This paper describes several common types of research studies in special education transition literature and the threats to their validity. It then describes how the evidential base may be broadened, how diverse sources of evidence can be combined to strengthen causal inferences, and the role of judgment within quasi-experimentation. The paper discusses issues internal to studies and to the methods used in conducting the studies, and discusses issues arising when attempts are made to use the results of the study with other groups or in other places. Threats to internal and external validity are examined. True experimental design is outlined, and then types of quasi-experimental designs are described, including one-group posttest-only design, one-group pretest-posttest design, comparison-group pretest-posttest design, prematched control group design, natural experiments, longitudinal research, cross-sectional research, case-study and single-subject designs, and meta-analysis. Technical and conceptual advances that provide a more significant basis for the interpretation and limitations of quasi-experimental designs are explored. Three considerations required in developing a compelling argument about the causal influence of an intervention are discussed: the analysis must provide a well-specified and credible rationale that links the causal mechanisms with outcomes; it must present evidence to substantiate the claim that the intervention is a plausible explanation for the observed outcome; and it must provide diagnostic assessments and establish the value of the information about purported causal mechanisms and rival explanations. (Contains approximately 30 references.) (JDD)