Item Response Theory for Scores on Tests Including Polytomous Items with Ordered Responses

Thissen, David; Pommerich, Mary; Billeaud, Kathleen; Williams, Valerie

doi:10.1177/014662169501900105

Cited by 169 publications

(165 citation statements)

References 9 publications

Supporting

Mentioning

163

Contrasting

Order By: Relevance

“…We used the item parameter estimates derived from the fixed-parameter calibration to construct a cross-walk table by applying expected a posteriori (EAP) summed scoring. 28,29 Cross-walk tables map simple raw summed scores from each legacy instrument to T-score values on the PROMIS Global Health metric.…”

Section: Analysesmentioning

confidence: 99%

See 1 more Smart Citation

Linking Physical and Mental Health Summary Scores from the Veterans RAND 12-Item Health Survey (VR-12) to the PROMIS® Global Health Scale

et al. 2015

View full text Add to dashboard Cite

BACKGROUND: Global health measures represent an attractive option for researchers and clinicians seeking a brief snapshot of a patient's overall perspective on his or her health. Because scores on different global health measures are not comparable, comparative effectiveness research (CER) is challenging. OBJECTIVE: To establish a common reporting metric so that the physical and mental health scores on the Veterans RAND 12-Item Health Survey (VR-12 © ) can be converted into scores on the corresponding Patient Reported Outcomes Measurement Information System (PROMIS ® ) Global Health scores. DESIGN: Following a single-sample linking design, participants from an Internet panel completed items from the PROMIS Global Health and VR-12 Health Survey. A common metric was created using analyses based on item response theory (IRT), producing score cross-walk tables for the mental and physical health components of each measure. The linking relationships were evaluated by calculating the standard deviation of differences between the observed and linked PROMIS scores and estimating confidence intervals by sample size. PARTICIPANTS: Participants (N=2025) were 49 % male and 73 % white; mean age was 46 years. MAIN MEASURES: Mental and physical health subscales of the PROMIS Global Health and the VR-12. The mean VR-12 physical component and mental component scores were 45.2 and 46.6, respectively; the mean PROMIS physical and mental health scores were 48.3 and 48.5, respectively. KEY RESULTS: We found evidence that the combined set of VR-12 and PROMIS items were relatively unidimensional and that we could proceed with linking. Linking worked better between the physical health than mental health scores using VR-12 item responses (vs. linking based on algorithmic scores). For each of the cross-walks, users can minimize the impact of linking error with modest increases in sample sizes.CONCLUSIONS: VR-12 scores can be expressed on the PROMIS Global Health metric to facilitate the evaluation of treatment, including CER. Extending these results to other common measures of global health is encouraged.

show abstract

Section: Analysesmentioning

confidence: 99%

“…28,29 Tables 3, 4, 5 and 6 show four cross-walk tables that associate VR-12 component scores with PROMIS Global Health scores. For both mental and physical health, we provide a summed VR-12 and an algorithmic VR-12 cross-walk.…”

Section: Score Cross-walk Tablesmentioning

confidence: 99%

Linking Physical and Mental Health Summary Scores from the Veterans RAND 12-Item Health Survey (VR-12) to the PROMIS® Global Health Scale

et al. 2015

View full text Add to dashboard Cite

show abstract

“…Thê θ estimates are obtained using the observed item responses, the item trace-line functions given their respective item parameters (ψ j ), and (potentially) prior distributional information about θ. Multiple methods exist for obtainingθ values, such as weighted and unweighted maximum-likelihood estimation (WLE and ML, respectively;Bock and Aitkin 1981;Warm 1989), Bayesian methods such as the expectation or maximum of the posterior distribution (EAP and MAP, respectively;Bock and Aitkin 1981), and several others which have seen less use in applied settings (e.g., EAP for sum scores; Thissen, Pommerich, Billeaud, and Williams 1995). ML estimation of θ for a given response pattern requires optimizing the likelihood function…”

Section: Predicting Latent Trait Scoresmentioning

confidence: 99%

Generating Adaptive and Non-Adaptive Test Interfaces for Multidimensional Item Response Theory Applications

Chalmers¹

2016

J. Stat. Soft.

View full text Add to dashboard Cite

Computerized adaptive testing (CAT) is a powerful technique to help improve measurement precision and reduce the total number of items required in educational, psychological, and medical tests. In CATs, tailored test forms are progressively constructed by capitalizing on information available from responses to previous items. CAT applications primarily have relied on unidimensional item response theory (IRT) to help select which items should be administered during the session. However, multidimensional CATs may be constructed to improve measurement precision and further reduce the number of items required to measure multiple traits simultaneously.A small selection of CAT simulation packages exist for the R environment; namely, catR (Magis and Raîche 2012), catIrt (Nydick 2014), and MAT (Choi and King 2014). However, the ability to generate graphical user interfaces for administering CATs in realtime has not been implemented in R to date, support for multidimensional CATs have been limited to the multidimensional three-parameter logistic model, and CAT designs were required to contain IRT models from the same modeling family. This article describes a new R package for implementing unidimensional and multidimensional CATs using a wide variety of IRT models, which can be unique for each respective test item, and demonstrates how graphical user interfaces and Monte Carlo simulation designs can be constructed with the mirtCAT package.

show abstract

“…Let us begin with the second part. Luckily, algorithms for converting between summed scores and EAP have already been developed [21,22], although one hundred-percent accuracy is not guaranteed. Most software packages that support IRT produce such conversion tables for each IRT-based model developed.…”

Section: From Summed Score To Expected a Posteriori (Eap) Scoresmentioning

confidence: 99%

“…Luckily, this does not apply to IRT. IRT is sample independent: What IRT calculates first is each item's properties; once the parameters are calculated in one group, the same parameters can be applied to other groups [21,27,28]. Thus, people who have the same level of safety attitudes present the same EAP scores, regardless of the group in they are categorized.…”

Section: Strengths Of Item Response Theory and Modeling Combinationmentioning

confidence: 99%

The Pure and the Overarching: An Application of Bifactor Model to Safety Attitudes Questionnaire

Jeong¹

2016

BBIJ

View full text Add to dashboard Cite

Although this article is long, we believe that it is worth reading. It answers the most fundamental question that patient safety professionals have asked: How do we measure the overall safety attitudes score of each person? Broadly speaking, we showed how to structuralize various safety-related traits in one's mind, using the Safety Attitudes Questionnaire-Korean version as an example. We applied the bifactor model, which explicitly contains a general safety attitudes dimension that governs all SAQ-K items as well as six original SAQ-K domains. The major finding regarding the model structure is that several items might not fall under a specific SAQ domain, although they are still largely governed by safety attitudes in general; yet the stress recognition domain might not be part of the general construct-namely, safety attitudes. However, the more important information that we intended to share was that the bifactor model can effectively take control of the seemingly inadequate items or domains and calculate domain-specific scores and general safety attitudes score by providing different weights for each item. Thus, we can obtain much more purified domains scores from the data, compared to the traditional mean score approach. In addition, the item response theory-based approach used in this article gives more solid theoretical strength in handling the ordinal data and also offers the possibility of adding or dropping items or even a domain, while still allowing a longitudinal analysiscomparing scores from different versions of safety attitudes questionnaire. In addition to these theoretical strengths, the bifactor model provides exceptional computational efficiency compared to the other models we have tried, thereby allowing us to unlock an extremely large and complex dataset that could not be analyzed thus far. We hope the approach introduced in this article can help all the patient safety professionals achieve more precise and valid information from their already collected safety attitudes data as well as data collected in the future, ultimately saving more lives. DisclaimerThe target audience of this article is those who actually use safety culture assessment tools in the healthcare setting, and we assumed they possess basic-to mediumlevel knowledge in statistics and psychometrics. Thus, we intentionally avoided showing complex formulae; instead, we tried to simplify, and sometimes oversimplify, the concepts introduced in this article to facilitate understanding. Terminology was also carefully chosen for those who did not major in statistics. In addition, we did not follow a typical research article style. We intentionally allocated many paragraphs usually reserved for the discussion section to the results section in order to facilitate understanding. Anyone curious about full mathematical descriptions are asked to please contact the authors. IntroductionA dozen of our previous articles on the Safety Attitudes Questionnaire-Korean version (SAQ-K) introduced several new approaches to analyze collected d...

show abstract

Item Response Theory for Scores on Tests Including Polytomous Items with Ordered Responses

Cited by 169 publications

References 9 publications

Linking Physical and Mental Health Summary Scores from the Veterans RAND 12-Item Health Survey (VR-12) to the PROMIS® Global Health Scale

Linking Physical and Mental Health Summary Scores from the Veterans RAND 12-Item Health Survey (VR-12) to the PROMIS® Global Health Scale

Generating Adaptive and Non-Adaptive Test Interfaces for Multidimensional Item Response Theory Applications

The Pure and the Overarching: An Application of Bifactor Model to Safety Attitudes Questionnaire

Contact Info

Product

Resources

About