item_analysis: Item Analysis

Description Usage Arguments Value

Description

Evaluates the standard deviation, correlation, difficulty, discrimination, and reliability of a set of test scores.

Usage

1
item_analysis(data_mat)

Arguments

data_mat

A numeric matrix of test scores with rows representing each examinee and columns representing each question. Elements should equal to 1 (correct) or 0 (incorrect).

Value

Returns a list of the following statistics.

question

Question names.

sd

Standard deviation of each question.

correlation

Pearson correlation coefficeint of each question with the average test score for each examinee. (Analogous to the point biserial correlation coefficient.)

difficulty

Average response for each question. High values indicate low difficulty.

discrimination

The ability of a question to differentiate between a high scoring student and a low scoring student. The statistic is bound between -1 and 1. The statistic for each question is caluclated using the top third scorers and the bottom third scorers. Questions with a large positive discrimination score indicates that examinees who performed well on this question also had high final test scores, and examinees who performed poorly on this question had low final test scores.

reliability

Calculated using the Kuder-Richardson Formula 20 (KR-20). Typically test reliability scores above 0.70 for classroom exams are preferred. (This is an overall statistic - not item-specific.)


imadmali/item.analysis documentation built on May 18, 2019, 3:44 a.m.