Premium
CROSS‐SECTIONAL VERSUS LONGITUDINAL DATA: AN EMPIRICAL COMPARISON OF MEAN DIFFERENCES IN ACADEMIC GROWTH
Author(s) -
HILTON THOMAS L.,
PATRICK CATHLEEN
Publication year - 1970
Publication title -
journal of educational measurement
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.917
H-Index - 47
eISSN - 1745-3984
pISSN - 0022-0655
DOI - 10.1111/j.1745-3984.1970.tb00689.x
Subject(s) - equating , longitudinal data , longitudinal study , cohort , cohort study , psychology , statistics , cross sectional study , cross sectional data , dropout (neural networks) , correlation , demography , mathematics , machine learning , sociology , computer science , rasch model , geometry
Three sources of data for studies of growth (matched longitudinal, unmatched longitudinal, and cross‐sectional) were compared using mean achievement test scores collected from 32,000 students tested repeatedly from 1961 to 1967. The principal results were that matched longitudinal data yielded significantly higher means than unmatched longitudinal data, and that a school's dropout rate was highly related to the discrepancy between cross‐sectional and matched longitudinal data ( r = .85). Seven sources of differences among the three types of data were considered: age differences, cohort differences, time changes, equating errors, retest effects, cohort change effects, and selection effects. Methodological guidelines for future research in studies of growth and change are presented.