Q1. A teacher records reading fluency as words read correctly per minute. This number is a:

• (A) Final judgement
• (B) Programme decision
• (C) Measure of fluency
• (D) Promotion rule

Q2. Deciding that a reading programme is worth expanding, based on fluency gains and costs, is:

• (A) Evaluation
• (B) Scoring only
• (C) Ranking only
• (D) Item writing

Q3. Which situation is assessment rather than a single measurement?

• (A) Counting correct answers on one quiz
• (B) Recording one test score
• (C) Timing one task
• (D) Combining quizzes, projects and observations to judge progress

Q4. A mid-course survey asks students what they find difficult so the teacher can adjust support. This is:

• (A) Summative grading
• (B) Formative feedback use
• (C) Final certification
• (D) External ranking

Q5. A final portfolio grade awarded at the end of a course is mainly:

• (A) Summative
• (B) Diagnostic only
• (C) Informal only
• (D) Placement only

Q6. Before teaching fractions, a teacher checks which students already understand equal parts. This check is:

• (A) Summative
• (B) Certifying
• (C) Ranking for prizes
• (D) Diagnostic and placement-related

Q7. A student passes because she scored at least 50 percent, the set standard for the unit. This is:

• (A) Norm referencing
• (B) Peer referencing
• (C) Criterion referencing
• (D) Random grading

Q8. A teacher reports that Ahmed scored higher than 90 percent of test takers nationally. This report uses:

• (A) Criterion referencing
• (B) Norm referencing
• (C) Diagnostic labelling
• (D) Formative scoring only

Q9. Which assessment is most likely to improve learning while it is happening?

• (A) A single final examination with no feedback
• (B) A ranking published once a year
• (C) A certificate issued at course end
• (D) Frequent low-stakes checks with feedback

Q10. A school uses test results, attendance and teacher reports together to decide support for a student. This decision is best described as:

• (A) A single measurement
• (B) A raw score
• (C) An evaluative judgement from multiple sources
• (D) An item analysis

Q11. Self-assessment helps learning mainly because students:

• (A) Avoid all teacher feedback
• (B) Compare their work with criteria and spot their own gaps
• (C) Receive a final grade sooner
• (D) Skip practice tasks

Q12. The key difference between measurement and evaluation is that evaluation:

• (A) Adds a judgement of value against a standard
• (B) Never uses numbers
• (C) Never uses tests
• (D) Only counts errors

Q13. Recording students home districts as categories for a survey uses which scale?

• (A) Ordinal
• (B) Nominal
• (C) Interval
• (D) Ratio

Q14. Ranking essays as excellent, good, satisfactory and weak gives data on which scale?

• (A) Nominal
• (B) Interval
• (C) Ratio
• (D) Ordinal

Q15. A score scale where a difference of 10 points means the same amount of change anywhere on the scale is:

• (A) Interval
• (B) Nominal
• (C) Ordinal
• (D) Categorical only

Q16. The length of a classroom in metres is measured on a:

• (A) Nominal scale
• (B) Ordinal scale
• (C) Ratio scale
• (D) Interval scale only

Q17. Which statistic is misleading if applied to nominal codes such as 1 for boys and 2 for girls?

• (A) Mode
• (B) Frequency count
• (C) Percentage in each group
• (D) Mean of the codes

Q18. A teacher orders students by test score and reports the middle student score. That middle value is the:

• (A) Mean
• (B) Median
• (C) Mode only
• (D) Range

Q19. On a ratio scale, which statement is valid?

• (A) Categories can only be counted
• (B) Order is absent
• (C) One quantity can be twice another
• (D) Zero is arbitrary

Q20. A Likert response from strongly agree to strongly disagree is best treated, strictly, as:

• (A) Ordinal
• (B) Ratio
• (C) Nominal
• (D) Count data

Q21. Which variable is ratio?

• (A) Number of correct answers
• (B) Favourite subject
• (C) Calendar year
• (D) Letter grade

Q22. The mean and standard deviation are most defensible for data that are at least:

• (A) Nominal
• (B) Ordinal only
• (C) Interval
• (D) Categorical only

Q23. Converting heights into ranks loses information about:

• (A) Order
• (B) Exact differences between heights
• (C) The tallest person
• (D) The shortest person

Q24. Zero marks on a test do not prove zero learning because test scores, in general, lack:

• (A) Nominal
• (B) Ordinal only
• (C) Interval with equal units
• (D) Ratio with a true zero for ability

Q25. A test of computer skill that only asks about keyboard shortcuts, ignoring files, safety and problem solving, suffers from:

• (A) High reliability
• (B) Good face validity only
• (C) Clear scoring
• (D) Poor content validity through under-representation

Q26. A candidate feels a test looks fair and relevant. This impression is:

• (A) Predictive validity
• (B) Construct validity
• (C) Face validity
• (D) Concurrent validity

Q27. A new test of teaching aptitude is compared with classroom observation ratings made the same term. This gives evidence of:

• (A) Predictive validity
• (B) Concurrent validity
• (C) Content validity only
• (D) Inter-rater reliability only

Q28. A screening test used to predict which trainees will complete a course is judged mainly by:

• (A) Predictive validity
• (B) Face validity only
• (C) Split-half reliability only
• (D) Item difficulty only

Q29. A measure of motivation should rise after a motivational programme and stay stable without one, if it has good:

• (A) Face validity only
• (B) Scoring reliability only
• (C) Construct validity
• (D) Test length

Q30. Expert review of a new test blueprint mainly establishes:

• (A) Content validity
• (B) Predictive validity
• (C) Test-retest reliability
• (D) Inter-rater reliability

Q31. A test is valid for selecting trainees but is then used to certify experienced teachers without new evidence. The problem is that:

• (A) Reliability is automatic
• (B) Face validity is enough
• (C) Content coverage is automatic
• (D) Validity does not transfer automatically to a new use

Q32. An English test written in very complex language for young learners will mainly threaten:

• (A) Test length
• (B) Validity, through construct-irrelevant difficulty
• (C) Scoring speed
• (D) Item count

Q33. A validity study with a very narrow, high-ability sample will probably show:

• (A) A higher validity coefficient
• (B) A lower validity coefficient due to restricted range
• (C) No effect on validity
• (D) Perfect prediction

Q34. Which combination is criterion-related?

• (A) Face plus content
• (B) Content plus construct
• (C) Split-half plus test-retest
• (D) Concurrent plus predictive

Q35. A reliable test that measures the wrong skill for a job is:

• (A) Consistent but invalid for that job decision
• (B) Automatically valid
• (C) Free from error
• (D) Useful without further evidence

Q36. The strongest validity evidence for a new theory-based trait measure comes from:

• (A) Appearance alone
• (B) Length alone
• (C) Multiple theory-based studies, including group differences and correlations
• (D) One easy item

Q37. A stopwatch that gives the same time for the same race on repeated checks is:

• (A) Valid for all races
• (B) Reliable
• (C) Easy to read only
• (D) Criterion-referenced

Q38. Giving the same spelling test in September and again in October, then correlating scores, estimates:

• (A) Test-retest reliability
• (B) Parallel-forms reliability
• (C) Inter-rater reliability
• (D) Content validity

Q39. Splitting a test into first-half and second-half scores and correlating them is a form of:

• (A) Test-retest method
• (B) Parallel-forms method
• (C) Split-half method
• (D) Predictive method

Q40. A test with many items all scored right or wrong can have internal consistency estimated by:

• (A) Face judgement only
• (B) Timing only
• (C) Item count only
• (D) KR-20

Q41. KR-21 will usually give a slightly lower estimate than KR-20 when item difficulties:

• (A) Vary across items
• (B) Are all equal
• (C) Are all zero
• (D) Are unknown

Q42. Two judges rate dance performances and agree closely after training on a rubric. This shows improved:

• (A) Content validity
• (B) Inter-rater reliability
• (C) Predictive validity
• (D) Item difficulty

Q43. If a test has SD 6 and reliability 0.64, its SEM is:

• (A) 2.4
• (B) 3.0
• (C) 4.8
• (D) 3.6

Q44. A test with a large SEM should be used with:

• (A) Complete confidence in single scores
• (B) No reporting at all
• (C) Caution, reporting a band around each score
• (D) Only rank orders

Q45. Which factor usually raises reliability?

• (A) Fewer, ambiguous items
• (B) Noisy conditions
• (C) More good items and clear directions
• (D) Rushed guessing

Q46. A reliability of 0.60 for a high-stakes selection test is generally:

• (A) Too low for high-stakes decisions alone
• (B) Perfect
• (C) Better than 0.90
• (D) Proof of validity

Q47. A test-retest gap of six months for a rapidly changing skill may underestimate reliability because:

• (A) Memory effects dominate
• (B) Items are forgotten
• (C) Scoring changes
• (D) True ability has changed in between

Q48. Standardising test administration helps reliability by:

• (A) Making items easier
• (B) Reducing irrelevant differences between sittings
• (C) Shortening the test
• (D) Removing all true differences

Q49. An item answered correctly by 90 of 180 students has difficulty:

• (A) 0.50
• (B) 0.90
• (C) 0.18
• (D) 0.25

Q50. An item with p = 0.05 is best described as:

• (A) Moderate
• (B) Easy
• (C) Ideal for all tests
• (D) Extremely difficult

Q51. Upper group 15 of 30 correct, lower group 3 of 30 correct. The discrimination index is:

• (A) 0.20
• (B) 0.40
• (C) 0.50
• (D) 0.60

Q52. A discrimination index of 0.00 means the item:

• (A) Is very difficult
• (B) Is very easy
• (C) Does not separate high and low scorers at all
• (D) Has a wrong key for sure

Q53. Which item should be revised first?

• (A) p = 0.55, D = 0.45
• (B) p = 0.20, D = -0.20
• (C) p = 0.50, D = 0.40
• (D) p = 0.60, D = 0.35

Q54. In distractor analysis, the key should ideally be chosen by:

• (A) Fewer high scorers than low scorers
• (B) Nobody at all
• (C) More high scorers than low scorers
• (D) Only the lowest group

Q55. A distractor chosen by 25 percent of students, mostly low scorers, is:

• (A) Working well
• (B) Dead
• (C) Mis-keyed for sure
• (D) Too easy

Q56. The best general difficulty range for items in a norm-referenced test is around:

• (A) 0.05 to 0.10
• (B) 0.90 to 0.95
• (C) 0.10 to 0.20 only
• (D) 0.30 to 0.70

Q57. An item with p = 0.82, D = 0.12 and two unused distractors is best handled by:

• (A) Retaining it unchanged
• (B) Doubling its marks
• (C) Using it as a model
• (D) Revising the item and distractors, or dropping it

Q58. Item analysis is most useful for:

• (A) Selecting and improving items for future tests
• (B) Setting the pass mark for a single class
• (C) Ranking schools by their average scores
• (D) Replacing classroom teaching with testing

Q59. A negative point-biserial for an item suggests that:

• (A) The item is too easy
• (B) The item has no distractors
• (C) Low scorers do better on the item than high scorers
• (D) High scorers tend to answer it correctly

Q60. A test developer keeps an item with p = 0.45, D = 0.50 and all distractors attracting some low scorers. This decision is sound because the item is:

• (A) Too difficult to use
• (B) Moderate, discriminating and well-distracted
• (C) Valid for essays only
• (D) Reliable without a trial