334859
Fear as a Predictor of Behavioral Health Outcomes: Using Instrumental Variables to Address Endogeneity
Methods: The data for these analyses were obtained from The Project on Human Development in Chicago Neighborhoods. Adolescents aged 11 to 17 years were included (n=2,329). Depression/anxiety was assessed using the Depression/Anxiety subscale of the Youth Self Report. Fear of neighborhood violence was the regressor of interest and a number of covariates were included as controls. A single-equation model was estimated first to provide a baseline for comparison with the instrumental variable estimation. Instrumental variables (IV) were used to assess whether the predictor of interest is causing the outcome rather than something unobservable, or that the outcome is actually causing the predictor of interest (i.e., reverse causality). A limited information maximum likelihood model was used as it performs better in smaller samples and with weaker instruments.
Results: According to the single equation model, the coefficient for mental health was positive and highly significant (β=1.33, p<0.0001). For each one unit increase in fear (in this case a one unit increase is a move from one level of fear to the next), anxiety/depression scores increased by 1.33 points. With a mean of 5.8, a 1.33 unit increase is fairly sizeable. The IV model tells a story similar to the single equation results though with a smaller effect size and insignificant p-value (β=1.05, p=0.46), which is not uncommon as IV estimation usually causes standard errors to inflate.
Conclusions: This analysis provides evidence that fear and mental health are not endogenous and that fear is predictive of worse mental health outcomes among adolescents.
Learning Areas:
Biostatistics, economicsOther professions or practice related to public health
Public health or related public policy
Public health or related research
Learning Objectives:
Discuss the issue of endogeneity in research. Evaluate the relationship between fear and behavioral health. Assess the validity of instrumental variables as a methodology.
Keyword(s): Methodology, Mental Health
Qualified on the content I am responsible for because: This work was done in part as a doctoral student in health services. I have done quite a bit of work on fear of violent crime and behavioral health outcomes both as a doctoral student and now as a faculty member. I continue to research fear of violent crime and continue to expand upon my projects in this area.
Any relevant financial relationships? No
I agree to comply with the American Public Health Association Conflict of Interest and Commercial Support Guidelines, and to disclose to the participants any off-label or experimental uses of a commercial product or service discussed in my presentation.