Classroom Assessment Ways to improve tests Improving Tests

Classroom Assessment Ways to improve tests

Improving Tests: Judgmental Procedures Check your tests and items: Do they… • adhere to item-writing guidelines? • lead to score-based inferences? • accurately reflect the content? • provide representative coverage of all content and learning targets? • provide fair and unbiased assessment?

Improving Tests: Empirically Based Procedures Difficulty index (p): p= = Number Getting Item Correct ——————— Total Number of Examinees R —— T

Improving Tests: Empirically Based Procedures Item-Discrimination Index (D). D = Correlation between an item and the total test score. – Indicates the extent to which an item tends to be answered correctly by those who score high on the test. – Can be positively discriminating or negatively discriminating, or non-discriminating.

Empirically Based Procedures: Discrimination Index Type of Item Proportion Correct Positive Discriminator High > Low Negative Discriminator High < Low Non-discriminator High = Low Popham gives instructions for calculating and interpreting D on pages 244 -247.

Empirically Based Procedures: Distracter Analysis • Particularly useful when an item needs revision. Item No. xx (p =. 4, D = -. 5) Alternatives A B C D* O Upper 15 students 3 7 2 2 0 Lower 15 students 3 4 4 3 1 Why did nearly half the upper 15 students select alternative B?

Empirically Based Procedure for Criterion-referenced Tests • Two general approaches: – Use a pretest and a posttest. D = p(Post) - p(Pre) – Use two groups of students. – One having received instruction. – One having not received instruction. D = p(Instructed) - p(Not-Instructed)

End