ERIC Number: ED164608
Record Type: RIE
Publication Date: 1978-Mar
Reference Count: 0
Precision for Simple Linear Models.
Porter, Andrew C.
A teaching aid appropriate for a beginning course on experimental design is presented. The aid is a numerical example which illustrates some of the theoretical interrelations among three competing design analysis strategies for estimating treatment effects in random assignment designs. The analysis strategies considered are analysis of variance (ANOVA) of the posttest, ANOVA of an index of response, and analysis of covariance (ANCOVA). The first relationship illustrated is that all three strategies estimate the same effects. Thus, choice of strategy may rest on precision and other factors not considered here, such as assumptions and robustness to violation of assumptions. The second relationship illustrated was precision; it was considered by comparison of mean square errors. Tables present a summary of the analysis strategies showing effects and error terms symbolically, as well as the numerical example itself. (Author/CP)
Publication Type: Reports - Research
Education Level: N/A
Authoring Institution: N/A
Note: Paper presented at the Annual Meeting of the American Educational Research Association (62nd, Toronto, Ontario, Canada, March 27-31, 1978)