Cutting evaluation costs by reducing sample size
Abstract
If sample size can be reduced without undermining validity of results, then the cost of evaluating Extension programs can be reduced. To test this hypothesis, three sample sizes at 1%, 3%, and 5% margins of error were drawn from the data of two doctoral dissertations. Comparison of the three sample sizes showed that evaluations with a descriptive purpose could drop sample size from a 1% to a 5% error margin, while evaluations with a comparative purpose could reduce sample size from 1% to 3% error margin without affecting validity of the results. As a result, it is conservatively estimated that the cost of data gathering and entry for a mail survey could be substantially reduced from $1,116 (1% error) to $348 (3% error) or $100 (5% error).
This paper has been withdrawn.