Health Sciences 3801A/B Lecture Notes - Lecture 8: Heteroscedasticity, Grip Strength, Homoscedasticity

• School: Western University
• Department: Health Sciences
• Course: Health Sciences 3801A/B
• Professor: Christopher Lee

Document Summary

Number of assumptions that underlie all calculations. Would avg out the same at some point. Not so diff that you would never think that they couldn"t be the same: violation of this assumption is called heterogeneity of variance: numbers will be diff in the long run. Homoscedasticity: an assumption in regression analysis, assumed that the residuals (points not on the line) have the same degree of variability at all values of x in the long run. Spread of scores is the same all the way along: violation of this assumption is called heteroscedasticity. Normality: assumed that scores are normally distributed around the means (categories) or predicted in the long run. Lack of t is normally distributed, most of the time people fall a little from the line and big differences are rare. Above are assumptions on which the probabilities used to test statistical signi cant are based.