SCALE DRIFT IN ON‐LINE CALIBRATION
Open accessAbstract
Recent advances in psychometrics and computer technology encourage the development of model‐based methods of individualized testing on a microcomputer, where each examinee receives short tests and the number of pretest items that can be administered is severely restricted. On‐line (i.e., data is collected on operational equipment) methods for calibrating pretest items in this setting face new challenges. In the context of adaptive testing, two LOGIST‐based methods of on‐line calibration were developed and are described. These two methods are tried out and compared in a simulation study using data for the CAT‐ASVAB.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Stocking's method A works well with limited sample sizes, removes the necessity of picking anchor items to link ability scales, and estimates general ability based on all available information instead of only the items being tested for DIF.
“Stocking's method A works well with limited sample sizes, as was the case for the 2019 sample; it removes the necessity of picking anchor items to link the ability scales of the two groups on the same metric; and it estimates general ability based on all available information, instead of only the items being tested for DIF.”
From No negative Flynn effect in France: Why variations of intelligence should not be assessed using tests based on cultural knowledge (2021) · cites this paper as Stocking, 1988
The Method B section says Method A does not use anchor items and summarizes adaptive-test information in an ability estimate, which supports those aspects of the claim. But the paper studies on-line calibration, not linking two groups for DIF analysis, and the Discussion reports that Method A’s estimation error increases across rounds. It does not establish that Method A works well with the cited limited sample size. (LLM, conf=0.88)
Experimental — accuracy not yet measuredItem parameter estimation used Stocking's method A.
“Item parameter estimation was performed using Stocking's method A (Stocking, 1988; see also Birnbaum, 1968;Ban et al., 2001).”
From No negative Flynn effect in France: Why variations of intelligence should not be assessed using tests based on cultural knowledge (2021) · cites this paper as Stocking, 1988
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
No references could be resolved yet.
Research works printed but not linked: 8
- Bock, R. D. (1987). Final report on on-line calibration. With P. ~. Holland, C. Davis, M. V. Levine, F. Samejima, and M. L. Stocking. In preparation. Davis, C. E. (1987). Final report on on-line calibration. With R. D. Bock, P. W. Holland, M. V. Levine, F. Samejima, and M. L. Stocking. In preparation. Hamming, R. W. (1962). Numerical methods for scientists and engineers. New York: McGraw-Hill.
- Levine, M. V. (1987). Final report on on-line calibration. With R. D. Bock, C. E. Davis, P. W. Holland, F. Samejima, and M. L. Stocking. In preparation. 23308516, 1988, 1, · DOI
- Lord, F. M. (1984). Simulation of on-line item calibration for the adaptive ASVAB. Paper presented at the Office of Naval Research Conference on Model-Based Measurement, Princeton, NJ.
- Lord, F. M., & Wingersky , M. S. (1985). Sampling variances and covariances of parameter estimates in item response theory. In D. J. Weiss (Ed.), Proceedings of the 1982 Item Response Theory/Computerized Adaptive Testing Conference (pp. 69-88). Minneapolis, MN: Computerized Adaptive Testing Laboratory, Department of Psychology, University of Minnesota.
- Mosteller, F., & Tukey, J. W. (1977). Data analysis and regression: A second course in statistics. Reading, MA: Addison-Wesley.
- Owen, R. (1975). A Bayesian sequential procedure for quantal response in the context of adaptive mental testing. Journal, of the American Statistical Association, 70, 351-356. 23308516, 1988, 1, · DOI
- Wingersky, M. S. (1983). LOGIST: A program for computing maximum likelihood procedures for logistic test models. In R. Hambleton (Ed.), ERIBC monograph on applications of item response theory (pp. 45-56). Vancouver, BC: Educational Research Institute of British Columbia.
- Wingersky, M. S., & Lord, F. M. (1984). An investigation of methods for reducing sampling error in certain IRT procedures. Applied Psychological Measurement, ~, 347-364. 23308516, 1988, 1, · DOI
Cited by
Sorted by citations
Gonthier, C., Grégoire, J., & Besançon, M. (2021) · Intelligence · doi:10.1016/j.intell.2020.101512 · Cited by 20 · see citing works
- Authors
- Stocking, M. L.
- Affiliations
- Educational Testing Service (US)
- Corresponding author
- Stocking, Martha L.
- Venue
- ETS Research Report Series
- Publisher
- Wiley
- Published
- 1988-06
- Volume
- 1988
- Issue
- 1
- ISSN
- 2330-8516 (ISSN-L)
- OpenAlex
- W1986364309
- Cited by
- 37
- References
- 0
- Citations by year
- 2026: 6 · 2025: 1 · 2024: 3 · 2022: 4 · 2021: 1 · 2020: 1 · 2019: 3 · 2014: 2 · 2013: 2 · 2012: 1
- FWCI
- 1.95
- Type
- journal article
- Language
- en
- Open access
- bronze
- doi.org
- Locations
- ETS Research Report Series (publishedVersion)
- Topics
- Social Sciences › Decision Sciences › Management Science and Operations Research › Psychometric Methodologies and Testing
- Physical Sciences › Computer Science › Information Systems › Educational Technology and Assessment
- Physical Sciences › Computer Science › Artificial Intelligence › Intelligent Tutoring Systems and Adaptive Learning
- Keywords
- scale drift, online calibration, adaptive testing, computerized adaptive testing, item calibration, individualized testing
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
