This is the group of test-takers whose scores are used as the context for interpreting any subsequent raw scores on the test. There are various ways of obtaining this sample.

(1) stratified sample --

(2) purposive sampling --

(3) incidental sample -- this is also called a convenience sample,
 

Norms

The distribution of the scores of that sample. They are usually reported in the form of one or more forms of transformed (converted) score. There are a number of types of norms available.
 

Types of Norms.

percentile norms

Age norms

Grade equivalent norms easily misinterpreted

national norm

national anchor norm,

subgroup norm

local norm

Remember, norms provide a context for interpreting the meaning of an individual test score.
 

Norm-referenced vs. Criterion-referenced Interpretation.
 

Correlation. A coefficient of correlation is the number that provides us with an index of the strength of the relationship between two things. Correlation never implies causation. Predictability is not the same as causality. Types of correlation coefficient : Pearson Product Moment, Spearman Rho, Eta.
 

Regression

Regression is drawing a line of "best fit" through a scatter-plot. Since once we do this we are only creating an estimate of what the person would get, we know that there is some error in there and the error in this case is called standard error of the estimate. The higher the correlation between the two variables, the lower and smaller the standard error of the estimate, because the more accurate will be the prediction.

Multiple Regression

In Multiple regression, you take a number of variables and use them in a combined way to predict a single variable.