You are on page 1of 1

Linear regression is generally used for purposes of prediction.

You use this sta tistic when you suspect that there might be a significant relationship between two variables, an d you want to use subject scores on one variable (the predictor variable) to predict subject s cores on the second variable (the criterion variable). This test is useful when you have computed the mean score on a criterion variabl e for just one sample, and wish to determine whether this mean is significantly different from a specif ied population value. PROC REG DATA=data-set-name ; MODEL criterion-variable = predictor-variable /options ; TITLE1 ' your-name '; RUN; Example : PROC REG DATA=D1; 33 MODEL KG_LOST = MOTIVAT / STB P; 34 TITLE1 'JANE DOE'; 35 RUN; STB - prints the standard regression coeffiecient. P - prints pridicted values on criterian variable for each observation.

You might also like