• Physical Therapy Korea
• /
• v.21 no.4
• /
• pp.56-66
• /
• 2014

For maintaining adequate psychometric properties when reducing the number of items from an instrument, item level psychometrics is crucial. Strategies such as low item correlation or factor loadings, using classical test theory, have traditionally been advocated. The purpose of this study is to describe the development of a new short form assessing the impact of low back pain on physical activity. Rasch measurement model has been applied to the International Classification of Functioning, Disability and Health Activity Measure (ICF-AM). One hundred and one individuals with low back pain aged 19-89 years (mean age: $48.1{\pm}17.3$) who live in the community were participated in the study. Twenty-seven items of lifting/carrying construct of the ICF-AM were analyzed. Ten items were selected from the construct to create a short form. Item elimination criteria include: 1) high or low mean square (out of the range: .6-1.4 for the fit statistics), 2) similar item calibrations to adjacent items, 3) person separation value, and item-person map for potential gap in person ability continuum. All 10 items of the short form fit to the Rasch model except one item (i.e., carrying toddler on back). Despite its high infit and outfit statistics (1.90/2.17), the item had to be reinstated due to potential gaps at the upper extreme of person ability level. The short form had a slightly better spread of person ability continuum compared to the entire set of item. The created short form separated individuals with low back pain into nearly 4 groups, while the entire set of items separated the individuals into 6 groups. The findings prompted multidimensional models for better explanation of the lifting/carrying domain. The item level psychometrics based on the Rasch model can be useful in developing short forms with rationally retained items.

• Journal of Environmental Science International
• /
• v.11 no.9
• /
• pp.829-842
• /
• 2002

The purpose of this study was to design a reliable and valid Likert-type scale to measure nuclear energy attitudes of various groups. 8 tests were applied to the construction of a 26-item long form, a 12-item middle form, and a 6-item short form of the scale. Three scales were field tested on 760 respondents representing a diverse range of subjects. The 26-item scale appeared to be reliable, content valid and construct valid. Scale reliability and homogeneity were evidenced by 0.95 coefficient alpha and positive interitem correlations ranging from 0.18 to 0.74. The results of factor analysis and known-groups comparison revealed that this scale had high validity. Reliability of the middle form and short form were evidenced by coefficient alpha of 0.92 and 0.90 respectively. The long and middle forms were strongly correlated, r-value of 0.98 and the long and short forms were also highly correlated, r-value of 0.93. Consequently, long form, middle form and short form scales were evidenced very reliable and valid in measuring nuclear energy attitudes of various groups in general. They can be used for a variety of purpose of measuring and testing nuclear energy attitude.

• Physical Therapy Korea
• /
• v.22 no.3
• /
• pp.61-70
• /
• 2015

To develop an effective and efficient measurement system for tracking changes of functional status across two measures, it is essential to integrate information and communicate scores across two measures. The lack of communication between two measures leads to score incompatibility. A potential solution would be the development of a crosswalk table between those measures. Prior to creating a crosswalk table, selecting common items between two measures is critical. By using the Oswestry low back pain disability questionnaire (Oswestry) and a short form measuring disability resulting from low back pain, item level statistics as well as differential item functioning (DIF) using the Rasch measurement were investigated. Eighty-two participants with known group validity were recruited. Based on the application of the Rasch measurement model, item difficulties across the two measures were logically and hierarchically ordered. Ceiling effects for both measures were detected, which were not be able to be effectively measured with the two measures. The DIF analysis across the two measures confirmed that five paired items were found to have DIF and five common items were selected for common items. Although five paired items function differently across the Oswestry and the short form, all items of both measures were well targeted study participants. The common items selected by the Rasch measurement model may be effective when creating a crosswalk table between the Oswestry and the short form.

• Journal of Environmental Science International
• /
• v.12 no.7
• /
• pp.763-774
• /
• 2003

This study was conducted to design long and short forms of reliable and valid Likert-type scales to measure environmental attitudes of Korean junior high school students. 8 tests were applied to the construction of a 32-item long form and 8-item short form of the scale. Two scales were field tested on 469 respondents consisting of male and female students in urban and rural areas and students participating in environmental programs. The 32-item scale was revealed to be highly reliable, content and construct valid. Reliability and homogeneity of this scale were evidenced by 0.92 coefficient alpha and positive interitem correlation values ranging from 0.10 to 0.52. Factor analysis and known-group comparison showed the sufficient validity of long form scale. Reliability and validity of short form scale were also evidenced by good results of 8 test analyses. The long and short forms were strongly correlated, ${\gamma}$-value of 0.90. Consequently, both long form and short form scales were evidenced very reliable and valid in measuring environmental attitudes of junior high school students in Korea. They can be useful for a variety of survey conditions and constraints and testing environmental attitude.

• Korean Journal of Child Studies
• /
• v.26 no.6
• /
• pp.73-88
• /
• 2005

This study aimed to develop a reliable and valid developmental test for toddlers and preschoolers. A trial form of the test with 230 items in 5 domains(cognitive, physical, language, social, and self-help) was constructed and administered to 200 children from 2 to 5-year-olds. Item analysis indicated that most items selected in cognitive, physical, and indicated domain were acceptable in terms of item response rate and item discrimination. However, some items in social and self-help domain did not show clear age trends in item response, but mostly were adequate in item discrimination. Based on item analysis, 112 items were selected for the final form : 24 items each for the cognitive, physical, and language domain; 20 items each for the social and self-help domain. Reliability indicated by internal consistency coefficients(KR-20) were high in all domains. Validity of the test was supported by inter-correlations among 5 domains(.61 to .93), and significant age differences in domain scores.

• English Language & Literature Teaching
• /
• v.18 no.1
• /
• pp.171-195
• /
• 2012

While noticing that there is insufficient interest in the assessment of EFL learners' vocabulary levels or sizes, the researchers developed two tests identical in form (Forms A and B) to assess the lexical knowledge of Korean university learners at the $1^{st}{\sim}10^{th}$ 1,000 word bands by adapting a pre-established vocabulary levels test (VLT). Of equal concern was to investigate if the VLT was equally a valid and reliable instrument to be used on measuring the lexical knowledge of EFL learners. The participants were 804 university freshmen enrolled in a General Education English Course from four different colleges. The learners were asked to respond to either Form A or B. While scores generally fell towards the lower frequency bands, multiple regression found the Korean College Scholastic Ability Test (CSAT) to be a significant variable for predicting the learners' vocabulary sizes. From a methodological perspective, however, noticeable differences between Forms A and B could be found with item response theory analysis. The findings of the study provide suggestions on how future VLT for testing EFL learners may have to be redesigned.

• Journal of the Korean Society of Clothing and Textiles
• /
• v.18 no.1
• /
• pp.130-139
• /
• 1994

This study was performed to provide the dress form for high school girls' clothing construction which can reflect the characteristics of their body. Two hundred and forty seven high school girls aged from 15 In 17 were directly measured. The results obtained from this study were as follows ; 1. High school girls' somatotype was similar to that of young adult women. They had the features of well developed upper trunk and well developed breadth, but less developed depth. 2. The correlation between length item and the other items was small enough to interpret that each item was independent. There was a Positive correlation between depth item and girth item. And, Observed correlation between weight and Rohrer's index and breadth, depth and girth item was much larger than expected, which showed that weight and Rohrer's index influenced on fatness factor. 3. Through factor analysis, six factors were extracted from anthropometric measurements. The first factor was representing upper body volumn which expressed the characteristics of somatotype better than any other factors, the second factor was breadth, the third factor was size, the fourth factor was length, the fifth was back shape and the sixth factor was shoulder shape. 4. For development of dress form for high school girls, directly anthoropometric measurements were used. The dress form constructed through this study represented the characteristics of high school girls, therefore this would be very useful to manufacture clothing for them in the future.

• Proceedings of the IEEK Conference
• /
• 2003.07d
• /
• pp.1371-1374
• /
• 2003

This paper designs and implements of Item pool-Based Program using Visual Basic, a sort of self- study system. This paper is implemented the system which makes and remakes the test items using item-pool data with different method from an ordinary fixed form of evaluation. This system helps teachers to make a various test type of items by searching data through the item-pool and accounting it to the test items. This system will also improve the learner's understanding of the problems and desire for self-solving a problem.

• The Journal of the Korea Contents Association
• /
• v.13 no.12
• /
• pp.321-327
• /
• 2013

Impaired naming difficulty is common in normal elderly as well as in patients with neurological impairment. The 60-item Korean version-Boston Naming Test(K-BNT) is one of the most commonly used test for measuring confrontational naming ability. However, age-related cognitive decline may make the elderly difficult concentrating during the 60-item test, therefore, item reduction of the K-BNT would improve test validity and reliability. Thus, the purpose of this study was to develop a short form of the K-BNT based on Item Response Theory(IRT). Considering item-fit index, sex factor, and item difficulty through Rasch analysis, the 15-item K-BNT(i.e., K-BNT-15) was developed. Via administration of the K-BNT-15, we observed age-related decline in naming ability and significantly different performance between the normal elderly and patients with mild cognitive impairment. This study demonstrates the utility of IRT for developing a short-form language evaluation tool. The K-BNT-15 can be effective as a language screening tool to differentiate between normal aging and pathological diseases.

• Journal of Korean Academy of Fundamentals of Nursing
• /
• v.16 no.4
• /
• pp.490-496
• /
• 2009

Purpose: The purpose of this study was to examine the validity of a single item scale by comparing a single item scale of health related quality of life to subscopes of the multi-item scales in the Medical Outcome Study Short Form-36 (SF-36), and Minnesota Living with Heart Failure Questionnaire (MLHFQ). Method: The data from 103 patients with heart failure were analyzed. The statistics program SPSS 12.0 used for descriptive statistics. Results: Significant correlations were found between the single item scale and subscopes of SF-36, for physical functioning (r= .42), role physical (r= .22), general health perception (r= .46), vitality (r= .40), social functioning (r= .20) and mental health (r= .51), but not for body pain (r= .12) and role emotional (r= .06). Physical (r= -.41), emotional (r= -.49) and total scores (r= -.49) of MLHFQ were significantly related to the single item scale. Conclusion: The single item scale of health related quality of life showed low to moderate correlation with multi-item scales. Even though the single item scale was correlated with several subscopes of multi-item scales, the correlation was not high, so we have to use caution when using the single item scale instead of multi-item scales.