@conference {2652, title = {Efficiency of Targeted Multistage Calibration Designs under Practical Constraints: A Simulation Study}, booktitle = {IACAT 2017 Conference}, year = {2017}, month = {08/2017}, publisher = {Niigata Seiryo University}, organization = {Niigata Seiryo University}, address = {Niigata, Japan}, abstract = {

Calibration of an item bank for computer adaptive testing requires substantial resources. In this study, we focused on two related research questions. First, we investigated whether the efficiency of item calibration under the Rasch model could be enhanced by calibration designs that optimize the match between item difficulty and student ability (Berger, 1991). Therefore, we introduced targeted multistage calibration designs, a design type that refers to a combination of traditional targeted calibration designs and multistage designs. As such, targeted multistage calibration designs consider ability-related background variables (e.g., grade in school), as well as performance (i.e., outcome of a preceding test stage) for assigning students to suitable items.

Second, we explored how limited a priori knowledge about item difficulty affects the efficiency of both targeted calibration designs and targeted multistage calibration designs. When arranging items within a given calibration design, test developers need to know the item difficulties to locate items optimally within the design. However, usually, no empirical information about item difficulty is available before item calibration. Owing to missing empirical data, test developers might fail to assign all items to the most suitable location within a calibration design.

Both research questions were addressed in a simulation study in which we varied the calibration design, as well as the accuracy of item distribution across the different booklets or modules within each design (i.e., number of misplaced items). The results indicated that targeted multistage calibration designs were more efficient than ordinary targeted designs under optimal conditions. Especially, targeted multistage calibration designs provided more accurate estimates for very easy and 52 IACAT 2017 ABSTRACTS BOOKLET very difficult items. Limited knowledge about item difficulty during test construction impaired the efficiency of all designs. The loss of efficiency was considerably large for one of the two investigated targeted multistage calibration designs, whereas targeted designs were more robust.

References

Berger, M. P. F. (1991). On the efficiency of IRT models when applied to different sampling designs. Applied Psychological Measurement, 15(3), 293\–306. doi:10.1177/014662169101500310

Session Video

}, keywords = {CAT, Efficiency, Multistage Calibration}, url = {https://drive.google.com/file/d/1ko2LuiARKqsjL_6aupO4Pj9zgk6p_xhd/view?usp=sharing}, author = {Stephanie Berger and Angela J. Verschoor and Theo Eggen and Urs Moser} } @conference {2669, title = {FastCAT {\textendash} Customizing CAT Administration Rules to Increase Response Efficiency}, booktitle = {IACAT 2017 Conference}, year = {2017}, month = {08/2017}, publisher = {Niigata Seiryo University}, organization = {Niigata Seiryo University}, address = {Niigata, Japan}, abstract = {

A typical pre-requisite for CAT administration is the existence of an underlying item bank completely covering the range of the trait being measured. When a bank fails to cover the full range of the trait, examinees who are close to the floor or ceiling will often never achieve a standard error cut-off and examinees will be forced to answer items increasingly less relevant to their trait level. This scenario is fairly typical for many patients responding to patient reported outcome measures (PROMS). For IACAT 2017 ABSTRACTS BOOKLET 65 example, in the assessment of physical functioning, many item banks ceiling at about the 50\%ile. For most healthy patients, after a few items the only items remaining in the bank will represent decreasing ability (even though the patient has already indicated that they are at or above the mean for the population). Another example would be for a patient with no pain taking a Pain CAT. They will probably answer \“Never\” pain for every succeeding item out to the maximum test length. For this project we sought to reduce patient burden, while maintaining test accuracy, through the reduction of CAT length using novel stopping rules.

We studied CAT administration assessment histories for patients who were administered Patient Reported Outcomes Measurement Information System (PROMIS) CATs. In the PROMIS 1 Wave 2 Back Pain/Depression Study, CATs were administered to N=417 cases assessed across 11 PROMIS domains. Original CAT administration rules were: start with a pre-identified item of moderate difficulty; administer a minimum four items per case; stop when an estimated theta\’s SE declines to \< 0.3 OR a maximum 12 items are administered.

Original CAT. 12,622 CAT administrations were analyzed. CATs ranged in number of items administered from 4 to 12 items; 72.5\% were 4-item CATs. The second and third most frequently occurring CATs were 5-item (n=1102; 8.7\%) and 12-item CATs (n=964; 7.6\%). 64,062 items total were administered, averaging 5.1 items per CAT. Customized CAT. Three new CAT stopping rules were introduced, each with potential to increase item-presentation efficiency and maintain required score precision: Stop if a case responds to the first two items administered using an \“extreme\” response category (towards the ceiling or floor for the in item bank, or at ); administer a minimum two items per case; stop if the change in SE estimate (previous to current item administration) is positive but \< 0.01.

The three new stopping rules reduced the total number of items administered by 25,643 to 38,419 items (40.0\% reduction). After four items were administered, only n=1,824 CATs (14.5\%) were still in assessment mode (vs. n=3,477 (27.5\%) in the original CATs). On average, cases completed 3.0 items per CAT (vs. 5.1).

Each new rule addressed specific inefficiencies in the original CAT administration process: Cases not having or possessing a low/clinically unimportant level of the assessed domain; allow the SE \<0.3 stopping criterion to come into effect earlier in the CAT administration process; cases experiencing poor domain item bank measurement, (e.g., \“floor,\” \“ceiling\” cases).

}, keywords = {Administration Rules, Efficiency, FastCAT}, url = {https://drive.google.com/open?id=1oPJV-x0p9hRmgJ7t6k-MCC1nAoBSFM1w}, author = {Richard C. Gershon} } @article {5, title = {Efficiency and sensitivity of multidimensional computerized adaptive testing of pediatric physical functioning}, journal = {Disability \& Rehabilitation}, volume = {30}, number = {6}, year = {2008}, note = {Allen, Diane DNi, PengshengHaley, Stephen MK02 HD45354-01/HD/NICHD NIH HHS/United StatesNIDDR H133P0001/DD/NCBDD CDC HHS/United StatesResearch Support, N.I.H., ExtramuralEnglandDisability and rehabilitationDisabil Rehabil. 2008;30(6):479-84.}, pages = {479-84}, edition = {2008/02/26}, abstract = {PURPOSE: Computerized adaptive tests (CATs) have efficiency advantages over fixed-length tests of physical functioning but may lose sensitivity when administering extremely low numbers of items. Multidimensional CATs may efficiently improve sensitivity by capitalizing on correlations between functional domains. Using a series of empirical simulations, we assessed the efficiency and sensitivity of multidimensional CATs compared to a longer fixed-length test. METHOD: Parent responses to the Pediatric Evaluation of Disability Inventory before and after intervention for 239 children at a pediatric rehabilitation hospital provided the data for this retrospective study. Reliability, effect size, and standardized response mean were compared between full-length self-care and mobility subscales and simulated multidimensional CATs with stopping rules at 40, 30, 20, and 10 items. RESULTS: Reliability was lowest in the 10-item CAT condition for the self-care (r = 0.85) and mobility (r = 0.79) subscales; all other conditions had high reliabilities (r > 0.94). All multidimensional CAT conditions had equivalent levels of sensitivity compared to the full set condition for both domains. CONCLUSIONS: Multidimensional CATs efficiently retain the sensitivity of longer fixed-length measures even with 5 items per dimension (10-item CAT condition). Measuring physical functioning with multidimensional CATs could enhance sensitivity following intervention while minimizing response burden.}, keywords = {*Disability Evaluation, Child, Computers, Disabled Children/*classification/rehabilitation, Efficiency, Humans, Outcome Assessment (Health Care), Psychometrics, Reproducibility of Results, Retrospective Studies, Self Care, Sensitivity and Specificity}, isbn = {0963-8288 (Print)0963-8288 (Linking)}, author = {Allen, D. D. and Ni, P. and Haley, S. M.} } @article {174, title = {Measurement precision and efficiency of multidimensional computer adaptive testing of physical functioning using the pediatric evaluation of disability inventory}, journal = {Archives of Physical Medicine and Rehabilitation}, volume = {87}, number = {9}, year = {2006}, note = {Haley, Stephen MNi, PengshengLudlow, Larry HFragala-Pinkham, Maria AK02 hd45354-01/hd/nichdResearch Support, N.I.H., ExtramuralResearch Support, Non-U.S. Gov{\textquoteright}tUnited StatesArchives of physical medicine and rehabilitationArch Phys Med Rehabil. 2006 Sep;87(9):1223-9.}, month = {Sep}, pages = {1223-9}, edition = {2006/08/29}, abstract = {OBJECTIVE: To compare the measurement efficiency and precision of a multidimensional computer adaptive testing (M-CAT) application to a unidimensional CAT (U-CAT) comparison using item bank data from 2 of the functional skills scales of the Pediatric Evaluation of Disability Inventory (PEDI). DESIGN: Using existing PEDI mobility and self-care item banks, we compared the stability of item calibrations and model fit between unidimensional and multidimensional Rasch models and compared the efficiency and precision of the U-CAT- and M-CAT-simulated assessments to a random draw of items. SETTING: Pediatric rehabilitation hospital and clinics. PARTICIPANTS: Clinical and normative samples. INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: Not applicable. RESULTS: The M-CAT had greater levels of precision and efficiency than the separate mobility and self-care U-CAT versions when using a similar number of items for each PEDI subdomain. Equivalent estimation of mobility and self-care scores can be achieved with a 25\% to 40\% item reduction with the M-CAT compared with the U-CAT. CONCLUSIONS: M-CAT applications appear to have both precision and efficiency advantages compared with separate U-CAT assessments when content subdomains have a high correlation. Practitioners may also realize interpretive advantages of reporting test score information for each subdomain when separate clinical inferences are desired.}, keywords = {*Disability Evaluation, *Pediatrics, Adolescent, Child, Child, Preschool, Computers, Disabled Persons/*classification/rehabilitation, Efficiency, Humans, Infant, Outcome Assessment (Health Care), Psychometrics, Self Care}, isbn = {0003-9993 (Print)}, author = {Haley, S. M. and Ni, P. and Ludlow, L. H. and Fragala-Pinkham, M. A.} }