Project Personality

Study 1

Summary

Schleider and Weisz (2018) randomly assigned 96 youth (aged 12-15) who showed symptoms of anxiety and depression to receive the single-session growth mindset intervention or the control intervention. Participants were surveyed at baseline and again at three, six, and nine months postintervention on their internalizing symptoms (depression and anxiety) and their perceived control over their behaviors and emotions.

Schleider and Weisz (2018) found that nine months after the intervention, compared to participants in the control condition, participants in the intervention condition showed significantly:

• Lower self- and parent-reported depression
• Lower parent-reported anxiety

Risk and Protective Factors

• Higher perceived behavioral control

Evaluation Methodology

Design:

Recruitment: The "skill-building" program was advertised at schools, after-school programs, and youth clinics. Interested parents were screened by phone. Child inclusion criteria were: 1) aged 12-15 years, and 2) either scoring ≥ 84th percentile on a psychological screening tool, receiving school-based accommodations for anxiety or depression, or seeking treatment for anxiety or depression within the previous three years. Exclusion criteria were psychosis, intellectual disability, developmental disorders, autism, and attempted suicide or suicidal ideation in the previous year. Of the 187 phone screens, 77 families did not meet inclusion criteria, and 14 declined to participate. The initial sample comprised the remaining 96 families who enrolled in the study.

Assignment: Children (n=96) were randomly assigned to the intervention (n=48) or control (n=48) condition via computerized assignment, so both participants and experimenters were blind to condition. Participants in the control condition received a computer-based session of supportive therapy (ST), designed to encourage youths to identify and express feelings.

Attrition: Of the 96 participants originally enrolled in the study and randomized to conditions, attrition rates at three, six, and nine months after the program session were 11.46%, 19.79%, and 26.04%.

Sample:

The mean age of participants in the intervention condition was 13.4 (SD = 1.58) and in the control condition was 13.3 (SD = 1.06). The intervention sample was 54.2% female, 75.0% Caucasian, and 14.6% Hispanic, and the control condition was 56.3% female, 70.8% Caucasian, and 12.5% Hispanic. A total of 58.3% of participants in the intervention condition and 52.1% of control participants had received prior treatment for anxiety or depression. At baseline, approximately 15% of children self-reported subclinical internalizing problems, and approximately 85% reported clinically elevated symptoms of depression, overall anxiety, or a specific anxiety disorder.

Measures:

Depressive symptoms, anxiety symptoms, perceived behavioral control, perceived emotional control, and personality mindsets were all measured using well-established self-report questionnaires.

Youth participants rated their depression using the Children's Depression Inventory, anxiety using the Screen for Child Anxiety and Related Disorders-Child version, behavioral control using the Perceived Control Scale for Children, emotional control using the Secondary Control Scale for Children, and personality mindsets through the Implicit Personality Theory Questionnaire. Parents rated their children's depression through the Children's Depression Inventory-Parent, and their children's anxiety through the Screen for Child Anxiety and Related Disorders-Parent. All measures had alphas of 0.75 to 0.93.

Analysis:

The researchers used hierarchical linear mixed models to analyze changes in outcome measures while accounting for repeated measures within individuals over time. Models included a random intercept and random slope for the individual, an autoregressive error structure, and employed full-information maximum likelihood (FIML) estimation with the missing-at-random (MAR) assumption for missing data, so that all available data were used in analyses and missing values were "implied." Covariates included youth age, gender, and family structure (single- vs. dual-parent home).

Intent-to-Treat: All available data were used in analyses.

Outcomes

Implementation Fidelity:

The researchers measured time to complete the intervention and control programs, understanding of program content, interest in material, and effort on activities, and found no significant differences between the two conditions. Youth in the intervention condition showed significantly greater increases in growth mindsets compared to youth in the control condition, supporting the effectiveness of the intervention.

Baseline Equivalence:

No significant differences were detected in baseline demographics or outcome variables between the two conditions. However, it is possible that Table 2 showed one significant baseline difference on parent-rated depression.

Differential Attrition:

Likelihood of attrition by 9-month follow-up did not differ by condition, baseline youth anxiety/depression, youth gender/age/race, family income, or parent education. However, attrition was more likely for youth living in single-parent homes than for those with partnered parents.

Posttest:

Nine months after baseline, participants in the intervention condition showed significantly reduced self- and parent-reported depression symptoms, parent-reported anxiety, and perceived behavioral control compared to participants in the control condition. Reductions in anxiety and perceived emotional control did not differ by condition.

Long-Term:

Not examined.

Study 2

Summary

Schleider et al. (2022) used a randomized controlled trial to compare a sample of 2,452 youth assigned to one of three conditions: (1) Project Personality; (2) Project ABC; and (3) control. Assessments at baseline, posttest, and three-month follow-up measured depression, anxiety, and COVID-19-related trauma symptoms, as well as perceived agency, hopelessness, and restrictive eating.

Schleider et al. (2022) found that, relative to the control group, students in the Project Personality intervention group showed significantly greater improvements in:

• Hopelessness (at posttest and three-month follow-up)
• Perceived agency (at posttest and three-month follow-up)
• Depressive symptoms (at three-month follow-up)
• Generalized anxiety symptoms (at three-month follow-up)
• COVID-19-related trauma symptoms (at three-month follow-up)
• Restrictive eating (at three-month follow-up)

Evaluation Methodology

Design:

Recruitment: The study was conducted from November 2020 to March 2021. Participants across all 50 states were recruited using Instagram advertisements, which linked to an eligibility screener. Youth inclusion criteria were: (1) 13-16 years old; (2) comfort reading and writing in English; (3) internet and computer, laptop, or smartphone access; and (4) endorsement of elevated depressive symptoms. To maintain youth confidentiality and minimize access barriers, parent permission to participate was not required. Of the 6,884 youth who completed eligibility screening, 3,851 were eligible and agreed to participate, though only 2,452 youth completed the baseline survey and were assigned to condition.

Assignment: Adolescents were randomly assigned to one of three conditions: (1) Project Personality (n = 813); (2) Project ABC (n = 821); or (3) active control (n = 818). Randomization was achieved using a Qualtrics-embedded randomizer; both adolescents and experimenters were blind to condition until data collection was complete. All three conditions utilized self-administered, online, single-session interventions that were length-matched. Project Personality teaches a growth mindset, whereas Project ABC teaches behavioral activation. Control youth received a supportive therapy program that encouraged them to express emotions to close others but did not teach specific skills.

Assessments/Attrition: Assessments occurred at baseline, post-intervention, and three-month follow-up. Of the 2,452 youth who completed the baseline assessment, 2,054 (84%) completed the posttest, and 1,766 (72%) completed the three-month follow-up (see Figure 2, p. 263).

Sample:

At baseline, 86.2% of youths reported clinically elevated depressive symptoms. The sample was 88.1% female, 3.8% American Indian, 12.6% Asian, 10.5% Black, 1.6% Native Hawaiian, 19.2% Hispanic, and 66.6% White. Approximately 80% identified as a sexual minority, and participating youths lived in all 50 states.

Measures:

Adolescents self-reported on measures of hopelessness and perceived agency at all three time points, as well as measures of mental health at baseline and three-month follow-up only; thus, all measures came from an independent source. Six outcomes were computed for analysis: (1) depressive symptoms; (2) generalized anxiety symptoms; (3) COVID-19-related trauma symptoms; (4) perceived agency; (5) hopelessness; and (6) restrictive eating. Depression symptom severity was the primary outcome; the remaining measures were all considered as secondary outcomes. The perceived agency and hopelessness intermediate outcomes represent risk and protective factors, whereas trauma symptoms and restrictive eating can be viewed as measures of mental health - other. In the current sample, scale derived alphas ranged from .73 (trauma symptoms) to .90 (anxiety symptoms).

Analysis:

Analyses used a series of linear regressions to test for intervention effects. Intervention condition was included as a categorical predictor variable, and baseline outcome controls were used as covariates. In addition, the false discovery rate (FDR) was applied to two-tailed tests of pre-registered outcomes (i.e., three-month depressive symptoms, post-intervention perceived agency and hopelessness).

Intent-to-Treat: All available data were used in the main analysis. Missing data were handled through imputation using "the expectation-maximization and bootstrapping algorithm implemented with Amelia II in R" (p. 266). The authors imputed as many datasets as there were percentage points of missing data for a particular outcome, rounding to the next-highest percentage (e.g., if 5.4% of data was missing on an outcome, six imputed datasets were created). Additionally, the sensitivity analyses, which used two alternative approaches to handling missing data (i.e., listwise deletion and multiple imputation), represent completers-only analyses, but in both cases the results were comparable to those of the main analysis, showing only minor differences with respect to secondary outcomes (see pp. 262-263).

Outcomes

Implementation Fidelity:

Approximately 84% of youth completed the program to which they were assigned in full. Adolescents rated all three single-session interventions as acceptable across all program feedback scale items, as indicated by the item-level means, all of which were above 3.5/5 (see Table 2, p. 261).

Baseline Equivalence:

The authors stated only that "no group differences emerged on demographic factors or baseline depressive symptoms" (p. 259) for the assigned sample, but no significance tests were reported. However, Table 1 (p. 260) presents counts for baseline demographic variables and means for baseline depressive symptoms, which appear similar across the treatment and control groups. The authors did not report findings from tests for other baseline outcomes beyond depressive symptoms (their primary outcome). 

Differential Attrition:

Attrition rates from baseline to three-month follow-up were similar in the intervention (27.1%) and control (26.2%) groups. There were no differences between completers and attritors, either during the intervention or at three-month follow-up, in baseline depressive symptoms, age, gender identity, sexual orientation, romantic/sexual attraction, or race. There were no tests of baseline-by-condition attrition. 

Posttest:

Relative to the control group, youth in the Project Personality intervention group showed significantly greater improvements in hopelessness and perceived agency from baseline to posttest, as well as depressive symptoms, generalized anxiety symptoms, COVID-19-related trauma symptoms, restrictive eating, hopelessness, and perceived agency from baseline to three-month follow-up. These significant effects all held in complete case sensitivity analyses, except for the model comparing change in COVID-19-related trauma symptoms from baseline to three-month follow-up, which fell to non-significance in the analysis using listwise deletion.

Long-Term:

Not examined.