PC PC Data Analysis and Interpretation 2 — Questions and Answers
Question 1: The standard error of measurement (SEM) is used by psychometricians to:
- Estimate the range of error around an observed score (Correct answer)
- Calculate the mean of a distribution
- Identify biased test items
- Determine passing cut scores
Correct answer: Estimate the range of error around an observed score
The SEM estimates the amount of error in an individual's observed score, helping construct confidence intervals around that score.
Question 2: Which analytical approach compares the performance of test items across different demographic subgroups to detect bias?
- Classical Test Theory analysis
- Differential Item Functioning (DIF) analysis (Correct answer)
- Rasch modeling
- Criterion-referenced scoring
Correct answer: Differential Item Functioning (DIF) analysis
Differential Item Functioning analysis statistically identifies items that perform differently for subgroups of equal ability, indicating potential bias.
Question 3: In Item Response Theory (IRT), the item discrimination parameter (a) indicates:
- The probability of guessing the correct answer
- How well an item differentiates between high and low ability examinees (Correct answer)
- The difficulty level of an item
- The number of response options
Correct answer: How well an item differentiates between high and low ability examinees
The discrimination parameter (a) reflects the slope of the item characteristic curve, showing how sharply performance differences correspond to ability differences.
Question 4: A negatively skewed score distribution in psychometric data indicates that:
- Most scores cluster at the low end of the scale
- Most scores cluster at the high end with a few low outliers (Correct answer)
- Scores are normally distributed
- The test has very low reliability
Correct answer: Most scores cluster at the high end with a few low outliers
A negatively skewed distribution has most scores concentrated at the higher end, with a tail extending toward lower scores.
Question 5: When conducting a multiple regression analysis with psychometric data, multicollinearity refers to:
- High correlations among predictor variables (Correct answer)
- Low reliability of the criterion variable
- Outliers in the dataset
- Non-normal distribution of residuals
Correct answer: High correlations among predictor variables
Multicollinearity occurs when predictor variables are highly intercorrelated, making it difficult to isolate each predictor's unique contribution.
Question 6: A psychometrician reviewing test data notices a bimodal score distribution. This most likely suggests:
- The test is highly reliable
- The sample may contain two distinct subpopulations (Correct answer)
- Items are optimally difficult
- Scores follow a normal distribution
Correct answer: The sample may contain two distinct subpopulations
A bimodal distribution with two distinct peaks often indicates the sample consists of two groups with different ability levels or characteristics.
PC PC Data Analysis and Interpretation 2
The standard error of measurement (SEM) is used by psychometricians to: