You are on page 1of 1

Statistics Cheat Sheet Types of Variables Nominal each value identified a category and the values can be reordered

d without any loss of meaning. Ex. Gender we can give men a value of 1 or 2 without losing meaning -restricted to bar charts and pie charts Ordinal each value identifies a category, but the order is important. There is no implication that there is a constant difference between values. Ex. Finishing position in an election we know which party came first but we do not know the magnitude of difference between them Scale each value identifies with an actual quantity, rather than a category, and therefore a constant difference between values is implied (i.e. all intervals represent an equal magnitude of difference e.g. 1cm). Ex. Height we know that someone who has a value of 170cm is 50cm taller than someone who has a value of 120cm -can use bar charts but beneficial to use histograms (recommended) or boxplots -when comparing scale variables across categories consider averages, variation/spread, shape Bivariate Analysis There are 4 possible bivariate scenarios: Independent Variable Categorical Dependent Variable Categorical Numeric Crosstabs, %s Graphic Clustered barchart Multiple Boxplot Scatterplot Multiple Boxplot Example Gender & social class Social class & ethnicity Gender & Income Social Class & BMI Age & Income Height and Weight Income & Social Class Rare consider categorising scale variable



Univariate statistical comparison Correlation statistics, Covariance Univariate statistical comparison

Scale Scale

Scale Categorical