Statistical power and outcomes

The scale of measurement of the outcome influences statistical power

The scale of measurement of the outcome has drastic influences on statistical power. The lack of precision and accuracy in categorical and ordinal measurement leads to increased sample sizes. This is because more observations of these outcomes are needed to detect the most precise and accurate measure of treatment effect. Categorical and ordinal outcomes also use less power non-parametric statistics to yield inferences.

​If at all possible, try to measure for outcomes at the continuous level. Continuous level measurement leads to more statistical power and smaller sample sizes because of the increased precision and accuracy associated with a "true zero" and being able to detect both the direction and magnitude of treatment effects. Continuous level measurement also allows you to use more power parametric statistics. 
The scale of measurement of outcomes impacts statistical power and the needed sample size.

What is the scale of measurement for the outcome?

The outcome represents numerical designations or categorical values that describe exposures, characteristics, phenomena, events, or group membership.
Categorical
The outcome variable is measured using an ordered numerical continuum, such as a Likert scale.
Ordinal
The outcome variable is an actual number that provides both a measure of distance and magnitude due to having a "true zero."    
Continuous
Back to Statistical Power
Research Engineer Home Page

Hire A Statistician - Statistical Consulting for Professionals

$100.00
Buy Now
DO YOU NEED TO HIRE A STATISTICIAN?
Eric Heidel, Ph.D. will provide statistical consulting for researchers, professionals, and organizations at $100/hour. Secure checkout is available with Stripe, Venmo, Zelle, or PayPal.
  • Statistical Analysis
  • Research Design
  • Sample Size Calculations
  • Diagnostic Testing and Epidemiological Calculations
  • Survey Design and Psychometrics