The Effect of the Dwelling Environment on Rural Elderly Cognition: Empirical Evidence from China

Abstract

Due to the falling birth rate and large-scale rural–urban migration, the ageing population in rural China is critical, and the quality of life of the rural elderly needs to be given more attention. In recent years, as an important part of new rural construction in China, the rural environment has been greatly improved, but the impact of such environmental improvement on the health of the rural elderly is not clear. Based on China Health and Retirement Longitudinal Study (CHARLS) data, this paper aims to assess the effect of the improvement of the indoor dwelling environment on cognition among the rural elderly by using a series of the difference-in-difference models (DID) under a counterfactual causality framework. The results showed that first, the improvement of both the toilet type and in-house shower/bath facility had a significant effect on the overall memory and delayed memory among the rural elderly in China, but the effect on immediate memory did not pass the significant test. Second, although access to running water could improve all types of memory, the effect was not significant. In summary, the indoor dwelling environment should be strengthened to ensure the process of healthy ageing in rural China.

1. Introduction

At present, China has entered the ageing stage, the speed of ageing is accelerating, and the elderly population is rapidly increasing. According to data from the Seventh National Population Census in 2021, the number of people over 60 years old is as high as 260 million, accounting for 18.70% of the total population, which is 5.44% higher than data from the Sixth Census. Under the trend of rising ageing levels, “healthy ageing” has become one of the most concerning topics in the public.

Cognitive ability is an important indicator to measure the health level of elderly individuals and refers to the ability of the human brain to process, store and extract information [1]. The decline in cognitive ability can cause mild cognitive impairment in the elderly and even develop into neurodegenerative diseases, such as Alzheimer’s disease [2]. As of 2013, the prevalence of mild cognitive impairment in the elderly population over 65 years old in China was 20.8%, and the prevalence of dementia was 5.14% [3]. In addition, the cognitive ability of elderly individuals is a key factor related to their quality of life, family harmony and social harmony. The decline in cognitive ability will lead to a decline in the learning, memory and decision-making ability of elderly individuals [4,5], increase the demand for treatment and rehabilitation and seriously affect individuals’ normal lives. Compared with the urban elderly population, the health condition of the cognitive ability of the rural elderly population is relatively low [6]. Specifically, the incidence of various cognitive disorders in rural areas is higher, and the rate of decline in cognitive ability is relatively faster [7]. Therefore, exploring the factors affecting the cognitive ability of the rural elderly and improving or delaying the decline in those cognitive abilities is an important issue that needs urgent attention and a solution.

For a long time, research on the cognitive ability of the elderly has been widely studied in academic circles, and scholars have mainly focused their research on the individual characteristics of the elderly and have made rich achievements. For example, the cognitive ability of elderly individuals declines with age [8], and the rate of decline is relatively slow among elderly individuals with a spouse compared to those without a spouse [9,10,11]. In recent years, scholars have generally used the above individual characteristics as controlled variables to explore the effects of other factors on the cognitive ability of elderly individuals. Among them, environmental factors have begun to attract the attention of scholars. Under certain conditions, the environment in which individuals live can have a positive effect on the reconstruction and growth of the nervous system, shaping people’s behavior and cognition [12,13]. Rural communities are the place where many older adults live in their later years. The improvement of the dwelling environment means the improvement of living conditions and the remodeling of the living habits of elderly individuals, which would stimulate the growth and development of their nervous system. However, there is not much literature on the influence of the environment on the cognitive ability of elderly individuals. In the few existing studies, the social factors in the environment are mainly discussed, such as community social capital [14], social exclusion [15] and social isolation [16].

In fact, compared with the social environment, the physical environment has a greater impact on the life and health in rural China with limited resources nowadays. Therefore, exploring the impact of the improvement of the physical dwelling environment on the cognitive ability of the rural elderly can both expand the scope and dimension of environmental health effects and provide a basis for improving the rural dwelling environment.

In view of this idea, this paper conducts an empirical study based on data from the China Health and Retirement Longitudinal Study to answer the following question: What is the impact of the dwelling environment on the cognitive ability of the rural elderly?

2. Literature Review

Over the past ten or so years, scholars have been exploring the causes of the decline in cognitive ability in the elderly and trying to find effective ways to alleviate and improve it. From combing the existing literature at home and abroad, it is found that numerous factors affect the cognitive ability of the elderly. These influencing factors can be divided into two aspects: individual factors and environmental factors.

2.1. Individual Factors and Cognitive Ability of the Elderly

In terms of individual factors, it is generally accepted in academic circles that the differences in cognitive abilities in the elderly are mainly manifested through the following aspects:

First, from the perspective of age, previous studies have found that the cognitive ability of elderly individuals showed a trend of continuous decline with age, and there would be accelerated decline at the end of life, which belongs to normal cognitive ageing [17]. Second, from a gender perspective, female elderly individuals have a faster rate of decline in cognitive ability and a greater risk of cognitive impairment than male elderly individuals [18,19,20,21]. Third, from the perspective of personality psychological traits, it was found that a more open personality promotes the development of cognitive abilities in the elderly [22], whereas anxiety and depression are associated with cognitive ability decline and deterioration in the elderly [23,24]. Fourth, from the perspective of socioeconomic status, some authors found that education and occupation are the common measures of an individual’s socioeconomic status, and studies have shown that education plays an important role in inhibiting cognitive ability decline. That is, elderly individuals with higher levels of education exhibit higher levels of cognitive health [25,26,27,28]. In addition, occupation is also significantly related to the cognitive ability of elderly individuals [29,30,31]. Management occupation and professional and technical occupation in the occupation category have a protective effect on cognitive ability [32]. Fifth, from the perspective of health-related behaviors, smoking and drinking were found to be negatively correlated with the cognitive ability of elderly individuals [33]. Active participation in healthy activities has a protective effect on the cognitive abilities of the elderly [34]. Specifically, daily leisure activities and intellectual activities can delay the decline in cognitive ability of elderly individuals [35,36].

2.2. Environment and Cognitive Ability of the Elderly

In addition to the personal characteristics and action choices at the micro level, more scholars have begun to explore the relationship between the environment and cognitive ability of the elderly on a more macro level. Most of the related academic studies are carried out from two aspects: social environment and physical environment.

In terms of social environmental factors, community social capital has a positive effect on the prevention of cognitive function impairment in elderly individuals; that is, the higher the level of community social capital, the better the cognitive ability of the elderly [37,38], and the reciprocal environment has the most significant impact. At the same time, some existing studies have found that rural elderly individuals with severe social exclusion are 23 times more likely to suffer from cognitive impairment than those without social exclusion [39]. The urban–rural dual structure mechanism, in which the household registration system is the core of the secondary division of various social resources, is also a key factor contributing to the difference in the cognitive ability of elderly individuals. The impact of this structural mechanism on elderly individuals’ health is mainly reflected in the inequality of infrastructure, social welfare and other aspects [40], thus increasing the economic burden of the elderly in rural areas, reducing the accessibility of their medical resources and affecting the cognitive ability of elderly individuals.

In terms of physical environmental factors, air pollution exposure can worsen the cognitive ability of elderly people, and biological and chemical air pollution in the dwelling environment can not only affect the respiratory system of elderly people, but it can also cause nervous system damage [41]. Cao et al. [42] further explored the relationship between household air pollution and the cognitive health of the elderly and found that the use of household fuel is also a risk factor for cognitive impairment in elderly individuals. In addition, the synergistic effect of psychosocial stress and air pollution on the health of the elderly was confirmed; that is, psychosocial stress is an important factor in determining susceptibility to environmental hazards [43]. The noise level of residential buildings may affect the cognitive ability of elderly individuals [44]. When road traffic noise in the residential area exceeds 45 decibels, it will cause a decline in sleep time and sleep quality and then affect the cognitive ability of the elderly [45].

In addition to the aforementioned natural geographic environment, the physical dwelling environment also plays a key role in the health status of individuals. Some scholars have found that the “toilet revolution” in rural areas plays an important role in protecting the overall health of rural people [46]. Running water is also an important factor affecting the cognitive health of adults. Compared with no access to running water, access to running water can improve the cognitive ability of rural adults [47]. However, it is worth noting that the above studies have not yet covered elderly individuals with more urgent cognitive health needs.

After reviewing the existing literature, it was found that academic circles generally pay attention to the health condition of the cognitive ability of the elderly and have produced rich research results. However, existing studies have the following two shortcomings. First, studies on the cognitive health of the elderly are mostly carried out at the individual level, and there is still a lack of research on the impact of the environment on cognitive ability. Second, in the few existing studies on environmental factors, scholars focus on either the social factors or the outdoor natural elements. Few studies have discussed the impact of the indoor physical dwelling environments closely related to the daily life of the elderly on their cognitive ability.

In summary, this paper aims to explore the impact of the indoor physical dwelling environment on the cognitive ability of the rural elderly, which is conducive to enriching the research perspective of their health condition and can provide empirical evidence for a deeper understanding of the impact of the environment on the elderly health.

3. Methods

3.1. Data Sources

The data used for this analysis were the China Health and Retirement Longitudinal Study (CHARLS). The CHARLS baseline survey was conducted in 2011 and carried out nearly every two years. To ensure sample representativeness, the CHARLS baseline survey covered 150 counties/districts and 450 villages/urban communities across mainland China, involving 17,708 individuals in 10,257 households, collectively reflecting the middle-aged and older Chinese population. A stratified (by per capita gross domestic product (GDP) of urban districts and rural counties) multistage (county/district–village/community–household) probability-proportional-to-size (PPS) random sampling strategy was adopted.

As a member of the Global Ageing Survey family, CHARLS collected various information, including personal and family, health status and function, cognition and depression, health care and insurance, work and retirement, pension and housing characteristics, etc. Specifically, the survey questionnaire includes the items about both memory and the changes in the indoor environment, which is exactly helpful to answer the question of this study. Furthermore, the poor rural environment has long existed, but it is only in recent years that the government has begun to pay more attention and devoted substantial resources to its reconstruction. Therefore, this study used the data from the latest two waves, namely 2015 and 2018, with the most sufficient information. After screening, the sample numbers aged 60 or above for the data analysis in this study were 5224 in 2015 and 4699 in 2018.

3.2. Variables and Measures

3.2.1. Dependent Variables: The Cognition

The dependent variable was the cognitive status of elderly individuals, which was operationalized into three components: immediate memory (immediate word recall (IWR)), delayed memory (delayed word recall (DWR)) and overall memory (overall word recall (OWR)). IWR indicates the number of words the respondent can immediately recall correctly from a 10-word list, with a minimum of 0 and a maximum of 10. DWR indicates the number of words the respondent can recall correctly from a 10-word list after a delayed time spent answering other survey questions, with a minimum of 0 and a maximum of 10. OWR is the respondent’s total word recall score, i.e., the sum of IWR and DWR, with a minimum of 0 and a maximum of 20.

3.2.2. Independent Variables: Improvement of the Dwelling Environment

Although the dwelling environment is a complex, multi-component system, the indoor physical environment is closely related to rural life and most affects the life and health in underdeveloped rural China. Therefore, the improvement of the rural indoor environment is an important part of China’s new rural construction and China’s rural revitalization project. Moreover, among the rural indoor environmental elements, the toilet, the water source and the bathing facilities are the most important in the daily life among the rural elderly [48,49]. Therefore, the improvement of the toilets, the water sources and the bathing facilities were selected as the independent variables in this study.

The first is the type of toilet, including two types: toilet with a seat and toilet without a seat. If the respondent had both, he was equal to selecting “toilet with a seat”. The choice of “toilet with seat” is recorded as 1, and the choice of “toilet without a seat” is recorded as 0. The second is whether the respondent’s residence has running water. If the respondent selects “yes”, it will be recorded as 1, and if he selects “no”, it will be recorded as 0. The third is whether the respondent has an in-house shower or bath facility. Corresponding to this question, there are three choices: “Concentration supply of hot water”, “Water heater installed by the household” and “no”. Based on the need for brevity and statistical analysis, the first two options are combined and defined as “there is an in-house shower or bath facility”, recorded as 1, and “no” is recorded as 0.

3.2.3. Control Variables

For the control variables, we included respondents’ demographics, physical health, socioeconomic status and residence arrangement. The demographic characteristics included age, gender (male = 1, female = 0) and marital status (the respondent lives with steady partner = 1, otherwise = 0). Physical health was measured by the respondent’s self-reported health (very poor = 1, poor = 2, fair = 3, good = 4, very good = 5). In the subsequent statistical analysis, based on the observation of distribution frequency and for the sake of brevity, this study recoded the initial five categories into three categories. That is, we combined very poor and poor into poor (recorded as 1), fair remained unchanged (recorded as 2), and we combined good and very good into good (recorded as 3). Socioeconomic status was evaluated by education and income. The education level is the highest degree of education level the respondent has finished when interviewed (under primary school = 1, the primary school = 2, fair = 3, junior middle school = 4, senior middle school or above = 5). In addition, income is the respondent’s income last year. Taking into account the non-normality of income distribution and drawing on the common practice of previous scholars, income was processed logarithmically in the subsequent statistical analysis. In addition, the residence arrangement of the rural elderly was considered and evaluated by asking if the respondent cohabitated with any child (yes = 1, no = 0). The specific meanings represented by each variable and the way they were coded are shown in

Table 1. The meaning of variables.

Variable	Variable Symbol	Variable Meaning
Interpreted variable	Overall word recall	The respondent’s total word recall score
	Immediate word recall	The respondent’s immediate word recall score
	Delayed word recall	The respondent’s delayed word recall score
Core explanatory variable	Toilet type	Whether the respondent’s toilet is with or without seat, yes = 1, no = 0
	Running water	Whether the respondent’s residence has running water, yes = 1, no = 0
	In-house shower/bath facility	Whether the respondent has in-house shower or bath facility, yes = 1, no = 0
Control variable	Age	The age of respondent
	Gender	The gender of respondent, male = 1, female = 0
	Marriage	Whether the respondent lives with steady life partner, yes = 1, no = 0
	Health	The respondent’s self-reported health, very poor = 1, poor = 2, fair = 3, good = 4, very good = 5
	Education	The education level of respondent, under primary school = 1, the primary school = 2, fair = 3, junior middle school = 4, senior middle school or above = 5
	Income	The respondent’s income from last year
	Coresidence	Whether the respondent cohabitated with any child, yes = 1, no = 0

.

3.3. Comparative Analyses

Before the formal statistical analysis, this study preliminarily compared the difference of the cognitive mean of the rural elderly before and after the environmental improvement.

Table 2. Comparative analyses before and after the environmental improvement.

	Toilet Type			Running Water			In-House Bath/Shower Facility
Variable	Mean1	Mean2	Difference	Mean1	Mean2	Difference	Mean1	Mean2	Difference
Overall word recall	5.154	5.353	−0.198 *	5.006	5.045	−0.039	4.911	5.117	−0.206 *
Immediate word recall	2.638	2.671	−0.033	2.592	2.574	0.018	2.556	2.575	−0.018
Delayed word recall	2.428	2.589	−0.162 **	2.336	2.375	−0.039	2.273	2.459	−0.187 ***
Age	68.681	68.948	0.267	68.956	68.866	0.090	69.534	68.736	0.798 ***
Gender	0.497	0.496	0.001	0.504	0.509	−0.005	0.485	0.495	−0.010
Education	1.564	1.595	−0.031	1.484	1.541	−0.057 **	1.487	1.539	−0.053 **
Marriage	0.777	0.784	−0.007	0.770	0.769	0.002	0.748	0.757	−0.010
Health	2.860	2.904	−0.044	2.833	2.827	0.006	2.811	2.867	−0.056 *
Coresidence	0.413	0.405	0.008	0.351	0.350	0.001	0.321	0.442	−0.121 ***
Ln_income	0.931	1.012	−0.081	0.731	0.991	−0.259 ***	0.762	1.099	−0.337 ***

Note: *, ** and *** indicate statistical significance at 10%, 5% and 1%, respectively.

shows the comparison results in terms of the three indoor environmental elements. In terms of the toilet type, the means of OWR, IWR and DWR of rural elderly increased by 0.198, 0.033 and 0.162, respectively. Meanwhile, the mean difference in OWR and DWR was significant. In terms of the water source, the means of OWR, IWR and DWR of rural elderly were not significantly different between the two groups. In terms of the bath/shower facility, the mean of OWR, IWR and DWR of rural elderly increased by 0.206, 0.018 and 0.187, respectively. In addition, the mean difference in OWR and DWR passed the significance test. On the whole, except for the water source, the changing of the toilet type and bath/shower facility are significantly good for OWR and DWR of the rural elderly.

3.4. Methodology

In this paper, we construct a series of difference-in-difference (DID) models to analyze the influence of the dwelling environment on the cognition of the rural elderly. The reasons for choosing DID are as follows: first, the CHARLS data selected in this paper are panel data, which meet the requirements of the DID model for data, and second, the DID model can largely avoid the endogeneity problem. Referring to other similar literature [50], the benchmark model is constructed as follows:

Y = α₀ + α₁ X × T + α₂ ΣβZ + δ + μ + ε

Y denotes the cognitive status of the rural elderly, i.e., IWR, DWR and OWR; X is the dwelling environment, i.e., toilet type, running water and in-house shower/bath facility; T is the time dummy variable; X × T denotes the interaction term between X and T; μ denotes the time fixed effect; ε is the random disturbance term; Z is the control variable; and δ is the environment fixed effect.

4. Results

To better explore the effect of the dwelling environment on cognition of the rural elderly, three models were constructed sequentially. Model 1 was constructed with only the interest variables, model 2 was constructed with the interest variables and control variables, and model 3 was constructed using the same variables as model 2 with the matched respondents from the propensity-matching score method.

4.1. Toilet Type and Cognition of the Rural Elderly

4.1.1. Balance Test after Propensity Score Matching of Toilet Type

Table 3. Balance test after propensity score matching of toilet type.

Variables	Mean Control	Mean Treated	Difference	T	p
Age	67.346	67.715	0.368	1.89	0.058 *
Gender	0.486	0.486	0.000	0.01	0.992
Under primary	0.623	0.613	−0.010	0.64	0.525
Primary	0.246	0.241	−0.005	0.35	0.727
Junior middle school	0.106	0.097	−0.009	0.95	0.341
Senior middle school or above	0.025	0.049	0.024	3.97	0.001 ***
Political	0.816	0.820	0.004	0.31	0.755
Residence arrangement	0.488	0.465	−0.022	1.41	0.159
Poor health	0.298	0.317	0.019	1.28	0.200
Fair health	0.505	0.486	−0.019	1.17	0.243
Good health	0.197	0.197	−0.000	0.02	0.983
Income	0.867	0.933	0.066	0.79	0.428

Note: * and *** indicate statistical significance at 10% and 1%, respectively.

shows the mean difference among the observable variables between the rural elderly who began to use the toilet with a seat and their counterparts who continued to use the squatting toilet after propensity score matching. It is clear that the difference between the two groups is not significant, except for age and senior middle school or above, which means that these two rural elderly groups are very similar before the change in toilet took place.

4.1.2. The Improvement Effect of Toilet Type on Cognition of the Rural Elderly

Table 4. The improvement effect of toilet type on cognition of the rural elderly.

Cognition	DID Model	Differences between Treatment and Control in 2015	Differences between Treatment and Control in 2018	Average Treatment Effects on Treated (ATT)	Standard Error	T	p
Overall word recall	Base DID	0.014	0.426	0.412	0.270	1.53	0.127
	Base DID + control variables	0.036	0.457	0.421	0.244	1.73	0.084 *
	PSM-DID	0.031	0.496	0.465	0.182	2.55	0.011 **
Immediate word recall	Base DID	0.023	0.093	0.070	0.131	0.54	0.593
	Base DID + control variables	0.025	0.111	0.086	0.120	0.71	0.476
	PSM-DID	0.022	0.113	0.091	0.088	1.04	0.299
Delayed word recall	Base DID	−0.004	0.319	0.323	0.163	1.98	0.048 **
	Base DID + control variables	0.014	0.333	0.320	0.151	2.12	0.034 **
	PSM-DID	0.006	0.332	0.326	0.110	2.97	0.003 ***

Note: *, ** and *** indicate statistical significance at 10%, 5% and 1%, respectively.

shows the results of the regression analysis of the effect of toilet type on the cognitive status of rural elderly individuals. Whether the results from model 1 or model 2, especially model 3 with more accuracy, the results consistently showed that elderly people who had a toilet with a seat had better overall and delayed memory than those who did not, and this difference was significant at the 5% and 1% levels, respectively. To be exact, the mean difference in the overall memory and the delayed memory greatly increased by 0.465 and 0.326, respectively. In other words, the improvement of the toilet environment in rural China can greatly improve elderly cognition. Meanwhile, the results of the three models show that the effect on the immediate memory of elderly individuals was not statistically significant, but the effect was still positive.

4.2. Water Source and Cognition of the Rural Elderly

4.2.1. Balance Test after Propensity Score Matching (PSM)

Table 5. Balance test after propensity score matching of running water.

Variables	Mean Control	Mean Treated	Difference	T	p
Age	67.629	67.612	−0.016	0.05	0.957
Gender	0.496	0.503	0.007	0.29	0.775
Under primary	0.665	0.637	−0.028	1.12	0.260
Primary	0.204	0.208	0.004	0.17	0.863
Junior middle school	0.106	0.122	0.016	0.97	0.330
Senior middle school or above	0.026	0.034	0.008	0.93	0.353
Political	0.803	0.798	−0.005	0.23	0.818
Residence arrangement	0.438	0.435	−0.003	0.11	0.910
Poor health	0.337	0.343	0.006	0.27	0.790
Fair health	0.484	0.475	−0.008	0.32	0.745
Good health	0.180	0.182	0.002	0.09	0.925
Income	0.857	0.914	0.057	0.43	0.669

shows the mean difference among the observable variables between the rural elderly who began to use running water and their counterparts who did not after propensity score matching. It is obvious that the difference among all the variables is not significant, which means that these two rural elderly groups are almost the same before the change in water source took place.

4.2.2. The Improvement Effect of Water Source on Cognition of the Rural Elderly

Table 6. The improvement effect of water source on cognition of the rural elderly.

Cognition	DID Model	Differences between Treatment and Control in 2015	Differences between Treatment and Control in 2018	Average Treatment Effects on Treated (ATT)	Standard Error	T	p
Overall word recall	Base DID	−0.051	0.089	0.139	0.284	0.49	0.624
	Base DID + control variables	−0.082	−0.034	0.047	0.258	0.18	0.854
	PSM-DID	−0.025	0.174	0.199	0.288	0.69	0.491
Immediate word recall	Base DID	−0.049	0.009	0.058	0.138	0.42	0.674
	Base DID + control variables	−0.068	−0.023	0.044	0.127	0.35	0.726
	PSM-DID	−0.049	0.035	0.084	0.140	0.60	0.551
Delayed word recall	Base DID	−0.019	0.029	0.048	0.173	0.28	0.781
	Base DID + control variables	−0.035	−0.049	−0.013	0.161	0.08	0.934
	PSM-DID	0.005	0.099	0.095	0.175	0.54	0.588

shows the results of the analysis of the effect of having or not having running water on the cognitive status of rural elderly individuals. The results of all three models indicate that the cognitive difference between the older adults who have running water and those who do not have running water is not statistically significant in terms of overall memory, immediate memory and delayed memory. However, we can still see that older adults with running water have better memory than those without running water. In other words, although there was no significant difference in the effect of changing the water source, the cognitive ability of the rural elderly with improved drinking water status was improved to some extent.

4.3. In-House Shower/Bath Facility and Cognition of the Rural Elderly

4.3.1. Balance Test after Propensity Score Matching of In-House Shower/Bath Facility

Table 7. Balance test after propensity score matching of in-house shower/bath facility.

Variables	Mean Control	Mean Treated	Difference	T	p
Age	67.466	67.399	−0.067	0.29	0.768
Gender	0.477	0.477	0.001	0.03	0.976
Under primary	0.639	0.636	−0.004	0.20	0.843
Primary	0.242	0.245	0.003	0.21	0.835
Junior middle school	0.089	0.088	−0.001	0.11	0.912
Senior middle school or above	0.029	0.031	0.001	0.22	0.829
Political	0.787	0.781	−0.006	0.37	0.713
Residence arrangement	0.440	0.465	0.025	1.32	0.187
Poor health	0.299	0.292	−0.007	0.38	0.700
Fair health	0.514	0.526	0.012	0.64	0.524
Good health	0.188	0.182	−0.005	0.37	0.714
Income	0.892	1.009	0.117	1.15	0.251

shows the mean difference among the observable variables between the rural elderly who decided to install in-house shower/bath facilities and their counterparts who did not after propensity score matching. The results are the same as those shown in Table 5. There are no significant differences among all the control variables. That is, these two elderly groups are similar before the installation of the in-house shower/bath facility occurred.

4.3.2. The Improvement Effect of Shower/Bath Facility on Cognition of the Rural Elderly

Table 8. The improvement effect of shower/bath facility on cognition of the rural elderly.

Cognition	DID Model	Differences between Treatment and Control in 2015	Differences between Treatment and Control in 2018	Average Treatment Effects on Treated (ATT)	Standard Error	T	p
Overall word recall	Base DID	−0.011	0.360	0.372	0.242	1.53	0.125
	Base DID + control variables	−0.174	0.216	0.390	0.222	1.76	0.079 *
	PSM-DID	−0.162	0.279	0.441	0.215	2.05	0.041 **
Immediate word recall	Base DID	−0.013	0.054	0.067	0.118	0.57	0.570
	Base DID + control variables	−0.088	−0.024	0.064	0.109	0.58	0.559
	PSM-DID	−0.093	0.008	0.101	0.104	0.97	0.330
Delayed word recall	Base DID	0.011	0.304	0.293	0.146	2.01	0.044 **
	Base DID + control variables	−0.081	0.232	0.313	0.136	2.30	0.021 **
	PSM-DID	−0.065	0.229	0.294	0.131	2.25	0.025 **

Note: *, ** indicate statistical significance at 10% and 5%, respectively.

shows the results of the analysis of the effect of indoor bathing facilities on the cognition of rural older adults. As shown in Table 4, the results of all three models show that overall memory and delayed memory were much better in older adults who had indoor bathing facilities than in those who did not have indoor bathing facilities, and the differences were significant at the 5% level. To be exact, the mean difference in the overall memory and the delayed memory greatly increased by 0.441 and 0.294, respectively. The difference in the effect of having an indoor bathing facility on the immediate memory of older adults was not significant, but the immediate memory of older adults with an indoor bathing facility was better than that of older adults without an indoor bathing facility. Therefore, in general, the improvement of indoor bathing facilities helped to improve the cognitive abilities of the rural elderly.

5. Discussion

The dwelling environment is critical in the process of healthy ageing, especially in rural China, which has comparatively poor infrastructure and serious population ageing, but it has not received enough attention. The present study was designed to determine the effect of the dwelling environment on rural elderly cognition, which will help to formulate more effective measures to prevent or delay the decline in the cognitive ability of the elderly in rural areas, thus promoting the cognitive health status of the elderly in other similar rural areas in the world. Based on data from the China Health and Retirement Longitudinal Study (CHARLS) in 2015 and 2018, this paper analyses the influence of the dwelling environment in terms of toilets, water sources and shower/bath facilities, which are directly connected with rural daily life, on the cognitive ability of the rural elderly by using a series of DID (difference in difference) models.

5.1. The Influence of Toilet Type on the Cognitive Ability of Rural Elderly Individuals

The study found that toilet type significantly affects the cognitive ability of rural elderly individuals. Compared with the traditional squatting toilet, the sitting toilet is more conducive to the development of the cognitive ability of elderly people. This result may be caused by two factors. On the one hand, sitting toilets in rural areas mostly benefit from the top-level design of the “toilet revolution” that is used in various countries. It takes into account health, safety and convenience needs [51]. For the initiative of changing toilets, rural elderly individuals mostly evaluate this as “more comfortable”, “more convenient” and “safer” [52]. Elderly people show high satisfaction, which helps to improve their happiness in life. At the same time, higher happiness plays an important role in protecting the cognitive health of elderly individuals. On the other hand, sitting toilets are beneficial in shaping health awareness and cultivating healthy behaviors in elderly individuals. The use of sitting toilets can encourage the rural elderly to realize the importance of their own hygiene and health management, thus gradually forming higher hygiene quality and health awareness in their daily lives. Accordingly, changes in health awareness and behavior will also benefit the development of cognitive ability.

It is worth noting that among the different variables selected to reflect the cognitive ability of older adults in this study, the effect of toilet type on the immediate memory ability of the rural elderly was not significant. This is also true for the enhancement of water security and in-house baths or shower facilities. This phenomenon may be related to the characteristics of immediate memory and the elderly neural system. On the one hand, immediate memory means a series of intensive neural activities that deal with collection, management and output within a very short time. On the other hand, with increasing age, the elderly’s physical function is continuously declining, thus the reaction speed of the neural system becomes slow, and immediate memory fades quickly [53]. Therefore, it is very difficult for the rural elderly to recall the words listed just now.

5.2. The Effect of Running Water on the Cognitive Ability of the Rural Elderly

This study shows that access to running water has no significant effect on the cognitive ability of the rural elderly. It was found that the pollution of drinking water sources caused by human and animal manure and industrial waste liquid in rural areas is the key factor that damages villagers’ health [54]. Therefore, the connection of running water has an obvious improvement effect on the health of rural residents [55,56]. However, this does not necessarily mean that access to running water can improve all aspects of health.

The reason why the environmental improvement in terms of water source had no significant effect may be as follows. First, in rural China, the people traditionally build a house near the water source, thus the rural elderly usually feel that it is convenient to obtain water. Therefore, the changing water source does not make a great difference to the rural elderly. Second, the improvement of the water source may take a comparatively long time to have an effect, but the longest interval between the two samplings in CHARLS is 3 years. It is highly possible that the timing of the changing water source in the respondent’s family is very close to the timing of sampling, thus the effect is not clear.

5.3. Effect of In-House Shower/Bath Facility on the Cognitive Ability of the Rural Elderly

This study found that having in-house shower/bath facilities is helpful for improving the cognitive ability of the rural elderly. Studies have shown that suffering from cardiovascular diseases, hypertension and other diseases will increase the probability of cognitive decline in elderly individuals [57]. During bathing, moderate water temperature can activate the whole vascular system of elderly individuals, which makes the performance of cardiac function more prominent [58,59], and proper hydrostatic pressure can also help the stability of individual blood pressure [60]. Therefore, bathing can reduce the probability of cognitive impairment by improving the status of cardiovascular diseases, hypertension and other diseases in rural elderly individuals. Compared with no in-house shower/bath facilities, having in-house shower/bath facilities allows rural elderly individuals to control the water temperature and water pressure more freely during bathing, and bathing frequency is relatively increased, which is more helpful for delaying the decline in their cognitive ability. In addition, bathing can improve the sleep quality of rural elderly individuals and thus improve their cognitive health [61].

5.4. Contribution and Limitations

This study contributes to the research on cognition of the rural elderly in China, a vulnerable group that deserves great attention in the process of healthy ageing. In terms of methodology, the use of difference-in-difference models can avoid the endogenous problem to some degree, thus inferring the real causal effect of the dwelling environmental improvement. Meanwhile, the data from CHARLS ensure the authority, representativeness and popularization of the conclusion. The findings help us strengthen the understanding of the conclusive effect of environmental changes; on the other hand, some aspects of rural elderly cognition are hard to change.

There are some limitations that need further improvement. First, the concept of cognition is not precisely operationalized. Although memory is a key part of cognition and this study used three memory-related indicators, these unidimensional indicators cannot represent this broad and inclusive concept well. Second, due to the limited space, this study did not explore the mechanism through which the improvement of the dwelling environment affects cognition of the rural elderly. With this part, we can understand how the environment affects cognition. Third, this study did not reveal subgroup differences in environmental effects. Because of the different levels of social and economic development in rural China, there are great differences between regions. Only when we have some clarity on how the environment affects the different groups can we formulate the targeted measures. Forth, there are still some confounding variables, especially the time-varying variables, which may confound the results.

Therefore, future studies need to operationalize this concept from more dimensions, deeply explore the pathways and further explore the heterogeneous effects of environmental improvement. In particular, the researchers need to pay more attention to and control other possible confounding factors and use more effective models to objectively evaluate the impact of environmental change on health.

6. Conclusions

This study has shown that the types of toilets and in-house shower/bath facilities all significantly affect the cognitive ability of the rural elderly. Compared with squatting toilets and no in-house shower/bath facilities, using sitting toilets and having in-house shower/bath facilities have greatly improved the cognitive ability of the rural elderly. In addition, although running water has no significant effect on the cognitive ability of the rural elderly, the results show that the cognitive ability of the rural elderly is improved after tap water is connected. The conclusion of this study has multiple implications for the field. First, in attitude, people should realize that there is an urgent need to improve the dwelling environment of the rural elderly and construct an elderly-friendly village, which improves not only quality of life and satisfaction, but also the neural system and cognitive ability, thus ensuring healthy population ageing.

Second, we should accelerate the construction speed of rural infrastructure, which is directly related to elderly daily life. To be specific, we should continue to promote the “toilet revolution” in rural areas as a whole and pay attention to the protective effect of more sanitary toilet types on the cognitive health of rural elderly people. Meanwhile, we must effectively use a variety of policy tools to encourage rural residents to invest in in-house shower/bath facilities and improve in-house bathing conditions. Of course, the change in water source cannot be neglected because it can improve the health of other aspects and quality of life among the rural elderly.