Over the past few decades, error feedback has received substantial attention in language pedagogy. Despite this, few studies have been carried out on the effect and benefits of various modalities of error feedback on the oral proficiency of the intermediate EFL learners. To this end, an experimental design was adopted and implemented with 80 female high school students who were divided into four groups, namely, the intra-error feedback group (Intra-EF group), the inter-error feedback group (Inter-EF group), the teacher error feedback group (TEF group), and the control group. Treatment sessions for the experimental groups included three sessions for completing the speaking tasks and three sessions for feedback conferences. The participants in all groups completed the same speaking activities under the same process of discussion, but received different modalities of error feedback in feedback sessions. Results of T-test, one-way ANOVA and Scheffe Post Hoc test revealed that different stages of peer error feedback (i.e., intra-and inter-error feedback), as well as teacher error feedback influenced the learners' speaking ability. Additionally, peer error feedback and teacher error feedback exerted different effects on their speaking ability.
The findings of previous research into the compatibility of stakeholders' perceptions with statistical estimations of item difficulty are not seemingly consistent. Furthermore, most research shows that teachers' estimation of item difficulty is not reliable since they tend to overestimate the difficulty of easy items and underestimate the difficulty of difficult items. Therefore, the present study aims to analyze a high stakes test in terms of heuristic (test takers' standpoint) and statistical difficulty (CTT and IRT) and investigate the extent to which the findings from the two perspectives converge. Results indicate that, 1) the whole test along with its sub-tests is difficult which might lead to test invalidity; 2) the respondents' ratings of the total test in terms of difficulty level are almost convergent with the difficulty values indicated by IRT and CTT, except for the two subtests where students underestimated the difficulty values, and 3) CTT difficulty estimates are convergent with IRT difficulty estimates. Therefore, it can be concluded that students' perceptions of item difficulty might be a better estimate of test difficulty and a combination of test takers' perceptions and statistical difficulty might provide a better picture of item difficulty in assessment contexts.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.