| Permanent household income is an important indicator to measure thesocio-economic status of a family; it’s also the preferred indicator for the evaluation of afamily’s poverty and living level. The data sources are mostly according to the head’sself-reported household income of the household. However, the self-reported familyincome always had poor credibility. Focus on the low credibility of the self-reportedhousehold permanent income by the housemaster in household consumption andsocio-economic surveys, and the problem that the self-reported houselhold permanentincome could not be compaired directly, such as the difference between the urban andrural income, this study took advantage of the data of the Fourth National Health ServicesSurvey in Shaanxi Province in2008and regarded the household permanent income whichcould not be measures accurately as a latent variable, the impacts indicator variables puton the self-reported household income as randome effects. By fitting DIHOPIT model to determine the cut-off points for each indicator variable and using methods of thehypothesis test to estimate the validity of each indicator variable to evaluate the wealthranks of household income between regions.The objectives of this research is to estimate the household permanent income usedHOPIT model and to evaluate the validity of indivator variables involved in the survey,build a latent variable which can indicate the household income rank to solve the difficultythat cross regional self-reported permanent income could not be compaired. At the sametime, Evaluated the model fitting validity after dimensionality reduction of the variablesinvolved in the survey, sdudied the household permanent income increased levels between2003and2008in Shaanxi Province. Selected the raletively common cut-off points as theindicator to divide the poor families and the rich ones and discussed the relationshipbetween income inequality and population health status. Built household incomeinequality index based on the calculation of Gini coefficient and studied its impact onpopulation health. Evaluated the validity of indicator variables in the HOPIT model, andprovided a theoretical base for the design of future household survey questionnaire.The main results of this research are as follows:1. The income level between40survey counties (districts) in Shaanxi Provincevaried greatly, the mean of self-reported permanent income of all survey households was￥17810±14634, it was right skewed distribution. After took the logarithm of the incomeits mean was9.448±0.903and obeyed normal distribution. Transformed all the surveyitems in2008to8indicator variables, it were: adobe house, housing area>90m2, one TV,two or more TVs, mobile phone, mobile phone and fixed phone, tap water, unprotectedtoilet. The distribution of the8indicator variables between different counties (districts)was different.2. The HOPIT model fitting results of all survey households in40survey counties(districts) in Shaanxi Province in2008showed well model fit, the validity of eachindicator variable all had statistical significance on the estimation of household permanentincome. According to the indicating validity to household permanent income, the8indicator variables from high to low were: two TVs, mobile phone and fixed phone, one TV, mobile phone, tap water, housing area>90m2, unprotected toilet, adobe house. Dividedthe8indicator variables into two subsets, and then fitted HOPIT model seperately.Compare with universal set, the results showed the HOPIT model fitted well afterdimension reduction, subset with less indicator variables could get comparative result. Thecorrelation coefficient between the two subset’s HOPIT model predictive values and thefull set were0.6956and0.8626respectively. On the basis of well stability of HOPITmodel estimated results after dimension reduction, the growth trend analysis results from2003to2008indicated that all40counties (districts) in Shaanxi Province had differentdegrees of growth over the5years. Langao County, Baihe County and Linyou had thefastest growth rate over the five years, meanwhile Weibin District, Fugu County andWeicheng District had the slowest growth rate.3. Regarded the phone type as the uniform cut-off point to devide high incomehousehold and low income household, the percentage of high income households wasmore than50%in the majority of the counties (districts). The self-reported permanentincome, HOPIT predicted value, prevalence rate, and admission rate had statisticalsignificance between diffetent income levels (α=0.10), the results showed theself-reported permanent income and HOPIT predicted value in high income householdswere all higher than low income hoseholds. The prevalence rate in high incomehouseholds was generally higher than low income households, but the admission rate inlow income households was higher. Made use of the household income inequality index(HII) and the residents standardized prevalence rate (SPR) to fit regression model, theresults showed that HII had adverse impacts on SPR (P<0.01), in addition the inpact HIIput on SPR didn’t disappear with the addition of other variables.The main conclusions of this research are as follows:1. The HOPIT model fitting results of all survey families in40counties (districts) inShaanxi Province in2008demonstrated that the model fitted well, each indicator variablehad statistically significant validity on the estimation of household permanent income, andthe model had well stability. 2. The HOPIT model fitted well after dimension reduction, comparative results couldbe obtained with fewer variable subsets compired with universal variable subsets, and themodel after dimension resuction had well stability and feasibility. The permanent incomeof every county (district) in Shaanxi Province had different growth from2003to2008;there was a greater growth of residents’ living standard.3. The higher income levels, the higher living standards and living quality, the higherhealth status, in other words lower prevalence rate and the higher the utilization rate ofhealth services. Household income inequality index (HII) had adverse impacts on theresidents standardized prevalence rate (Standard Prevalence Rate SPR)(P<0.01), inaddition the inpact HII put on SPR didn’t disappear with the addition of other variables.The results supported the hypothesis that income inequality caused adverse impact onpopulation health. It showed that with the widening of income gap, it would aggravate thepopulation health situation, this sduty made up the void that income inequality indicatorcould not be calculated. In order to gradually reduce the income gap, reduce the poor-richpolarization to achieve the goal of common prosperity, we need to take further policiesand measures to reduce income inequality and improve the population health situation. |
PDF Full Text Request