Marginal maximum likelihood estimation of item parameters: Application of an EM algorithm

Bock, R. Darrell; Aitkin, Murray

doi:10.1007/bf02293801

Cited by 1,949 publications

(1,684 citation statements)

References 1 publication

Supporting

Mentioning

1,574

Contrasting

Unclassified

Order By: Relevance

“…This IRT model, generated using the BILOG-MG statistical program (Scientific Software International, 2003), yields marginal maximum likelihood estimates (Bock and Aitkin, 1981;Harwell et al 1988) of two parameters: the b or threshold parameter and the a or discrimination parameter. The a parameter measures the ability of a criterion to discriminate people who are higher on the continuum and those who are lower on the continuum.…”

Section: Resultsmentioning

confidence: 99%

The role of alcohol consumption in future classifications of alcohol use disorders

Saha

Stinson

Grant

2007

Drug and Alcohol Dependence

198

201

View full text Add to dashboard Cite

Background-Item response theory (IRT) was used to determine whether DSM-IV alcohol abuse and dependence and consumption criteria were arrayed along a continuum of severity.Methods-Data came from a large, nationally representative sample of the U.S. adult population.Results-DSM-IV alcohol abuse and dependence criteria formed a continuum of alcohol use disorder severity along with the drinking 5+/4+ at least once a week in the past year criterion. Criteria were invariant across sex, race-ethnicity, and age subgroups. Conclusion-The drinking 5+/4+ high-risk drinking pattern was identified as a suitable criterion for future classifications of DSM-IV alcohol use disorder. Some dependence criteria were among the least severe criteria, and some abuse criteria were among the most severe, findings that question the validity of DSM-IV abuse and dependence categories as distinct entities and that do not support the assumption of abuse as prodromal to dependence. Physical dependence and addiction were identified as defining elements of the continuum. Further research examining their dimensional properties and relationships to high-risk drinking patterns appears warranted. An approach highlighting a more important role of consumption in future classifications of alcohol use disorder defined broadly to encompass all alcohol-related harm, including addiction and physical dependence is discussed.

show abstract

Section: Resultsmentioning

confidence: 99%

The role of alcohol consumption in future classifications of alcohol use disorders

Saha

Stinson

Grant

2007

Drug and Alcohol Dependence

198

201

View full text Add to dashboard Cite

show abstract

“…Marginal maximum likelihood estimation with Bock and Aitkin's [48] formulation of the EM (Expectation-Maximization) algorithm [49] is implemented in ConQuest to estimate all the parameters in the MRCMLM (n; l and R) simultaneously, so that measurement errors in h are directly taken into account. Based on the assumption of conditional independence among items and persons, the probability of a response vector x conditioned on the random quantities h is ðA:4Þ…”

Section: Appendix Parameter Estimation Procedures For the Mrcmlmmentioning

confidence: 99%

Validating, Improving Reliability, and Estimating Correlation of the Four Subscales in the WHOQOL-BREF using Multidimensional Rasch Analysis

et al. 2006

View full text Add to dashboard Cite

Objective: This study examined the construct validity, and improved the test reliability and the estimation accuracy for the correlation between domains of the WHOQOL-BREF using multidimensional Rasch analysis. Method: A total of 13,083 adults were administered the 28-item WHOQOL-BREF Taiwan version, which consists of 4 subscales (domains). The multidimensional form of the partial credit model was used to examine the fit of the 4 subscales. For comparison, each subscale individually was also fitted to the unidimensional partial credit model. Standard item fit statistics and analysis of differential item functioning (DIF) were used to check model-data fit. Results: After excluding 2 overall items and deleting 7 DIF items, the remaining items of each subscale in the WHOQOL-BREF constituted a single construct. The test reliabilities and correlations between domains obtained from the multidimensional approach, (0.82-0.86) and (0.79-0.89), respectively, were much higher than those obtained from the unidimensional approach, (0.67-0.75) and (0.53-0.65), respectively. Conclusion: The 19-item WHOQOL-BREF measures more succinct latent traits than the original design. The multidimensional approach yields not only more accurate estimates for the correlation between domains but also substantially higher reliabilities, than the standard unidimensional approach.

show abstract

“…an equicorrelation structure. Specifically, it is not an explicit function of the covariance matrix of the random coefficients, although its form is sometimes derived from considering a particular US model and integrating over the random coefficients to obtain the marginal distribution (see for example Bock and Aitkin (1981)). …”

Section: Two-level Binary Response Modelmentioning

confidence: 99%

Improved Approximations for Multilevel Models with Binary Responses

Goldstein¹,

Rasbash²

1996

Journal of the Royal Statistical Society. Series A (Statistics

410

270

View full text Add to dashboard Cite

Marginal maximum likelihood estimation of item parameters: Application of an EM algorithm

Abstract: estimation of item parameters, EM algorithm, item analysis, latent trait, dichotomous factor analysis, Law School Aptitude Test (LSAT),

Cited by 1,949 publications

References 1 publication

The role of alcohol consumption in future classifications of alcohol use disorders

The role of alcohol consumption in future classifications of alcohol use disorders

Validating, Improving Reliability, and Estimating Correlation of the Four Subscales in the WHOQOL-BREF using Multidimensional Rasch Analysis

Improved Approximations for Multilevel Models with Binary Responses

Contact Info

Product

Resources

About