The research literature is replete with relatively small scale studies investigating various theories of reading. This research is often exploratory rather than confirmatory in nature. To avoid such limitations, samples from a large data base were used to examine causal models based on the Goodman theory of reading and the concepts of process and product comprehension. Two separate causal models were created using the following variables: graphic similarity, sound similarity, acceptability with prior text, acceptability with following text, acceptability or correction, the retelling score for the text read, the Comprehensive Tests of Basic Skills Vocabulary and Comprehension scores, and the Degrees of Reading Power (DRP) scores. The sample for one model consisted of 448 Grade Two to Grade Five students enrolled in a Chapter I program in the southwestern United States. The sample for the second model consisted of only the Grade Four and Grade Five students from the larger sample since DRP scores were not available for other students. In the hypothesized models graphic and sound similarity, and acceptability with prior and following were posited to load on a latent factor representing the unitary dimension of reading which in turn was posited to be causally related to process and product variables. Process variables included the acceptable or corrected score (RMI comprehending score) and the DRP score. Product variables included the CTBS scores and the retelling score. Product variables were posited to be dependent upon process variables. Analysis was conducted using the LISREL program. For both models causal pathways were dropped between process and product comprehension variables but were retained between the latent variables of reading and process and product comprehension. In addition, the DRP score loaded on product comprehension more than on process comprehension while retelling loaded more on process comprehension than product comprehension. Variables relating to the Goodman theory appeared to be more interrelated than those external to the theory. Implications for the uses of causal modeling and the constructs of process and product comprehension are discussed.