Health Sciences 3801A/B Lecture Notes - Lecture 4: Multivariate Normal Distribution, Ordinal Data, Standard Deviation
2 Feb 2019
School
Department
Course
Professor
Document Summary
There are different coefficients that are appropriate for different levels of measurement. The two levels of categorical measurement are nominal and ordinal. Continuous data can be interval or ratio. Interval = does not have a meaningful zero. Ratio = meaningful zero, the absence of the characteristic. Typically means that there are no negative values. Reported when you have two interval/ratio variables. One variable must be interval/ratio and one must be dichotomous. Both variables must be measured on a dichotomous scale. A categorical variable that can only have two categories. Almost every correlation can also be done w/ pearson"s r. Pearson"s r and spearman"s rho have their own value table. If males are dummy-coded into a number that is higher than females then you will have a positive point-biserial correlation. The coefficient sign is based on how you"ve organized it. Assumes normality, bivariate normal distribution (we assume that distribution is normal across the relationship b/w the two variables, linearity.
