Samples of the student population in high school constitute the backbone of this study. They were constructed out of two series of student records. The first series was composed of student enrollment lists and graduation lists of the Somerville High School and Latin High School. The enrollment lists recorded the date of enrollment, the student's name, home address, date of birth, age, previous school attended, and the name of the parent. The graduation lists listed graduates by name; they recorded the colleges entered; in the 1870s and 1880s they ranked graduates by class standing. The second series, running from 1895 to 1905, listed students in the English High School and came from two sources. One was the high school yearbook which listed entrants and graduates by name and home address; the other was a file of student career records that contained information on family background and course work.

The primary data on family characteristics of students was derived by linking student records with the manuscript schedules of the U. S. census for Somerville. Students entering high school in the years clustering around the decadal census years 1850, I860, 1870,1880, and 1900 were traced to the manuscript schedules. Matches between students and households listed in the census schedules were made according to the name and age of a student, the name of the parent, and household address when available in the 1880 and 1900 schedules. The task of record linkage was performed through simple manual inspection. The availability of the Soundex indexing system for the 1880 and 1900 schedules eased the tedium of this job.