A factorial study is usually concerned with two kinds of inference: psychometrical and statistical. Psychometrical inference is the generalization of results to some universe of tests representing the psychological domain under study. Statistical inference is the generalization to some population of examinees. The image factor analytic model is designed for both types of inference and is related to Guttman's image theory. A maximum likelihood solution is developed for this model. The number of factors question is discussed. The method is illustrated using Harman's 13 psychological tests.