Statistical Approach to Test Construct Validity and Obtain Weights for the Children's Dietary Life Recognition and Practice Index

Constructs with seven latent evaluation indicators and 18 observable survey questions were developed by food and nutrition experts to calculate a food safety recognition and practice index for children. The purpose of this study was to suggest statistical approaches to test construction validity on the constructs, obtain weights of the evaluation indicators, and develop questionnaires to calculate a children's food recognition and practice index. Survey data of 2,400 elementary fifth grade students were used as empirical results. Test validity was evaluated by exploratory factor analysis and confirmed to be highly significant by confirmatory factor analysis [i.e., linear structural relations (LISREL) analysis]. Standardized path coefficients of the LISREL analysis were suggested based on weights, and the weights were compared using the AHP and Delphi methods.

The Trend of Korean Nursing Research with the LISREL / 간호학회지

This study was conducted by analyzing all 33 articles based on the LISREL, published from January 1991 to March 1999 in Korea. The analyses consisted of the publication date of articles, principal dependent variables, subjects of the research, adequacy of sampling, adequacy of research purposes and results, accordance between theoretical model and hypothetical model, fit measures, theoretical base of model modification, and adequacy of conclusion. The results were as follows : The thesis of 33 articles in total were outnumbered as 25 (75.8%) to 8 (24.2%) research articles. As for a sex classification of the subjects, 45.5% of the research were conducted around a female group of subjects, while 54.5% were done for both sex, The range of the sample size was 105 to 803, and the average was 259 subjects. A single theoretical variable was measured for each measurement variable, any difference between variables was hardly found in 8 articles (24.2%), and 19 articles (57.6%) did not consider any measurement error. To analyze if the representative has been considered while collecting the data, most data were collected by a convenient sampling. Seven articles (21.2%) were seen with a sign of a representative. Questionnaires were used in a majority (31 articles) of the data collecting process. Only 2 articles (6.1%) were measured with a physiologic index simultaneously. 14 articles (42.2%) were centered on theory development, 10 articles on theory synthesis, and 9 articles on theory test. The research purposes and results were consistent in 25 articles (75.5%) and 8 articles (24.2%) were inconsistent. The quality of life and health promotion behavior were the concepts most frequently studied as a dependent variable, and 7 articles centered on them. In applied theories a health promotion model was used on 4 articles (12.1%), while role theory and stress-coping models were in 3 articles respectively. The articles were analyzed to see if the hypothetical model was elaborated and tested by the theoretical model. Twenty-five articles proved to be rationale for the inconsistencies. Also, 56.5% proposed hypotheses were supported among the subject articles, and 30 articles (90.0%) suggested a revised model. Path coefficient (17 articles) and theoretical adequacy (17 articles) were the standards mostly used. In conclusion, the principal factors were obtained from the research are to be considered as the principes of LISREL application. First, a model has to be established on a theoretical base rather than empirical results dependent on the data. The results are also required to be globally interpreted. Secondly, at least 200 samples are necessary to satisfy the need. Third, more than 3 measurement variables are to be adjusted to a single theoretical variable; the measurement errors must be suggested as well. Finally, normal distribution characteristics of the data and the estimation method need to be reported. Based on the research result, the follows are suggested; Systematic criteria on the LISREL application and procedure need to be developed Agreement form is required to report the results of research using the LISREL

Structural Modeling of Stress, Life Style and Health Status in Industrial Employees / 대한산업의학회지

OBJECTIVES: The purpose of this study was to examine stress affecting the life style(such as alcohol consumption, smoking, exercise, sleep disturbance, obesity), and to determine the relationships among the stress, life style and health status of in industrial employees. METHODS: This study utilized the data from health survey and laboratory test(such as liver enzymes and lipid profiles) in 463 Industrial Employees. Psychosocial wellbeing index was used in stress scales. The data were analyzed using SAS version 6. 11, LISREL version 8. 12i and PRELIS 2, 12i. The analytic methods for study were chi-square analysis, anova and covariance structural analysis. RESULTS: There was no significant relationship between socio-demographic characteristics and stress. The older person, married, higher education level, higher economic status and researcher group included the good life style. The higher streas groups included the more alcohol consumption, more smoking, less exercise and more sleep disturbance. There was no significant relationship between stress and health status(liver enzymes and lipid profiles) Alcohol consumption, less exercise and obesity had a significant relationship with abnormal liver enzymes. Alcohol consumption, smoking, less exercise, sleep disturbance and obesity had a significant relationship with abnormal lipid profiles. Based on the findings, the structural model of stress, life style and health status was established and then consider to socio-demographic characteristics the covariance structural analysis was used. The higher stress scores included the bad life style. The bad life style included the lower health status. Regardless stress, obesity was significant relationship with health status. CONCLUSIONS: Control of stress was initiative clues in health-promotive behaviors. Less alcohol consumption, no smoking, regular exercise, no sleep disturbance and not obese were health-promotive behaviors. Stress influenced the life style, and then life style influenced the health status.