2. 您的位置
3. 首页
4. 会议论文
5. Testlet effect on IRT-based DIF methods:Bayesian testlet response theory approach

Testlet effect on IRT-based DIF methods:Bayesian testlet response theory approach

来源 :北京师范大学 | 被引量 : 0次 | 上传用户：resway

【摘 要】

：

The purpose of this study is to examine the testlet effect on IRT-based DIF methods,since these traditional methods do not take Local Item Dependence(LID) caused by testlets into account.

【作 者】

：

Burhanettin (O)zdemir 

【机 构】

：

Hacettepe University,Turkey

【出 处】

：

北京师范大学

【发表日期】

：

2015年5期

下载到本地 , 更方便阅读

下载此文 赞助VIP

声明 : 本文档内容版权归属内容提供方 , 如果您对本文有版权争议 , 可与客服联系进行内容授权或下架

论文部分内容阅读

The purpose of this study is to examine the testlet effect on IRT-based DIF methods,since these traditional methods do not take Local Item Dependence(LID) caused by testlets into account.

其他文献

Flawed intuitions about power

Because of relatively small effect and sample sizes,many psychological studies are underpowered.Even if researchers understand the importance of well-powered research designs,their intuitions about po

会议

A Monte Carlo evaluation of parametric and permutation-based methods for two-group multivariate mean

A Monte Carlo study was carried out to examine the performance of two-group multivariate means comparison under the conditions of homogeneous and heterogeneous variance-covariance matrices across two

会议

A probability-based effect size(AG)robust to multivariate non-normality and heterogeneity of covaria

This study develops an effect size(ES;AG ) that is robust to violations of data assumptions – multivariate normality and homogeneity of covariance matrices – in Multivariate Analysis Of VAriance(MANOV

会议

Using LASSO penalization for explanatory IRT:An application on instructional covariates for mathemat

LASSO penalization can be used for models with very high numbers of covariates.The regression parameters are penalized and more of them become zero as the penalty increases,thereby resulting in covari

会议

An alternative to post-hoc model modification in confirmatory factor analysis:The Bayesian Lasso

As a commonly used tool for operationalizing measurement models,Confirmatory Factor Analysis(CFA) requires strong assumptions that may lead to poor fit of the model to real data.The post-hoc modificat

会议

Detection of treatment-subgroup interactions in clustered datasets

Identification of subgroups of patients for which treatment A is more effective than treatment B,and vice versa,is of key importance to the development of personalized psychological medicine.

会议

Differential item functioning in Rasch models with recent methods:An application to educational data

Many procedures have been developed to investigate DIF in questionnaire response data,among others the wellknown Likelihood-Ratio test and the Mantel-Haenszel statistic(see Magis et al.(2010) for a co

会议

Anchor item selection methods in DIF analysis

Differential Item Functioning(DIF) occurs when individuals from different populations show different probabilities of correctly responding to an item with the same latent trait.If DIF items exist in t

会议

Methods to enhance small-sample DIF estimation:Implications for minimum sample size requirements and

The Mantel-Haenszel(MH) procedure is a widely used nonparametric statistical method to measure the amount of DIF present in an item.For operational programs at ETS,at least 200 members are required fo

会议

A procedure to select more invariant items for the anchor in tests of differential item functioning

Differential Item Functioning(DIF) occurs when items behave differently for examinees at the same ability level that are in different groups.Choosing a proper group of invariant items as anchors is an

会议