Concise Notes: Test Construction & Table of Specification

Validity: the test measures what it intends to measure; content aligns with taught lessons.
Reliability: consistency of scores across retests or raters; e.g., Test-Retest method.
Fairness: use gender-sensitive language; avoid offense to any student.
Scorability: presence of a clear scoring key to enable consistent scoring.
Administrability: ease of administration (methods, time, complexity, materials) and low cost.
Adequacy: content coverage and diversity of item types to reflect the topic scope.
Simplicity: clear, student-friendly words and instructions; straightforward presentation.

Purpose: blueprint to ensure coverage of topics and to determine weight/importance of each topic.
Components (typical):
- Topics
- Objectives
- Content areas
- No. of items
- Percentage allocation
- Time/days allocated
One-Way Table of Specification (TOS):
- Crosses topics with item counts or percentages to ensure coverage by topic.
Two-Way Table of Specification (TOS):
- Crosses topics with cognitive levels (e.g., Knowledge, Comprehension, Application, Analysis).
- Used to ensure distribution of items across both topics and levels of learning.
Example format for item distribution:
- For a test with total items N and level/topic proportions p{t,l}, the number of items is n{t,l} = p_{t,l} \cdot N.
- Total across all cells satisfies \sum{t,l} n{t,l} = N.
Simple example for a 50-item test with level proportions rac{3}{8}, \frac{2}{8}, \frac{2}{8}, \frac{1}{8}:
n = 50 \times \frac{3}{8} = 18.75 \approx 19,
50 \times \frac{2}{8} = 12.5 \approx 13,
50 \times \frac{2}{8} = 12.5 \approx 13,
50 \times \frac{1}{8} = 6.25 \approx 6.
(Rounding adjusts to sum to 50.)

Claire and the TOS: follows comprehensiveness (ensures all tackled topics are covered).
Benedick and gender-sensitive language: fairness in assessment.
Item 4: validity (measures what it intends to measure).
Item 5: scorability (clear scoring key promotes ease of scoring).
Item 6: reliability (test-retest consistency).
Item 7: administrability (ease of administration; consider cost, time, complexity).
Item 8: adequacy (range of item samples for outcome determination).
Item 9: simplicity (familiar words and student-friendly instructions).
Item 10: validity factors – EXCEPT: length of the test does not directly affect validity (it affects reliability and practicality more).

Total items: N. Proportion for a topic/level: p{t,l}. Number of items in a cell: n{t,l} = p_{t,l} \cdot N.
For a single-dimension (One-Way) TOS:
nt = pt \cdot N.
Two-Way TOS example (Topics × Cognitive Levels): cross-tab with
\sum{t,l} n{t,l} = N.
For example distribution across levels in a 50-item test:
- Level proportions: \left(\frac{3}{8}, \frac{2}{8}, \frac{2}{8}, \frac{1}{8}\right)
- Item counts: n = 50 \times \frac{3}{8} \approx 19, \; 50 \times \frac{2}{8} \approx 13, \; 50 \times \frac{2}{8} \approx 13, \; 50 \times \frac{1}{8} \approx 6.

Theory: High-quality assessments ensure validity, reliability, fairness, practicality, and administrability.
Practice: Use a Table of Specification to blueprint coverage and weights.
Writing: Use clear, student-friendly language and a precise scoring key.
Distribution: Allocate items across topics and cognitive levels; adjust to total item count.
Evaluation: Consider the impact of test length on reliability and practicality, not validity.

One-Way TOS: Focus on topic coverage and item counts/percentages per topic.
Two-Way TOS: Add cognitive levels to the table to ensure balanced coverage of knowledge, understanding, and higher-order thinking across topics.

The Table of Specification serves as a blueprint for item construction and helps decide what to include and weight for each topic.
Group activities in the transcript illustrate how to allocate items across groups and level of difficulty for larger item banks.