Average Effects Based on Regressions with a Logarithmic Link Function: A New Approach with Stochastic Covariates
- 25 Downloads
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.
Keywordsnegative binomial regression model average treatment effects stochastic covariates count data
- Greene, W. (2007). Econometric analysis. Upper Saddle River, NJ: Prentice-Hall.Google Scholar
- Helseth, S. A., Waschbusch, D. A., Gnagy, E. M., Onyango, A. N., Burrows-MacLean, L., Fabiano, G. A., et al. (2015). Effects of behavioral and pharmacological therapies on peer reinforcement of deviancy in children with ADHD-only, ADHD and conduct problems, and controls. Journal of Consulting and Clinical Psychology, 83(2), 280–292. https://doi.org/10.1037/a0038505.CrossRefPubMedGoogle Scholar
- Hittner, J. B., Owens, E. C., & Swickert, R. J. (2016). Influence of social settings on risky sexual behavior. SAGE Open, 6(1). https://doi.org/10.1177/2158244016629187.
- Johnson, P. O., & Neyman, J. (1936). Tests of certain linear hypotheses and their application to some educational problems. Statistical Research Memoirs, 1, 57–93.Google Scholar
- Kröhne, U. (2009). Estimation of average causal effects in quasi-experimental designs: Non-linear constraints in structural equation models. Unpublished doctoral dissertation, Friedrich-Schiller-University Jena, Germany.Google Scholar
- Layland, E. K., Calhoun, B. H., Russell, M. A., & Maggs, J. L. (2018). Is alcohol and other substance use reduced when college students attend alcohol-free programs? Evidence from a measurement burst design before and after legal drinking age. Prevention Science, 1–11. https://doi.org/10.1007/s11121-018-0877-6.
- Long, J. S. (1997). Advanced quantitative techniques in the social sciences. Regression models for categorical and limited dependent variables (Vol. 7). Thousand Oaks, CA: Sage Publications.Google Scholar
- Lord, D. (2006). Modeling motor vehicle crashes using Poisson-gamma models: Examining the effects of low sample mean values and small sample size on the estimation of the fixed dispersion parameter. Accident Analysis and Prevention, 38(4), 751–766. https://doi.org/10.1016/j.aap.2006.02.001.CrossRefPubMedGoogle Scholar
- Macdonald, K., Germine, L., Anderson, A., Christodoulou, J., & McGrath, L. M. (2017). Dispelling the myth: Training in education or neuroscience decreases but does not eliminate beliefs in neuromyths. Frontiers in Psychology, 8, 1314. https://doi.org/10.3389/fpsyg.2017.01314.CrossRefPubMedGoogle Scholar
- Molenberghs, G., Fitzmaurice, G., Kenward, M. G., Tsiatis, A., & Verbeke, G. (Eds.). (2014). Handbook of missing data methodology. Boca Raton, FL: CRC Press. https://doi.org/10.1201/b17622.
- Muench, F., van Stolk-Cooke, K., Kuerbis, A., Stadler, G., Baumel, A., Shao, S., et al. (2017). A randomized controlled pilot trial of different mobile messaging interventions for problem drinking compared to weekly drink tracking. PLoS ONE, 12(2), e0167900. https://doi.org/10.1371/journal.pone.0167900.CrossRefPubMedGoogle Scholar
- Nagengast, B. (2006). Standard errors of ACE estimates: Comparing adjusted group means against the adjusted grand mean. A simulation study. Unpublished diploma thesis, Friedrich-Schiller-University Jena, Germany.Google Scholar
- Pelham, W. E, Jr., Fabiano, G. A., Waxmonsky, J. G., Greiner, A. R., Gnagy, E. M., Pelham, W. E, I. I. I., et al. (2016). Treatment sequencing for childhood ADHD: A multiple-randomization study of adaptive medication and behavioral interventions. Journal of Clinical Child & Adolescent Psychology, 45(4), 396–415. https://doi.org/10.1080/15374416.2015.1105138.CrossRefGoogle Scholar
- R Core Team. (2018). R: A language and environment for statistical computing [Computer software manual]. Vienna, Austria. Retrieved October 06, 2018, from http://www.R-project.org/.
- Rosset, S., & Tibshirani, R. J. (2018). From fixed-X to random-X regression: Bias-variance decompositions, covariance penalties, and prediction error estimation. Journal of the American Statistical Association. https://doi.org/10.1080/01621459.2018.1424632.
- Schrimshaw, E. W., Antebi-Gruszka, N., & Downing, J., M J. (2016). Viewing of internet-based sexually explicit media as a risk factor for condomless anal sex among men who have sex with men in four US cities. PLoS ONE, 11(4). https://doi.org/10.1371/journal.pone.0154439.
- Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs for generalized causal inference. Boston, MA: Houghton Mifflin.Google Scholar
- Steyer, R., & Nagel, W. (2017). Probability and conditional expectation. Somerset, NJ: Wiley.Google Scholar