If nearly everyone answers nearly every item correctly, the items are too easy, so the level of difficulty needs reconsideration.
Identifying a definition from alternatives is recall of information, which is the Knowledge level.
Using a learned principle in a new situation is Application.
Breaking an argument into parts and showing how they relate is Analysis.
Combining principles to build something original is Synthesis.
A discrimination index near 1.00 means the item sharply separates higher- from lower-achieving students.
A distractor chosen more by high achievers than low achievers may be misleading or flawed.
A good distractor attracts more low achievers than high achievers.
Consistent scores indicate reliability, scoring free from examiner bias indicates objectivity, and low cost and ease of administration indicate usability.
When different examiners score identical answers differently, scoring is influenced by personal judgment, so objectivity is compromised.
Rejecting a test for excessive time, equipment and cost questions its usability (practicality).
A higher difficulty percentage means more students answered correctly, so the 65% item is easier than the 25% item.
Items with difficulty between 30% and 70% are generally considered desirable, neither too easy nor too hard.
A difficulty value of 85% means most students answered correctly, so the item is easy, not very difficult.
A discrimination index of 0.20 is below the commonly accepted range, so the item discriminates inadequately.
54 of 60 is 90% correct; a high difficulty percentage means most students answered correctly, so the item is relatively easy.
Agreement between independent scorers is inter-rater reliability.
Comparing performance on two equivalent versions of a test is parallel-forms (equivalent-forms) reliability.
Checking consistency between two halves of one test is split-half reliability.
Consistency alone does not guarantee validity: a test can consistently measure the wrong construct, so it is reliable but may lack validity.
Measuring an unintended construct consistently is reliable but not valid, which separates the two concepts.
Difficulty percentage = (correct responses / total students) × 100 = 18/60 × 100 = 30%.
Content validity concerns whether a test adequately covers the whole content domain; drawing from one chapter undermines it.
Face validity is the surface impression that a test appears to measure what it should.