Advanced Ocularmetrics: Graphing Multiple Time Series Residual Data
Educational Studies (CGU)
The desirability of randomized experiments (see Gilbert, Light, & Mosteller, 1975) and the usual need for quasi-experimental designs (because real programs rarely can be administered randomly) are staples of the evaluation literature. Campbell (1969) and others long have advocated the need for acquisition and analysis of longitudinal data in such evaluations.
Today's graduate students dutifully study trend data from the famous Connecticut crackdown on speeding and the British Breathalyser experiment to learn how graphs can be interpreted to shed light on possible program effects. Inevitably, the data portrayed represent zero-order effects. An outcome is plotted over time, or several graphs portray multiple-time series for the treatment and control groups. Techniques for quantifying changes in these graphs at the time of an intervention (e.g., see Box & Tiao, 1968) were developed early in the evolution of quasiexperimental methods.
This paper presents a new technique in which covariates are controlled and eval- It is suggested in this paper, as in previous articles (see, e.g., Campbell, 1969, 1976; Brown, Durbin & Evans, 1975), that frequently program effects are suggested, and can be illustrated, in graphs the analyst or policymaker can "eyeball." Some have humorously labeled this method "ocularmetrics." Lincoln Moses and Donald Campbell made significant and valuable contributions to the development of this methodology. I also want to thank Dale Berger, Mark Lipsey and Arthur Warga for their helpful comments, and Judith Guthrie and Ronald Karpf for programming assistance. uation data still are presented graphically in a manner that makes sense to policymakers. An illustration is provided using data drawn from an evaluation by the author of a major National Science Foundation (NSF) university funding program (Drew, 1975).
©1983 Sage Publications
Drew, D.E. (1983). Advanced ocularmetrics: Graphing multiple time series residual data. Educational Evaluation and Policy Analysis, 5(1), 97-105.