Children's pronunciations of vowel clusters in synthetic words were analyzed in relation to common English words containing the same vowel clusters. Subjects were 436 elementary-school students of both high and low reading levels from a suburban, an urban, and a rural community. Independent variables were grade level, sex, reading level, community type, response type (principle and secondary), and vowel cluster (ai, au, ay, ea, ie, oa, oo, ou, and ow). Dependent variables were difference scores between subsets' principle and secondary pronunciations of vowel clusters and the principle and secondary pronunciation proportions of vowel clusters on a 1963 modification of the Thorndike Frequency Count (type corpus) and on the 1967 Kucera and Francis computational analysis of modern American English (token corpus). Statement and discussion of the problem, development of the test instrument, and study procedures are reported in Part 1. Results and conclusions of the study are given in Part 2, RE 003 627, and appendixes are found in Part 3, RE 003 628. Tables are included. (VJ)