Effects of residential indoor environments on occupant satisfaction and performance

ABSTRACT

With vast influences of Chinese industrial transformation, also largely impacted by the global challenges, such as climate change and pandemic, the home becomes as a place for occupants to perform their work tasks. To that end, this study concentrates on the effects of residential indoor environments, especially the factor of indoor environment quality (IEQ) and indoor greenery (IG), on occupants’ residential satisfaction and performance. By conducting a post-occupancy evaluation (POE) survey with the perspective of user experience (UX), 764 samples were collected, of which 528 are valid for statistical analysis. The method of structural equation modeling (SEM) was adopted to test the hypotheses. The results show that both indoor environment quality and indoor greenery have a positive effect on residential satisfaction and occupant performance, however, the former is more influential. In addition, residential satisfaction is a partial mediator, which demonstrates the mechanism: residential indoor environment mainly affects satisfaction, then occupant performance. The findings of this study provide important evidence from China in the period of the post-pandemic era that combines the unique characteristics of urbanization, gaining more insight for researchers to understand the relationship between the built environment of homes and their occupants, revealing specific design implications for practitioners.

KEYWORDS:

• Indoor environment quality
• indoor greenery
• post-occupancy evaluation
• residential satisfaction
• occupant performance

1. Introduction

Encountering the modern challenges of climate change, pandemics, regional conflicts, and industrial transformation, homes become more crucial to support the tasks which are used to be performed in the office. Working at home (WAH) has multiple potential benefits including reducing the emission of the greenhouse gas during the workers’ routine transportation (Kitou and Horvath 2003; Nilles et al. 1976); saving the consumption of energy in centralized commercial or educational buildings (Dimartino and Wirth 1990); and, balancing the needs of occupants’ work and life (Nakrošienė, Bučiūnienė, and Goštautaitė 2019), which can even provide more job opportunities for disability. However, the knowledge of the influences of the residential indoor environment on occupant performance (OP) is limited, which not only restrains the understanding of the relationship between the built environment and its occupants for researchers but also constructs the barrier for practitioners to design a supportive home for its residents to ensure their satisfaction and performance.

Both indoor environment quality (IEQ) factors and indoor greenery (IG) can significantly influence occupant performance. On the one hand, previous studies have demonstrated their effects separately (Al Horr et al. 2016), which exhibit indoor built environment can affect satisfaction and performance in different types of conditions (Luo et al. 2023; van den Bogerd et al. 2020; Vasquez et al. 2022); on the other hand, IG is intertwined with nearly all aspects of IEQ, for plants can directly influence indoor air quality (Katoch and Kulshrestha 2022), thermal comfort (Tan et al. 2018), acoustic comfort (Moya et al. 2018). However, evidence from China is very insufficient, especially in the circumstance of residential indoor environments, which limits the understanding of its effects on occupant satisfaction and performance in the specific region with unique characteristics of urbanization (Li, Li, and Wang 2022; Liu and Lo 2022; Xu et al. 2022) that may cause differences in the results and generalizability of the former researches.

Hence, combining the perspective of the post-occupancy evaluation (POE) and user experience (UX) theory, this study focuses on the effects of residential indoor environments, including IEQ and IG, on occupant satisfaction and performance. It contributes to the knowledge of human and indoor built environment relationship (Ganesh et al. 2021), and also responds to the current need of working at home in the post-pandemic era after China opened up the social mobility, which provides important evidence to the state-of-art. The following sections are organized as followed: Section 2 introduces the directly related literature that concentrates on this topic and puts forward the hypotheses; Section 4 illustrates the measurement of each factor, sample characteristics, and analysis procedures; Section 5 shows the results into three parts, including reliability and validity test, correlation analysis, and hypothesis test; Section 6 discusses the result, limitation and future research, and design implications; Section 7 concludes the main findings of this study.

2. Literature review

2.1. POE and UX theory

A balance of the built environment energy consumption and human perception can achieve a sustainable life-cycle of the building (Basbagill et al. 2013; O’Brien et al. 2017). Post-occupancy evaluation provides an opportunity for engineers to gain feedbacks on both sides that may assist the reflection of the current condition and impact future design and practices (Gill et al. 2010; Menezes et al. 2012; Whitehouse et al. 2001). Studies demonstrated that conducting POE surveys can help to understand the influence of residential indoor environments on occupants. For instance, Umishio et al. (Umishio et al. 2022) analyzed the officer workers’ subjective evaluation of the indoor environment quality during the period of working from home in 2020 in Japan and illustrated that work productivity can be improved by decreasing the negative influence of indoor air pollution. Similarly, the result of another questionnaire survey by Bergefurt et al. (Bergefurt et al. 2022) showed that noise is the major distraction that reduces occupant performance. Nevertheless, in the field of human-computer interaction (HCI) (Fischer 2001), user experience theory (Hassenzahl and Tractinsky 2006), defined as the user’s before, during, and after-use responses to the use of a system and product, puts more concentration on the end-user (Xu and Zhang 2022). This theory provides a two-factor model that includes pragmatic and hedonic factors (Deng et al. 2010). Previous studies have adopted UX surveys via online questionnaires to investigate the built environment, as reported by Koyaz and Ünlü (Koyaz and Ünlü 2022). They applied a holistic approach to not target certain buildings, but the exact experience from different perspectives of occupants in various work environments, which can improve the generalizability of the findings. In short, the common aspect of the two theories is that they all have emphasized the system (i.e., indoor environment) and the occupant (i.e., user) spontaneously to improve the quality and utility of the built environment, as a system, through each occupancy period.

Combining the POE and UX theory can generate a new perspective of understanding. However, few studies have integrated them to analyze the residential indoor environment, and none of them focuses on the sample from China. Thus, this study tries to reduce the gap between the two theories, by providing empirical evidence.

2.2. Indoor environment quality

Former researchers have reached a consensus on the physical factors of IEQ, including indoor air quality (IAQ), lighting environment, acoustic environment, and thermal comfort (Geng et al. 2019). Each factor has been demonstrated to have a significant effect on occupants’ evaluation of the occupancy of the built environment, also the productivity, especially in the commercial and educational indoor environment (Bavaresco et al. 2019; Ortiz and Bluyssen 2022; Tsoulou et al. 2021). In recent years, studies have started to focus on the residential indoor environment. For instance, Salamone et al. (Salamone et al. 2021) conducted an online survey in Italy and demonstrated that not all of the occupants are positively affected by the thermal environment, but IAQ has a significant effect on both consequence of satisfaction and occupant performance. However, both studies by Guo et al. (Guo et al. 2022) and Awada et al. (Awada et al. 2021) reported that IAQ reveals no significant impacts on occupant performance. Though the fact that IEQ factors can significantly influence occupants has been examined in different built environments, the reports of evidence in terms of the residential indoor environment are different from each country, which may be due to the various sample characteristics of each country. Therefore, it is necessary to collect and compare the evidence around the world, which can increase the knowledge of the relationship between this type of built environment and its occupants. In light of the connection and synchroneity of each factor, previous studies have adopted the holistic subjective evaluation of the IEQ, representing the IEQ factors with a single latent variable, as illustrated by Zhang and Tu (Zhang and Tu 2021) who focused on green buildings and Zhang (Zhang 2019) who targeted on university library buildings. Both studies, using path analysis, found a significant direct influence of IEQ on satisfaction which is a partial mediator. Hence, considering the indoor environment complexity without segregation may provide a potential understanding of the balance of the influences of each factor in the residential indoor environment, which may also generate practical design implications for engineers in a holistic view.

2.3. Indoor greenery

As one of the eight main factors of physical environment illustrated by Al Horr et al. (Al Horr et al. 2016), biophilia (i.e., the love of nature (Fromm 1973; Kellert and Wilson 1993)) can influence occupant perception in not only occupants’ satisfaction and performance but also anxiety and stress (Yin et al. 2020). The biophilic design of the indoor environment includes mainly three parts (Sachs 2022): direct experience, indirect experience, and experience of space. Accordingly, indoor greenery, or phytophilia (i.e., the love of plants (Tifferet and Vilnai-Yavetz 2016)), as an independent factor, is one of the biophilic design aspects that have been widely investigated in the condition of the office indoor environment (Lei, Yuan, and Lau 2021; Yeom, Kim, and Hong 2021; Yin et al. 2019). In addition, studies reported that indoor greenery can directly influence indoor environment quality, which, as a result, exhibits an interrelated effect on occupants (Raji, Tenpierik, and van den Dobbelsteen 2015). As concluded by Wang et al. (Wang, Er, and Abdul-Rahman 2014), indoor plants can improve indoor air quality and stabilize the humidity and temperature, which all contribute to occupant performance enhancement. In short, studies, focusing on office environments, show the significant positive effect of indoor greenery (i.e., indoor plants) on workers; however, few studies focus on this effect potentially on the occupants’ satisfaction and performance at home with consideration of the effect IEQ in the same research model.

2.4. Satisfaction and performance

According to Toftum et al. (Toftum, Andersen, and Jensen 2009), occupant performance is defined as the outcome of occupants’ cognitive ability, including concentration, memory, and thinking, which are the core function to perform any professionalized work tasks. Processes of various tasks in different indoor environments may be affected by completely different elements (Kang, Ou, and Mak 2017). Besides the significant effects of IEQ and IG that have been mentioned above, satisfaction and performance, individually as a hedonic and pragmatic factor of UX theory, are affected by multiple factors, such as home infrastructure (including furniture, equipment, etc.) (Tunk and Kumar 2022), internet connection (Yang, Kim, and Hong 2021), and the number of residents (Feng and Savani 2020). Studies demonstrated that the subjective evaluation of the occupants on the built environment occupancy (i.e., satisfaction) has a significant positive effect on the perceived performance (Frontczak et al. 2012). For instance, Zhang (Zhang 2019), adopting the method of structural equation modeling, focused on the library indoor environment and demonstrated that satisfaction is a complete mediator of the effect of IEQ on occupant performance. Similarly, Kamaruzzaman et al. (Kamaruzzaman et al. 2011) conducted a case study of six historic buildings in Malaysia. They illustrated that to achieve sustainable development of the built environment, “user satisfaction” towards the multiple aspects of the physical environment should be emphasized to increase productivity. In short, in different situations of indoor environments, occupant performance can be improved by increasing satisfaction towards the occupancy of the built environment. Another common aspect of many previous studies, focusing on occupant performance at home, is that they were conducted during the special period of the global pandemic, which gives a similar base to compare different evidence around the world, however, limits the generalizability of the results in the post-pandemic period when occupants have adopted a mixed style of workplace choice, as illustrated by (British Council of Office). Therefore, this study provides evidence that may supplement the knowledge of the relationship between residential satisfaction (RS) and occupant performance, also their antecedents, and contributes to the practice of designing a supportive home for occupants to achieve better performance.

3. Hypothesis

Based on the theory and literature reviewed above, the hypotheses of the relationship between each main factor (i.e., occupant performance, residential satisfaction, indoor environment quality, and indoor greenery) can be drawn as followed, which constructed the research model, as shown in the following Figure 1. We also hypothesized that the total days of working at home can influence the mechanism because occupant experience may vary according to time spent in the residential indoor environments, which can alter their post-occupancy evaluation outcomes.

Figure 1. Research model.

H1:

IEQ has a positive effect on RS;

H2:

IEQ has a positive effect on OP;

H3:

IG has a positive effect on RS;

H4:

IG has a positive effect on OP;

H5:

RS has a positive effect on OP;

H6:

RS is a complete mediator.

H7:

Total days of working at home can influence the mechanism.

The research model illustrates the hypothetical relationship of the latent variables (i.e., main factors) that are measured by the observed variables illustrated in the following section 4. More specifically, in Figure 1, the arrow line along with the hypothesis mark (e.g., H1, H2, etc.) indicates the effect of one variable on another. For instance, the H1 arrow line means that indoor environment quality has a positive effect on residential satisfaction. It should be noticed that residential satisfaction, as a mediator, is being pointed by indoor environment quality and indoor greenery and pointing to occupant performance as a dependent variable.

4. Method

4.1. Measurement

This study combines the approach of POE and UX survey, based on Zhang (Zhang 2019) and Koyaz and Ünlü (Koyaz and Ünlü 2022). By consulting the experts in the field of the built environment, statistics, and user experience, also having discussed with several occupants from different domains of the profession who are commonly working at home, the initial structure of the survey was confirmed. Accordingly, the questionnaire was constructed into five parts, including the measurement of indoor environment quality, indoor greenery, residential satisfaction, occupant performance, and control variables. To examine the quality of the questionnaire, a pilot study was conducted, further checking the reliability and validity of the latent variables. The final measurement of the dependent, mediator, independent and control variables is shown as followed:

Indoor environment quality: the measurement of IEQ is based on Bergefurt et al. (Bergefurt et al. 2022) and Zhang (Zhang 2019). Seven items that encompass indoor air quality, natural light, humidity, artificial light, noise, ventilation, and temperature are measured through on a 7-point Likert scale, from 1 – “extremely disagree” to 7 – “extremely agree”. Item questions from this latent variable apply a similar pattern, such as “I feel comfortable with the indoor air quality at home.”

Indoor greenery: the measurement of IG is based on Mousavi Samimi and Shahhosseini (Mousavi Samimi and Shahhosseini 2020) and Dzhambov et al. (Dzhambov et al. 2021). Three items are adopted, including plant types, the number of plants, and plant colors. An example item question is as followed: “How many types of plants are there in your residence.”

Residential satisfaction: the measurement of RS is based on Zhang and Tu (Zhang and Tu 2021). Three items are measured with a 7-point Likert scale, from 1 – “extremely unsatisfied” to 7 – “extremely satisfied”. For example, the question is asked: ”How satisfied do you feel towards the quality of the residence as a place to live.”

Occupant performance: the measurement of OP is based on Zhang (Zhang 2019) and Wargocki and Wyon (Wargocki and Wyon 2017). Four items, including concentration, memory, thinking, and overall performance, are evaluated by a 7-point Likert scale, from 1 – “extremely disagree” to 7 – “extremely agree”. The sample item question is as followed: “I can concentrate on my task at home.”

Control variables: the control variables include age, gender, province, education level, self-reported health, number of people at home, size of residence, sleep quality, normal workspace, internet connection, the abundance of furniture and equipment for work tasks, total days a year of working at home in normal condition, task type (Kang, Ou, and Mak 2017).

4.2. Sampling

Considering the POE and UX approach as mentioned above, the sample from China suits the criteria of balancing the generalizability and heterogeneity in light of the characteristics of a relatively broad geographical range with a unified economics and culture environment. The survey was conducted online to collect anonymous samples around the country from March to April 2023, after the recovery of the pandemic and opening-up of the social mobility. Respondents who completely finished the survey questionnaire and successfully submitted can receive three CNY as compensation for their time and effort.

A total of 7548 questionnaires were sent out through major social media platforms in China, and 764 of them were collected. To ensure the quality of the sample, five steps of collecting and filtering were taken: (1) random sampling was performed through different platforms. Equipment and IP address were double-checked during the submission of each sample to avoid repeated submission of the questionnaire; (2) the minimum threshold amount of samples (i.e., 10–25 times that of the measurement items (Bergefurt et al. 2022; Gorsuch 1983)) and minimum survey time cost, based on the pilot study, was confirmed; (4) an inverted 7-point Likert scale was implemented to identify the invalid samples that were suspected as internal inconsistency; (5) any respondent who spent less than 1 month a year, in total, working at home were not qualified in the sample. As a result, 528 valid samples that are above the threshold of 425 (i.e., 25 times of measurement items) were confirmed and utilized in the analysis process. The demographic information is shown in Table 1. Nearly 55% of the sample are female residents. More than half of the respondents are undergraduate and below in education level. Among the valid samples, other characteristics are also shown in Figure 2, including age, size of residence, normal workspace at home, task types, total days of working at home in normal conditions, and the number of people who share the same residence with the respondent.

Figure 2. Characteristics of the sample.

Note: Ind.-Independent; NFS-No function separation; Both items of workspace and task type are multiple choices that are limited from 1 to 3 selection for respondent.

Table 1. Demographics.

Item	n (%) or Mean±SD
Gender
Female	288 (54.55%)
Male	240 (45.45%)
Education level
Undergraduate and below	345 (65.34%)
Graduate	176 (33.33%)
Ph.D.	7(1.33%)
Self-reported health	5.25±1.19
7-Good	85 (16.10%)
6	137 (25.95%)
5	181 (34.28%)
4	88 (16.67%)
3	27 (5.11%)
2	8 (1.51%)
1-Bad	2 (0.38%)
Internet connection	5.59±1.27
7-Good	145 (27.16%)
6	159 (30.11%)
5	141 (26.71%)
4	47 (8.90%)
3	25 (4.73%)
2	4 (0.76%)
1-Bad	7 (1.33%)
Furniture & equipment	5.35±1.25
7-Sufficient	117 (22.16%)
6	122 (23.11%)
5	168 (31.82%)
4	84 (15.91%)
3	27 (5.11%)
2	8 (1.51%)
1-Not sufficient	2 (0.38%)
Sleep quality	5.23±1.28
7-Good	109 (20.64%)
6	110 (20.83%)
5	160 (30.30%)
4	106 (20.08%)
3	36 (6.82%)
2	2 (0.38%)
1-Bad	5 (0.95%)

The main body (n = 413) of the sample is of the age between 20s to 40s who are the main labor force in present China (Liu, Qi, and Liu 2022). Most home sizes of the respondents are less than 150 m². Independent bedroom is the main workspace (n = 371) where occupants may perform reading, writing, and calculation, as the top three main task types. More than half of the sample (n = 337) are in a family size of 3–4 people, which may partly reflect the shift of easing the one-child policy since 2015 (He, Li, and Han 2023).

Moreover, over 68% of the respondents spend at least 3 months working at home in normal conditions. Among the main group of this category of WAH: Group 1 includes 165 (31%) of the respondents spending a total of 1–2 months working at home; and Group 2 includes 154 (29%) of them spending 3–4 months. Both groups are utilized in the further analysis of the hypothesis test, while other groups are limited in the sample size.

4.3. Analysis procedures

Both IBM SPSS 25.0 and AMOS 21.0 were used to process the statistical analysis. The reliability and validity of each latent variable were examined by Cronbach’s alpha coefficient and confirmatory factor analysis (CFA) which both illustrate the representation between latent variables (i.e., OP, RS, IEQ, and IG) and observed variables (i.e., item questions). After the observed variables combined into the latent variable, the Pearson correlation test was conducted to test the correlation of each latent variable. After confirming the above test results, based on the research model (i.e., shown in Figure 1), AMOS was used to construct a structural equation model that was then further tested by the model fit indicators to identify the fitness of the hypothesized relationship.

5. Results

5.1. Reliability and validity test

The Cronbach’s alpha coefficient of each latent variable individually is 0.879 (Indoor environment quality), 0.868 (Indoor greenery), 0.933 (Residential satisfaction), and 0.867 (Occupant performance), all above the minimum threshold of 0.7. The confirmatory factor analysis was conducted to examine whether the latent variables can represent the observed variables accordingly. The result of the confirmatory factor analysis is shown in Table 2. The thresholds of the critical model fit indices are as follows: χ²/df <3, root mean square residual (RMR) <0.08, goodness-of-fit index (GFI) >0.9, adjusted goodness-of-fit index (AGFI) >0.9, comparative fit index (CFI) >0.9, normed fit index (NFI) >0.9, incremental fit index (IFI) >0.9, root mean square error of approximation (RMSEA) <0.08. Accordingly, the model fit indices are all within the recommended value: χ²/df = 2.455, GFI = 0.940, AGFI = 0.919, CFI = 0.972, NFI = 0.953, IFI = 0.972, RMR = 0.058, RMSEA = 0.053. In addition, each observed variables’ standard regression weight is higher than 0.5. The threshold of construct reliability (CR) is greater than 0.7; the average variance extracted (AVE) is greater than 0.5. Therefore, the results of the reliability and validity test indicate that latent variables can represent the related observed variables. The value of each latent variable is determined by the average value of the observed variables. For instance, the value of IG is obtained by the average value of the number of plants, type of plants, and color of plants; and the value of IEQ is obtained by the average value of the related seven observed variables accordingly. Taken together, the latent variables are qualified for the next step of the analysis.

Table 2. Confirmatory factor analysis results.

Latent	Observed	Estimate	S.E.	C.R.	S.R.W.	CR	AVE
IEQ	Indoor Air	1			0.737	0.886	0.527
	Natural Light	1.067	0.062	17.308	0.770
	Humidity	1.040	0.064	16.296	0.727
	Artificial Light	1.027	0.061	16.789	0.748
	Noise	1.005	0.08	12.620	0.569
	Ventilation	1.004	0.06	16.643	0.741
	Temperature	1.128	0.065	17.296	0.769
IG	Number of plants	1			0.858	0.880	0.713
	Type of plants	0.901	0.037	24.063	0.942
	Color of plants	0.822	0.043	18.980	0.717
RS	Home satisfaction_1	1			0.851	0.935	0.829
	Home satisfaction_2	1.126	0.038	30.001	0.931
	Living satisfaction	1.077	0.035	30.725	0.946
OP	Concentration	1			0.826	0.878	0.647
	Memory	1.124	0.05	22.560	0.834
	Think	1.134	0.045	25.377	0.921
	Overall	0.848	0.058	14.601	0.601

Note: IEQ-Indoor environment quality; IG-Indoor greenery; RS-Residential satisfaction; OP-Occupant performance; S.E.-Standard error; C.R.-Critical ratio; S.R.W.-Standard regression weight; CR-Construct reliability; AVE-Average variance extracted.

5.2. Correlation analysis

The Pearson correlation test is adopted as the first step to analyze the relationship between residential satisfaction, occupant performance, indoor environment quality, and indoor greenery before the hypotheses test. The result of the correlation analysis is shown in Table 3, along with the mean value and standard deviation of each latent variable.

Table 3. Pearson correlation analysis results.

	IEQ	IG	RS	OP	Mean	SD
IEQ	1				5.320	0.944
IG	0.169**	1			2.311	1.121
RS	0.670**	0.236**	1		5.600	1.079
OP	0.505**	0.176**	0.563**	1	4.963	1.208

Note: ** is at the significant level of 0.01 (two-tailed).; SD-Standard deviation.

As expected, all latent variables are significantly correlated (p < 0.01). Especially, the correlation between indoor environment quality and residential satisfaction is higher than other correlation relationships. Hence, based on the results of the reliability and validity test, and correlation analysis, the research model can be further examined.

5.3. Hypothesis test

The analysis method of structural equation modeling, conducted through IBM AMOS, is applied to test the research model, which not only enables the spontaneous process of the reflection between observed and latent variables but also involves more than one dependent latent variable. The maximum likelihood method was used. Two steps of path analysis, the total effect model and research model, were conducted, as shown in Figure 3.

Figure 3. Research model path analysis results.

Note: *** is at the significant level of 0.001; ** is at the significant level of 0.01; * is at the significant level of 0.05. The dotted gray arrow line means a statistically insignificant path.

The total effect model demonstrates the direct relationship between each independent variable, indoor environment quality and indoor greenery, and the dependent variable, i.e., occupant performance. As shown in (a) of Figure 3, both IEQ (β = 0.538, p < 0.001) and IG (β = 0.127, p < 0.01) have a significant positive effect on occupant performance. The model fit indices of the total effect model are within the recommended value: χ²/df = 2.740, GFI = 0.948, AGFI = 0.926, CFI = 0.968 NFI = 0.950, IFI = 0.968, RMR = 0.063, RMSEA = 0.057. Hence, by confirming the total effect, the second step of path analysis can proceed with the mediator of residential satisfaction, as shown in (b), (c), and (d).

Both IEQ (β = 0.701, p < 0.001) and IG (β = 0.117, p < 0.01) have a significant positive effect on residential satisfaction, verifying hypotheses H1 and H3. However, compared to the total effect model, the direct influence of independent variables on dependent variables is reduced. It is worth to noticing that hypothesis H6 is unverified, which indicates that residential satisfaction is a partial mediator. Meanwhile, residential satisfaction (β = 0.405, p < 0.001) has a higher significant positive effect on occupant performance than IEQ (β = 0.255, p < 0.001) and IG (β = 0.081, p < 0.05). Therefore, hypotheses H2, H4, and H5 are verified. The model fit indices of the research model are within the recommended value: χ²/df = 2.181, GFI = 0.948; AGFI = 0.929, CFI = 0.977, NFI = 0.959, IFI = 0.978, RMR = 0.058, RMSEA = 0.047.

To test hypothesis H7, both Group1 (n = 165) and Group2 (n = 154) of WAH with 1–2 and 3–4 months, accordingly, are used based on the research model. The other groups with limited sample sizes are not included. As shown in (c) and (d) in Figure 3, IEQ has a universal positive effect on RS (p < 0.001); however, IG is not a statistically significant variable in these two groups. More specifically, in the model of group 1, RS is a complete mediator on the one hand, which is comprised of the direct effect of IEQ on RS (β = 0.599, p < 0.001) and RS on OP (β = 0.555, p < 0.001). On the other hand, in the model of group 2, IEQ has a positive effect on both RS (β = 0.785, p < 0.001) and OP (β = 0.393, p < 0.01), along with a positive effect of RS on OP (β = 0.321, p < 0.01), which indicates a partial mediator. Furthermore, the model fit indices of the (c) model are as follows: χ²/df = 1.378, GFI = 0.903; AGFI = 0.868, CFI = 0.976, NFI = 0.919, IFI = 0.976, RMR = 0.077, RMSEA = 0.048. In parallel, the model fit indices of the (d) model are as follows: χ²/df = 1.240, GFI = 0.909; AGFI = 0.874, CFI = 0.984, NFI = 0.925, IFI = 0.984, RMR = 0.076, RMSEA = 0.040. Though most of the fit indices are within the recommended value, both their index of AGFI is lower than, but near, the threshold (0.9).

The total final results of the hypotheses test are summarized in Table 4.

Table 4. Hypotheses test results.

Hypothesis	result
H1: IEQ has a positive effect on RS	Verified
H2: IEQ has a positive effect on OP	Verified
H3: IG has a positive effect on RS	Verified
H4: IG has a positive effect on OP	Verified
H5: RS has a positive effect on OP	Verified
H6: RS is a complete mediator	Unverified
H7: Total days of working at home can influence the mechanism.	Verified

6. Discussion

6.1. The residential indoor environment and its occupant

In the last sections, seven hypotheses of the research model have been processed through the statistical analysis that illustrates the results of each hypothetical relationship between indoor environment quality, indoor greenery, residential satisfaction, and occupant performance, along with the effect of the total days of WAH.

Most of the hypotheses have been verified to have significant positive effect. For instance, the result of indoor environment quality can positively influence residential satisfaction and occupant performance, confirming hypotheses H1 and H2, consistent with the previous study in different types of built environments (Zhang and Tu 2021). Another study conducted in Japan also reported a positive relationship between office indoor air quality and occupant satisfaction, however, its results were collected in the specific period of the global pandemic (Umishio et al. 2022). The comparison suggests that this effect commonly exists in various types of indoor built environments. Similarly, indoor greenery, as the factor that is highly related to much of the indoor environment quality sub-factors and occupants’ daily routine, can also positively affect the mediator and dependent factor, confirming the hypotheses H3 and H4. However, what needs to be addressed is that, as shown in Figure 3, indoor environment quality is more influential than indoor greenery, both in the condition of total effect and indirect effect. This result may be caused by the stimulus of indoor environment quality that can be experienced by occupants, which is more comprehensive than indoor greenery (Geng et al. 2019). Nevertheless, a former study has demonstrated a positive relationship between occupant satisfaction and performance (Zhang 2019), which concentrated on the educational workplace (i.e., university library), finding that occupant satisfaction is a complete mediator. A slightly different phenomenon is observed in the results of this study, which verified hypothesis H5 and partially H7 with a focus on the sub-groups. Another study that adopted user experience theory has investigated a virtual indoor environment considering different work tasks, also demonstrating the relationship between satisfaction and performance (Xu and Zhang 2022), which shows that this effect is identified in both physical and virtual indoor environments. In addition, by comparing the total effect model and research model that includes residential satisfaction as a mediator, it has been found that, though a significant direct effect of independent factors on occupant performance still exists, the main effect is realized through residential satisfaction, then affecting occupant performance. Thus, hypothesis H6 is unverified. In other words, a better subjective evaluation of indoor environment quality and indoor greenery can increase the occupants’ perception of residential satisfaction level which can directly improve the performance.

A time-dependent effect has been found that the total days of WAH can influence the mechanism of occupants’ subjective evaluation, which verifies hypothesis H7. This effect should be considered together with the research model in general (i.e., as shown in the (b) of Figure 3), which demonstrates that the longer time spent in a residential indoor environment for performing work tasks, the more influence the indoor greenery has on both residential satisfaction and occupant performance. The direct effect of residential indoor environments on occupant performance also tends to increase. This may cause by the characteristics of the plants that vary much slower through seasons for occupants to observe and also may be due to the higher interaction possibility increased by the exposure time.

From the perspective of the user experience, the residential indoor environment and its occupant can be considered as an integrated system and its user (Wilson 2014). The interface of the residential indoor environment includes but is not limited to indoor environment quality and indoor greenery, which, as expected, have a significant positive effect on the hedonic and pragmatic factor model of UX. Therefore, a post-occupancy evaluation of the systems’ interface and its relationship with occupants’ satisfaction and performance can provide more insight for both researchers and practitioners. It is also important to take into account time-dependent factors that may alter the overall experience of the occupants.

6.2. Limitations and future research

This study investigates the occupants’ subjective evaluation in residential indoor environments, of which some limitations may exist and future research can take further investigation. First, the sample size, though sufficient in the level of statistics and research goal, can be expanded to include more respondents, which is helpful to develop more precise relationship estimates in the structural equation model; however, the current result may still stand, considering the relatively unified condition of evidence from China. More evidence should be gathered in light of different longitude of total days of working at home, especially the occupants who spend more than half a year working in the residential indoor environment. Second, as the common weakness of the questionnaire survey, this study acquires occupant performance through the self-reported items, including concentration, memory, thinking, and overall performance, which may cause inaccuracy in light of subjective perception. Therefore, the next step in future research on the relative topic should concentrate on field study or lab experiments to gain more understanding of the relationship between the residential indoor environment and occupant performance. Third, this study has not concentrated on the occupant behavior that may be more diversified compared to office or educational use indoor environment, for the multiple occupancies of the built environment at home. Another concern also reveals from the findings that occupants may interact with indoor greenery. Hence, future research should also address the occupants’ interaction in changing the “interface” of the residential indoor environment.

6.3. Design implications

Based on the findings of this study, design implications can be drawn into three points. First, indoor environment quality should be considered as the main path to improve satisfaction and occupant performance when compared with indoor greenery, which means, in light of the limited resource of the designer and occupant, regeneration strategy after POE or initial design should focus first on the indoor air quality, noise, thermal comfort, and illumination level, then on the number, type, and color of the plants at home. Second, considering the multiple occupancies of the residential indoor environment that supports work and life, designs should take the function of both uses into account to provide a sufficient home for future possibilities. Third, a development perspective of design should be adopted, because occupants may have different perceptions of the indoor environment when time spent working at home accumulated. This implication reminds practitioners of architectural design to follow up the practice of their client changing working pattern that requires an adjustment in the design process of refurbishment, etc. Therefore, the dynamic condition of the built environment and its occupant requires the practitioners to adopt the POE method in this relatively new field of residence, who may also need pre-occupancy evaluation (Pre-OE) (Tseng and Giau 2022) to identify the specific indoor environment quality and indoor greenery perception on their client’s satisfaction and performance.

7. Conclusion

The conclusion can be drawn as followed: in this study, (1) the residential indoor environment, including indoor environment quality and indoor greenery, can positively affect residential satisfaction and occupant performance; (2) residential satisfaction is a partial mediator and has a positive effect on the occupant performance; and (3) comparingly, the sequence of influence among each factor that significantly affects occupant performance is residential satisfaction, indoor environment quality, indoor greenery; moreover, (4) occupants change their perception of indoor environment quality and indoor greenery when more time is spent in working at home.

Acknowledgements

We sincerely thank Professor Ke Li, Dream Lee, Shuai Fang, Alexandra Lee, and Zheya Cao for their inspirations; we also thank the reviewers for their insightful comments; we appreciate all the respondents in the survey for their time and effort.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Notes on contributors

Leiqing Xu

Leiqing Xu is a Professor at Tongji University and a registered architect. He is specialized in the field of urban studies, environmental psychology and behavior study, and urban generation.

Zhubai Zhang

Zhubai (Mutsing) Zhang is a Ph.D. candidate at Tongji University. He loves black coffee and dark chocolate, no sugar.