ALTHOUGH THE MERITS OF VARIOUS QUANTITATIVE TECHNIQUES FOR EVALUATION RESEARCH CAN BE DEBATED, CAUSAL MODELING, WHEN INTEGRATED INTO A CONCEPTUAL SCHEME INVOLVING CONSTRUCT VALIDATION, REPRESENTS A USEFUL TECHNIQUE FOR THIS FIELD. A CAUSAL MODEL CAN BE DEFINED AS THE REPRESENTATION OF A SUBSTANTIVE THEORY BY A STRUCTURAL MODEL AND A MEASUREMENT MODEL. THE STRUCTURAL MODEL REPRESENTS THE INTERRELATIONS AMONG CONSTRUCTORS THROUGH MATHEMATICAL EQUATIONS, WHEREAS THE MEASUREMENT MODEL REPRESENTS THE INTERRELATIONS BETWEEN CONSTRUCTS AND MANIFEST VARIABLES THROUGH MATHEMATICAL EQUATIONS. ONE MAJOR CONTRIBUTION OF CAUSAL MODELING METHODS MAY BE THE ATTRIBUTION OF CAUSE IN NONEXPERIMENTAL SETTINGS. WHILE SUCH CAUSAL EXPLANATION IS MOST GENERALLY ASSOCIATED WITH CLASSICAL EXPERIMENTAL METHODS INVOLVING RANDOMIZATION, THE OCCURRENCE OF ETHICAL, LEGAL, AND PRACTICAL CIRCUMSTANCES MAKE CLASSICAL METHODS DIFFICULT IN SOME AREAS OF EVALUATION RESEARCH. AN OVERVIEW OF RECENTLY DEVELOPED MODELS THAT CAN BE APPLIED TO THE PROBLEM OF MAKING CAUSAL INFERENCES IN NONEXPERIMENTAL DATA IS PRESENTED, AND TWO BROAD CATEGORIES OF MODELS DISCUSSED: (1) THOSE APPROPRIATE FOR QUANTITATIVE DATA AND (2) THOSE DESIGNED FOR QUALITATIVE DATA. DESCRIPTIONS ARE GIVEN OF SEVERAL CAUSAL MODELS FOR QUANTITATIVE DATA INCLUDING CONFIRMATORY FACTOR ANALYSIS, THREE-MODE FACTOR ANALYSIS, FACTOR ANALYSIS WITH STRUCTURED MEANS, STRUCTURAL EQUATION MODELS, AND ANALYSIS OF COVARIANCE WITH MEASUREMENT MODEL. CAUSAL MODELS FOR QUALITATIVE DATA, WHICH REQUIRE COMPUTER ASSISTANCE, ARE DISCUSSED INCLUDING THE FACTOR ANALYSIS MODEL, DICHOTOMOUS REGRESSION MODEL, STRUCTURAL EQUATION MODEL, AND THE LOG-LINEAR MODEL. A CASE STUDY APPLIES STRUCTURAL EQUATION MODELS TO THE PROBLEM OF EVALUATING THE EFFECT OF A SUMMER HEAD START PROGRAM ON FIRST GRADERS' COGNITIVE ABILITY. TABULAR DATA AND A LIST OF REFERENCES ACCOMPANY THE PAPER. (MJW)