Many social programs and programs for prevention of drug use are designed to affect a wide variety of targets, including individuals, families and neighborhoods, and organizations such as schools, companies, or hospitals. The nature of the intervention and the design of the particular study determine the choice of the appropriate unit of analysis in assessments of outcome. When the units of assignment and units of observation differ from one another, that is, when clusters of persons rather than persons are assigned at random to treatments, analyses performed at lower levels in the study hierarchy provide inefficient estimates of parameters and often lead to inappropriate significance tests. The present goal was to illustrate the applications of linear mixed models for evaluating statistically the effectiveness of programs.
Get full access to this article
View all access options for this article.
References
1.
AitkinM. A.LongfordN. T. (1986) Statistical modeling issues in school effectiveness studies. Journal of the Royal Statistical Society, Series A, 149, 1–43.
2.
CarvajalS. C.BaumlerE.HarristR. B.ParcelG. S. (2001) Multilevel models and unbiased tests for group based interventions: Examples from the safer choices study. Multivariate Behavioral Research, 36, 185–205.
3.
GoldsteinH. (2002) Multilevel statistical models. (3rd ed.) London: Edward Arnold.
4.
GrahamJ. W.MarksC.HansenW. B. (1991) Social influence processes affecting adolescent substance use. Journal of Applied Psychology, 76, 291–298.
5.
HarvilleD. A. (1977) Maximum likelihood approaches to variance component estimation and to related problems. Journal of the American Statistical Association, 72, 320–338.
6.
HendersonC. R. (1975) The best linear unbiased estimation and prediction under a selection model. Biometrics, 31, 423–447.
7.
KishL. (1967) Survey sampling. (2nd ed.) New York: Wiley.
8.
KreftI. G. G. (1997) The interactive effect of alcohol prevention program in high school classes: An illustration of item homogeneity scaling and multilevel analysis techniques. In BryantK. J.WindleM.WestS. G. (Eds.), The science of prevention: Methodological advances from alcohol and substance abuse research. Washington, DC: American Psychological Association. Pp. 251–277.
9.
KrullL. J.MacKinnonD. P. (2001) Multilevel modeling of individual and group level mediated effects. Multivariate Behavioral Research, 36, 249–277.
10.
LairdN. M.WareJ. H. (1982) Random-effects models for longitudinal data. Biometrics, 38, 963–974.
11.
LittellR. C. (2002) Analysis of unbalanced mixed model data: A case study comparison of ANOVA versus REML/GLS. Journal of Agricultural, Biological, and Environmental Statistics, 7, 472–490.
12.
LittellR. C.MillikenG. A.StroupW. W.WolfingerR. D. (1996) SAS®System for mixed models. Cary, NC: SAS Institute, Inc.
13.
LongfordN. T. (1990) VARCL: Software for variance component analysis of data with nested random effects (maximum likelihood). Princeton, NJ: Educational Testing Service.
14.
MaasC. J. M.SnijdersT. A. B. (2003) The multilevel approach to repeated measures for complete and incomplete data. Quality and Quantity, 37, 71–89.
15.
MurrayD. M. (2001) Statistical models appropriate for designs often used in group-randomized trials. Statistics in Medicine, 20, 1373–1385.
16.
RaudenbushS. W. (1997) Statistical analysis and optimal design for cluster randomized trials. Psychological Methods, 2, 173–185.
17.
RaudenbushS. W.BrykA. S. (2002) Hierarchical linear models: Applications and data analysis methods. (2nd ed.) Thousand Oaks, CA: Sage.
18.
RaudenbushS. W.LiuX. (2000) Statistical power and optimal design for multisite randomized trials. Psychological Methods, 5, 199–213.
19.
RindskopfD.SaxeL. (1998) Zero effects in substance abuse programs: Avoiding false positives and false negatives in the evaluation of community-based programs. Evaluation Review, 22, 78–94.
20.
SAS Institute. (2001) SAS/STAT software: Version 8.2 (TS M0). Cary, NC: SAS Institute, Inc.
21.
ShadishW. R. (2002) Revisiting field experimentation: Field notes for the future. Psychological Methods, 7, 3–18.
22.
ShadishW. R.CookT. D.CampbellD. T. (2002) Experimental and quasi-experimental designs for generalized causal inference. Boston, MA: Houghton Mifflin.
23.
SingerD. J. (1998) Using SAS PROG MIXED to fit multilevel model, hierarchical models, and individual growth models. Journal of Educational and Behavioral Statistics, 24, 323–353.
24.
ZimmermanD. L.Núñez-AntónV. (2001) Parametric modeling of growth curve data: An overview. Test, 10, 111–186.
