Working from a common data base and hypothesized model, this paper demonstrates and compares the EQS and LISREL computer program strategies in the analysis of a second-order factor model. Program similarities and differences are noted with respect to: (1) preliminary analyses of the data; (2) treatment of data that are not multivariately normal; (3) assessment of overall model fit; (4) identification of parameter misspecification; (5) post hoc model-fitting; and (6) tests for multigroup invariance. Data comprise scores on the Beck Depression Inventory for 658 (337 males; 321 females) nonclinical adolescents. Issues addressed should be of substantial interest to those unfamiliar with the two problems and/or the methodological procedures presented. Includes one figure and three tables. (Contains 37 references.) (Author/SLD)