Abstract
This paper focuses on the impact of selection bias in the context of extended, community-based prevention trials that attempt to “unpack” intervention effects and analyze mechanisms of change. Relying on dose-response analyses as the most general form of such efforts, this study provides two examples of how selection bias can affect the estimation of treatment effects. In Example 1, we describe an actual intervention in which selection bias was believed to influence the dose-response relation of an adaptive component in a preventive intervention for young children with severe behavior problems. In Example 2, we conduct a series of Monte Carlo simulations to illustrate just how severely selection bias can affect estimates in a dose-response analysis when the factors that affect dose are not recorded. We also assess the extent to which selection bias is ameliorated by the use of pretreatment covariates. We examine the implications of these examples and review trial design, data collection, and data analysis factors that can reduce selection bias in efforts to understand how preventive interventions have the effects they do.
Similar content being viewed by others
Notes
Although necessary to examine what happened in second grade, it should be noted that biases might be introduced by stratifying on a post-randomization variable, such as promotion to second grade. The children in the intervention group who were promoted after receiving a year of intensive Fast Track services might be quite different than the children in the control group who were promoted without such services.
Additional analyses that excluded the large percentage of children who transferred schools revealed the same pattern of findings as the analyses presented here; therefore it is unlikely that compositional differences between children receiving different doses of peer pairing were related to residential mobility only.
As a frame of reference to help understand the magnitude of the confounding on the response, consider the case of the linear regression of a standardized response \( \left( {{Y_{std}} = \left[ {Y - \bar{Y}} \right]/{s_Y}} \right) \) on a single standardized predictor \( \left( {{X_{std}} = \left[ {X - \bar{X}} \right]/{s_X}} \right) \). The regression model for standardized variables is intercept free: \( {Y_{std}} = {\beta_{std}}*{X_{std}} + \varepsilon \), where ε∼N(0,1). For this model, the standardized coefficient (β std )2 is equal to R2 (Neter et.al. 1996). Under this frame of reference, a standardized coefficient of −0.02 corresponds to an R2 value of (−0.02)2 = 0.0004, and a standardized coefficient of −0.15 corresponds to an R2 value of (−0.15)2 = 0.0225.
For each simulation, we also fit a non-parametric model that included one predictor for every value of cumulative dose, which fit the data perfectly. We regressed Y on a series of polynomials that were orthogonal in cumulative dose, to avoid problems with collinearity. Bias was of similar magnitude to that reported in each simulation, suggesting that the bias illustrated in these simulations is due solely to selection bias, not problems with lack of fit in the models.
References
Achenbach, T. M. (1991). Manual for the Teacher’s Report Form and 1991 profile. Burlington: University of Vermont Department of Psychiatry.
Barber, J. S., Murphy, S. A., & Verbitsky, N. (2004). Adjusting for time-varying confounding in survival analysis. Sociological Methodology, 34, 163–192.
Bierman, K. L., Greenberg, M. T., & Conduct Problems Prevention Research Group. (1996). Social skills training in the Fast Track program. In R. D. Peters & R. J. McMahon (Eds.), Preventing childhood disorders, substance abuse, and delinquency (pp. 65–89). Thousand Oaks, CA: Sage.
Bodnar, L. M., Davidian, M., Siega-Riz, A. M., & Tsiatis, A. A. (2004). Marginal structural models for analyzing causal effects of time-dependent treatments: An application in perinatal epidemiology. American Journal of Epidemiology, 159, 926–934.
Box, G. E. P., Hunter, W. G., & Hunter, J. S. (1978). Statistics for experimenters. An introduction to design, data analysis, and model building. New York: Wiley.
Bray, B., Almirall, D., Zimmerman, R., Lynam, D., & Murphy, S. A. (2006). Assessing the total effect of time-varying predictors in prevention research. Prevention Science, 7, 1–17.
Cicchetti, D., & Hinshaw, S. P. (2002). Prevention and intervention science: Contributions to developmental theory. Development and Psychopathology, 14, 667–671.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.
Coie, J. D., & Dodge, K. A. (1988). Multiple sources of data on social behavior and social status in the school: A cross-age comparison. Child Development, 59, 815–829.
Cole, S. R., & Hernán, M. A. (2008). Constructing inverse probability weights for marginal structural models. American Journal of Epidemiology, 168, 656–664.
Collins, L. M., Murphy, S. A., & Bierman, K. A. (2004). A conceptual framework for adaptive preventive interventions. Prevention Science, 5, 185–196.
Collins, L. M., Murphy, S. A., Nair, V. N., & Strecher, V. (2005). A strategy for optimizing and evaluating behavioral interventions. Annals of Behavioral Medicine, 30, 65–73.
Collins, L. M., Murphy, S. A., & Strecher, V. (2007). The Multiphase Optimization Strategy (MOST) and the Sequential Multiple Assignment Randomized Trial (SMART): New methods for more potent ehealth interventions. American Journal of Preventive Medicine, 32, S112–S118.
Conduct Problems Prevention Research Group. (1992). A developmental and clinical model for the prevention of conduct disorders: The FAST Track program. Development and Psychopathology, 4, 509–527.
Conduct Problems Prevention Research Group. (1999). Initial impact of the Fast Track prevention trial for conduct problems: I. The high-risk sample. Journal of Consulting and Clinical Psychology, 67, 631–647.
Conduct Problems Prevention Research Group. (2002). Evaluation of the first 3 years of the Fast Track prevention trial with children at high risk for adolescent conduct problems. Journal of Abnormal Child Psychology, 30, 19–35.
Dimidjian, S., Hollon, S. D., Dobson, K. S., Schmaling, K. B., Kohlenberg, R. J., Addis, M. E., et al. (2006). Randomized trial of behavioral activation, cognitive therapy, and antidepressant medication in the acute treatment of adults with major depression. Journal of Consulting and Clinical Psychology, 74, 658–670.
Dobson, K. S., Hollon, S. D., Dimidjian, S., Schmaling, K. B., Kohlenberg, R. J., Gallop, R., et al. (2008). Randomized trial of behavioral activation, cognitive therapy, and anti-depressant medication in the prevention of relapse and recurrence of major depression. Journal of Consulting and Clinical Psychology, 76, 468–477.
Domitrovich, C. E., & Greenberg, M. T. (2000). The study of implementation: Current findings from effective programs that prevent mental disorders in school-aged children. Journal of Educational and Psychological Consultation, 11, 193–221.
Durlak, J. A., & DuPre, E. P. (2008). Implementation matters: A review of research on the influence of implementation on program outcomes and the factors affecting implementation. American Journal of Community Psychology, 41, 327–350.
Feinstein, A. L. (1991). Intention to treat policy for analyzing randomized trials: statistical distortions and neglected clinical challenges. In J. Cramer & B. Spilker (Eds.), Patient compliance in medical practice and clinical trials (pp 359–370). New York: Raven.
Hall, R. C. W. (1995). Global assessment of functioning: A modified scale. Psychosomatics: Journal of Consultation Liaison Psychiatry, 36, 267–275.
Heckman, J. (1976). The common structure of statistical models of truncation, sample selection, and limited dependent variables and a simple estimator for such models. Annals of Economic and Social Measurement, 5, 475–492.
Hernán, M. A., Brumback, B., & Robins, J. M. (2000). Estimating the causal effect of zidovudine on CD4 count with a marginal structural model for repeated measures. Statistics in Medicine, 21, 1689–1709.
Hill, J. L., Brooks-Gunn, J., & Waldfogel, J. (2003). Sustained effects of high participation in an early intervention for low birth-weight premature infants. Developmental Psychology, 39, 730–744.
Jacobson, N. S., Schmaling, K. B., Holtzworth-Munroe, A., Katt, J. L., Wood, L. F., & Follette, V. M. (1989). Research-structured vs. clinically flexible versions of social learning-based marital therapy. Behaviour Research and Therapy, 27, 173–180.
Kreuter, M., Farrell, D., Olevitch, L., & Brennan, L. (2000). Tailoring health messages: Customizing communication with computer technology. Mahwah, NJ: Erlbaum.
Lavori, P. W., Dawon, R., & Roth, A. J. (2000). Flexible treatment strategies in chronic disease: Clinical and research implications. Biological Psychiatry, 48, 605–614.
Lyons-Ruth, K., & Melnick, S. (2004). Dose-response effect of mother-infant clinical home visiting on aggressive behavior problems in kindergarten. Journal of the American Academy of Child and Adolescent Psychiatry, 43, 699–707.
Mortimer, K. M., Neugebauer, R., van der Laan, M., & Tager, I. B. (2005). An application of model-fitting procedures for marginal structural models. American Journal of Epidemiology, 162, 382–388.
Murphy, S. A., Oslin, D., Rush, A. J., & Zhu, J. for MCATS. (2006). Methodological challenges in constructing effective treatment sequences for chronic disorders. Neuropsychopharmacology, 32, 257–262.
Neter, J., Kutner, M. H., Nachtsheim, C. J., & Wasserman, W. (1996). Applied linear statistical models (4th ed.). New York: McGraw-Hill.
Pocock, S. J., & Abdalla, M. (1998). The hope and hazards of using compliance data in randomized controlled trials. Statistics in Medicine, 17, 303–317.
Robins, J. M. (1999). Association, causation, and marginal structural models. Synthese, 121, 151–179.
Robins, J. M., Hernán, M. A., & Brumback, B. (2000). Marginal structural models and causal inference in epidemiology. Epidemiology, 11, 550–560.
Rohrbach, L. A., Graham, J. W., & Hansen, W. B. (1993). Diffusion of school-based substance abuse prevention program: Predictors of program implementation. Preventive Medicine, 22, 237–260.
Rosenbaum, P. R. (1984a). The consequences of adjustment for a concomitant variable that has been affected by the treatment. Journal of the Royal Statistical Society, Series A, 147, 656–666.
Rosenbaum, P. R. (1984b). From association to causation in observational studies: The role of tests of strongly ignorable treatment assignment. Journal of the American Statistical Association, 79, 41–48.
Rosenbaum, P. R. (2002). Observational studies (2nd ed.). New York: Springer.
Rosenbaum, P. R., & Rubin, D. B. (1983). The central role of the propensity score in observational studies for causal effects. Biometrika, 70, 41–55.
Rubin, D. B. (1997). Estimating causal effects from large data sets using propensity scores. Annals of Internal Medicine, 127, 757–763.
Seitz, V., Apfel, N. H., & Rosenbaum, L. K. (1991). Effects of an intervention program for pregnant adolescents: Educational outcomes at two years postpartum. American Journal of Community Psychology, 19, 911–930.
Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs for generalized causal inference. New York: Houghton Mifflin.
Silverman, W. K. (2006). Shifting our thinking and training from evidence-based treatments to evidence-based explanations of treatments. In Balance: Society of Clinical Child and Adolescent Psychology Newsletter, 21.
Trochim, W. M. (2006). The research methods knowledge base (2nd ed.). Retrieved November 10, 2009, from http://www.socialresearchmethods.net/kb/expfact.php.
Wilkinson, L., & The Task Force on Statistical Inference on Statistical Inference. (1999). Statistical methods in psychology journals: Guidelines and explanations. American Psychologist, 54, 594–604.
Winship, C., & Mare, R. D. (1992). Models for sample selection bias. Annual Review of Sociology, 18, 327–350.
Author information
Authors and Affiliations
Consortia
Corresponding author
Additional information
Members of the Conduct Problems Prevention Research Group are, in alphabetical order, Karen L. Bierman, Pennsylvania State University; John D. Coie, Duke University; Kenneth A. Dodge, Duke University; Mark T. Greenberg, Pennsylvania State University; John E. Lochman, University of Alabama; Robert J. McMahon, University of Washington; and Ellen E. Pinderhughes, Tufts University.
Rights and permissions
About this article
Cite this article
McGowan, H.M., Nix, R.L., Murphy, S.A. et al. Investigating the Impact of Selection Bias in Dose-Response Analyses of Preventive Interventions. Prev Sci 11, 239–251 (2010). https://doi.org/10.1007/s11121-010-0169-2
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11121-010-0169-2