Effect size is the degree of treatment effect in experimental research and is considered in relationship to power, significance criterion, and sample size. Effect size is useful in assessing meaningfulness of statistically significant research results. Its use as a discriminator between "meaningful" and "trivial" treatment effects is considered.…

Two problems are addressed in relation to agricultural education research: (1) Is the Type I error rate affected when nonindependent multiple statistical tests are made? And (2) does statistical significance automatically mean that the results are of practical importance? (LRA)

The authors state that multiple regression is a powerful method of statistical analysis, provides a strength of relationship index, and should replace analysis of variance (ANOVA) in educational research. They also discuss the coding of categorical variables and available computer programs for multiple regression. (Author/MH)

Pattern changes in science education are analyzed for the years 1962-77 in relation to pure science subjects studied by students. Numbers are analyzed of boy/girl entrants involved in (1) a balanced science course and (2) at least one pure science subject. (CS)

Macready and Dayton introduced two probabilistic models for mastery assessment based on an idealistic all-or-none conception of mastery. Alternatively, an application of latent trait theory to mastery testing is proposed (a three parameter logistic model) as a more plausible model for test theory. (Author/CTM)

The method and means by which statistics can raise a prima facie case of Title VII violation are analyzed. A standard is identified that can be applied to determine whether a statistical disparity is sufficient to shift the burden to the employer to rebut a prima facie case of discrimination. (LBH)

Presents two experiments which demonstrate that language-generalization tests should tend to decrease Type 2 errors. (AM)

The research summarized and reported contradicts the view that schooling makes little difference and that its improvement in quality and equality of opportunity lies beyond our reach. (Author)

A 13-item Student Description of Instruction Questionnaire developed at Bronx Community College (New York) was subjected to a series of reliability, stability, and validity studies based on data collected over a period of four semesters. Indices suggested moderate to good reliability and stability, depending on department. (JG)

Examines superintendents' attitudes toward community participation at the advisory and control levels in four areas of educational policy-making. (Author)

The accuracy of statistics based on performance assessments that represent percentages of students reaching standards is explored using data from a large-scale performance assessment, the Maryland School Performance Assessment Program. Results with students in grades 3, 5, and 8 support the accuracy of pooling results to produce the statistics.…

The relationship between formal planning and the performance of university student teams competing in a comprehensive business simulation is examined. Subjects were 182 teams of business students organized into 32 competitive sets. Statistical testing indicated that planning teams performed significantly better than nonplanning teams. (Author)

Past attempts to estimate peer group effects on socioeconomic or academic outcomes may have used insufficiently flexible techniques that stressed the mean of peer characteristics without considering their overall distribution. Using Philippines data, this paper demonstrates an estimation model accounting for the shape of distribution. Failure to…

Discusses how responses to multiple-choice items could be interpreted, demonstrates how responses on constructed-response items can be analyzed, and examines interactions between item characteristics and student responses. Uses information, specifically items and student responses, from the Third International Mathematics and Science Study…