CRIM 320 Lecture Notes - Lecture 5: Conduct Disorder, Kurtosis, Null Hypothesis
Document Summary
Tesing assumpion of normality: 2 components to normality. Formula: 95% ci (aka 95% conidence interval) If range foes not encompass 0 the variable is signiicantly skewed and/or. Example kurtoic: normality for skewness technically should do for skew and kurt. (same formula) 95% ci(skew) = skewness +/- 1. 96 (se stat) Posiive (upper bound: 95%ci(skew) = -1. 35 + 0. 43, = -0. 92. Negaive (lower bound: 95%ci(skew) = -1. 35 0. 43, = -1. 78. If you don"t have both on one side and + on another, it is very skewed. If it does not encompass zero = skewed/kurtoic. No, it is negaively skewed and leptokuric. *you have to have both skewness & kurtosis normal to have a normal distribuion. Remedy for non-normally distributed coninuous variables: natural log transformaion. Therefore if the variable has 0 you must add a constant to each case in the variable. *if newly transformed variable from logging is worse than the one before, go back to.