체육측정평가 : 우수 배드민턴 선수의 형태학적 비율

강상조(KangSang-Jo),박기현(ParkKi-Hyun),이미숙(LeeMi-Sook),최희남(ChoiHee-Nam) 한국체육학회 1998 한국체육학회지 Vol.37 No.2

The present study was attempted to describe the morphological parameters that characterize elite badminton players in order to examine the efficacy of using elite player`s morphological characteristics as a model for success. Data were collected on 26 female elite badminton players who selected from three different groups such as 4 olympic medalists, 8 national representative players, and 14 business team players. In order to evaluate the possibility of employing estimates of regional active tissue development (corrected diameters) as indicators of morphological prerequisites in badminton players, the skinfold-corrected diameters were measured in addition to the anthropological parameters such as height, weight, and % fat. These variables were analyzed by ANOVA and Phantom z-scores procedures. The olympic medalists had a slight tendency to be taller and lighter than the business team players. The thigh skinfold thickness and the estimated % of body fat of the olympic medalists had a significantly (p<.01) lower than those of the business team players. The % of fat mass relative to total height were significantly (p<.01) increased from the Olympic medalists (23.4%) to the business team players (25.4%). The Phantom z-scores for the relative fat mass were significantly (p<.01) increased with the athletic performance level. The results indicate that weight, thigh girth, skinfold thickless of thigh and calf, % of body fat and muscle mass are important characteristics prerequisites for success at elite levels of badminton players.

체육측정평가학 : 각종 탁구 경기대회의 비중도: 어떻게 결정할 것인가?

강상조(KangSang-Jo) 한국체육학회 2005 한국체육학회지 Vol.44 No.3

Almost all the sport federations have operated the weight system which gives different points according to the level of the games. Many attempts have been made to determine the appropriate standard for evaluating the quality level of various national and international sport games. The present study was an attempt to calibrate the weights associated with the different quality of sport games being held in Table-tennis event by using the Rasch rating scale model and to test the validity of the perceived weights of different games suggested by experts. Total 47 high school Table-tennis teams (male team = 22, female team = 25) participated in at least one game out of 7 major national games were selected for weighing the quality of the games. Tournament records were used as indicator of performance level for the selected teams in each game. The results showed that the Spearman's correlations between the game weights suggested by experts and derived by Rasch model were ranging from 0.354 to 0.825. The game weights suggested by experts were not consisted of those derived from Rasch analysis. The quality level of the games differed among the 7 different games. For example, ‘National games of men & women's table tennis championships’ had the highest quality measure (logit = -1.25) while ‘National men & women's Junior Championships of table tennis’ had the least quality measure (logit = 1.97) on the basis of 2004 tournament records of 7 major games for male. The results from a Rasch model analysis make it possible for a college board to identify the relative quality of games.

측정평가 : 번역된 운동심리 측정척도의 문항편파성 검증: 국가간 비교

강상조(KangSang-Jo) 한국체육학회 2003 한국체육학회지 Vol.42 No.2

When tests are belated into different language and culture, the measurement equivalence of the adapted instrument should be assessed, because the original and translated instruments may not be equivalent. The most of researdi in sport psychology has examined a white, college-and, and male individuals. As a result, conclusions from the research may not be applicable to all individuals of the other race, Many studies investigating the difference between ethnic groups have been done without examining differential item functioning (DIF) of the scale/lest used in the studies. DIF refers to unequal probabilities of a correct response to a test item by examinees from two groins, formed according to some demographic variable, such as gender or ethnicity, when controlled for their ability differences. The present study was to examine culture-related DIF for the translated instrument by applying administered to a total of 1,130 high school students from Seoul metropolitan area in Koea and Urbana-Champaign area in USA. It contained 19 items which have 6 response categories, raring from representing always to 1 representing never. Graded response model were used to examine whether a polytomous IRT model fit the data Unidimensionality assumption was examined by principal component factor analysis. Five DIF detection methods, such Likelihood ratio test, Lord’s chi-square, and Raju’s two area measures as IRT method and MH procedure as non-IRT method were used to test DIF of the BIAQ items between Cultural groups. Five methods identified 6 to 13 items with significant DIF, Raju's UA test detected the most items with significant DIF, followed by Rasch method Lord’s chi-square MH, and Raju's SA. Two items (I fest for a day or longer” and I get dressed up or made up) among the total items with significant DIF over culture were identified as having significant DIF across more than four methods. IRT and MH-based DIF analyses demonstrated to be an useful procedure in evaluating the measurement equivalence of instruments over the different cultures. Suggestions for future research are provided.

측정평가 : 체밀도 추정공식의 차이가 체지방추정 결과에 미치는 영향

강상조(KangSang-Jo),박재현(ParkJae-Hyeon),김혜진(KimHea-Jin) 한국체육학회 2003 한국체육학회지 Vol.42 No.2

The purpose of this study was to examine influences of estimated body density equations on the results of relative body fat. For this purpose, we used an computer analysis method called Monte Card computer simulation and 65 body density equations were collected. Among these equations, we used only 45 equations which are reported SEE and one of the data(subject C) which bad full profile which takes 29 region anthropometry measurement. For the raw data, Monte Carol computer simulation procedure attends to create 46,000 pseudo data by using 100 random variables which repeated 10 times from 45 equations. Random sampling is performed 10 times from these data which were categorized according to gender, age, race, professional-sports experience, and caliper. The result has it that the range of Relative body fat from different equations on a single person was about 16%-37%. In conclusion, the most influential variable was age, the second was gender, the third was ethnic and the next was caliper. But this study figure out that the experience of athlete has no influence on these equations. For the following studies using Montecarol computer simulation, we suggest the data for the analysis include extended peculation of data

채밀도 추정을 위한 생체 전기저항 분석법의 타당도일반화

강상조(KangSang-Jo),김기범(KimKee-Bum) 한국체육학회 2000 한국체육학회지 Vol.39 No.1

Validity generalization is a method of establishing the generalizability of the validity of a test by examining the results of all previous validity studies of the test of interest. In this paper, the validity generalization model was modified for application to predict validity studies of bio-electrical impedance. A literature search was conducted to identify studies in which bio-electrical impedance test scores were correlated with underwater weighting. The predict validity of bio-electrical impedance test was not shown to be generalizable across all situations. In this study, the percentage of variance accounted for by the artifacts was total 35.8%, men 37.8%, women 17.6%, boys 12.4%, girls 18.5%. Thus, the 75% decision rule was not met. However, a model is in place for future use. A new validity study should be carried out on boys, girls and old people particularly. This new study will represent new data and should be added to the prior information to form the posterior distribution. The posterior distribution can be used in a subsequent validity generalization study. And the reason why regarding low power by stating that studies underestimate the power of validity generalization, since this study only included quantifiable artifacts. Therefore, next study will include nonquantifiable artifacts, such as typographical and corxputational errors and criterion contamination. To date, use of validity generalization as a model to examine validity has not been widely applied for exercise science research. Validity generalization holds considerable premise as a tool for exercise scientists to use in integrating knowledge in the area of criterion-related validity.

특수체육 : 체력요인(體力要因)에 의한 정신지체아(精神遲滯兒) 집단(集團)간의 판별분석(判別分析)

한민규(HanMinKyu),강상조(KangSang-Jo) 한국체육학회 2001 한국체육학회지 Vol.40 No.1

The present study attempted to evaluate the differences of physical fitness among three group categorized mentally retarded children and to investigate the physical fitness structures of the three different groups. The subject were classified into three groups; educable (group I), trainable (group II), and down syndrome (group III). The subjects consisted of 596 mentally retarded children selected from special school. The 7 physical fitness tests were selected from motor-related fitness components. Multiple discriminant analysis was used to determine if the selected physical fitness tests were sensitive to differentiate among the three groups and successfully classify individual students in the appropriate performance levels. MANOVA showed that there was a significant difference (p $lt; .01) among the three different groups in the selected 7 physical fitness test mean vectors. Discriminant function analysis showed that two functions were extracted, but first function was significant. The first was a flexibility and cardio-endurance factor represented by sit & reach and 1000m run that separated the group I and the group II, III. Classification analysis was used to evaluate the accuracy of the discriminant analysis. The result showed that 48.09% of the three different mentally retarded children groups were correctly classified into their respective groups by the weighted discriminant scores. Is was concluded that the one discriminant function provided a valid model for evaluating differences among the three different mentally retarded groups.

체육측정평가 : 성인 남성의 건강체력수준의 평가: 판별분석의 적용

이미숙(LeeMi-Sook),강상조(KangSang-Jo),田中喜代次(KiyojiTanaka) 한국체육학회 1998 한국체육학회지 Vol.37 No.1

The purposes of this study were; 1) to discriminate the three different groups categorized by physical health status and physical activity level using discriminant function analysis, 2) to examine the degree of relationship between physical health status and physical fitness and the contribution of each element of physical fitness to the physical health status of middle-aged men. This study was achieved by classifying the original subject pool into three different experimental subgroups based on physical examination results and self-reported activity levels; a normal group (I), a regularly exercise group (II), and an obese group (III). The health status was assessed by physicians. Physical fitness tasks were selected from health-related physical fitness components; maximal oxygen uptake (VO2max; X1), grip strength (X2), body fat (X2), and trunk flexion from a standing position (X4). Discriminant function analysis was applied to determine if the selected physical fitness tasks were sensitive to differentiate among the three groups and successfully classify individual in the appropriate physical activity levels and physical health status. Correct discriminant probabilities of multiple discriminant functions to discriminate the I, II,and III group were 86.6%. These results suggested that there was a relatively strong relationship between physical health status and the level of physical fitness. The contribution of each element of physical fitness to physical health status was evaluated by correlation between the discriminant score representing physical health status, and performance on each element of the physical fitness test. It was concluded that discriminant function provided a valid model for evaluating differences among the three different physical health status and physical activity levels. Therefore, our data suggests that commonly used measures of health-related physical fitness can be utilized as a valid index of physical health status in exercise and health science.

다국면 Rasch 모형에 의한 운동수행 평가의 평정오차 분석

김미예(KimMi-Ye),강상조(KangSang-Jo) 한국체육학회 2003 한국체육학회지 Vol.42 No.6

The purpose of this study was to examine rating errors influencing on judgemental scoring, and to suggest the way of providing feedback for judges as a graphical format using Many-facet Rasch analysis. The data of lady’s figure skating events at the 2002 winter Olympics was used to compare the rankings when four methods of analysis are used. Methods are simple adding score, trimmed mean, median rank, and Many-facet Rasch analysis. Using Many-facet Rasch analysis, rating error-severity or leniency, halo effect, central tendency, restriction of range, and reliability or agreement were examined. Judge-by-athletes interaction was also calculated to extract feedback information. The results presented that ranking from the each methods were vary, and rating error can be detected in many facet Rasch analysis using the indexes such as mean square error, separation, and reliability, and also using the chi-square test. The information obtained from many-facet Rasch analyses could be used effectively to provide feedback for judges in training and certification program.

측정평가 : 단일변인 중복면적: 개념과 추정방법

박재현(ParkJae-Hyeon),강상조(KangSang-Jo) 한국체육학회 2003 한국체육학회지 Vol.42 No.4

The purpose of this study was to introduce on univariate Overlap Zone (OZ) and applicate on MS-Excel programing for OZ. OZ is mean that two distributions are the degree to overlapping. Two distributions can differ from each other if the mean of one is dislocated relative to the other, if the variability of one is markedly different from the other. It would like to be able to quantify in OZ for a selection pressure, there are some difficult on OZ calculation because it is need to calculate by iteration over 10 or more for the intersection of two curves. The easyOZ by Excel is program for compute OZ based on the MS-Excel by Microsoft. This OZ and Easy OZ would allow us to understand more completely the relative importance of different anthropometric variables, make informed judgments about talent identification criteria, and compare selection pressures in population subgroups such as males vs females, or national-level player vs international-level player.