LET Elementary Assessment of Learning — Constructing, Administering and Analyzing TestsRevision Notes

Exam Tips

• LET items on MCQ construction often show you a flawed item and ask WHICH RULE was violated. Practice naming the specific rule, not just saying 'the item is bad.'
• Memorize the three 'avoid' keywords for true-false: ABSOLUTES signal false; QUALIFIERS signal true.
• For matching type, remember the rule: MORE responses than premises. This is a frequently tested detail.
• When a stem ends with 'a' or 'an,' check whether this creates a grammatical clue pointing to one option — this is a classic LET trap.
• The LET may present an item with four options, one of which is much longer and more qualified. The flaw is that the key stands out by length — this violates option homogeneity.
• Remember: 'all of the above' is problematic because partial knowledge (knowing two of four options are correct) leads to a correct response, undermining the test's validity.

Key Points

• Selected-response items include Multiple Choice Questions (MCQ), True-False, and Matching Type — the formats most commonly used in Philippine elementary classrooms and the LET itself.
• An MCQ has two parts: the STEM (the problem or question) and the OPTIONS (one correct answer called the KEY, plus wrong choices called DISTRACTORS).
• The stem must state a single, clear problem. Ideally, a student can answer the question before reading the options.
• Place as much wording as possible in the stem to keep options short and parallel.
• Avoid using negatives (NOT, EXCEPT) in stems. If unavoidable, CAPITALIZE or BOLD the negative word to make it visible.
• Distractors must be PLAUSIBLE — they should reflect common student errors or misconceptions, not obviously wrong answers.
• All options must be HOMOGENEOUS in content, grammar, and approximate length. The key should not stand out by being the longest or most detailed option.
• Avoid grammatical clues: the article 'a' or 'an' in the stem should not point to a specific option.
• Avoid absolute words (always, never, all, none) in distractors — test-wise students eliminate these on sight.
• Use 'all of the above' and 'none of the above' sparingly; 'all of the above' allows partial knowledge to earn full credit.
• Randomize the position of the key across items so there is no predictable pattern (e.g., always 'C').
• For True-False items: test ONE idea per statement; avoid specific determiners. Words like 'always, all, never, none' usually signal FALSE; words like 'sometimes, generally, usually, often' usually signal TRUE.
• For Matching Type: keep both columns HOMOGENEOUS (same category); provide MORE responses than premises so the last answer cannot be guessed by elimination; keep lists short (5–8 premises); place shorter items on the right (responses column).
• Write clear directions specifying the basis for matching and whether responses may be reused.

Definitions

Section Title

Writing Selected-Response Items

Common Mistakes

• Writing stems that are incomplete or that do not fully communicate the problem before the student reads the options.
• Making the correct answer noticeably longer or more detailed than the distractors — students learn to pick the longest option.
• Using 'a' or 'an' at the end of a stem when only one option grammatically fits — this is an unintentional give-away (e.g., 'A synonym is an ___' when only option A starts with a vowel).
• Including implausible distractors that even the weakest student can eliminate immediately.
• Using absolute determiners (always, never) in distractors, making them easy to eliminate for test-wise examinees.
• In matching type, using the same number of premises and responses — this allows the last pair to be matched by elimination.
• In true-false items, testing two ideas in one statement — if one part is true and one is false, the item is confusing and unfair.
• Placing the key in the same position repeatedly (e.g., always option B), creating a pattern that savvy students can exploit.

Exam Tips

• LET items may ask you to distinguish between restricted-response and extended-response essays — know that RESTRICTED limits both content and form; EXTENDED allows freedom.
• Remember the three reliability strategies for essay scoring: (1) prepare rubric/model answer FIRST, (2) score item-by-item across all papers, (3) score anonymously.
• Analytic rubric = detailed, criterion-by-criterion; Holistic rubric = single overall score. The LET may ask which is better for formative vs. summative use.
• For completion items, the mnemonic is: BLANK NEAR THE END, EQUAL BLANK LENGTHS, NO GRAMMAR CLUES.
• If the LET asks about the purpose of anonymous essay scoring, the answer is to reduce HALO BIAS, not cheating prevention.

Key Points

• Constructed-response items require students to produce their own answer rather than select from given options. They include Completion (short answer) and Essay items.
• COMPLETION items: require a single, brief, correct answer. Place the blank NEAR THE END of the statement so the problem is clear before the blank appears. Use EQUAL-LENGTH blanks so length does not give away the answer. Avoid grammatical clues like 'a' or 'an' before the blank.
• ESSAY items measure higher-order thinking — synthesis, evaluation, and organization — that objective items cannot fully capture.
• RESTRICTED-RESPONSE essays limit both the content and form of the answer. They are more focused, easier to score reliably, and can sample more content across a test. Example: 'List three causes of the Cry of Pugad Lawin and explain each in one sentence.'
• EXTENDED-RESPONSE essays give students freedom to organize their ideas and demonstrate higher-order thinking. They are richer in terms of what they reveal about student thinking but are harder to score consistently. Example: 'Evaluate the impact of the K-12 curriculum reform on the teaching and learning of Mother Tongue in Grades 1-3.'
• To score essays RELIABLY, always prepare a MODEL ANSWER or RUBRIC before scoring begins.
• Score ONE item across ALL papers before moving to the next item — this is called the POINT METHOD or ITEM-BY-ITEM scoring and keeps your standard consistent.
• Score ANONYMOUSLY where possible to reduce HALO BIAS (the tendency to let your general impression of a student affect your score on a specific item).
• Use an ANALYTIC RUBRIC for detailed criterion-by-criterion feedback or a HOLISTIC RUBRIC for a single overall judgment.
• Analytic rubrics are better for formative assessment; holistic rubrics are faster for summative assessments.

Definitions

Section Title

Writing Constructed-Response Items

Common Mistakes

• Placing the blank at the BEGINNING of a completion item instead of near the end — students don't know the context of the blank when they encounter it first.
• Using blanks of different lengths in completion items, giving students clues about the expected answer.
• Using 'a' or 'an' before a blank in a completion item — this reveals whether the answer starts with a vowel or consonant.
• Writing essay items without preparing a rubric or model answer in advance — this leads to inconsistent scoring.
• Scoring all items for one student before moving to the next student — this invites halo bias and inconsistent standards.
• Writing extended-response essays when restricted-response would better match the learning objective and available scoring time.
• Failing to define scoring criteria for essay items, leaving room for subjective or biased grading.

Formulas

Exam Tips

• Memorize the guessing correction formula: Score = R − W/(k − 1). Practice with 3-option (k=3) and 4-option (k=4) items.
• For k=4 (standard 4-option MCQ): the penalty per wrong answer = 1/3 (approximately 0.33 points).
• For k=5 (5-option MCQ): the penalty per wrong answer = 1/4 (0.25 points).
• LET question type: 'What should a teacher do to minimize test anxiety during administration?' — acceptable answers include clear instructions, comfortable setting, encouraging tone, and adequate time.
• The sequence of test assembly: group by type → arrange easy to hard → write clear directions → prepare answer key BEFORE printing.
• The LET may ask: 'In the correction for guessing formula, what happens to omitted items?' Answer: They are NOT penalized — neither added nor subtracted.

Key Points

• Proper test assembly ensures clarity, fairness, and accurate measurement. Poor assembly introduces construct-irrelevant variance — students lose points due to confusing layout, not lack of knowledge.
• GROUP ITEMS BY TYPE: all true-false items together, all MCQs together, all essay items together. Mixed formats confuse students.
• ARRANGE ITEMS FROM EASY TO DIFFICULT within each section. This builds student confidence and reduces test anxiety, particularly important for Grades 1-6 pupils in Philippine elementary schools.
• Write CLEAR DIRECTIONS for each section: specify how to answer, how points are earned, and how much time is allocated.
• Keep an item and all its options ON THE SAME PAGE — never split an item across two pages.
• Space items for readability. Crowded tests increase cognitive load and reduce validity.
• Prepare the ANSWER KEY and scoring plan BEFORE printing and distributing the test.
• For test administration: provide a COMFORTABLE, WELL-LIT, QUIET setting. This is not just good practice — it reflects the teacher's ethical responsibility under the Code of Ethics for Professional Teachers and child welfare principles connected to RA 7610.
• Give CLEAR ORAL AND WRITTEN INSTRUCTIONS and state the time limit explicitly at the start.
• Minimize TEST ANXIETY through positive reinforcement during instructions and fair proctoring.
• PROCTORING: prevent cheating through seating arrangement (spacing students apart) and active monitoring — without making students feel they are being treated as suspects.
• Manage TIMING fairly so all learners have a reasonable chance to complete the test.
• CORRECTION FOR GUESSING FORMULA: Score = R − W / (k − 1), where R = number of right answers, W = number of wrong answers, k = number of options per item. OMITTED items are NOT counted as wrong.
• Worked example: 50-item, 4-option test; student answers 40 correctly, 9 wrong, 1 omitted. Score = 40 − 9/(4−1) = 40 − 3 = 37.
• The logic of the guessing correction: on a 4-option item, a pure guesser has a 1-in-4 chance of being right. Every 3 wrong answers statistically 'came with' one lucky correct guess, so 3 wrong = deduct 1 point.
• Most Philippine classroom tests do NOT use the guessing correction, but the LET tests both the computation and the rationale behind it.

Definitions

Section Title

Assembling and Administering the Test

Common Mistakes

• Splitting a test item across two pages — students may not notice the continuation and answer incompletely.
• Writing directions AFTER printing the test — directions must be planned and pre-tested for clarity.
• Preparing the answer key AFTER students have already taken the test — this can lead to key errors and compromised objectivity.
• Failing to announce the time limit at the start — students have a right to know how much time they have.
• In the correction for guessing formula, counting OMITTED items as wrong — omissions are not penalized.
• Computing guessing correction as W/(k) instead of W/(k−1) — the denominator is k MINUS 1, not k.

Formulas

Exam Tips

• The LET frequently asks: 'A test item had p = 0.85. What does this mean?' Answer: The item is VERY EASY — 85% of students got it right. This item should be revised because it barely discriminates.
• Memorize the five difficulty bands: 0.00–0.20 Very Difficult | 0.21–0.40 Difficult | 0.41–0.60 Moderately Difficult | 0.61–0.80 Easy | 0.81–1.00 Very Easy.
• The ideal p is approximately 0.50. Items near this value show the strongest discrimination.
• Quick computation tip: always check — does your p fall between 0 and 1? If not, you made a math error.
• When the LET gives you upper/lower group data: add both groups' correct responses, then divide by the total count in both groups.

Key Points

• Item analysis is conducted AFTER scoring to evaluate the quality of individual test items. It answers the question: 'Did this item do its job?'
• The DIFFICULTY INDEX (p) measures the PROPORTION of examinees who answered an item CORRECTLY.
• Formula: p = Number of Correct Responses / Total Number of Examinees
• p ranges from 0.00 to 1.00. HIGHER p means EASIER item (more students got it right). LOWER p means HARDER item (fewer students got it right). This is counterintuitive — memorize it.
• Difficulty bands: 0.00–0.20 = Very Difficult; 0.21–0.40 = Difficult; 0.41–0.60 = Moderately Difficult (Ideal range); 0.61–0.80 = Easy; 0.81–1.00 = Very Easy.
• The IDEAL DIFFICULTY is approximately p = 0.50, where the item best separates strong from weak learners.
• Items with p above 0.85 (very easy) or below 0.15 (very difficult) contribute very little discrimination and should be revised.
• When computing p using only the UPPER and LOWER groups: p = (correct in upper group + correct in lower group) / total students in both groups.
• Worked example (full class): 40 students, 30 correct. p = 30/40 = 0.75 → EASY. The item is retainable but not highly discriminating.
• Worked example (upper/lower groups only): Upper group (10 students): 8 correct; Lower group (10 students): 3 correct. p = (8+3)/20 = 11/20 = 0.55 → Moderately Difficult. Good item — right in the ideal band.

Definitions

Section Title

Item Analysis: Difficulty Index

Common Mistakes

• Confusing a HIGH p with a HARD item — HIGH p means MORE students got it right, so it is an EASY item.
• Saying that p = 1.00 is the 'best' item — it actually means EVERYONE got it right, so it discriminates nobody and should be revised.
• Forgetting to divide by the TOTAL number of students in BOTH groups when computing p from upper/lower group data.
• Using only the upper group's correct responses in the p formula instead of combining both groups.
• Confusing the difficulty index range brackets — 0.41–0.60 is MODERATELY DIFFICULT (the ideal), not 'moderate' from 0.50–0.70.

Formulas

Exam Tips

• LET computation pattern: You are given upper group and lower group responses for one item. Compute p and D. Practice until you can do this in under 2 minutes.
• Memorize the five D categories: ≥ 0.40 = Excellent; 0.30–0.39 = Good; 0.20–0.29 = Fair; ≤ 0.19 = Poor; Negative = Defective.
• If the LET asks 'What is the FIRST action when D is negative?' — Answer: CHECK THE ANSWER KEY for miskeying.
• Distractor analysis question type: 'A distractor was chosen by 8 students in the upper group and 1 in the lower group. What does this suggest?' Answer: The item may be miskeyed or ambiguous — upper-group students are picking a wrong answer more than lower-group students.
• The relationship between p and D: items near p = 0.50 tend to have the highest D values. Items near p = 0.00 or p = 1.00 tend to have D near 0.00 — they cannot discriminate because there is no variability in responses.
• Quick mnemonic for distractor analysis: GOOD distractor = more LOW-group students. FUNCTIONAL KEY = more HIGH-group students. REVERSED = check the key.

Key Points

• The DISCRIMINATION INDEX (D) measures how well an item SEPARATES high scorers (upper group) from low scorers (lower group).
• Formula: D = (CU − CL) / n, where CU = correct in upper group, CL = correct in lower group, n = number of students in ONE group (they must be equal in size).
• D ranges from −1.00 to +1.00.
• A POSITIVE D means more students in the upper group answered correctly — this is the desired outcome. The item is 'working.'
• A NEGATIVE D is a RED FLAG: the lower group outperformed the upper group on this item. This usually means the item is MISKEYED (wrong answer listed as the key) or the item is seriously flawed.
• D = 0.00 means the item does not discriminate at all — equal numbers in both groups answered correctly.
• Discrimination bands: 0.40 and above = Very Good/Excellent → Retain; 0.30–0.39 = Good → Retain with possible minor improvement; 0.20–0.29 = Fair/Marginal → Needs improvement; 0.19 and below = Poor → Revise or reject; Negative = Defective → Discard; verify key first.
• The 27% convention: Kelley's research showed that using the top 27% and bottom 27% maximizes the statistical power of the discrimination analysis. In a class of 40, each group = approximately 11 students (0.27 × 40 = 10.8 ≈ 11). In smaller classes, teachers may use the upper and lower halves.
• DISTRACTOR ANALYSIS: After computing D, examine how each WRONG option performed. A GOOD distractor attracts MORE lower-group students than upper-group students. A NON-FUNCTIONAL distractor is chosen by NOBODY — it should be replaced with a more plausible option. A distractor chosen by MORE upper-group than lower-group students signals AMBIGUITY or MISKEYING — recheck the answer key.
• The KEY should always be chosen by more upper-group than lower-group students — mirroring positive discrimination.
• COMPLETE WORKED EXAMPLE: Key = C. Upper group (n=10): A=1, B=0, C=8, D=1. Lower group (n=10): A=3, B=2, C=3, D=2. Difficulty: p = (8+3)/20 = 0.55 → Moderately Difficult. Discrimination: D = (8−3)/10 = 5/10 = 0.50 → Very Good. Distractors A and D attract more lower-group students (functioning). B attracts 2 lower-group students and 0 upper-group (functioning but could be strengthened). VERDICT: RETAIN the item.
• The built-in tension: very easy items (high p) and very hard items (low p) both have limited room to discriminate, which is why items near p = 0.50 typically show the strongest D values.
• WHAT TO DO with flagged items: RETAIN if p = 0.30–0.70 AND D ≥ 0.30; REVISE if D = 0.20–0.29 or non-functional distractors; REJECT/REWRITE if D ≤ 0.19 or negative D; VERIFY KEY whenever the upper group underperforms on the keyed response.

Definitions

Section Title

Item Analysis: Discrimination Index and Distractor Analysis

Common Mistakes

• Dividing by the TOTAL number of students in BOTH groups instead of ONE group when computing D — the denominator is n (one group's size), not 2n.
• Forgetting that a NEGATIVE D means the item might be MISKEYED, not necessarily that the item topic is wrong. Always verify the key first.
• Concluding that an item is 'excellent' based only on D without checking p — an item with D = 0.50 but p = 0.10 (only 10% correct) is too hard for most students and still needs revision.
• Calling a distractor 'good' because many students chose it, without checking whether it attracted MORE lower-group than upper-group students.
• Confusing the discrimination index range (−1 to +1) with the difficulty index range (0 to 1).
• Using different group sizes for upper and lower groups in the D formula — both groups MUST be equal in size.