RELIABILITY

• Alternate Forms/Parallel Forms

• Test-Retest Reliability

Measures of Stability

• Split Half Reliability

• KR20

• Inter-scorer Reliability

Measures of Internal Consistency

Parallel Forms

The mean and variance of both observed scores from the tests are equal.

Alternate Forms

Are designed to be equivalent with respect to the variables such as content level and difficulty

Estimate of Inter-item Consistency

Evaluates the internal consistency of items.

Pearson’s R

Used when both scores are ordinal, ratio, or interval.

Spearman Rho

Used when we are correlating two set of ranks (ordinal)

Point Biserial

Used when one is true dichotomous variable and one is continuous.

True Dichotomy

Dichotomy wherein there are only two possible categories.

Artificial Dichotomy

Dichotomy wherein there are other possibilities in a certain category

Reactivity

Being evaluated increases performance

Drift

Moving away from what was once learned, suggests that observers should be retrained at a certain point in time

Contrast Effect

The tendency to rate the same behavior differently when observations are repeated in the same context.

Expectancies

Tendency for results to be influenced by what test administrators expect to find

Rosenthal Effect

The test administrator’s expected positive results influence the result of the test

Golem Effect

Negative expectation from the test administrators decreases performance