The term criterion refers to the dependent variable and predictor typically refers to the independent variable. Criterion variables attempts to measure presumed effect while the predictor variable is the one that causes an effect.

Collected data can be categorized into types of variables and scales of measurement. As mentioned before, a variable is something that changes. A Continuous Variable can have many values and is on a continuum. For example, weight is a continuous variable since there are infinite possibilities.  Some variables can’t have unlimited possible values, they are considered Discrete Variables. These variables have only a set of possible values. For example, number of balls in a jar has a limit to what possible values might be. Number of symptoms of a diagnosis has a specific limit. Some variables are considered Discrete Dichotomous Variables because they can only have one of two possibilities. Flipping a coin can result in either heads or tails.

When a data point belongs to a distinct category and can neither be truly measured or ordered, it is considered to be part of a Nominal Scale(e.g. city of birth). You can count the frequency of a nominal scale (e.g. number of males). If a data can be ordered and one data point is less or more than the other, it is considered to be part of a Ordinal Scale (e.g. grades in a class). In ordinal scales, there’s no information about the difference between two points. So if one person got an A and another got a B, we don’t know what their actual percentage grade was. Nominal and Ordinal are also called Non-Parametric Scales. Examples of Parametric Scales include interval and ration scales. Whenever data points have equal intervals between each point (e.g. temperature) without a true zero (0 degrees does not mean there’s no heat), it is considered as an Interval Scale. Interval scales can be added or subtracted. Whenever a scale has an absolute zero point (e.g. weight), it is considered a Ratio Scale. These can be added, subtracted, divided, or multiplied. These can have qualities or ordinal scales (can have ranks) and interval scales (can have equal intervals).