Rasch analysis of the help seeking scale

Yi ZENG, Magdalena Mo Ching MOK

Research output: Contribution to conferencePapers

Abstract

Empirical Research Background: The Help Seeking questionnaire is developed by Magdalena Mo Ching Mok, The research is supported by a UGC CERG Grant. It is part of a larger questionnaire on self-directed learning of secondary students. In this survey, the researchers hope to find the secondary students’ attitudes towards seeking help. Empirical Research Aims: We hope to do analysis of the questionnaire quality using Rasch Model. The main objective is to spell out the quality of this test through the Rasch analysis of reliability, fit statistics, measure estimates of person and items in this research. Empirical Research Sample: In this survey, there are random samples of 65 secondary students. There are 35 boys and 30 girls. The age range is between 13 and 16. Empirical Research Method: In this survey, the 65 secondary students are asked to answer 10 items, each item has four categories to choose: strongly disagree, disagree, agree, and strongly agree. We label these categories as: strongly disagree (1), disagree (2), agree (3), and strongly agree (4). The students are required to ion that best describes their situation. There is no right or wrong answers. We use Bond&Fox Steps 2.0 software to do Rasch analysis. Empirical Research RASCH: From the analysis of Item-Person Map, you can understand the estimate of the respondents‘ attitude and the difficulty of items in logits and their relationship; Summary Statistics give us the value of fit, difficulty/ability estimation and error, reliability indices; in which, the mean attitudes of the students (58.93) and the mean difficulty of the items (50). The spread in, or variation of fit scores for persons (infit t SD=0.84 and outfit t SD=0.84), for items (infit t SD=0.2 and outfit t SD=0.25). And also there show the person reliability 0.68 (separation=1.46) and the item reliability 0.43 (separation=0.87). Item-Category Structure shows the quality of the category structure; DIF Measure indicates the difference of the estimates between boys and girls; Empirical Item-Category Measures tell the agree items located in order from the easiest to the most difficult. From all the analysis, we can understand how the estimates of parameters meet the expectation of Rasch Model and how to improve the test quality. Empirical Research Results: In this analysis, we can conclude that there need more multilevel hierarchical items to support the precise and reliable estimates of the students’ attitudes and also the items estimate. This research has an acceptable reliability for person, and the precision of the error estimates are also in an acceptable range. The person fit value is in the right track. The items estimate is not reliable and not fit. The category structure meet the Rasch Model’s expectation, that is to say, the numbers of the responds are optimal. From the DIF analysis, there is no indicator of difference among genders. In empirical Item-Category Measures, there shows disagree and agree answers are more likely to be responded than strongly disagree and strongly agree, this may indicate that there are not enough students with strongly positive or negative attitude, or maybe there need more items that could broaden the scope of attitude. Empirical Research Conclusions: From the quality analysis of seeking help questionnaire, we can understand how the estimates of parameters meet the expectation of Rasch Model and how to improve the test quality. Here we can get lots of information about problem item or person performance through statistics analysis and do more work to analyze and understand the situation, then seek solutions to keep better fit to Rasch Model. This would be a continuous and repetitive work. From the quality analysis of seeking help questionnaire, we can understand how the estimates of parameters meet the expectation of Rasch Model and how to improve the test quality. Here we can get lots of information about problem item or person performance through statistics analysis and do more work to analyze and understand the situation, then seek solutions to keep better fit to Rasch Model. This would be a continuous and repetitive work. From the quality analysis of seeking help questionnaire, we can understand how the estimates of parameters meet the expectation of Rasch Model and how to improve the test quality. Here we can get lots of information about problem item or person performance through statistics analysis and do more work to analyze and understand the situation, then seek solutions to keep better fit to Rasch Model. This would be a continuous and repetitive work.
Original languageEnglish
Publication statusPublished - 2009

Citation

Zeng, Y., & Mok, M. M. C. (2009, July). Rasch analysis of the help seeking scale. Paper presented at the Pacific Rim Objective Measurement Symposium 2009 (PROMS 2009) Hong Kong, The Hong Kong Institute of Education, China.

Fingerprint Dive into the research topics of 'Rasch analysis of the help seeking scale'. Together they form a unique fingerprint.