e., intentional) and emotionally reactive (i.e., one’s willingness to use PBS based on the social situation; Gerrard et al., 2008). Studies reveal that intentions and willingness tend to be positively associated, but when examined simultaneously, they each account for significant unique variance in predicting behavioral outcomes (e.g., Mallett et al., 2015; Scaglione et al., 2015). Scaglione and colleagues (2015) observed women’s intentions and willingness were positively correlated with each other (r = .78), but each accounted for unique and significant proportions of variance in predicting future sexual assault PBS use. The role of intentions in PBS use is consistent with prior literature that has shown positive health behaviors (e.g., healthy eating) tend to be more intentional or planned compared to health risk behaviors (e.g., substance use), which tend to be more reactive (Gerrard et al., 2008; Pomery et al., 2009). Studies such as these tend to examine associations between intentions/willingness assessed at one time and behavioral outcomes assessed a month to several months later. The limitation of this approach is that it does not account for situational changes in decision making or PBS use. The current study attempted to fill this gap by examining event-level intentions and willingness to use PBS and behavioral outcomes on days when women might be at increased risk for victimization.

Ecological momentary assessment (EMA) is an event-level methodology that allows for repeated measurement of events, experiences, or behaviors as they occur in real time and in participants’ natural environments (Shiffman et al., 2008). The EMA approach provides a method to examine the answers to more nuanced process-based questions, such as: (a) Does the influence of decision-theoretical constructs (intentions and willingness) on PBS use vary from day to day? (b) Does the influence vary within a day leading up to a drinking event? (c) Is the influence of within- or between-day variability affected by individuals’ typical decision-making patterns? Answers to these questions could provide insight into factors that decrease risk on a given day and inform the optimal timing and theoretical targets for interventions. For example, observed variability in the influence of intentions on PBS use from one day to the next might suggest existing risk reduction programs could be enhanced by encouraging women to consistently plan for what PBS they might use when they go out. Alternatively, observing variability within a day (e.g., if PBS use becomes more likely as intentions increase throughout the day) might suggest the need for supplemental event-specific interventions that make PBS use seem appealing right before or throughout a night out. Understanding these processes can help inform the nature of prevention efforts to increase overall PBS use, empower women to think about which PBS may be most beneficial in different situations, and subsequently reduce harm.

The Current Study

The current study used a 14-day EMA design to examine decision-making constructs associated with sexual assault PBS and drinking PBS. To understand the decision-making processes women use on days when they could be at increased risk for victimization, the current study examined PBS used only on drinking days. Further, drinking days are plausible and easily identifiable targets for intervention, both globally and at the event-level. Research suggests that, on average, women use more sexual assault PBS when they have greater intentions and willingness to do so (Scaglione et al., 2015). The current study aimed to explore these associations at the event level for both sexual assault and drinking PBS by answering the following research questions (RQ): (a) Do women use more sexual assault PBS and drinking PBS on days when their intentions and willingness to do so are higher than normal (e.g., within-person/between-day effects)? (b) Do women use more sexual assault PBS and drinking PBS on days when their intentions and willingness to do so increase throughout the day, relative to when they decrease or stay the same (e.g., within-day effects)? (c) Do event-level associations between PBS use and decision-theoretical constructs vary based on one’s average overall levels of intentions and willingness (e.g., between-person by within-person interaction effects)? These hypothesized associations were tested separately for sexual assault and drinking PBS because they may require unique intervention approaches for increased use. The effects of intentions and willingness were examined simultaneously (e.g., the effect of willingness controlling for one’s average level of intentions) because previous research has identified both as being integral and complementary in the behavioral decision-making process (Mallett et al., 2015; Scaglione et al., 2015). Answers to these questions will provide critical insight into how typical risk reduction approaches might be enhanced to address HED and event-level risk.

Method

Recruitment

A random sample of 750 women was selected from the university registrar’s database of incoming students at a large, public northeastern university. Students received personalized introductory letters and email invitations describing the study and containing a URL and PIN for accessing a consent form and brief web-based screening survey. Non-respondents received up to seven personalized email reminders. Approximately, 58% (n = 436) enrolled and completed the screening, resulting in a response rate comparable to other studies using similar web-based recruitment methods with new students transitioning into the university (e.g., Mallett et al., 2015; Turrisi et al., 2013). Over half of enrolled students met eligibility criteria (54%; n = 235), which included (a) having a 4G network smartphone and (b) reporting at least one heavy drinking episode in the past month (i.e., consuming four or more drinks on a single occasion; Hingson et al., 2017). Ineligible students were paid $5 and removed from the participant list. Eligible students were directed to the baseline assessment and paid $15 for their time. Students also earned a $5 rapid completion bonus if they completed the survey within 3 days of the initial invitation.

EMA Study Procedures

Following baseline assessment, two-thirds of participants (n = 156) were randomly selected for the EMA study; the remaining third was assigned to a reactivity control group. Students assigned to the current study were directed to a scheduling tool and invited to continue in the study by signing up for a required training on Snap Mobile, a mobile phone application used for daily data collection. Research staff made up to two phone calls and sent three additional emails to follow up with eligible students who did not schedule electronically or to reschedule participants who missed their scheduled session; participants also received text message reminders the night before their selected training session. Approximately, 54% (n = 84) of those invited, or 19% of all screened participants, attended training and enrolled in the EMA study. There were no significant differences between students who enrolled and students who did not enroll on age, race/ethnicity, baseline drinking, or PBS use (all ps > .05).

The current study employed a 14-day event- and signal-contingent EMA protocol. Upon waking each morning, participants logged into their Snap Mobile app to complete a short survey of the previous day’s drinking and PBS use. At the end of this morning survey, participants also completed their first prospective momentary assessments of intentions and willingness to use PBS on that day. Two additional momentary assessments of intentions and willingness occurred randomly between 2:00 pm and 6:00 pm and between 7:00 pm and 11:00 pm each day. Text message reminders were sent at 9:00 am on weekdays/10:00 am on weekends, and again at 12:30 pm for daily diaries (the survey closed at 1:00 pm daily); reminders for the afternoon and evening momentary assessments were sent 30 minutes before each survey period closed. Participants received $30 for participation, with a $20 bonus if they achieved at least 90% compliance on the daily morning assessments. To further encourage compliance, participants were entered into a lottery for each survey they completed; five lottery winners were randomly selected at the end of the study, with prizes ranging from $50 to $250. All study procedures were reviewed and approved by the university’s Institutional Review Board.

Participants

Only the days on which participants reported drinking were included in analyses, as the current study aimed to understand PBS use on drinking days. The final sample included 56 participants who reported at least 1 drinking event during the 14-day assessment period with corresponding momentary data from the prior day. The mean age of the sample was 18.05 years (SD = 0.23). Most participants identified as Caucasian (82.1%), although other racial/ethnic backgrounds were represented (10.7% African American, 5.4% multiracial, 5.5% Hispanic). All participants lived on campus, and many intended to join intramural/club sports (55.4%) or sororities (42.9%) in their first semester. At baseline, participants reported consuming an average of 8.71 (SD = 6.06) drinks per week.

Measures

PBS

If participants reported drinking the day/night before, within their daily waking recall assessments, they were asked to indicate which, if any, PBS they used (1 = Yes; 0 = No). For both types of PBS, briefer versions of established scales were used to keep daily assessments short and minimize response burden (Stone et al., 1991).

The sexual assault PBS measure included six items from the Dating Self-Protection Against Rape Scale (Moore & Waterman, 1999). Items were selected based on preliminary work identifying the types of behaviors women tend to use most frequently (Scaglione et al., 2015). Sample items included, “Letting a friend or family member know where you are,” and “Walking home with a trusted friend.” A summed score was created to indicate the number of sexual assault PBS used on each drinking occasion.

The drinking PBS measure included six items from the Protective Behavioral Strategies Scale (Martens et al., 2005). Items with factor loadings of at least 0.65 (Martens et al., 2007) were selected from each subscale, including Stopping/Limiting Drinking (e.g., determined in advance not to exceed a set number of drinks), Manner of Drinking (e.g., avoided trying to keep up with or out drink others), and Serious Harm Reduction (e.g., tried to know where my drink was at all times). A summed score was created to indicate the number of drinking PBS used on each drinking occasion.

Behavior-specific Intentions

Each of the three daily momentary assessments asked participants about the activities they intended to engage in later that day. They were asked, “Do you ‘intend’ to engage in any social activities today/tonight?” (1 = Yes; 0 = No) and “Do you ‘intend’ to drink?” (1 = Yes; 0 = No). If drinking intentions were affirmed, participants also indicated which, if any, PBS from the lists above they intended to use (1 = Yes; 0 = No). Summed scores were created to indicate the number of sexual assault PBS and drinking PBS participants intended to use within each drinking occasion.

Behavior-specific Willingness

Similar to intentions, participants were presented with willingness questions specific to each behavior. They were asked, “Are you ‘willing’ to engage in social activities tonight, should the opportunity arise?” (1 = Yes; 0 = No) and “Are you ‘willing’ to drink?” (1 = Yes; 0 = No). If affirmed, participants indicated which sexual assault and drinking PBS they were willing to use (1 = Yes; 0 = No). Summed scores indicated the number of sexual assault PBS and drinking PBS participants were willing to use on each occasion.

Analytic Strategy

A series of four multilevel regression models were estimated in SAS (v. 9.4) to examine within-person/between-day effects (RQ1), within-day effects (RQ2), and cross-level interaction effects (RQ3; between-person X within-person/day) of decision-theoretic constructs (intentions and willingness) on daily PBS use. Prior to analysis, waking recall assessments of the prior day’s PBS use were lagged by one day and merged to match with the previous day’s momentary assessments of intentions and willingness. Because predictor variables were assessed at a lower level (within day) than the outcome variables (between day), the lower level data (three momentary assessments of intentions and willingness within a day) were aggregated (Schwartz & Stone, 1998). Within-day aggregation was conducted in two ways to answer unique research questions. First, all three momentary assessments were averaged to create a day-level mean for intentions and willingness to use sexual assault and drinking PBS, respectively (RQ1). The second aggregation method created behavior-specific change scores (Allison, 1990) by subtracting intention and willingness values reported at the first assessment within each day from the value reported at the third assessment within each day (RQ2). Positive values indicated an increase in intentions or willingness throughout the day, whereas negative values indicated a decrease. Missing data within the final analytic sample were handled using Full Information Maximum Likelihood (Walls et al., 2006), and outcome-specific unconditional means models confirmed the utility of the intended multilevel approach.

The first set of models examined whether daily intentions and willingness to use PBS predicted actual PBS use on a given drinking day. Each behavioral outcome (sexual assault PBS and drinking PBS) was regressed onto behavior-specific day-level means of intentions and willingness. For example, sexual assault PBS use was regressed onto day-level means of intentions and willingness to use sexual assault PBS. To separate within- and between-person effects, predictors were group-mean centered at Level 1 (i.e., centered around the individual’s mean) and grand-mean centered at Level 2 (i.e., centered around the sample mean). A second set of models examined whether within-day changes in intentions and willingness to use PBS predicted PBS use on a given drinking day. Again, using separate models for sexual assault PBS and drinking PBS, each behavioral outcome was regressed onto behavior-specific within-day changes in intentions and willingness to use PBS at Level 1 and grand-mean centered intentions and willingness to use PBS at Level 2. Within each model, we estimated both main effects (RQ1 & RQ2) and within-domain cross-level interactions (e.g., Level 1 intentions × Level 2 intentions; Level 1 willingness × Level 2 willingness; RQ3).

Results

Participants provided 368 total assessments (M = 6.57/person). Intraclass correlation coefficients from the unconditional means models indicated .85 of the variability in sexual assault PBS and .78 of the variability in drinking PBS occurred between-persons, leaving 15% and 22% of the variability, respectively, to be accounted for at the event-level. Results from the multilevel models predicting both sexual assault and drinking PBS are included in Table 1. For each outcome, results of RQ1 (between-day effects of intentions and willingness) are presented in the top panel and results of RQ2 (within-day changes in intentions and willingness) are presented in the bottom panel. The cross-level interaction terms within each panel (Rows 6, 7, 12, and 13) represent results for RQ3. Each will be discussed in turn.

Table 1.

Observed Effects by Outcome.

		Coefficient (SE) for Each Outcome
Row		Number of Sexual Assault PBS Used	Number of Drinking PBS Used
RQ 1 (Daily Average of Intentions and Willingness)
1	Intercept, γ₀₀	2.76 (0.12)***	2.90 (0.17)***
2	Between-person intentions, γ₀₁	0.34 (0.12)**	0.30 (0.16)
3	Between-person willingness, γ₀₂	0.50 (0.13)***	0.38 (0.15)*
4	Within-person (daily average) intentions, γ₁₀	0.10 (0.10)	0.18 (0.25)
5	Within-person (daily average) willingness, γ₂₀	−0.16 (0.09)	−0.70 (0.55)
6	Between-person intentions × within-person intentions, γ_{11 (RQ3)}	−0.05 (0.06)	0.12 (0.13)
7	Between-person willingness × within-person willingness, γ_{21 (RQ3)}	−0.03 (0.08)	0.58 (0.29)*
RQ 2 (Within-Day Changes in Intentions and Willingness)
8	Intercept, γ₀₀	2.76 (0.13)***	2.74 (0.19)***
9	Between-person intentions, γ₀₁	0.37 (0.12)**	0.28 (0.17)
10	Between-person willingness, γ₀₂	0.47 (0.14)***	0.39 (0.16)*
11	Within-day change in intentions, γ₁₀	−0.07 (0.05)	0.12 (0.11)
12	Within-day change in willingness, γ₂₀	0.15 (0.06)**	−0.02 (0.12)
13	Between-person intentions × within-day change in intentions, γ_{11 (RQ3)}	−0.03 (0.04)	−0.09 (0.07)
14	Between-person willingness × within-day change in willingness, γ_{21 (RQ3)}	−0.10 (0.05)*	0.09 (0.08)

Note. ***p < .001. **p < .01. *p < .05.

RQ = Research question; RQ1 is addressed via effects reported in the top panel of the table; RQ2 is addressed via effects reported in the bottom panel, and RQ3 is addressed via the interaction effects reported in the last two rows of each panel.

RQ1: Do Women Use More PBS on Days When Their Intentions and Willingness to Do So Are Higher Than Normal?

On any given drinking day, assuming average levels of PBS intentions and willingness, women used 2.76 (SE = 0.12) sexual assault PBS and 2.90 (SE = 0.17) drinking PBS (Table 1, Row 1). The main effects of daily intentions and willingness to use PBS on PBS use were not significant for either outcome (Rows 4 and 5). Thus, there was no evidence to suggest women used more sexual assault PBS or drinking PBS on days when their intentions and willingness to do so were higher than normal.

RQ2: Do Women Use More PBS on Days When Their Intentions and Willingness to Do So Increase Throughout the Day?

Consistent with the first set of models, on any given drinking day, and assuming average intentions and willingness, women used 2.76 (SE = 0.13) sexual assault PBS and 2.74 (SE = 0.19) drinking PBS (Table 1, Row 8). There was a positive association between within-day changes in willingness and sexual assault PBS use (Row 12), such that at average levels of intentions, each 1-unit increase in willingness was associated with a 0.15 (SE = 0.06) increase in PBS use. In other words, women reported their highest sexual assault PBS use on days when their willingness to use sexual assault PBS increased throughout the day. Within-day changes in intentions to use sexual assault PBS were not significantly associated with sexual assault PBS use (Row 11). There were no significant associations between within-day changes in intentions and willingness and drinking PBS use (Rows 11 and 12).

RQ3: Do Event-level Associations Between PBS Use and Decision-theoretical Constructs Vary Based on One’s Average Overall Levels of Intentions and Willingness?

Sexual Assault PBS

Across both models, and consistent with global theories of behavior change (e.g., Gerrard et al., 2008), the between-person effects of both intentions and willingness to use sexual assault PBS were positively associated with number of PBS used (Table 1, Rows 2, 3, 9, and 10). Individuals who reported greater intentions and willingness to use sexual assault PBS, on average, also reported using a greater number of sexual assault PBS, compared to individuals who reported lower average intentions and willingness to use sexual assault PBS.

Building on this work, RQ3 examined whether event-level associations between PBS use and decision-theoretical constructs vary based on one’s average overall levels of intentions and willingness. The between-person by daily average interactions were not significant (Rows 6 and 7), suggesting the non-significant event-level associations described above (RQ1) were consistent across all women in the sample; they did not vary based on one’s typical levels of intentions and willingness. However, the between-person by within-day change interaction was significant for willingness to use sexual assault PBS (Row 14), suggesting the positive association between within-day changes in willingness and sexual assault PBS use (RQ2 above) might vary depending on one’s typical willingness to use sexual assault PBS.

To probe the nature of this interaction, the interaction was graphed in Figure 1. Separate lines are used to indicate the within-day association among individuals who reported (a) average (mean), (b) above-average (1 SD above the mean), and (c) below-average (1 SD below the mean) levels of overall willingness to use sexual assault PBS. The interaction was further probed by calculating the simple slopes and region of significance for each of the three between-person levels of sexual assault PBS willingness (Preacher et al., 2006). Women who had average and below-average levels of overall willingness to use sexual assault PBS reported increased sexual assault PBS use on days when their willingness to use those PBS increased throughout the day (p < .01). This association was not significant for women with overall above-average levels of sexual assault PBS willingness; these women used the highest number of sexual assault PBS regardless of any within-day fluctuations in willingness.

Figure 1.

Note. SA-PBS = Sexual assault protective behavioral strategies.

Drinking PBS.

Across both models, the between-person association (Level 2) between willingness to use drinking PBS and PBS use was significant (Table 1, Rows 3 and 10), suggesting that individuals who, on average, reported willingness to use more drinking PBS also reported using a greater number of drinking PBS over the period of assessment. There was a significant cross-level interaction of between-person by daily average willingness to use drinking PBS (Row 7), suggesting that even though no event-level association between willingness and drinking PBS use was found when we examined main effects (RQ1 above), there may be an event-level association for certain individuals. Further probing (see Figure 2) revealed individuals who, on average, reported the lowest overall level of drinking PBS willingness (i.e., below-average) decreased their PBS use on days when they reported more than their typical amount of willingness to use drinking PBS (p < .05). The simple slopes for the other between-person levels of drinking PBS willingness (i.e., average and above-average) trended in the expected direction but were not significant. No other significant interaction effects were observed; intentions did not predict drinking PBS use at any level.

Figure 2.

Note. D-PBS = Drinking protective behavioral strategies.

Discussion

The current study aimed to enhance existing risk reduction programs by examining event-level variability in the decision-making processes that influence PBS use using an EMA study design. First, we examined whether women used more sexual assault and drinking PBS on days when their intentions and willingness to do so were higher than normal. Findings did not support this hypothesis. This was surprising, as previous work indicates women regularly use a number of sexual assault PBS already (Scaglione et al., 2015), suggesting they are generally aware of social and environmental risks. One notable difference from this prior work was that our study focused solely on women in their first semester of college. These women may still be learning about social and environmental risk factors, suggesting a need to understand how PBS use and motivations for use differ between first year students and their more senior counterparts. At the event-level, specific contextual factors may impact women’s risk perceptions and, in turn, their decisions to use sexual assault PBS on a given day. Women tend to be less willing to use sexual assault PBS in situations where they are socializing with someone they know because they do not think of that person as a potential perpetrator (Clay-Warner, 2002). Students may also be less willing to use drinking PBS if they have competing intentions to drink, as drinking PBS focus on limiting alcohol consumption. This effect might be heightened by affiliation with groups that support a heavy drinking culture (e.g., athletic teams, Greek life). Future event-level studies of PBS decision making should examine the role of contextual factors, such as with whom and in what environments women socialize, and between-person differences that might further explain event-level processes.

Next, we asked whether women used more sexual assault and drinking PBS on days when their intentions and willingness to do so increased throughout the day relative to days when their intentions and willingness decreased or stayed the same. Whereas no significant findings were observed for drinking PBS, women reported their highest sexual assault PBS use on days when their willingness to use sexual assault PBS increased throughout the day. This might suggest that women adjust their sexual assault PBS as they coordinate with friends or learn more about their plans through the day (e.g., where they are going, who they will see). Future research should explore what factors impact shifts in sexual assault PBS willingness, as well as between-person differences in who adjusts their willingness and PBS use versus those who do not. Interventions might be enhanced by encouraging women to learn as much as possible about the events they plan to attend (e.g., where they are going, what could happen in that setting) so they can proactively consider which PBS they might be willing to use in each situation.

The third research question examined whether event-level associations between PBS use and willingness and intentions varied based on individuals’ average levels of willingness and intentions. Consistent with global theories of behavior change and prior PBS-specific studies (e.g., Gerrard et al., 2008; Scaglione et al., 2015), we did observe significant between-person effects such that women who reported more willingness to use both types of PBS, on average, also endorsed greater use. The between-person effect of intentions was significant for sexual assault PBS only, indicating women who reported higher average intentions were more likely to use PBS such as walking home with a friend. The impact of these between-person effects on event-level decision making varied based on the PBS type. For sexual assault PBS, a significant between-person by within-day change interaction was observed. Women who had below-average and average levels of willingness to use sexual assault PBS reported the highest actual use on days when their willingness to use sexual assault PBS increased; this pattern was not observed for women with above-average levels of overall willingness. Examination of Figure 1 suggests we may be observing a ceiling effect among above-average sexual assault PBS users. These women are regularly using sexual assault PBS at high rates with little room to increase. Findings suggest women who typically are less willing to use sexual assault PBS may benefit from event-level types of intervention to increase their willingness prior to higher risk occasions involving alcohol or potential sexual encounters (Heron & Smyth, 2010), and that existing prevention efforts could focus on positively reinforcing the PBS use that is already happening.

When examining cross-level interactions for drinking PBS, a different pattern of findings emerged. We observed a significant between-person by daily average interaction for willingness to use drinking PBS; women who tended to have the lowest overall willingness to use drinking PBS actually decreased their drinking PBS use on days when they reported more willingness to use them. There are several possible explanations for this finding that should be explored in future work. Perhaps these women are less successful with implementing these behaviors because they may be less familiar with using them or simply do not like to use them. It is also plausible that despite being more willing to use drinking PBS in advance of a drinking event, they may find themselves in an environment where they perceive none of their peers are using them, or that none of their peers would approve of their use. These women may be particularly affected by both drinking and PBS-specific normative influences (Lewis et al., 2009). Either way, the findings are somewhat concerning because they highlight a lack of planning to use drinking PBS, and a lack of follow-through even when willingness is high. Interventions may need to strengthen the use of drinking PBS on a global level before event-level interventions are considered, and additional research may be needed to better understand normative influences and barriers associated with drinking PBS use.

Notably, intraclass correlations revealed less event-level variability in PBS use than we expected to observe (15% for sexual assault PBS; 22% for drinking PBS), which may highlight several ways in which our study design could be strengthened in subsequent research. First, to minimize participant burden we chose to focus on college students’ most commonly used PBS, yet decision-making processes may be different for less common PBS, especially if those are PBS women only use out of necessity (e.g., if a situation escalates and requires forceful physical resistance; Edwards et al., 2014). Expanding the number and types of PBS measured may capture the full range of strategies women choose to implement. More nuanced understanding of how women combine PBS and which PBS women use in different contexts would also better inform EMA measurement, and ultimately prevention (Sell et al., 2018). Second, post-hoc comparison of EMA participants to a reactivity control group suggested women may have increased PBS use in response to frequent assessment. While this is encouraging when considering the utility of momentary/event-level interventions to increase PBS use, it could have also resulted in reduced variability, or even over-reporting, of PBS use in daily assessments. Finally, we aimed to understand decisions leading up to drinking events, necessarily limiting analysis to drinking days. However, not all sexual assaults involve alcohol, and a woman’s decision not to drink on a given day or in a specific context may be an intentional PBS. We may capture more event-level variability by examining a broader range of PBS and decisions to use PBS on both drinking and non-drinking days.

Implications for Prevention

Previous research has called for the integration and increased use of both sexual assault PBS and drinking PBS to enhance existing sexual assault prevention efforts (Gilmore et al., 2015; Sell et al., 2018). By examining women’s decisions to use (or not use) PBS on drinking days, the current study serves as an initial step in providing an evidence-based way to help women utilize a “toolkit” of strategies that will help them to feel empowered and in control of their environment, while also reducing the likelihood they will be perceived by perpetrators as potential victims. Overwhelmingly, the current study highlights the need to increase intentions to use drinking PBS. Practitioners may consider using existing alcohol risk reduction strategies (e.g., motivational interviewing [MI; Dimeff et al., 1999]; personalized normative feedback [PNF; Lewis et al., 2014]; parent-based intervention [PBI; Turrisi et al., 2013]) and sexual assault risk reduction programs targeting heavy drinkers (e.g., Gilmore et al., 2015) to encourage women to think through common social scenarios and plan for which PBS they would use to minimize potential risk. Conversations that address barriers to using drinking PBS, like competing intentions to drink, might be enhanced by providing alternative strategies that can be used even if intentions to drink are high (e.g., alternating alcoholic and non-alcoholic beverages vs. limiting to three or fewer drinks). Senn and colleagues (2015) have successfully used similar approaches to increase women’s use of forceful verbal and physical sexual assault resistance strategies in their Enhanced Assess, Acknowledge, Act (EAAA) program.

Findings also revealed that intentions and willingness to use PBS operate at both between-person and within-person levels and may require different strategies to achieve behavior change. To address between-person differences, individuals who are more willing to use PBS may benefit from highlighting successful outcomes from prior PBS use to reinforce positive decisions. Those who are less willing to use PBS may require more intensive efforts to increase their willingness to use PBS, perhaps by first increasing their confidence in their ability to effectively use PBS (Scaglione et al., 2015). A similar approach to target within-person differences might be to reinforce intentions/willingness throughout the day, during an actual drinking event, or even prior to a potential sexual encounter, via text messages or interactive mobile apps. Because women are generally willing to use sexual assault PBS (Scaglione et al., 2015) and may increase use in response to frequent mobile assessment, interventions that reach them in the moments when they need to be used (or immediately before) may increase use during higher-risk occasions. Prevention scientists might consider supplementing efficacious individual-level interventions (e.g., MI, PNF, PBI, EAAA) with adaptive mobile components that aim to increase or maintain high levels of event-level PBS use. Event-level evaluation of PBS use following program exposure would provide valuable insight into how to structure supplemental intervention messaging, while capturing program impacts on PBS use that are not captured using more traditional longitudinal evaluation methods.

Finally, practitioners and scientists alike might consider the utility of both sexual assault and drinking PBS in their prevention programs and research. Previous research shows women use more sexual assault PBS on days when they drink (Sell et al., 2018). This suggests sexual assault PBS may be particularly appealing and more likely to be used among drinkers. These strategies may enhance alcohol reduction programs’ impact on sexual assault outcomes (e.g., Clinton-Sherrod et al., 2011; Testa et al., 2010a). Further, sexual assault risk reduction programs are often sought out by women with prior victimization experience (Senn et al., 2015), who may be at increased risk for heavy drinking (Testa et al., 2010b). Whereas most programs address alcohol as a situational risk factor, more robust discussion of drinking PBS may be useful for these women. Regardless of the setting, discussions of PBS use should be coupled with reminders that these strategies aim to reduce risk, but may not completely eliminate it. It is equally important to highlight that if women experience sexual assault even after using PBS, that does not imply the assault was their fault; sexual assault is always caused by a perpetrator. In a continued effort to shift the responsibility of sexual assault prevention away from potential victims, practitioners should implement both alcohol and sexual assault risk reduction programs in combination with efficacious programs that reduce community-level risk (e.g., bystander intervention; Orchowski et al., 2016) and perpetration behaviors (Orchowski et al., 2018).

Limitations and Future Directions

Although careful controls were in place throughout the study, certain limitations remain. Our sample size was sufficient to address the proposed research questions, but findings would be strengthened from replication in a larger, more diverse sample of women who engage in different patterns of drinking (abstainers, light, moderate, and heavy). Including light and non-drinkers in future research would be beneficial considering they may still be at risk when they are in the presence of others who are drinking. Decision-making patterns also should be examined among different groups of students we know to be at increased risk for alcohol-related problems and/or sexual victimization (e.g., LGBTQ, international, racial/ethnic minority and first-generation students; those involved in Greek Life or with a history or prior victimization), as they may have unique prevention needs. Although a significant portion of our sample indicated intentions to join Greek Life, our data do not indicate who actually joined, and we were underpowered for the three-way interactions required to reliably examine group differences. Although EMA compliance was high (>90%), initial response and enrollment rates could be improved, perhaps by offering online/virtual training or higher incentives. To minimize respondent burden, the current study examined a limited set of theoretical constructs (i.e., intentions and willingness). Additional work is needed to examine contextual and intrapersonal decision-making factors (e.g., norms, attitudes, expectancies, perceived situational risk) at the event level to better understand malleable constructs that can be targeted by interventions. Further, we were unable to link PBS use to sexual assault outcomes because there was limited endorsement of unwanted sexual experiences on any of the 14 measured occasions. Although these associations are well-established using cross-sectional and non-diary longitudinal designs (Gilmore et al., 2018; Mallett et al., 2015), linking PBS use and sexual assault outcomes at the event-level would uniquely inform prevention. Such a study would likely require a longer EMA data collection period (e.g., 30 days) and an event-contingent assessment method to capture outcomes tied to specific PBS use. Finally, our data were exclusively self-reported. The intrusive nature of repeated self-report data collection has been identified as a potential source of bias (e.g., reactivity to study procedures) and a threat to ecological validity (Ram et al., 2017). This work could be enhanced by incorporating unobtrusive data collection techniques, such as video diaries, biosensors, or geolocation data, which have the potential to eliminate self-report biases, provide robust monitoring of alcohol consumption, and add to our contextual understanding of PBS use (e.g., how decisions to use PBS are made at different levels of intoxication; whether decisions vary across social contexts or physical locations).

Conclusion

The current study was the first to examine event-level decisions to use both sexual assault and drinking PBS. The findings demonstrated decision-making constructs related to using sexual assault and drinking PBS among college women vary between people and at the event level, and unique intervention approaches may be needed to increase specific types of PBS. Findings from the study also highlight the need for both global and event-level intervention approaches to be tailored to women based on their current willingness and intentions to utilize PBS rather than using a one-size-fits-all approach.

Footnotes

Author’s Note

This research was supported by the National Institutes of Health National Institute on Alcohol Abuse and Alcoholism grant F31 AA022274 awarded to Nichole Scaglione while at Pennsylvania State University. The authors would like to thank Dr Liana Hone and Ms Jessica Cohen for their invaluable feedback on earlier drafts of this manuscript.

Declaration of Conflicting Interests

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the National Institutes of Health National Institute on Alcohol Abuse and Alcoholism grant F31 AA022274 awarded to Nichole Scaglione while at Pennsylvania State University.

ORCID iD

Nichole M. Scaglione

Author Biographies

Nichole M. Scaglione, PhD, CHES, is an assistant professor in the Department of Health Education and Behavior at the University of Florida. Her work is aimed at preventing sexual assault and substance use in diverse adolescent and young adult populations. She is interested in the decision-making processes that influence the use of protective behaviors that could be used in prevention efforts to reduce sexual assault.

Rob Turrisi, PhD, is a professor at Penn State University with a joint appointment in the Department of Biobehavioral Health and the Prevention Research Center. Dr Turrisi’s work focuses on developing and evaluating parent and individual-level prevention efforts to reduce alcohol abuse and alcohol-related consequences for college students and emerging adults.

Nichole M. Sell, PhD, is a postdoctoral fellow at the Clinical and Research Institute on Addictions at the University at Buffalo. Her recent research focuses on examining the role of alcohol, sexual risk behaviors, and protective behavioral strategies (PBS) in women’s experiences of sexual victimization. In her future work, Nichole hopes to develop approaches to enhance PBS’ efficacy in reducing negative alcohol-related sexual outcomes (e.g., sexual assault, unwanted or regretted sexual experiences).

Kimberly A. Mallett, PhD, is a research professor in the Prevention Research Center at Penn State University. Her research focus is on the application of brief, motivationally enhanced behavioral interventions to prevent (a) harms resulting from high-risk drinking, and (b) the development of skin cancers.

Michael J. Cleveland, PhD, is currently an associate professor of Human Development at Washington State University. His research examines how social contextual factors influence health behaviors, with emphasis on etiology and prevention of alcohol and other substance use among emerging adults. Dr Cleveland is particularly interested in adapting evidence-based interventions to better serve the needs of diverse groups of youth during the high-risk transition out of high school.

References

Abbey

, Ross

L. T.

, McDuffie

, & McAuslan

(1996). Alcohol and dating risk factors for sexual assault among college women. Psychology of Women Quarterly , 20, 147–169.

Abbey

, Zawacki

, Buck

P. O.

, Clinton

A. M.

, & McAuslan

(2004). Sexual assault and alcohol consumption: What do we know about their relationship and what types of research are still needed? Aggression and Violent Behavior , 9(3), 271–303.

Allison

P. D.

(1990). Change scores as dependent variables in regression analysis. Sociological Methodology , 20, 93–114.

Clay-Warner

(2002). Avoiding rape: The effects of protective actions and situational factors on rape outcome. Violence and Victims , 17(6), 691–705.

Clinton-Sherrod

, Morgan-Lopez

A. A.

, Brown

J. M.

, McMillen

B. A.

, & Cowell

(2011). Incapacitated sexual violence involving alcohol among college women: The impact of a brief drinking intervention. Violence Against Women , 17(1), 135–154.

Dimeff

L. A.

, Baer

J. S.

, Kivlahan

D. R.

, & Marlatt

G. A.

(1999). Brief alcohol screening and intervention for college students: A harm reduction approach . Guilford Press.

Edwards

K. M.

, Probst

D. R.

, Tansill

E. C.

, Dixon

K. J.

, Bennett

, & Gidycz

C. A.

(2014). In their own words: A content-analytic study of college women’s resistance to sexual assault. Journal of Interpersonal Violence , 29(14), 2527–2547.

Fisher

B. S.

, Cullen

F. T.

, & Turner

M. G.

(2000). The sexual victimization of college women . Department of Justice, Bureau of Justice Statistics.

George

W. H.

, & Stoner

S. A.

(2000). Understanding acute alcohol effects on sexual behavior. Annual Review of Sex Research , 11(1), 92–124.

10.

Gerrard

, Gibbons

F. X.

, Houlihan

A. E.

, Stock

M. L.

, & Pomery

E. A.

(2008). A dual-process approach to health risk decision making: The prototype willingness model. Developmental Review , 28(1), 29–61.

11.

Gidycz

C. A.

(2018). Sexual assault risk reduction: Current state and historical underpinnings. In L. M. Orchowski & C. A. Gidycz (Eds.), Sexual assault risk reduction and resistance (pp. 3–14). Academic Press.

12.

Gidycz

C. A.

, Orchowski

L. M.

, Probst

D. R.

, Edwards

K. M.

, Murphy

, & Tansill

(2015). Concurrent administration of sexual assault prevention and risk reduction programming: Outcomes for women. Violence Against Women , 21(6), 780–800.

13.

Gilmore

A. K.

, Lewis

M. A.

, & George

W. H.

(2015). A randomized controlled trial targeting alcohol use and sexual assault risk among college women at high risk for victimization. Behaviour Research and Therapy , 74, 38–49.

14.

Gilmore

A. K.

, Maples-Keller

J. L.

, Pinsky

H. T.

, Shepard

M. E.

, Lewis

M. A.

, & George

W. H.

(2018). Is the use of protective behavioral strategies associated with college sexual assault victimization? A prospective examination. Journal of Interpersonal Violence, 33 (17), 2664–2681. http://doi.org/10.1177/0886260516628808

15.

Heron

K. E.

, & Smyth

J. M.

(2010). Ecological momentary interventions: Incorporating mobile technology into psychosocial and health behaviour treatments. British Journal of Health Psychology , 15(1), 1–39.

16.

Hingson

R. W.

, Zha

, & White

A. M.

(2017). Drinking beyond the binge threshold: Predictors, consequences, and changes in the US. American Journal of Preventive Medicine , 52(6), 717–727.

17.

Krebs

, Lindquist

, Berzofsky

, Shook-Sa

, & Peterson

(2016). Campus climate survey validation study: Final technical report . Department of Justice, Bureau of Justice Statistics. https://www.bjs.gov/content/pub/pdf/ccsvsftr.pdf

18.

Lewis

M. A.

, Patrick

M. E.

, Litt

D. M.

, Atkins

D. C.

, Kim

, Blayney

J. A.

, Norris

, George

W. H.

, & Larimer

M. E.

(2014). Randomized controlled trial of a web-delivered personalized normative feedback intervention to reduce alcohol-related risky sexual behavior among college students. Journal of Consulting and Clinical Psychology , 82(3), 429–440.

19.

Lewis

M. A.

, Rees

, & Lee

C. M.

(2009). Gender-specific normative perceptions of alcohol-related protective behavioral strategies. Psychology of Addictive Behaviors , 23(3), 539–545.

20.

Mallett

K. A.

, Turrisi

, Cleveland

M. J.

, Scaglione

N. M.

, Reavy

, Sell

N. M.

, & Varvil-Weld

(2015). A dual-process examination of alcohol-related consequences among first-year college students. Journal of Studies on Alcohol and Drugs , 76(6), 862–871.

21.

Martens

M. P.

, Ferrier

A. G.

, Sheehy

M. J.

, Corbett

, Anderson

D. A.

, & Simmons

(2005). Development of the protective behavioral strategies survey. Journal of Studies on Alcohol , 66(5), 698–705.

22.

Martens

M. P.

, Pederson

E. R.

, LaBrie

J. W.

, Ferrier

A. G.

, & Cimini

M. D.

(2007). Measuring alcohol-related protective behavior strategies among college students: Further examination of the Protective Behavior Strategies Scale. Psychology of Addictive Behaviors , 21(3), 307–315.

23.

Moore

C. D.

, & Waterman

C. K.

(1999). Predicting self-protection against sexual assault in dating relationships among heterosexual men and women, gay men, lesbians, and bisexuals. Journal of College Student Development , 40(2), 132–140.

24.

Moorer

K. D.

, Madson

M. B.

, Mohn

R. S.

, & Nicholson

B. C.

(2013). Alcohol consumption and negative sex-related consequences among college women: The moderating role of alcohol protective behavioral strategies. Journal of Drug Education , 43(4), 365–383.

25.

Neilson

E. C.

, Gilmore

A. K.

, Pinsky

H. T.

, Shepard

M. E.

, Lewis

M. A.

, & George

W. H.

(2018). The use of drinking and sexual assault protective behavioral strategies: Associations with sexual victimization and revictimization among college women. Journal of Interpersonal Violence , 33(1), 137–158.

26.

Orchowski, L.

, Barnett, N.

, Berkowitz

, Borsari

, Oesterle

, & Zlotnick

(2018). Sexual assault prevention for heavy drinking college men: Development and feasibility of an integrated approach. Violence Against Women , 24(11), 1369–1396.

27.

Orchowski, L.

, Berkowitz

, Boggis

, & Oesterle

(2016). Bystander intervention among college men: The role of alcohol and correlates of sexual aggression. Journal of Interpersonal Violence , 31(17), 2824–2846.

28.

Orchowski

L. M.

, Edwards

K. M.

, Hollander

J. A.

, Banyard

V. L.

, Senn

C. Y.

, & Gidycz

C. A.

(2020). Integrating sexual assault resistance, bystander, and men’s social norms strategies to prevent sexual violence on college campuses: A call to action. Trauma, Violence, & Abuse , 21(4), 811–827.

29.

Orchowski

L. M.

, Gidycz

C. A.

, & Raffle

(2008). Evaluation of a sexual assault risk reduction and self-defense program: A prospective analysis of a revised protocol. Psychology of Women Quarterly , 32(2), 204–218.

30.

Parks

K. A.

, & Fals-Stewart

(2004). The temporal relationship between college women’s alcohol consumption and victimization experiences. Alcoholism, Clinical and Experimental Research , 28(4), 625–629.

31.

Pearson

M. R.

(2013). Use of alcohol protective behavioral strategies among college students: A critical review. Clinical Psychology Review , 33(8), 1025–1040.

32.

Pomery

E. A.

, Gibbons

F. X.

, Reis-Bergan

, & Gerrard

(2009). From willingness to intention: Experience moderates the shift from reactive to reasoned behavior. Personality and Social Psychology Bulletin , 35(7), 894–908.

33.

Preacher

K. J.

, Curran

P. J.

, & Bauer

D. J.

(2006). Computational tools for probing interactions in multiple linear regression, multi-level modeling, and latent curve analysis. Journal of Educational and Behavioral Statistics , 31(4), 437–448.

34.

Ram

, Brinberg

, Pincus

A. L.

, & Conroy

D. E.

(2017). The questionable ecological validity of ecological momentary assessment: Considerations for design and analysis. Research in Human Development , 14(3), 253–270.

35.

Scaglione

N. M.

, Hultgren

B. A.

, Reavy

, Mallett

K. A.

, Turrisi

, Cleveland

M. J.

, & Sell

N. M.

(2015). Do students use contextual protective behaviors to reduce alcohol-related sexual risk? Examination of a dual-process decision-making model. Psychology of Addictive Behavior , 29(3), 733–743.

36.

Scaglione

N. M.

, Turrisi

, Mallett

K. A.

, Ray

A. E.

, Hultgren

, & Cleveland

(2014). How much does one more drink matter? Examining effects of event-level alcohol use and previous sexual victimization on sex-related consequences. Journal of Studies on Alcohol and Drugs , 75(2), 241–248.

37.

Schwartz

J. E.

, & Stone

A. A.

(1998). Strategies for analyzing ecological momentary assessment. Health Psychology , 17(1), 6–16.

38.

Sell

N. M.

, Turrisi

, Scaglione

N. M.

, Cleveland

M. J.

, & Mallett

K. A.

(2018). Alcohol consumption and use of sexual assault and drinking protective behavioral strategies: A diary study. Psychology of Women Quarterly , 42(1), 62–71.

39.

Senn

C. Y.

, Eliasziw

, Barata

P. C.

, Thurston

W. E.

, Newby-Clark

I. R.

, Radtke

H. L.

, & Hobden

K. L.

(2015). Efficacy of a sexual assault resistance program for university women. New England Journal of Medicine , 372(24), 2326–2335.

40.

Shiffman

, Stone

A. A.

, & Hufford

M. R.

(2008). Ecological momentary assessment. Annual Review of Clinical Psychology , 4, 1–32.

41.

Stone

A. A.

, Kessler

R. C.

, & Haythomthwatte

J. A.

(1991). Measuring daily events and experiences: Decisions for the researcher. Journal of Personality , 59(3), 575–607.

42.

Testa

, Hoffman, J.

, Livingston, J.

, & Turrisi

(2010a). Preventing college women’s sexual victimization through parent based intervention: A randomized controlled trial. Prevention Science , 11(3), 308–318.

43.

Testa

, Hoffman

J. H.

, & Livingston

J. A.

(2010b). Alcohol and sexual risk behaviors as mediators of the sexual victimization–revictimization relationship. Journal of Consulting and Clinical Psychology , 78(2), 249.

44.

Testa

, & Livingston

J. A.

(2009). Alcohol consumption and women’s vulnerability to sexual victimization: Can reducing women’s drinking prevent rape? Substance Use & Misuse , 44(9–10), 1349–1376.

45.

Turrisi

, Mallett

K. A.

, Cleveland

M. J.

, Varvil-Weld

, Abar

, Scaglione

, & Hultgren

(2013). Evaluation of timing and dosage of a parent-based intervention to minimize college students’ alcohol consumption. Journal of Studies on Alcohol and Drugs , 74(1), 30–40.

46.

Ullman

S. E.

(2007). A 10-year update of review and critique of empirical studies of rape avoidance. Criminal Justice and Behavior , 34(3), 411–429.

47.

Walls

T. A.

, Jung

, & Schwartz

J. E.

(2006). Multilevel models for intensive longitudinal data. In T. A. Walls & J. L. Schafer (Eds.), Models for intensive longitudinal data (pp. 3–37). Oxford University Press.