TAPAS Forced-Choice Methodology 1 — Questions and Answers
Question 1: What is the fundamental structure of a forced-choice item on TAPAS?
- A single statement rated on a 1-5 scale
- Two statements from different personality dimensions paired together, requiring a preference choice (Correct answer)
- A true/false question about behavior
- A multiple-choice question with four options
Correct answer: Two statements from different personality dimensions paired together, requiring a preference choice
Each TAPAS forced-choice item presents two statements from different personality dimensions, and the respondent must indicate which statement is more self-descriptive. This pairwise comparison format is the core of the forced-choice methodology. By comparing statements across dimensions rather than rating them independently, the format reduces several types of response bias.
Question 2: Why are the two statements in a TAPAS forced-choice pair matched on social desirability?
- To make the test more difficult
- To prevent respondents from identifying and choosing the more socially acceptable option (Correct answer)
- To reduce the number of items needed
- Social desirability matching is not actually done in TAPAS
Correct answer: To prevent respondents from identifying and choosing the more socially acceptable option
Matching statements on social desirability ensures that both options in a pair are equally attractive or equally neutral. When both statements are equally socially desirable, respondents cannot simply choose the 'better sounding' option and must instead make a genuine preference judgment. This is a key anti-faking mechanism in the TAPAS forced-choice design.
Question 3: What psychometric model underlies the scoring of TAPAS forced-choice items?
- Classical test theory only
- The Thurstonian item response theory (IRT) model for forced-choice data (Correct answer)
- Simple percentage scoring
- The Rasch model exclusively
Correct answer: The Thurstonian item response theory (IRT) model for forced-choice data
TAPAS uses a Thurstonian item response theory model specifically designed for forced-choice data. This model estimates the latent trait levels for each personality dimension while accounting for the comparative nature of forced-choice responses. Developed by researchers including Drasgow and colleagues, it overcomes the traditional limitations of ipsative scoring that plagued earlier forced-choice instruments.
Question 4: What is 'ipsative scoring' and why was it historically a problem for forced-choice personality measures?
- Ipsative scoring produces absolute ability levels
- Ipsative scoring produces within-person relative scores that cannot be meaningfully compared across individuals (Correct answer)
- Ipsative scoring is the most accurate form of personality measurement
- Ipsative scoring only applies to cognitive tests
Correct answer: Ipsative scoring produces within-person relative scores that cannot be meaningfully compared across individuals
Ipsative scoring occurs when forced-choice item scores are calculated such that each person's total across dimensions is constrained to be the same. This means scores only indicate relative strengths within a person, not absolute standing compared to others. For selection purposes, between-person comparisons are essential, making traditional ipsative scoring problematic for forced-choice instruments.
Question 5: How does TAPAS overcome the ipsative scoring problem that limited earlier forced-choice instruments?
- By converting to a Likert scale
- By using the Thurstonian IRT model to estimate normative (between-person comparable) scores from forced-choice data (Correct answer)
- By ignoring ipsativity
- By adding a separate normative section to the test
Correct answer: By using the Thurstonian IRT model to estimate normative (between-person comparable) scores from forced-choice data
TAPAS overcomes the ipsative scoring limitation through the Thurstonian IRT model, which provides normative score estimates from forced-choice data. This statistical model separates the individual trait levels from the comparative format, producing scores that can be meaningfully compared across individuals. This breakthrough made forced-choice measurement viable for high-stakes selection contexts.
Question 6: In a TAPAS forced-choice pair, what happens when a respondent is high on both dimensions represented?
- The item is skipped
- The respondent must still choose, and the IRT model accounts for this by modeling individual trait levels on both dimensions (Correct answer)
- Both statements are marked as endorsed
- The item is scored as incorrect
Correct answer: The respondent must still choose, and the IRT model accounts for this by modeling individual trait levels on both dimensions
Even when a respondent is high on both dimensions represented in a pair, they must still choose which statement is more self-descriptive. The Thurstonian IRT model handles this by modeling the probability of choosing one statement over the other as a function of the difference between the respondent's trait levels on the two dimensions. Over many pairs, the model can accurately estimate absolute levels on both dimensions.
