Program evaluation

Research could compare outcomes between classes with large, small ratios

As one of my duties under the Comprehensive Development Plan, I was responsible for program evaluation.

This effort was controversial, but I believe the results, as feedback, gave planners and government a basis for affirming or sensibly modifying the size and direction of programs as the plan unfolded.

Briefly, program evaluation applies the methods of the social sciences to a dynamic situation, as a "natural experiment," or creating the conditions for a natural experiment.

Currently in education, we have a situation where many claims are made about class size and the relative value of studentteacher contact.

Usually it is claimed that a smaller class size results in better student achievement, and the latter factor (student contact) is sometimes claimed to be the single most important factor in student achievement.

The government is claiming that even with some proposed cuts the student-teacher ratio here is as good as any.

The teachers meanwhile seem to want it both ways, rejecting declining enrolments as a proper justification for reducing the number of teachers. (Guardian, June 25)

I suggest that a program evaluation could help to inform this debate. Using the current Standardized Testing results, a researcher could make comparisons on outcomes between classes with large and small ratios.

Four "natural experiments" suggest themselves, to discover to what extent class size is a determining factor in student achievement, and even what is the optimum class size.

Within a given school, identify a class significantly above, and one significantly below the average, and compare the student achievement levels.

Between any two similar schools, or among several schools, compare Standard Test scores between large and small classes.

Between a rural school with small classroom enrolments, and an urban school with a large teacher ratio, measure the same achievement variable.

Between the French system with a student-teacher ratio of 8.7:1, and the English system with 13.(something):1, make the same comparison.

The researcher would need to control or create comparisons for many variables to ensure the classes were similar, e.g. Immigrant cultural differences, eliminate outrider scores, and ensure teachers had similar qualifications.

With skill and co-operation, these could be done and a valid study be constructed.

The result could be an informed and agreeable system for regulating student-teacher ratios and budget establishment during an expected long period of transition.

I hope some education student, union policy analyst, or government budget analyst will pick up on this suggestion.