Studied by 0 people
Looks like no one added any tags here yet for you.
Continuous variable
infinite number of values (age)
Dichotomous
categorical with 2 levels (true/false)
Pearson’s r
continuous predictor, continuous outcome
age, grade
Phi
dichotomous predictor, dichotomous outcome
sex, pass/fail
Point bi-serial
dichotomous predictor, continuous outcome
sex, grade
Bi-serial
artificially dichotomous, continuous outcome
pass/fail (grade spectrum), age
Spearman’s rho, Kendalls tau
ordinal predictor, ordinal outcome
class standing, rankings
Why should you never artificially dichotomize your variables?
You lose variance and impact your analysis which could lead to misleading results
What do levels of measurement indicate?
Amount of quantitative information extracted from the variable
→ basis of operationalization, not the construct (may not always reflect the nature of the variable)
