Q1. A teacher counts spelling errors in each essay and reports the counts. This step is:

• (A) Evaluation
• (B) Measurement
• (C) Diagnosis
• (D) Certification

Q2. A school board judges a literacy programme as successful after comparing results with its targets. This judgement is:

• (A) Scoring
• (B) Scaling
• (C) Item writing
• (D) Evaluation

Q3. Which tool best supports assessment of speaking skills?

• (A) Attendance register only
• (B) Textbook page count only
• (C) Observation of oral presentations with criteria
• (D) Counting library books only

Q4. A quiz used only to give students feedback and guide revision, with no grade recorded, is:

• (A) Formative
• (B) Summative
• (C) Certifying
• (D) Selective

Q5. End-of-year board examinations that decide promotion are mainly:

• (A) Diagnostic
• (B) Formative
• (C) Informal
• (D) Summative

Q6. A teacher finds through a quick check that most students can add fractions but not subtract them. The next step of re-teaching subtraction shows:

• (A) Diagnostic use of assessment
• (B) Final certification
• (C) Norm referencing
• (D) External ranking

Q7. A student score is reported as above the 80th percentile. This interpretation is:

• (A) Criterion-referenced
• (B) Mastery-based only
• (C) Norm-referenced
• (D) Diagnostic only

Q8. A practical skills test requires students to perform each step correctly to pass. This pass decision is:

• (A) Norm-referenced only
• (B) Criterion-referenced
• (C) Peer-referenced only
• (D) Random

Q9. Which assessment practice gives the fullest picture of learning?

• (A) One test on one day
• (B) One oral question
• (C) Multiple methods over time, including tests, projects and observation
• (D) One homework task

Q10. Students review each other drafts using a checklist. This is an example of:

• (A) Standardised testing
• (B) Peer assessment
• (C) Final grading
• (D) Placement testing

Q11. Measurement without evaluation can tell a teacher:

• (A) What score was obtained, but not whether it is good enough
• (B) Whether a programme should continue
• (C) Whether a student should be promoted
• (D) Whether a course is excellent

Q12. A teacher uses weekly short tests to monitor progress and adjust support for weak students. This is best called:

• (A) One-off certification
• (B) External examination
• (C) Final summative grading
• (D) Continuous formative assessment

Q13. Sorting students into science, arts and commerce streams for records uses which scale?

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

Q14. A music competition ranks performers gold, silver and bronze. This ranking is:

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

Q15. Test scores expressed as standard scores with equal units but no true zero are treated as:

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

Q16. The mass of chemicals used in an experiment, measured in grams, is on a:

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

Q17. For data in categories such as blood groups, the appropriate average is the:

• (A) Mode
• (B) Mean
• (C) Median of codes
• (D) Standard deviation of codes

Q18. A claim that 20 degrees Celsius is twice as hot as 10 degrees is wrong because Celsius is:

• (A) Nominal
• (B) Ordinal
• (C) Ratio
• (D) Interval, with an arbitrary zero

Q19. Which data can be ordered but should not be averaged as if gaps are equal?

• (A) Heights in centimetres
• (B) Weights in kilograms
• (C) Class ranks
• (D) Times in seconds

Q20. The difference between first and second place may be tiny, while the gap between second and third may be large. This illustrates a limit of:

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

Q21. A survey asks teachers to rate workload as low, medium or high. These ratings are:

• (A) Ratio
• (B) Ordinal
• (C) Nominal
• (D) Interval with a true zero

Q22. Which operation is valid for interval data?

• (A) Saying one score is twice another
• (B) Counting only
• (C) Comparing differences between scores
• (D) Ordering categories only

Q23. A researcher wants to report the coefficient of variation, which divides the standard deviation by the mean. This statistic is defensible only for data on which scale?

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

Q24. A teacher reports the median class position rather than the mean position because positions are:

• (A) Ordinal, so the median is safer
• (B) Ratio, so the mean is unsafe
• (C) Nominal, so no average works
• (D) Interval, so the median is wrong

Q25. An Urdu test that uses only poetry questions, although the syllabus also covers prose and grammar, lacks:

• (A) Reliability
• (B) Scoring speed
• (C) Content validity
• (D) Face appeal only

Q26. A test that looks like a proper maths test to parents and students, with no further evidence, has:

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

Q27. A test of leadership is studied to see if its scores match leadership theory, including differences between trained and untrained groups. This is evidence for:

• (A) Construct validity
• (B) Face validity only
• (C) Inter-rater reliability
• (D) Split-half reliability

Q28. A college compares new admission scores with current first-semester grades. This checks:

• (A) Content validity
• (B) Face validity
• (C) Test-retest reliability
• (D) Concurrent validity

Q29. An entry test is used to forecast training success after six months. The relevant validity is:

• (A) Content validity
• (B) Predictive validity
• (C) Face validity
• (D) Internal consistency

Q30. A history test includes many questions on one small topic and none on others. The main threat is to:

• (A) Scoring reliability
• (B) Test length
• (C) Content validity
• (D) Item difficulty alone

Q31. A test intended for Grade 10 is given to Grade 6 students and fails to predict their later grades. The likely issue is that validity depends on:

• (A) The population and use, not the test alone
• (B) The number of items in the test alone
• (C) The time limit used for the test alone
• (D) The difficulty of one item alone

Q32. If a test measures test anxiety as well as maths skill, its maths scores contain:

• (A) Pure true variance
• (B) No error at all
• (C) Only content error
• (D) Construct-irrelevant variance

Q33. The best first step to support content validity when building a new test is to:

• (A) Prepare a table of specifications from objectives
• (B) Print the test in colour
• (C) Add more easy items
• (D) Shorten instructions

Q34. A validity coefficient near zero between a test and its intended criterion suggests the test:

• (A) Is highly reliable
• (B) Covers content well
• (C) Looks professional
• (D) Does not predict or reflect that criterion

Q35. Which evidence belongs to criterion-related validity?

• (A) Expert ratings of coverage
• (B) Student opinions on appearance
• (C) Correlation with job performance ratings
• (D) Agreement between two markers

Q36. A test developer finds that a reasoning test correlates 0.70 with another reasoning test and 0.10 with a clerical speed test. This supports:

• (A) Face validity only
• (B) Construct validity through convergent and discriminant patterns
• (C) Split-half reliability only
• (D) Content validity only

Q37. A test that gives nearly the same score when repeated under similar conditions is:

• (A) Reliable
• (B) Valid for all uses
• (C) Easy
• (D) Criterion-referenced

Q38. Administering the same test twice with a suitable gap and correlating scores estimates:

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

Q39. Two parallel forms given close together mainly control for:

• (A) Inter-rater disagreement
• (B) Memory of specific items
• (C) Content coverage
• (D) Face appeal

Q40. A split-half correlation of 0.60 for half a test, corrected by Spearman-Brown for full length, will usually become:

• (A) Lower than 0.60
• (B) Exactly 0.60
• (C) Higher than 0.60
• (D) Zero

Q41. KR-21 is preferred over KR-20 when the test developer:

• (A) Has essay items only
• (B) Needs judge agreement
• (C) Has two administrations
• (D) Wants a quick estimate assuming equal item difficulty

Q42. A detailed answer key and marking guide for short-answer items mainly improves:

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

Q43. If SD is 12 and reliability is 0.91, the SEM is closest to:

• (A) 3.6
• (B) 6.0
• (C) 9.0
• (D) 12.0

Q44. Two scores 70 and 74 on a test with SEM 5 should be interpreted as:

• (A) Clearly different abilities
• (B) Proof of cheating
• (C) Possibly the same true score, given measurement error
• (D) Proof of different tests

Q45. Which practice most harms reliability?

• (A) Clear timing
• (B) Ambiguous stems and unclear directions
• (C) Standard scoring
• (D) Enough testing time

Q46. A test with items that all measure different, unrelated topics will likely show:

• (A) Low internal consistency
• (B) High internal consistency
• (C) Perfect predictive validity
• (D) No measurement error

Q47. Reliability coefficients range, in practice, from:

• (A) 0 to 10
• (B) -1 to 0 only
• (C) -1.00 to +1.00, like a correlation
• (D) 0.00 to 1.00

Q48. A longer testing time, up to a reasonable limit, usually helps reliability by:

• (A) Adding fatigue only
• (B) Reducing the number of items answered with care
• (C) Reducing guessing pressure
• (D) Making items harder

Q49. The difficulty index is also called:

• (A) Discrimination power
• (B) p-value of the item
• (C) Reliability coefficient
• (D) Validity coefficient

Q50. An item answered by 120 of 300 students has p equal to:

• (A) 0.12
• (B) 0.25
• (C) 0.40
• (D) 0.60

Q51. An item with p = 0.92 is most useful as:

• (A) A strong discriminator
• (B) A mastery check only
• (C) A difficult challenge item
• (D) An easy warm-up or mastery item, not a discriminator

Q52. Discrimination index D is calculated as:

• (A) (Upper correct minus lower correct) divided by group size
• (B) Total correct divided by total students
• (C) Upper correct plus lower correct
• (D) Number of distractors divided by items

Q53. Upper 18 of 20 correct, lower 8 of 20 correct. D equals:

• (A) 0.20
• (B) 0.40
• (C) 0.50
• (D) 0.90

Q54. A D value of -0.30 should lead a developer to:

• (A) Keep the item unchanged
• (B) Use it for top students only
• (C) Increase its marks
• (D) Investigate and likely revise or delete it

Q55. An item with D between 0.20 and 0.29 is generally:

• (A) Excellent
• (B) Acceptable, but could be improved
• (C) Completely useless
• (D) Negative

Q56. A distractor that attracts mainly strong students is a warning that the option may be:

• (A) Ambiguous or partly correct
• (B) Too obviously wrong
• (C) Too short
• (D) Placed last

Q57. Which distractor is non-functioning?

• (A) One chosen by 30 percent of low scorers
• (B) One chosen by 20 percent of all students
• (C) One chosen mainly by low scorers
• (D) One chosen by almost nobody

Q58. After analysis, an item has p = 0.35, D = 0.38 and all options used. The best decision is to:

• (A) Delete it at once
• (B) Retain it
• (C) Make it much easier
• (D) Change the key at random

Q59. Item analysis helps a test developer to:

• (A) Improve items and build a better item bank
• (B) Avoid all future testing
• (C) Rank schools directly
• (D) Set the pass mark without judgement

Q60. A positive point-biserial for an item means students who do well overall tend to:

• (A) Skip the item
• (B) Choose distractors
• (C) Answer the item correctly
• (D) Take longer on the item