Method for measuring external influencing factors of ordinal data structure and its application

Objective To explore a method for measuring external influencing factors of ordinal data structure.Methods The agreement index was introduced to measure the external influencing factors of ordinal data structure based on information measurement theory and the application of the index was illustrated with an example.Results Information on health status and its related factors among 17 577 middle aged and elderly persons living in 145 communities and 305 villages were extracted from the database of The China Health and Retirement Longitudinal Study(CHARLS).The number of the middle aged and elderly persons with very good,good,fair,poor,and very poor health were 1 338 (7.61%),2 864 (16.29%),8 169 (46.48%),4 723 (26.87%),and 483 (2.75%),respectively.The average agreement index for health status for all the middle aged and elderly persons was 0.57±0.08 and the average agreement index for the persons living in villages was significantly lower than that for the persons living in urban communities (0.55±0.07 vs.0.61±0.06,t=-8.889;P<0.000 1).The results of multivariate linear regression analysis showed that the middle aged and elderly persons living in villages providing the pension to the elderly aged ≥ 65 years,with sewer system,and with outdoor physical exercise equipments were more likely to have better health status;the results also revealed that the urban middle aged and elderly persons living in rainless regions and living in communities with outdoor physical exercise equipments were more likely to have better health status.Conclusion Agreement index could be applied to measure external influencing factors of ordinal data structure in research data analysis.