Tests of Simple Slopes in Multiple Regression Models with an Interaction: Comparison of Four Approaches

Interactions between latent variables in count regression models

In psychology and the social sciences, researchers often model count outcome variables accounting for latent predictors and their interactions. Even though neglecting measurement error in such count regression models (e.g., Poisson or negative binomial regression) can have unfavorable consequences like attenuation bias, such analyses are often carried out in the generalized linear model (GLM) framework using fallible covariates such as sum scores. An alternative is count regression models based on structural equation modeling, which allow to specify latent covariates and thereby account for measurement error. However, the issue of how and when to include interactions between latent covariates or between latent and manifest covariates is rarely discussed for count regression models. In this paper, we present a latent variable count regression model (LV-CRM) allowing for latent covariates as well as interactions among both latent and manifest covariates. We conducted three simulation studies, investigating the estimation accuracy of the LV-CRM and comparing it to GLM-based count regression models. Interestingly, we found that even in scenarios with high reliabilities, the regression coefficients from a GLM-based model can be severely biased. In contrast, even for moderate sample sizes, the LV-CRM provided virtually unbiased regression coefficients. Additionally, statistical inferences yielded mixed results for the GLM-based models (i.e., low coverage rates, but acceptable empirical detection rates), but were generally acceptable using the LV-CRM. We provide an applied example from clinical psychology illustrating how the LV-CRM framework can be used to model count regressions with latent interactions.

Parenting-Related Social Networking Site Use and Psychological Distress in Parents of Infants: Cross-sectional Study Exploring the Moderating Effects of Loneliness and Parenting Anxiety

Background

In the digital age, social networking sites (SNSs) have revolutionized the approach to parenting. These platforms, widely used to access parenting information and support, affect parents both positively and negatively, with negative effects potentially increasing for those experiencing loneliness or anxiety.

Objective

This study examined the relationship between SNS use and psychological distress among parents of young children, controlling for the moderating effects of loneliness and parenting anxiety. We hypothesized that higher SNS use correlates to greater psychological distress, particularly among parents with elevated levels of loneliness or parenting anxiety.

Methods

A cross-sectional survey design using a closed web-based questionnaire was employed. Participants included 429 parents (205 mothers and 224 fathers) of children aged 0-3 years recruited through a web-based survey company in Japan. The majority of the participants were couples, with some living with extended family members. The sample also encompassed individuals in cohabiting partnerships and single parents. The survey included measures of psychological distress, loneliness, parenting anxiety, frequency of SNS use for parenting, and covariates. Analytical models to explain psychological distress included interactions between loneliness or parenting anxiety and SNS use, individually for both fathers and mothers.

Results

For mothers, a significant interaction effect was determined only between parenting anxiety and SNS use (b=0.247, SE 0.091; P=.008). Meanwhile, for fathers, significant interaction effects were observed for both loneliness (b=0.324, SE 0.127; P=.012) and parenting anxiety (b=0.144, SE 0.069; P=.038) with SNS use. A simple slope analysis for mothers indicated that SNS use was related to psychological distress only at higher levels of parenting anxiety (b=0.304, SE 0.090, β=.317; P<.001). Among fathers, SNS use was associated with psychological distress at higher levels of either parenting anxiety (b=0.330, SE 0.069, β=.346; P<.001) or loneliness (b=0.390, SE 0.098, β=.409; P<.001).

Conclusions

The study concluded that the relationship between SNS use and psychological distress among parents of young children is moderated by loneliness and parenting anxiety. The findings highlight the need for tailored approaches to help parents manage SNS use, particularly focusing on those with higher levels of loneliness and parenting anxiety. It is imperative that health professionals provide nuanced guidance to parents on SNS use, considering individual psychological factors and potential gender differences in the impact of SNSs on mental well-being.

Average treatment effects on binary outcomes with stochastic covariates

When evaluating the effect of psychological treatments on a dichotomous outcome variable in a randomized controlled trial (RCT), covariate adjustment using logistic regression models is often applied. In the presence of covariates, average marginal effects (AMEs) are often preferred over odds ratios, as AMEs yield a clearer substantive and causal interpretation. However, standard error computation of AMEs neglects sampling-based uncertainty (i.e., covariate values are assumed to be fixed over repeated sampling), which leads to underestimation of AME standard errors in other generalized linear models (e.g., Poisson regression). In this paper, we present and compare approaches allowing for stochastic (i.e., randomly sampled) covariates in models for binary outcomes. In a simulation study, we investigated the quality of the AME and stochastic-covariate approaches focusing on statistical inference in finite samples. Our results indicate that the fixed-covariate approach provides reliable results only if there is no heterogeneity in interindividual treatment effects (i.e., presence of treatment-covariate interactions), while the stochastic-covariate approaches are preferable in all other simulated conditions. We provide an illustrative example from clinical psychology investigating the effect of a cognitive bias modification training on post-traumatic stress disorder while accounting for patients' anxiety using an RCT.

Resting heart rate moderates the relationship between parental emotion socialization and callous-unemotional traits in children

Although empirical findings have indicated that both familial and neurobiological risk factors contribute to the development of callous-unemotional (CU) traits in children, relatively few studies have investigated how these two factors interact to influence these traits. The current study focused on the combined effects of parental emotion socialization and child's resting heart rate on CU traits. Parents of Chinese children (N = 166) completed the Coping with Children's Negative Scale when children were 9.39 years old (SD = 0.92), while children's resting heart rate data were collected when they were 10.21 years old (SD = 0.72). When they were 11.15 years old (SD = 0.67), parents completed the Inventory of Callous-Unemotional Traits Short-Form. Results showed that parental supportive emotion socialization was negatively associated with CU traits and Callous behaviors in particular. In addition, resting heart rate moderated the relationship between parental emotion socialization and child's CU traits. Findings provide further evidence that an interdisciplinary approach that combines both psychosocial and biological factors is essential to further our understanding of CU traits in youth.

Accommodating a Latent XM Interaction in Statistical Mediation Analysis

Statistical mediation analysis is used in the social sciences and public health to uncover potential mechanisms, known as mediators, by which a treatment led to a change in an outcome. Recently, the estimation of the treatment-by-mediator interaction (i.e., the XM interaction) has been shown to play a pivotal role in understanding the equivalence between the traditional mediation effects in linear models and the causal mediation effects in the potential outcomes framework. However, there is limited guidance on how to estimate the XM interaction when the mediator is latent. In this article, we discuss eight methods to accommodate latent XM interactions in statistical mediation analysis, which fall in two categories: using structural models (e.g., latent moderated structural equations, Bayesian mediation, unconstrained product indicator method, multiple-group models) or scoring the mediator prior to estimating the XM interaction (e.g., summed scores and factor scores, with and without attenuation correction). Simulation results suggest that finite-sample bias is low, type 1 error rates and coverage of percentile bootstrap confidence intervals and Bayesian credible intervals are close to the nominal values, and statistical power is similar across approaches. The methods are demonstrated with an applied example, syntax is provided for their implementation, and general considerations are discussed.

Inducing consumers to use calorie information: a multinational investigation

OBJECTIVE

We identify individuals who set daily intake budgets and examine if an intervention making people estimate their calorie intake up to a certain point in the day helps those setting daily budgets to regulate their calorie intake for the remainder of the day, after high prior consumption.

DESIGN

We conducted an online experiment in five countries: Australia, China, Germany, India, and the UK (n = 3,032) using a 2 (setting calorie budget: yes vs. no, measured) x 2 (intervention: intake reminder vs. control, manipulated) between-subjects design, with the amount of prior consumption measured. Participants were contacted in the afternoon. Those in the intervention condition were asked to estimate their prior calorie intake on that day.

MAIN OUTCOME MEASURES

We measured the individual characteristics of those who set daily calorie budgets and the intended calorie intake for the remainder of the day.

RESULTS

Among people who set daily calorie budgets, the intervention reduced intended calorie intake for the remainder of the day by 176 calories if they had already consumed a high amount of calories that day.

CONCLUSION

A timely intervention to estimate one's calorie intake can lower additional intended calorie intake among those who set daily calorie budget.

Shared Leadership and Improvisation: Dual Perspective of Cognition-Affection

Improvisation is an effective way to cope with rapid changes and obtain unexpected opportunities in a complex environment. Based on the cognitive-affective system theory, this study investigates the dual mediating role of cognitive flexibility and emotional intelligence between shared leadership and improvisation and the moderating role of promotion focus. We used multilevel and multi-sourced data to test the theoretical model and used a social network approach to measure shared leadership in teams. Our sample was comprised of 40 teams and 240 team members. The empirical findings indicated that cognitive flexibility and emotional intelligence mediated the relationship between shared leadership and improvisation; promotion focus moderated the relationship between shared leadership and improvisation, and the mediation effect via cognitive flexibility. This study contributes to expanding on improvisation research from the perspective of shared leadership and incorporating both the cognitive and the emotional process of the generation of improvisation into a theoretical framework from a compound perspective, which will open the black box for the mediation mechanism from shared leadership to improvisation. Furthermore, promotion focus is introduced into the research and creatively corresponds to the cognition-affection mediation mechanism, which expands the applicable scope of the regulatory focus theory.

Parents' attitudes towards the content of sex education in the USA: Associations with religiosity and political orientation

While most parents support their adolescents receiving school-based sex education, there is variability in which sex education topics receive the most support from parents. Conservative political orientation and greater religiosity have been independently associated with parents' lack of support for school-based sex education; however, no studies have examined the intersection of these two factors. The three goals of this study were to: 1) identify how specific sexual education topics cluster together to form content areas; 2) examine if religiosity and political orientation are uniquely associated with these content areas; and 3) examine if political orientation moderates the relationship between parents' religiosity and their perceived importance for the specific sex education content areas. Participants were a national sample of 881 US parents. The sex education topics clustered into three content areas: Factual Knowledge (e.g., STI transmission), Practical Skills (e.g., how to access condoms), and Pleasure and Identity (e.g., pleasurable aspects of sex). Politically conservative and more religious parents reported the lowest perceived importance for each content area. Importantly, these main effects were qualified by a significant interaction: parents who reported both political conservativism and high levels of religiosity reported the lowest perceived importance for these three content areas being taught.

The protective role of secure attachment in the relationship between experiences of childhood abuse, emotion dysregulation and coping, and behavioral and mental health problems among emerging adult Black women: A moderated mediation analysis

To determine the extent to which secure attachment moderates the effects of previous child abuse history on the intermediate variables (putative mediators) of emotion dysregulation and coping, which, in turn, influence adult behavioral health and mental health problems. Black women (N = 440, M age = 20.33, SD = 1.88) were selected from the baseline data collection of a large, randomized trial. Study participants had consumed alcohol, had had unprotected sex in the last 90 days, and either reported abuse prior to age 18 or no lifetime history of abuse. Women completed measures of sociodemographics, abuse history, attachment security, coping, emotion dysregulation, psychological functioning, risky sexual behavior, and substance use problems. At low attachment security, the conditional indirect effects of childhood abuse through the intermediate variable, coping, were statistically significant for all dependent variables except proportion condom use and perceived stress. At high attachment security, none of the conditional indirect effects through coping achieved statistical significance. High attachment security also mitigated the conditional indirect effects of childhood abuse through the intermediate variable, emotion dysregulation, reducing the magnitude of the relationship with trait anger, depression, marijuana problems, and perceived stress by about 50%. These results demonstrate the potential mitigating effects of secure attachment on the relationship between childhood abuse history and select behavioral and mental health problems through the intermediate variables of coping and emotional dysregulation. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Family Dysfunction and Cyberchondria among Chinese Adolescents: A Moderated Mediation Model

Cyberchondria has become a severe health problem and a significant public concern. In addition to the impacts that cyberchondria involves, individual psychological and behavioral factors have been identified. However, the role of family function and the mediating and moderating mechanisms underlying these relations are not understood well, especially among adolescents. Based on family functioning and cognitive-behavioral theory, this study sought to examine whether family dysfunction was associated with cyberchondria, and a moderated mediation model was prepared as a means of exploring whether health anxiety was a mediator of relationships between family dysfunction and cyberchondria, as well as whether optimism moderated these mediating processes. A total of 2074 Chinese adolescents (mean = 15.08 years, SD = 1.79) reported their demographic information, family dysfunction, health anxiety, optimism, and cyberchondria. The findings showed that family dysfunction was positively related to cyberchondria. Moreover, health anxiety partially mediated the relationship between family dysfunction and cyberchondria. Finally, optimism moderated the interplay among health anxiety and cyberchondria. Consistent with the expectancy-value models, this positive relationship was weaker for adolescents with a higher level of optimism. These results suggest that it is vital to simultaneously consider individual and family factors as a means of understanding adolescent cyberchondria when performing cyberchondria intervention programs.

Mother and child hair cortisol during the COVID-19 pandemic: Associations among physiological stress, pandemic-related behaviors, and child emotional-behavioral health

The current study assessed the associations between pandemic-related stressors and physiological stress, as indexed by hair cortisol concentration (HCC), for mothers and their children (N = 180) aged 5-14-years old (M = 8.91). The associations between maternal HCC and children's HCC and children's behavioral adjustment were also examined. Mothers reported on COVID-19-related behaviors and children's adjustment, and both mother and child participants collected and mailed hair samples between August and November of 2020. Results indicated that higher maternal HCC was correlated with living in a more urban environment, job loss, working from home, exposure to pandemic-related news, and social isolation. Child HCC was correlated with family job loss and social isolation. Mother HCC and child HCC were significantly associated, and this association was moderated by child age; younger children's HCC was more strongly associated with mothers' HCC than older children's HCC. Finally, maternal HCC was associated with greater child internalizing symptoms, but was not associated with children's externalizing symptoms. Child HCC was not associated with child behavior.

Treatment effects on count outcomes with non-normal covariates

The effects of a treatment or an intervention on a count outcome are often of interest in applied research. When controlling for additional covariates, a negative binomial regression model is usually applied to estimate conditional expectations of the count outcome. The difference in conditional expectations under treatment and under control is then defined as the (conditional) treatment effect. While traditionally aggregates of these conditional treatment effects (e.g., average treatment effects) are computed by averaging over the empirical distribution, a recently proposed moment-based approach allows for computing aggregate effects as a function of distribution parameters. The moment-based approach makes it possible to control for (latent) multivariate normally distributed covariates and provides more reliable inferences under certain conditions. In this paper we propose three different ways to account for non-normally distributed continuous covariates in this approach: an alternative, known non-normal distribution; a plausible factorization of the joint distribution; and an approximation using finite Gaussian mixtures. A saturated model is used for categorical covariates, making a distributional assumption obsolete. We further extend the moment-based approach to allow for multiple treatment conditions and the computation of conditional effects for categorical covariates. An illustrative example highlighting the key features of our extension is provided.

The role of trait empathy in the processing of observed actions in a false-belief task

Empathic brain responses are characterized by overlapping activations between active experience and observation of an emotion in another person, with the pattern for observation being modulated by trait empathy. Also for self-performed and observed errors, similar brain activity has been described, but findings concerning the role of empathy are mixed. We hypothesized that trait empathy modulates the processing of observed responses if expectations concerning the response are based on the beliefs of the observed person. In the present study, we utilized a false-belief task in which observed person’s and observer’s task-related knowledge were dissociated and errors and correct responses could be expected or unexpected. While theta power was generally modulated by the expectancy of the observed response, a negative mediofrontal event-related potential (ERP) component was more pronounced for unexpected observed actions only in participants with higher trait empathy (assessed by the Empathy Quotient), as revealed by linear mixed effects analyses. Cognitive and affective empathy, assessed by the Interpersonal Reactivity Index, were not significantly related to the ERP component. The results suggest that trait empathy can facilitate the generation of predictions and thereby modulate specific aspects of the processing of observed actions, while the contributions of specific empathy components remain unclear.

Illusion of control affects ERP amplitude reductions for auditory outcomes of self-generated actions

The reduction of neural responses to self-generated stimuli compared to external stimuli is thought to result from the matching of motor-based sensory predictions and sensory reafferences and to serve the identification of changes in the environment as caused by oneself. The amplitude of the auditory event-related potential (ERP) component N1 seems to closely reflect this matching process, while the later positive component (P2/ P3a) has been associated with judgments of agency, which are also sensitive to contextual top-down information. In this study, we examined the effect of perceived control over sound production on the processing of self-generated and external stimuli, as reflected in these components. We used a new version of a classic two-button choice task to induce different degrees of the illusion of control (IoC) and recorded ERPs for the processing of self-generated and external sounds in a subsequent task. N1 amplitudes were reduced for self-generated compared to external sounds, but not significantly affected by IoC. P2/3a amplitudes were affected by IoC: We found reduced P2/3a amplitudes after a high compared to a low IoC induction training, but only for self-generated, not for external sounds. These findings suggest that prior contextual belief information induced by an IoC affects later processing as reflected in the P2/P3a, possibly for the formation of agency judgments, while early processing reflecting motor-based predictions is not affected.

Parental Attachment and Problematic Internet Use among Chinese Adolescents: The Moderating Role of Gender and Grit

Problematic Internet use (PIU) is currently becoming a more serious public health concern, representing a deleterious effect on adolescent adaptive emotional and behavioral patterns. Given the prevalence of PIU and its deleterious impact on adolescents' optimal functioning, it is valuable to investigate the risk and protective factors of PIU. Guided by a socio-ecological framework, the current study examines the associations of paternal attachment and maternal attachment with PIU among Chinese adolescents. Furthermore, this study investigates whether adolescents' gender and grit moderate this association. A total of 2677 Chinese adolescents (56.5% girls; Mage = 15.56; SD = 1.57) was involved in this study. Adolescents were uniformly instructed to complete a battery of self-reported questionnaires. The results of linear regression analyses showed that paternal attachment and maternal attachment security were negatively related to PIU. Moreover, moderation analyses revealed that higher levels of grit buffered against boys' PIU in the context of paternal attachment security and girls' PIU in the context of paternal attachment insecurity. The current study suggests that parental attachment security plays an important role in mitigating the likelihood of Chinese adolescents' PIU. Moreover, the buffering role of grit in PIU varies by the levels of paternal attachment security, depending on the adolescents' gender.

Analysis of Covariance in Randomized Experiments with Heterogeneity of Regression and a Random Covariate: The Variance of the Estimated Treatment Effect at Selected Covariate Values

Researchers detecting heterogeneity of regression in a treatment outcome study including a covariate and random assignment to groups often want to investigate the simple treatment effect at the sample grand mean of the covariate and at points one standard deviation above and below that mean. The estimated variances of the simple treatment effect that have traditionally been used in such tests were derived under the assumption that the covariate values were fixed constants. We derive results appropriate for a two-group experiment that instead presume the covariate is a normally distributed random variable. A simulation study is used to confirm the validity of the analytical results and to compare error estimates and confidence intervals based on these results with those based on assuming a fixed covariate. Discrepancies between estimates for fixed and random covariates of the variability of treatment effects can be substantial. However, in situations where the extent of heterogeneity of regression is like that typically reported, presuming the covariate is random rather than fixed will generally result in only a modest increase in estimated standard errors, and in some circumstances can even result in a smaller estimated standard error. We illustrate the new methods with an empirical data set.

Intergroup Contact, COVID-19 News Consumption, and the Moderating Role of Digital Media Trust on Prejudice Toward Asians in the United States: Cross-Sectional Study

BACKGROUND

The perceived threat of a contagious virus may lead people to be distrustful of immigrants and out-groups. Since the COVID-19 outbreak, the salient politicized discourses of blaming Chinese people for spreading the virus have fueled over 2000 reports of anti-Asian racial incidents and hate crimes in the United States.

OBJECTIVE

The study aims to investigate the relationships between news consumption, trust, intergroup contact, and prejudicial attitudes toward Asians and Asian Americans residing in the United States during the COVID-19 pandemic. We compare how traditional news, social media use, and biased news exposure cultivate racial attitudes, and the moderating role of media use and trust on prejudice against Asians is examined.

METHODS

A cross-sectional study was completed in May 2020. A total of 430 US adults (mean age 36.75, SD 11.49 years; n=258, 60% male) participated in an online survey through Amazon's Mechanical Turk platform. Respondents answered questions related to traditional news exposure, social media use, perceived trust, and their top three news channels for staying informed about the novel coronavirus. In addition, intergroup contact and racial attitudes toward Asians were assessed. We performed hierarchical regression analyses to test the associations. Moderation effects were estimated using simple slopes testing with a 95% bootstrap confidence interval approach.

RESULTS

Participants who identified as conservatives (β=.08, P=.02), had a personal infection history (β=.10, P=.004), and interacted with Asian people frequently in their daily lives (β=.46, P<.001) reported more negative attitudes toward Asians after controlling for sociodemographic variables. Relying more on traditional news media (β=.08, P=.04) and higher levels of trust in social media (β=.13, P=.007) were positively associated with prejudice against Asians. In contrast, consuming news from left-leaning outlets (β=-.15, P=.001) and neutral outlets (β=-.13, P=.003) was linked to less prejudicial attitudes toward Asians. Among those who had high trust in social media, exposure had a negative relationship with prejudice. At high levels of trust in digital websites and apps, frequent use was related to less unfavorable attitudes toward Asians.

CONCLUSIONS

Experiencing racial prejudice among the Asian population during a challenging pandemic can cause poor psychological outcomes and exacerbate health disparities. The results suggest that conservative ideology, personal infection history, frequency of intergroup contact, traditional news exposure, and trust in social media emerge as positive predictors of prejudice against Asians and Asian Americans, whereas people who get COVID-19 news from left-leaning and balanced outlets show less prejudice. For those who have more trust in social media and digital news, frequent use of these two sources is associated with lower levels of prejudice. Our findings highlight the need to reshape traditional news discourses and use social media and mobile news apps to develop credible messages for combating racial prejudice against Asians.

Conditional Direction Dependence Analysis: Evaluating the Causal Direction of Effects in Linear Models with Interaction Terms

Direction dependence analysis (DDA) makes use of higher than second moment information of variables (x and y) to detect potential confounding and to probe the causal direction of linear variable relations (i.e., whether x → y or y → x better approximates the underlying causal mechanism). The "true" predictor is assumed to be a continuous nonnormal exogenous variable. Existing methods compatible with DDA, however, are of limited use when the relation of a focal predictor and an outcome is affected by a moderator. This study presents a conditional direction dependence analysis (CDDA) framework which enables researchers to evaluate the causal direction of conditional regression effects. Monte-Carlo simulations were used to evaluate two different moderation scenarios: Study 1 evaluates the performance of CDDA tests when a moderator affects the strength of the causal effect x → y. Study 2 evaluates cases in which the causal direction itself (x → y vs y → x) depends on moderator values. Study 3 evaluates the robustness of DDA tests in the presence of functional model misspecifications. Results suggest that significance tests compatible with CDDA are suitable in both moderation scenarios, i.e., CDDA allows one to discern regions of a moderator in which the causal direction is uniquely identifiable. An empirical example is provided to illustrate the approach.

Empathy as a buffer: How empathy moderates the emotional effects on Preschoolers' sharing

This study experimentally investigated how empathy moderated the influence of external emotional stimuli on preschoolers' subsequent sharing behaviour. Children aged 4-6 (N = 218) were randomly presented with one of the video clips that elicited sad, happy, or neutral emotion, and were then asked to play a dictator game with puppet partners. Results revealed that compared to the neutral condition, children with higher empathy increased sharing after induced happiness but did not change sharing after induced sadness; by contrast, children with lower empathy decreased sharing after induced sadness. It appears that empathy may enhance the positive effect of happiness while reducing the negative effect of sadness on sharing behaviour. These findings provide experimental evidence in early childhood to support approaches that emphasize the role of immediate emotions in moral decision-making; critically, such emotional effects vary with individual differences in dispositional empathy.

Race moderates the effects of Motivational Interviewing on smoking cessation induction

OBJECTIVE

Health disparities necessitate exploration of how race moderates response to smoking cessation treatment. Data from a randomized clinical trial of Motivational Interviewing (MI) for smoking cessation induction were used to explore differential treatment response between African American (AA) vs Non-Black (NB) smokers.

METHODS

Adult tobacco smokers (138 AA vs 66 NB) with low desire to quit were randomly assigned to four sessions of MI or health education (HE). Outcomes (e.g., quit attempts) were assessed 3- and 6-months.

RESULTS

There was evidence of a Race by Treatment interaction such that MI was less effective than HE in AA smokers. Mean Cohen's d for the interaction effect was -0.32 (95% CI [-0.44, -0.20]). However, the race interaction could be accounted for by controlling for baseline relationship status and communication preference (wants directive approach).

CONCLUSIONS

MI may be less effective for smoking cessation induction in AA vs NB smokers when compared to another active and more directive therapy. The differential response between races may be explained by psychosocial variables.

PRACTICE IMPLICATIONS

MI may not be an ideal choice for all African American smokers. Patients' relationship status and preference for a directive counseling approach might explain disparities in response to MI treatment.

Alcohol demand moderates brief motivational intervention outcomes in underage young adult drinkers

INTRODUCTION

The Alcohol Purchase Task (APT), a behavioral economic measure of alcohol's reinforcing value (demand), has been used to predict the effects of Brief Motivational Intervention (BMI) on alcohol use outcomes. However, it is not known whether BMI may be more or less efficacious, relative to control, among those with different levels of alcohol demand prior to treatment.

METHODS

Non college-attending young adults (N = 150) reporting past-month heavy drinking were randomized to a single in-person session of BMI or a relaxation training control (REL). The BMI included delivery of personalized feedback and focused on developing discrepancy between the young adults' goals and their current pattern of alcohol use. At baseline, participants completed assessments of alcohol use and the APT. Drinking levels were re-assessed at 6 weeks and 3 months post-intervention. Demand indices derived from the APT were examined as moderators of treatment effects on follow-up drinking after covarying for baseline alcohol use.

RESULTS

Two of four APT demand indices - intensity and Omax - moderated treatment outcomes. Relative to REL, BMI led to greater reductions in total number of drinks consumed and drinks per drinking day among participants with higher baseline alcohol demand. This association was not observed among participants with lower levels of alcohol demand.

CONCLUSIONS

These results demonstrate that BMI may be particularly beneficial for those with a high reinforcing value of alcohol. The mechanism for this effect is unclear, and determining the process by which BMI confers increased benefit for these individuals is a fruitful area for future work.

Probing curvilinear-by-linear interactions when the predictors are randomly sampled

Curvilinear effects and curvilinear-by-linear interactions are hypothesized, tested, and probed in various fields of the social and behavioral sciences. When such effects are found, researchers traditionally have tested the simple slope, and more recently have utilized the Johnson-Neyman technique, often at the values of the sample statistics of the predictor(s). Both approaches implicitly assume that values on the predictors have been sampled according to a fixed sampling plan. More typically in social and behavioral research, however, values on the predictors can be more appropriately considered as randomly sampled from a multivariate population distribution. I show analytically and through simulation that for regression models containing a curvilinear-by-linear interaction, fixed and random regression models produce the same estimates but different standard errors of the simple slope at values of the sample statistics of the predictors. When values on the predictors are randomly sampled, treating them as fixed when testing the simple slope or generating conditional Johnson-Neyman confidence bands at values of the sample statistics of the predictors can lead to inflated Type I error rates and inaccurate coverage rates. Recommendations for researchers are provided, and directions for future research are discussed.

Average Effects Based on Regressions with a Logarithmic Link Function: A New Approach with Stochastic Covariates

Researchers often use regressions with a logarithmic link function to evaluate the effects of a treatment on a count variable. In order to judge the average effectiveness of the treatment on the original count scale, they compute average treatment effects, which are defined as the average difference between the expected outcomes under treatment and under control. Current practice is to evaluate the expected differences at every observation and use the sample mean of these differences as a point estimate of the average effect. The standard error for this average effect estimate is based on the implicit assumption that covariate values are fixed, i.e., do not vary across different samples. In this paper, we present a new way of analytically computing average effects based on regressions with log link using stochastic covariates and develop new formulas to obtain standard errors for the average effect. In a simulation study, we evaluate the statistical performance of our new estimator and compare it with the traditional approach. Our findings suggest that the new approach gives unbiased effect estimates and standard errors and outperforms the traditional approach when strong interaction and/or a skewed covariate is present.