First and foremost, variables are chosen to answer the research question. The primary predictor and outcome variables should be derived directly from the research question. Secondary questions based on demographic, predictor, and confounding variables are always feasible when understanding phenomena.
The way that researchers measure for variables has a drastic impact on their ability to detect precise and accurate treatment effects. While categorical and ordinal variables are important and prevalent in empirical research, researchers must always be cognizant of the increased sample size needed to detect clinically meaningful treatment effects measured at these levels.
So, if at all possible, try to measure for predictor, confounding, and outcome variables at a continuous level because it leads to increased statistical power, increased precision and accuracy, and decreased sample size.
Click on the Scales of Measurement button to learn about categorical, nominal, ordinal, interval, ratio, count, and continuous level measurement.
Click on the Types of Variables button to learn about demographic, independent, control, dependent, predictor, confounding, and outcome variables.
The way that researchers measure for variables has a drastic impact on their ability to detect precise and accurate treatment effects. While categorical and ordinal variables are important and prevalent in empirical research, researchers must always be cognizant of the increased sample size needed to detect clinically meaningful treatment effects measured at these levels.
So, if at all possible, try to measure for predictor, confounding, and outcome variables at a continuous level because it leads to increased statistical power, increased precision and accuracy, and decreased sample size.
Click on the Scales of Measurement button to learn about categorical, nominal, ordinal, interval, ratio, count, and continuous level measurement.
Click on the Types of Variables button to learn about demographic, independent, control, dependent, predictor, confounding, and outcome variables.
Categorical, nominal, ordinal, interval, ratio, count, and continuous scales of measurement
Demographic, independent, control, dependent, predictor, confounding, and outcome variables
Reliability, stability, consistency, and confidence in measurement
Validity, utility, interpretability, and generalizability in measurement
