Testing Research Hypotheses Using Multiple Linear Regression

Multiple regression is becomingmore wide­ly used as the statistical technique for answering research hypotheses. This is so for several reasons: 1) the technique is extreme­ly versatile; 2) the computer has made the technique more available to researchers; and 3) texts such as the authors’ earlier work are making the technique more available to re­searchers. The statistical technique of mul­tiple regression allows the inclusion of numerous continuous (quantitative) and categorical (qualitative) variables in the prediction of some criterion.

Appendixes contain a multiple regression computer program and data on which the problems are based; a discussion of the simi­larities and differences between analysis of variance and multiple regression; and a computer program providing the regression solution to natural language research hy­potheses.

Keith McNeil is President of Educational Monitoring Systems, Ann Arbor, Michigan.

Judy McNeil is Research Associate with High/Scope Education Research in Ypsi­lanti.

Francis J. Kelly is Professor of Guid­ance and Educational Psychology at Southern Illinois University.