Method

87

and mental health; namely individuals with a lower socioeconomic status have a higher possibility of suffering from mental health problems (Gong et al., 2012); Or in other words, people with the lower socioeconomic status had the higher likelihood of being with mental illness (Mavrinac et al., 2009; Sani et al., 2010). Specifically, individuals with higher incomes, compared to those with lower incomes, have better mental health (Huijts et al., 2010; Theodossiou & Zangelidis, 2009); as do those who are employed (Baron-Epel & Kaplan, 2009). In addition, concerning the relationship between education and mental health, one study conducted among the Japanese population showed that there is a significant linear correlation between the education level of the female and their mental health, while such an association does not exist in the Japanese male (Honjo et al., 2006).

3.1.2.3 Summary of the literature and the objective

In summary, research about the LTC issues in China has only recently begun. Most research is qualitatively focusing on comparing the national LTCI systems in different countries to verify the necessity and feasibility of the implementation of a LTCI system in China, while a few are quantitative studies related to the LTC issues. What is more, there are no studies exploring the relationship between socioeconomic status, mental health and the need for LTC. Compared with studies undertaken by Chinese researchers focusing on LTC issues in China, early research on the LTC issues conducted by foreign scholars involved more substantial total numbers, and are therefore more comprehensive and detailed.

Specifically, the previous research that examined the relationship between socioeconomic status and mental health is relatively substantial, followed by the research focusing on the relationship between socioeconomic status and need for LTC, while fewer studies clarified the association between the mental health and the need for LTC. However to date, there is still no research that clearly articulates the relationship between socioeconomic status, mental health and need for LTC at the same time.

Based on the summary of the previous research, the purpose of this study can be summarized as follows:

1) To investigate the extent that the need for LTC of the Tibetan urban elderly is satisfied;

2) To explore the structural relationship between the socio-economic status, mental health and the need for LTC among the Tibetan urban elderly.

88

Based on the literature review and the objective of the research, the following hypotheses are proposed:

Hypothesis 1: There is a strong positive correlation between socioeconomic status and mental health among the Tibet urban elderly. That is, a better socioeconomic status of the Tibet urban elderly would generally predict a better status of their mental health.

Hypothesis 2: A negative relationship would be observed between mental health and the need for LTC among the Tibetan urban elderly, which means a better mental health of the Tibet urban elderly would result in the lower need for LTC.

Hypothesis 3: Socioeconomic status does not only affect the need for LTC directly, but also exerts an indirect effect on the need for LTC through mental health. Here, the direct and indirect effects should be negative.

3.2.2 Study location and subjects

Tibet Autonomous Region has one prefecture-level city, six regions and 73 counties. A prefecture-level city is also the capital of the Tibet Autonomous Region – Lhasa city. In addition, a county-level city (Shigatse City) is included in the Shigatse region. Seven counties and one region (named Chengguan District, including seven sub-district offices and 28 community committees) are under the jurisdiction of Lhasa;

while Shigatse City includes 10 townships and two sub-district offices (10 community committees). Data used in this study were collected from late July to late August 2009, while subjects included in the survey were the urban elderly aged 60 years or over in the two cities of the Tibet Autonomous Region -- Lhasa and Shigatse.

3.2.3 Sampling method

Firstly, we collected the list of all community committees in Lhasa and Shigatse, including 28 communities in the former city and 10 in the latter city. Secondly, these 38 communities were then arranged in descending order based on their population numbers. Thirdly, 9 of the 28 communities in Lhasa and4 of the 10 communities in the city of Shigatse (total of 13 communities) were selected at random using a cluster sampling method. Those persons aged 60 years old and above from these 13 communities were selected as study participants. Specifically, there were 1,437 older persons in Lhasa, and 571 older adults in Shigatse, with a total number of 2,008 elderly people in the 13 communities from these two cities. One questionnaire was issued to each study participant, and a total of 1,836 completed questionnaires were returned, resulting in a valid response rate of 91.4 %.

This study employed both the staff of the community committee and college students whose mother language is Tibetan as questionnaire investigators. All investigators were able to speak both the Tibetan and Chinese, and were able to easily

89

alternate between these two languages, in order to ensure that the language translation had no impact on the authenticity and correctness of the questionnaire. Meanwhile, before the start of the survey, investigators undertook unified training to increase their knowledge and understanding about the purpose and content of the survey, the structure of the questionnaire, and skills and mastery on conducting a survey.

During the survey, the investigator mainly conducted the face-to- face questionnaire survey in households using a uniform questionnaire, and accurately recorded the responses of the elderly participants. In some communities, if it was possible, the respondents were gathered together to complete the questionnaire.

3.2.4 Measurement of the variables

The indicators of socioeconomic status, in the present study, are education and household income. The answers for education included: ①illiterate; ②elementary school; ③junior middle school ; ④senior middle school; ⑤vocational college and above. Options for the household incomes were: ①less than 1000 RMB; ② 1000-2999 RMB; ③3000-5999 RMB; ④6000 RMB and above.

The indicators for mental health used in this study are from the “Three Health Factors”, which was created by Prof Hoshi, resulting in a total of nine indicators to evaluate mental health, physical health and social health separately (Hoshi et al., 2012;

Hoshi, 2012; Hoshi & Sakurai, 2012; Kubo, 2012). The “Three Health Factors” index system firstly originated from the analysis of a database about the health conditions of the Tokyo elderly, and was also found to be equally applicable to evaluate the elderly from other cities in Japan, such as the Hanno City in Saitama Prefecture (Inoue, 2012).

Three indicators of mental health were included: (1) How is your health condition this year?; (2) Is your health status as good as last year?; and (3) Are you satisfied with your current life? In the present study, questions 1 and 3 will be used as the two indicators of mental health.

As stated previously, the quantitative studies that focused on the LTC issues were relatively few in China, while the empirical research that studied the LTC issues were relatively more, and the indicators for the need for LTC were different from one to the other in researches conducted in some developed countries. In this study, three questions will be used to measure the need for LTC: 1) the caring time of the first care provider; 2) the caring time of the second care provider; 3) the caring time of the third care provider. The same options are given to these three questions, namely: ① care just when needed; ② 2-3 hours per day; ③ half day per day; ④ almost all day long; ⑤ others.

3.2.5 Statistical approach

90

SPSS 17.0 was used to describe the basic distribution of the socioeconomic status, mental health and the need for LTC among the Tibetan elderly.

Amos version 17.0 statistical software package for Windows was used to perform SEM and to obtain the maximum-likelihood estimates of model parameters and provide goodness-of-fit indices. In other words, whether the structural relationship between socioeconomic status, mental health and the need for LTC among the Tibetan elderly is corresponded with the hypotheses stated previously will be examined using Amos.

Assessment of the model fitness calculates how the proposed model might be consistent with the data. Maximum-likelihood estimation is used to estimate the best-fitting model in this study. χ² test was not used to assess model fitness because when sample sizes is large, a non-significant chi-square is rarely obtained (Bentler &

Bonett, 1980; Jöreskog et al., 1981) and it was too dependent on sample size to be useful in many situations (Fan et al., 1999). While using maximum likelihood estimation in SEM, if sample size is large (n> 250) and independent latent variables is included, TLI (also known as NNFI) and IFI could be used as the fitness statistics.

Because TLI and IFI are relatively unaffected by sample size, they were used in the refinement of the model in present study. In addition, CFI is not too sensitive to sample size (Fan et al., 1999) and could adjust the value of TLI and IFI (Rong, 2010), it was also used as the fitness indices in this study. Another commonly used index, RMSEA was also reported and used as the fitness statistics. The model is regarded as good when TLI>0.90 (Hu & Bentler, 1999), IFI value close to 1 (Bentler & Bonett, 1980); CFI value >0.90 (Rong, 2010) and RMSEA <0.05 (Browne & Cudeck, 1989).

Table 3.1 Descriptive Analysis of the Subjects

n(%)

Lahsa Shigatse Total

Gender Male 528(40.3) 203(38.6) 731(39.8)

Female 782(59.7) 323(61.4) 1,105(60.2)

Age 60-69 749(57.2) 329(62.5) 1,078 (58.7)

70-79 432(33.0) 157(29.8) 589(32.1)

MoreThan80 129(9.8) 40(7.6) 169(9.2)

Ethnic group

Tibetan 1,292(98.6) 526(100.0) 1,818(99.0)

Han 8(0.6) 0(.0) 8(0.4)

Hui 7(0.5) 0(.0) 7(0.4)

Menba 2(0.2) 0(.0) 2(0.1)

Naxi 1(0.1) 0(.0) 1(0.1)

Total 1,310(100.0) 526(100.0) 1,836(100.0)

91