Not just "big" data: Importance of sample size, measurement error, and uninformative predictors for developing prognostic models for digital interventions.

There is strong interest in developing a more efficient mental health care system. Digital interventions and predictive models of treatment prognosis will likely play an important role in this endeavor. This article reviews the application of popular machine learning models to the prediction of treatment prognosis, with a particular focus on digital interventions. Assuming that the prediction of treatment prognosis will involve modeling a complex combination of interacting features with measurement error in both the predictors and outcomes, our simulations suggest that to optimize complex prediction models, sample sizes in the thousands will be required. Machine learning methods capable of discovering complex interactions and nonlinear effects (e.g., decision tree ensembles such as gradient boosted machines) perform particularly well in large samples when the predictors and outcomes have virtually no measurement error. However, in the presence of moderate measurement error, these methods provide little or no benefit over regularized linear regression, even with very large sample sizes (N = 100,000) and a non-linear ground truth. Given these sample size requirements, we argue that the scalability of digital interventions, especially when used in combination with optimal measurement practices, provides one of the most effective ways to study treatment prediction models. We conclude with suggestions about how to implement these algorithms into clinical practice.

Development of a brief online intervention to address aggression in the context of emotion-related impulsivity: Evidence from a wait-list controlled trial.

Trait-like tendencies to respond impulsively to emotion, labelled emotion-related impulsivity, are robustly related to aggression. We developed and tested an online intervention to address emotion-related impulsivity and aggression. The 6-session intervention focused on behavioral techniques shown to decrease arousal and aggression, supplemented with implementation intentions and smartphone prompts to facilitate skills transfer into daily life. First, we piloted the intervention in-person with 4 people. Then, 235 participants were randomly assigned to take the online intervention immediately or after a wait-list period; those in the waitlist were then invited to take part in the intervention. Participants completed the self-rated Feelings Trigger Action Scale to assess emotion-related impulsivity, the interview-based Modified Overt Aggression Scale and the self-rated Buss Perry Aggression Questionnaire. Participants who took part in the treatment completed daily anger logs. Attrition, as with other online programs, was high; however, treatment completers reported high satisfaction, and outcomes changed more rapidly during treatment than waitlist across all key outcome indices. In analyses including all participants who took part in the treatment (immediate or delayed), we observed moderate-to-large treatment gains, which were maintained as of the 3-month follow-up assessment. This work supports the usefulness of an intervention for addressing emotion-related impulsivity and aggression.

A brief online intervention to address aggression in the context of emotion-related impulsivity for those treated for bipolar disorder: Feasibility, acceptability and pilot outcome data.

Although aggression is related to manic symptoms among those with bipolar disorder, new work suggests that some continue to experience elevations of aggression after remission. This aggression post-remission appears related to a more general tendency to respond impulsively to states of emotion, labelled emotion-related impulsivity. We recently developed the first intervention designed to address aggression in the context of emotion-related impulsivity. Here, we describe feasibility, acceptability, and pilot data on outcomes for 21 persons who received treatment for bipolar disorder and endorsed high levels of aggression and emotion-related impulsivity. As with other interventions for aggression or bipolar disorder, attrition levels were high. Those who completed the intervention showed large changes in aggression using the interview-based Modified Overt Aggression Scale that were sustained through three months and not observed during wait list control. Although they also showed declines in the self-rated Buss-Perry Aggression Questionnaire and in self-rated emotion-related impulsivity as assessed with the Feelings Trigger Action Scale, these self-ratings also declined during the waitlist control. t Despite the limitations, the findings provide the first evidence that a brief, easily disseminated intervention could have promise for reducing aggression among those with bipolar disorder.

The relationship between entrepreneurial intent, gender and personality.

PURPOSE: We examine gender differences in personality traits of people with and without entrepreneurial intent to assess whether women who intend to become entrepreneurs exhibit particular tendencies that can be fostered. DESIGN/METHODOLOGY/APPROACH: Participants completed an online battery of well-established questionnaires to cover a range of personality traits relevant to entrepreneurship and gender. Participants also answered items concerning intent to become an entrepreneur. A factor analysis of personality traits produced four factors (Esteem and Power, Ambition, Risk Propensity, and Communal Tendency, the latter reflecting Openness and Cooperation, without Hubris). We constructed four parallel regression models to examine how gender, entrepreneurial intent, and the interaction of gender with intent related to these four personality factor scores. FINDINGS: Participants who endorsed a desire to become an entrepreneur reported higher Ambition. Women with entrepreneurial intentions endorsed higher levels of Communal Tendency than men with entrepreneurial intent. Those without entrepreneurial intent did not show gender differences in Communal Tendency. IMPLICATIONS: Current findings suggest that men and women who intend to become entrepreneurs share many traits, but women with entrepreneurial intent show unique elevations in communal tendencies. Thus, a worthwhile locus for intervention into the gender disparity in self-employment may be providing space for and acknowledgement of prosocial motivation and goals as one successful route to entrepreneurship. ORIGINALITY/VALUE: Given the underutilized economic potential of women entrepreneurs, there is a fundamental need for a rich array of research on factors that limit and promote women's entry into entrepreneurship. Current findings indicate that personality may be one piece of this puzzle.