A Survey on Nocturnal Air Conditioner Adjustment Behavior and Subjective Sleep Quality in Summer

Abstract

Sleep is a critical physiological process for the mental and physiological restoration of people. The air conditioning usually serves as a common approach to maintain or improve sleep quality. However, available data are still limited regarding the actual sleep quality under different air conditioning modes, which leads to insufficient evidence to support the optimization of the temperature control strategies of air conditioners. To address this gap, an online questionnaire survey was carried out to identify the adjustments of air conditioners during nocturnal sleep, as well as the subjective sleep quality of residents in the summer. A total of 571 valid responses were collected from participants across various age groups, genders, and climatic regions in China through the online surveys that considered several aspects of sleep and air conditioner usage. Pearson’s Chi-square test was used to detect the differences between items in surveys. The results indicated that 74.6% of respondents used air conditioners to regulate their sleep environments in summer, with a preferred temperature of approximately 26 °C. Gender difference had a limited contribution to the adjusting behaviors of air conditioners (χ² = 3.83, p = 0.281), while age played a significant role (χ² = 20.06, p = 0.018). On the contrary, sleep-related adjusting behaviors of the air conditioner were more influenced by subjective factors such as concerns about being awakened by cold or heat. Nonetheless, over 50% of respondents reported experiencing thermal disturbances during sleep, including awakenings by either cold or heat, regardless of the adjustments (χ² = 20.3, p = 0.002). Furthermore, 68.7% of respondents reported their preference for dynamic temperature adjustments during sleep. Findings revealed that the age and subjective aspects were critical for the adjusting behaviors of air conditioners during sleep, and the dynamic air conditioning control was preferred more by users. This study provided empirical evidence to support the optimization of air conditioning modes and the development of adaptive, dynamical sleeping air conditioning systems.

1. Introduction

Humans usually spend approximately one-third of their lives on sleep that plays a critical role in both physical and psychological restoration, as well as cognitive functions [1,2]. However, according to the “2024 China Healthy Sleep White Paper”, 81% of respondents reported suffering from insomnia [3]. Therefore, improving sleep quality has become an increasingly urgent priority. It has been well documented that the indoor thermal environments during sleep significantly affect people’s sleep quality [4,5]. The use of air conditioners could create comfortable thermal environments, thereby effectively enhancing sleep quality and even mental health [6,7]. Previous studies showed that the residential air conditioning cooling demand was particularly high during the summer [8,9], with the highest demand during nocturnal hours, specifically from 23:00 to 09:00 [10]. This indicated that the air conditioner, as a critical tool for the indoor thermal environment regulation during sleep, played an important role in improving the sleep quality of residents in buildings.

Previous studies on indoor sleep thermal environment regulation primarily focused on two main directions, namely the constant temperature and variable temperature environments. As for the constant temperature aspect, it has been proposed that exposure to heat or cold environments would affect sleep stages [11], which sparked increasing interest in thermal neutrality and related studies. Cao et al. [12] carried out an experiment on sleep quality under different thermal conditions, finding that the thermal neutral temperature for sleep was 26 °C during summer. Lan et al. [13] claimed that the air temperature during nighttime sleep maintained at 26 °C could reduce the sleep latency and enhance the duration of the deep sleep (slow wave sleep). Additionally, a field study conducted in Shanghai by Xu et al. identified the thermal neutral temperature for nocturnal sleep in summer as 24.8 °C [14]. However, it has increasingly become evident that the sleep quality is highly sensitive to air temperature fluctuations [13] and that maintaining a constant thermal environment may raise concerns about comfort and health [15,16]. Consequently, the attention shifted to the studies on variable temperature environments for sleep.

There have been quite a few studies on the variable temperature sleeping environments in recent years. Lan et al. [17] devised three nocturnal thermal conditions in a laboratory, including the constant temperature (26 °C) condition, the rising–falling temperature condition (25–26–27–28–27–26 °C), and the falling–rising condition (28–27–26–27–28 °C). The results indicated that subjects experienced prolonged sleep onset time and lower subjective sleep quality under the rising–falling temperature condition, whereas the falling–rising condition delayed the onset of rapid eye movement (REM) sleep, without affecting overall sleep quality significantly. Togo et al. [18] conducted a comparative study to watch the difference in sleep between the constant temperature condition (29.5 °C) and variable temperature condition (29.5–27.5–29.5 °C) in a laboratory. The results revealed that the variable temperature condition reduced the required time to reach minimum core body temperature, and increased slow-wave sleep (SWS) duration while reducing the REM sleep duration, but the overall sleep quality for both conditions showed little difference.

Moreover, another study showed that some major air conditioning brands in China incorporated variable temperature control modes for nighttime sleep [19] in their products. However, a survey involving 554 respondents in Hong Kong indicated that approximately 60% of participants woke up due to thermal discomfort during their sleep, even when the air conditioner was operating [20]. This suggested that current air conditioning systems in actual operation failed to ensure the thermal comfort of people throughout the entire sleep, although the “sleep mode” of air conditioners was incorporated and used.

Based on the literature reviews, it could be concluded that the available data were still insufficient to support the effectiveness of these control modes on the improvement of occupants’ sleep quality, despite the availability of multiple operational modes in air conditioners. In addition, individual requirements for dynamic temperature adjustments during nighttime sleep need further investigation, although there have been a few related laboratory studies that might have lower ecological validity. In order to acquire the empirical data regarding the air conditioner usage in buildings, some surveys on real residents are needed.

For the purpose of addressing these gaps and providing a basis for the development of control strategies for the sleeping environment, behavioral data of air conditioner usages in real sleeping environments in buildings were investigated through online questionnaire surveys to explore occupants’ behavioral adjustments of the air conditioner usage, and to reveal the actual regulation effectiveness of the current temperature control modes of air conditioners, and to assess their impacts on sleep quality. The novelty of this study lies in that the online survey data were from both genders, all age groups, and different buildings in different locations, which implies that the results would have better ecological validity than those from laboratory studies. Findings of this study could provide support for the development and optimization of air conditioning strategies.

2. Methodology

This study adopted a questionnaire-based analytical methodology to explore air conditioning adjustment behaviors and the demand for dynamic temperature regulation in bedroom environments during summer nighttime sleep. The questionnaire consisted of 21 items, the majority of which were structured as multiple-choice questions. Specifically, questions pertaining to air conditioning usage habits and adjustment behaviors upon waking during sleep permitted multiple responses, thereby facilitating a more comprehensive understanding of user preferences and behavioral tendencies.

2.1. Questionnaire Design and Validation

The questionnaire survey was designed in two stages, and the questionnaires were the same in each stage. The first one was carried out on 11 August 2023, while the second one was administered on 16 August. The reason for surveying twice was due to concerns about a high invalid response rate in the first survey.

The specific questionnaires were developed and distributed online, with data collected from participants across diverse regions of China. It was structured in alignment with the research objectives and comprised three primary sections. (1) Background Information: This section collected demographic data of participants, including age, gender, health status, housing type, and residential location. (2) Air Conditioning Usage Behavior and Sleep Characteristics: Participants were surveyed regarding their air conditioning usage patterns, preferred temperature settings, concerns on being awakened by thermal discomfort (i.e., feeling excessively cold or hot) during sleep, actual occurrences of cold- or heat-induced awakenings, and the adjustment measures typically applied in response to such discomforts. (3) Demands for Dynamic Temperature Adjustment: This section explored participants’ expectations on dynamic changes in indoor temperature during sleep, including whether they preferred such changes and their expected temperature variation patterns during the night’s sleep.

The objective of this survey was to obtain the behavioral data of air conditioner usage, rather than obtaining one specific feature with a few different questionnaires. As a result, it did not make sense to calculate Cronbach’s α to test the validity. Instead, the core of the validity of these questionnaires lies in whether the same group of participants could provide stable responses to the same questionnaire at different times. Therefore, the Test–Retest Reliability method would be the most suitable way to test the validity of these questionnaires. We then calculated the Kappa coefficients of survey results between these two stages, finding that all questionnaires had the Kappa coefficients greater than 0.7, which represented a reliable validity of these questionnaires.

2.2. Participants

Before the online survey, the required sample size was calculated with the online sample size calculator [21], considering that this study did not include any hypothesis testing and modeling. During the calculation, the confidence level was set as 95%, and the margin of error was set as 0.05, which has been recommended by similar studies [22,23]. The population proportion was set as 0.5, a recommended level when uncertain. The results indicated that 385 or more surveys were needed to have a confidence level of 95% that the real value would fall within ±5% of the surveyed value. In addition, the expected recovery rate was set as 0.8, while the expected effective response rate was set as 0.9, both recommended as default values. Therefore, a total of 534 questionnaires are needed.

Consequently, a total of 660 questionnaires were distributed twice (on 11 August and 16 August 2023, respectively), and finally, 571 valid responses were collected for both times, which was much larger than the minimum required sample size of 385, which in turn ensured the reliability of the survey results of this study. In addition, the gender distribution of the participants was approximately balanced, with a nearly 1:1 male-to-female ratio. Respondents were instructed to recall their typical air conditioning usage behaviors during their sleep in the summer, and to complete the questionnaire based on their experiences.

A convenience sampling method was adopted due to the exploratory nature of this study [24,25]. The findings of this study were not generalizable to the broader population and should be interpreted as preliminary insights that warrant further investigation with a more representative sample.

Figure 1 presents the age distribution of respondents and the climatic zones of their residential locations, including Hot Summer and Warm Winter regions (HSWW), Hot Summer and Cold Winter regions (HSCW), Cold regions, and Severely Cold regions. The survey encompassed participants from diverse age groups and a wide range of climate zones, thereby ensuring broad representativeness and enhancing the reliability of the collected data. Notably, the majority of respondents fell within the age ranges of 18–30 and 31–50 years, which ensures adequate representation of these key adult demographic groups.

2.3. Data Analysis

SPSS Statistics 27 was used to conduct the statistical analysis of the collected data, with the statistical significance level of 0.05. Due to the large sample size in this study, the Kolmogorov–Smirnov test was applied to examine the normality of data, which produced significance levels of all data below 0.05. It meant that all data failed to conform to the normal distribution, thus a non-parametric analysis method would be used.

Given that the survey data were categorical variables, and the sample size was much larger than 40 with all expected frequencies greater than 5, Pearson’s Chi-square test was used to examine whether significant differences existed among variables (items of surveys) and to analyze correlations between them. As a non-parametric test, the Pearson’s chi-square test is particularly appropriate for analyzing associations among two or more categorical variables, thereby enabling an effective evaluation of distributional differences across groups with varying characteristics in different survey dimensions. The Chi-Square Statistic χ² quantifies the overall difference between the observed (O) and expected (E) frequencies of data. It was calculated in terms of the equation below. A larger χ² indicates a greater difference between observed and expected data, providing stronger evidence against the hypothesis that states the variables are independent.

$${\chi }^2={\displaystyle \sum \frac{{(O_{ij}-E_{ij})}^2}{E_{ij}}}$$

(1)

where O_ij means the observed value, and E_ij means the expected value.

If significant differences were detected, post hoc analysis would be further conducted using the analysis of adjusted residuals method that could eliminate the Type-1 error, as recommended by a few similar studies [26,27]. If the absolute residuals were larger than 1.96, the difference between the two variables was statistically significant. The calculation equation of the adjusted residual is given in the equation below.

$${\mathrm{Adjusted} \mathrm{Residual} (\mathit{Res}}_{adj})={\displaystyle \frac{O_{ij}-E_{ij}}{\sqrt{E_{ij}\cdot (1-p_{i+})\cdot (1-p_{+j})}}}={\displaystyle \frac{s_{ij}}{\sqrt{(1-p_{i+})(1-p_{+j})}}}$$

(2)

where Res_adj means the adjusted residual, O_ij means the observed value, E_ij means the expected value, p_i+ means the marginal proportion of the i-th row, p_+j means the marginal proportion of the j-th column, and the s_ij means the raw residual (O_ij − E_ij).

In addition, to enhance the clarity and visual representation of the findings, Origin 2021 was employed for data visualization, with bar charts as the primary form of figures.

3. Results

3.1. Behaviors of Air Conditioner Usages During Summer Nighttime Sleep

This study examined the primary behaviors that respondents used to regulate their sleeping environments during summer nighttime sleep. Considering the extended duration of the summer season, the impact of outdoor climate on indoor thermal conditions, and the likelihood that individuals may utilize multiple regulation methods during sleep, the questionnaire included multiple selections to capture the full range of participants’ preferred sleeping environment adjustment approaches.

As shown in Figure 2, only 3.3% of respondents did not use any measures to regulate the thermal environments during summer nighttime sleep. Among those who did, 74.6% reported using air conditioners. These results suggested that the majority of individuals used air conditioners or fans to regulate the sleeping environments during summer nights, with air conditioning being the most prevalent method. Consequently, the subsequent analysis in this study focused on the subgroup of air conditioning users to further investigate their adjustment behaviors and sleep quality during summer nighttime sleep.

3.1.1. Modes of Air Conditioners and Temperature Setting Behavior

Firstly, the usage of different air conditioning modes during summer nighttime sleep was analyzed, as shown in Figure 3. These modes include the following: timing mode (TM), keeping the air conditioner on (AC ON), turning off before sleep and turning on when woken up by heat (OFF-ON), and turning on before sleep and turning off when woken up by cold (ON-OFF). According to the data, among respondents who used air conditioning during summer nights, 41.5% used the “TM”, which accounted for the highest proportion. This was followed by the “AC ON” mode, accounting for 29.8%. The “OFF-ON” and “ON-OFF” modes were less frequently used. Notably, the “ON-OFF” mode was the least adopted. Furthermore, when the “AC ON” mode was categorized into sleep mode and fixed-temperature operation, the results revealed that 74.8% of respondents used the fixed-temperature operation during summer nighttime sleep, whereas 25.2% used the sleep mode.

As shown in Figure 4, 28.2% of respondents set the air conditioner temperature to 26 °C during summer nighttime sleep, with 11.5% setting it to 25 °C. These findings suggested that respondents generally preferred to set the air conditioner temperature close to the thermally neutral one during summer nighttime sleep.

3.1.2. Gender Differences in Air Conditioning Usage Behavior

Given that physiological and psychological differences between genders may influence air conditioning usage behaviors, a comparative analysis of air conditioning usage during summer nighttime sleep across genders was carried out. However, statistical analysis results of the Chi-square test showed that the differences in air conditioner usage behaviors between genders had no significant difference (χ² = 3.83, p = 0.281).

As shown in Figure 5, among male respondents, 36.5% selected the “TM” (Res_adj = −1.9), 31.8% reported “AC ON” (Res_adj = 0.8), 22.9% preferred “OFF-ON” (Res_adj = 1.0), and only 8.9% chose “ON-OFF” (Res_adj = 0.6).

Furthermore, female respondents showed comparable patterns in their selections of air conditioning modes relative to male respondents. The results showed that 45.7% of the females selected “TM” (Res_adj = 1.9), 28.2% reported “AC ON” (Res_adj = −0.8), 18.8% preferred “OFF-ON” (Res_adj = −1.0), and only 7.3% chose “ON-OFF” (Res_adj = −0.6).

Post hoc analysis indicated that the absolute adjusted residuals (Res_adj) of all these percentages were less than 1.96, which meant that no significant differences in each air conditioner usage behavior were detected between genders.

3.1.3. Age Differences in Air Conditioning Usage Behavior

The air conditioning usage behaviors across different age groups during summer nighttime sleep are shown in Figure 6. The Chi-square test results showed that age produced significant differences in air conditioner usage behaviors (χ² = 20.06, p = 0.018).

Post hoc analysis detected these specific differences. Among respondents under the age of 18, 63.6% opted for “AC ON” (**Res_adj = 3.6), while 31.8% selected the “TM” (Res_adj = −1.0). For the 18–30 age group, 44.8% chose the “TM” (Res_adj = 1.8) compared to 27.0% who preferred “AC ON” (Res_adj = −1.6). Among respondents aged 31–50, 38.1% selected the “TM” (Res_adj = −0.9), whereas 30.1% preferred “AC ON” (Res_adj = 0.1) and 23.0% preferred “OFF-ON” (Res_adj = 0.7). For individuals over 50 years old, 38.1% selected the “OFF-ON” mode (*Res_adj = 2.0), while 28.6% opted for the “TM” (Res_adj = −1.2), and another 28.6% preferred “AC ON” (Res_adj = −0.1).

These post hoc analysis results above indicated that participants under 18 had a significantly higher percentage preferring to keep the air conditioner on continuously than expected (Res_adj > 2.58), and participants over 50 had a significantly higher percentage preferring the mode “OFF-ON” than expected (Res_adj > 1.96). Other age groups had no significant differences among the air conditioner usage behaviors (Res_adj < 1.96).

Additionally, the figure illustrates that the proportion of respondents who selected the “OFF-ON” mode gradually decreased with decreasing age. This observation suggests that younger respondents tended to proactively use air conditioners during the falling-asleep stage to create a more thermally comfortable sleeping environment.

3.2. Subjective Sleeping Concerns During Summer Nighttime Sleep

As presented in Section 3.1, various air conditioning modes were adopted by respondents during summer nighttime sleep. However, these behaviors are not only influenced by environmental factors but are also closely linked to individual subjective concerns. Therefore, subjective sleep quality and psychological responses under different air conditioning modes were investigated.

3.2.1. Subjective Sleep Concerns Under Different Air Conditioning Modes

The relations between the subjective concerns of participants and air conditioner usage behaviors were analyzed in this section. The Chi-square test results showed that the subjective concerns varied significantly among air conditioner usage behaviors (χ² = 20.32, p = 0.002).

Post hoc analysis was also conducted to detect the specific differences. As shown in Figure 7, the level of concern about being awakened due to cold or heat varied significantly among respondents, depending on the air conditioning modes used. Those who adopted the “OFF-ON” mode had the highest level of sleeping concerns, with 21.6% reporting no concern (**Res_adj = −3.1), while 42.0% worried about being awakened by cold (*Res_adj = 2.4) and 36.4% showed concerns about being awakened by the heat (Res_adj = 0.8). Judging from the adjusted residuals, this air conditioner mode was associated with a significantly higher percentage of participants with concerns about being awakened by the cold than expected, while a significantly lower percentage of those with no concern than expected.

Moreover, respondents adopting the “AC ON” mode reported the lowest level of concern, with 48.0% indicating no concerns (**Res_adj = 3.5), and 24.4% reported concerns about being awakened by the heat (*Res_adj = −2.4), while 27.6% worried about being awakened by the cold (Res_adj = −1.1). Judging from the adjusted residuals, this air conditioner mode was associated with the significantly highest percentage of participants with no concern than expected, as well as a significantly lower percentage of participants with concerns about being awakened by the heat than expected.

Furthermore, looking into the two subcategories of the Mode “AC ON”, namely the constant temperature mode and the sleep mode, some different results were obtained. The Chi-square test results showed that the subjective concern levels had no significant difference between these two modes (χ² = 2.7, p = 0.259). Conclusively, the selection of air conditioning modes had notable influences on users’ psychological aspects during summer nighttime sleep. Judging from the results above, the sleep mode under continuous operation appeared to be comparatively more effective in reducing sleep concerns related to disturbances by thermal discomfort.

3.2.2. Actual Sleep Disturbances Under Different Air Condition Modes

The actual sleep disturbances reported by participants under these four air condition modes were analyzed in this study. The Chi-square test results showed that the actual sleep disturbances varied significantly among these air conditioner modes (χ² = 39.21, p < 0.001). Post hoc analysis was also conducted to detect the specific differences.

As illustrated in Figure 8, respondents adopting the “OFF-ON” mode had the worst sleep quality, with only 21.6% reporting uninterrupted sleep without cold- or heat-induced awakenings, which was significantly lower than expected (*Res_adj = −1.98). And 42.0% of participants reported having been awakened by heat during their sleep, which was significantly higher than expected (**Res_adj = 3.3).

Moreover, under the “timing” mode, 33.9% of participants reported having been awakened by the heat, which was significantly higher than expected (*Res_adj = 2.3), while only 23.7% reported no sleep disturbances due to heat or cold, which was significantly lower than expected (*Res_adj = −2.3).

In addition, the post hoc results also showed that significantly more participants reported having been awakened by the cold (**Res_adj = 2.6), while significantly fewer participants reported having been awakened by the heat (**Res_adj = −2.6), when they selected the “ON-OFF” mode.

In addition, under the “AC ON” mode, 42.5% of respondents reported no experience of disturbance by thermal discomfort, which was significantly higher than expected (**Res_adj = 3.7). And only 15.0% percent of participants reported having been awakened by the heat, which was significantly lower than expected (**Res_adj = −3.9).

Furthermore, sleep disturbances under the two subcategories of the “AC ON” mode, namely the constant temperature mode and the sleep mode, were compared. The Chi-square test results showed that the sleep disturbances had no significant difference between these two modes (χ² = 5.67, p = 0.129).

Additionally, post hoc results also showed that more participants were awakened by heat (Res_adj = 0.8) or cold (Res_adj = −1.7) than expected under the sleep mode, although the increase was not statistically significant. This finding suggests that the current settings of air conditioner sleep mode remain suboptimal and do not yet fully satisfy users’ demands for thermal comfort and stability throughout the nighttime sleep period.

3.2.3. Gender and Age Differences in Subjective Sleep Concerns

Differences in sleep concerns due to gender and age were analyzed in this study. The Chi-square test results showed that the sleep concerns had no significant correlations with gender (χ² = 1.3, p = 0.521). However, the results of the Chi-square test showed that the sleep concerns varied significantly among age groups (χ² = 29.1, p < 0.001).

Post hoc analysis results are provided in Figure 9. Only 4.5% of respondents under 18 expressed concerns about being woken up by the cold (**Res_adj = −2.8), which was significantly lower than expected. And 72.7% of these participants reported no concerns (**Res_adj = 3.7), which was significantly higher than expected. For those aged 18–30, 28.9% of them were concerned about being awakened by the cold (*Res_adj = −2.3), which was significantly lower than expected. For the 31–50 age group, 43.4% of them expressed concern about being awakened by the hot (**Res_adj = 2.8), which was significantly higher than expected. However, only 22.1% had no sleep concerns (**Res_adj = −3.5), which was significantly lower than expected. For respondents over 50, there were no significant differences between the reported and the expected results. In total, 14.3% expressed concerns about being awakened by the cold (Res_adj = −1.7), 38.1% were worried about being woken up by the heat (Res_adj = 0.5), and 47.6% of them reported no sleep concerns (Res_adj = 1.2).

These results indicate that participants under 18 reported the lowest levels of concern about heat or cold during sleep, whereas the 31–50 age group reported the highest levels of concern, particularly with respect to heat-induced awakenings. For the 18–30 age group, the percentage of those who worried about being awakened by the heat was overestimated.

3.2.4. Age and Gender Difference in Actual Sleep Disturbances

Differences in actual sleep disturbances due to gender and age were analyzed in this study. The Chi-square test results showed that sleep disturbances had no significant correlations with gender (χ² = 1.413, p = 0.702). However, the results of the Chi-square test showed that the sleep disturbances varied significantly among age groups (χ² = 20.087, p = 0.017).

Post hoc analysis results are presented in Figure 10; for respondents under 18, 9.1% reported having been awakened by cold, 18.2% by heat, 13.6% by either cold or heat, while 59.1% reported no disturbances during sleep, which was significantly higher than expected (**Res_adj = 3.1). For the 18–30 age group, 28.9% reported being awakened by either heat or cold, which was significantly higher than expected (*Res_adj = 2.1). But the percentages of other kinds of sleep disturbances in this age group were only slightly less than expected. For those aged 31–50, 24.8% reported being awakened by cold (**Res_adj = 2.7), which was significantly higher than expected. In addition, 31.9% of this group reported having been awakened by heat, and 20.4% by either cold or heat. Only 23.0% reported no disturbances, which was slightly lower than expected, indicating that this group was the most affected by nocturnal thermal disturbances. For respondents over 50, no significant differences were observed in sleep disturbances between the reported results and the expected results.

Conclusively, respondents under 18 exhibited both the lowest levels of subjective sleep concerns and the actual sleep disturbances during sleep. In contrast, individuals aged 31–50 demonstrated the highest sensitivity to nocturnal thermal conditions, both in terms of subjective concerns and actual awakenings, suggesting that their sleep quality was most adversely impacted by the thermal environment.

3.3. Dynamic Temperature Needs During Summer Nighttime Sleep

The results above revealed that respondents expressed concerns about, and experienced actual awakenings due to, excessive cold or heat during nighttime sleep in summer. These phenomena also varied across age groups. To further investigate individual-specific thermal environment needs during sleep, the questionnaire included items addressing participants’ preferences for temperature adjustments in sleeping environments. The results are presented in Figure 11.

As illustrated, 68.7% of respondents expressed a preference for dynamic temperature variations during sleep, while only 31.3% preferred the temperature to be constant. This finding indicated a clear demand for dynamic thermal environments among the majority of respondents. However, as shown in the results in these former sections, currently available sleep modes of air conditioners have not yet been able to fully meet users’ requirements for their nighttime sleep in summer.

Therefore, when developing air conditioner control strategies for sleep, more consideration should be given to users’ dynamic temperature preferences in order to improve sleep quality, thus enhancing sleep health.

4. Discussion

4.1. Sleep Mode and Dynamic Temperature Adjustment

Based on the survey results of this study, the majority of participants demonstrated a clear preference for dynamic temperatures during summer nighttime sleep. Supporting evidence from prior research also indicates that prolonged exposure to steady-state thermal environments compromises individuals’ adaptability to environmental changes. For instance, individuals who stayed in consistently heated environments during winter tended to have increased sensitivity to cold [28]. Similarly, those who relied intensively on air conditioners for cooling in summer often showed attenuated tolerance for naturally ventilated conditions and diminished heat adaptability [29].

The findings from this study suggest that steady-state thermal environments created by air conditioners were suboptimal for promoting human sleep quality and sleep health. In contrast, dynamic temperature fluctuations might be a better choice because they can enhance occupants’ acceptance of warmer thermal conditions, improve perceived thermal comfort, and strengthen physiological adaptability [30]. This adaptive mechanism could improve subjective thermal satisfaction and enable the use of higher setting temperatures in air-conditioning system design. This, in turn, would contribute to energy savings in buildings [31,32].

Unfortunately, as reported in Section 3.2.1 and Section 3.2.2, both sleep concerns and sleep disturbances showed little differences between the constant temperature mode and the sleep mode. This might indicate the inadequacies in the actual capacity of the sleep mode of air conditioners. Most commercial air conditioners usually incorporate the sleep mode. However, it seems that no actual contributions to sleep quality are made by the sleep mode of air conditioners. The reason could be attributed to the fact that an effective and reliable “temperature variation model” for air conditioners during sleep has not yet been fully developed. More solid theoretical foundations and empirical support from users are needed to produce the dynamic temperature setting models. And more factors have to be accounted for, e.g., as reported in Section 3.2 of this study, age played a significant role in sleep concerns, sleep disturbances, and air conditioning modes. Therefore, this study provides direct references for the development and improvement of sleep modes in air conditioners.

4.2. Timing Mode and Subjective Sleep Quality

This survey also confirmed that the timing mode was the most commonly used air conditioning control strategy during summer nighttime sleep. The primary reasons for selecting this mode included respondents’ lack of confidence in the current temperature control models of the sleep mode and concerns that these modes might fail to meet nighttime thermal requirements for sleep. In addition, the timing mode requires users to preset the air conditioner’s on/off schedule prior to sleep, which depends on accurately predicting the duration of sleep. However, due to variations in individual circadian rhythms and daytime workloads, actual sleep durations often fluctuated, making it difficult for users to determine the optimal activation and deactivation times. This misalignment may result in nighttime thermal discomfort or unnecessary energy consumption [19]. And this may also be the critical reason for the actual sleep disturbances reported in Section 3.2.2 of this study.

Further analysis of the survey data revealed that respondents using the timing mode reported higher subjective sleep concern on heat-related awakenings and experienced more frequent disruptions due to heat. These findings underscore the limitations of the timing mode in regulating nocturnal thermal environments and highlight the need for further optimization of existing sleep modes with respect to temperature control and thermal comfort maintenance.

4.3. Air Conditioning Modes and Gender, Age Differences

The results of this study indicate no statistically significant differences based on gender in the selection of air conditioning modes during summer nighttime sleep. However, existing research suggests that males and females exhibit distinct thermal preferences during sleep. Specifically, females are regarded as having higher mean skin temperatures and therefore prefer warmer sleeping environments than males [33,34]. The contradictory result could only be explained by the methods used to achieve these results. Studies claiming the gender differences were usually from the lab research scenarios, while our results were from the real living and sleeping environments, and thus had a higher ecological validity. Consequently, results regarding the gender differences in this survey could provide references for studies on sleep mode improvements.

Furthermore, age represented a statistically significant factor that affected both thermal perception and sleep behaviors. Prior studies had demonstrated that aging was associated with a decline in basal metabolic rate [35], reduced thermal sensitivity [36], and weakened thermoregulatory capacity [37]. These changes also led to altered thermal thresholds and narrower preferred temperature ranges [38,39]. These physiological changes were reflected in sleeping characteristics such as delayed sleep onset, altered sleep inertia, and fundamental differences in sleep architecture relative to younger individuals [40].

This study identified significant age-related differences in air conditioning usage behaviors and subjective sleep concerns in Section 3.2.3, aspects that are seldom considered in the design of air conditioning modes. The reason for the age differences could be attributed to variations in lifestyle patterns and psychological stress levels. Individuals aged 31–50 are typically subjected to dual pressures from career development and family responsibilities, which may enhance their awareness of sleep quality and increase their sensitivity to potential sleep disturbances [41,42].

Based on these findings, it is recommended that gender- and age-specific thermal comfort characteristics be integrated into the design and optimization of air conditioning systems. The implementation of such personalized and adaptive control strategies can improve nighttime sleep quality and, in return, contribute to greater energy efficiency.

4.4. Limitations and Future Remarks

Several limitations existed in this study. Firstly, our data were derived from retrospective self-reports without synchronized environmental or physiological measurements, which made it impossible to conduct a direct comparison with established thermal comfort models such as the PMV/PPD model that usually requires detailed inputs of specific environmental or physiological variables. In addition, although our findings indicated that sleep mode did not outperform constant temperature, the data did not allow us to determine the underlying reasons. The underperformance could reflect suboptimal default algorithms, mismatched parameters for local climates or bedding, or user misunderstanding and override of the intended profiles. Without concurrent device logs and microclimate measurements, these possibilities cannot be distinguished. Moreover, while age-related differences in air conditioner usage and thermal awakenings were observed, our questionnaire did not include covariates such as health status, hormonal levels, BMI, or long-term acclimatization. As a result, the role of chronological aging cannot be separated from potential confounders, and causal interpretation is limited. Finally, other environmental variables, such as the relative humidity, acoustic levels, and the CO₂ concentrations, were not included in this survey, which might also attenuate the reliability of our findings.

Future research is supposed to combine large-scale surveys with concurrent in-house monitoring, device data, and simple physiological measures to enable model-based validation, explain behavioral differences more robustly, and refine algorithms for personalized, adaptive sleep cooling strategies.

5. Conclusions

This study used a questionnaire-based approach to investigate the air conditioner usage behaviors and preferences for dynamic temperature variation in air-conditioned environments during summer nighttime sleep. The results indicated that 74.6% of respondents used the air conditioner to manage the thermal environment during sleep, with an average preferred comfort temperature of approximately 26 °C. Although 41.5% of respondents reported employing the timer mode, this strategy was found to be insufficient in meeting thermal comfort expectations, both from a sleep concern and actual sleep disturbance perspective. In contrast, continuous air conditioner operation mode and the sleep mode led to a lower percentage of sleep concerns and disturbances.

Additionally, no significant differences were observed between male and female respondents in the selection or usage of air-conditioning modes. Age-related differences in air conditioning behavior were significant. Respondents under 18 preferred to use continuous air conditioner operation mode throughout the night, whereas those over 50 tended to switch off the air conditioner before sleep and switch it back on upon waking due to heat. Individuals aged 31–50 demonstrated the highest sensitivity to nighttime thermal environments, reporting increased rates of sleep disruption and heightened concerns regarding temperature-related disturbances.

Furthermore, sleep-related air conditioning behaviors varied among individuals due to a range of interacting factors. Nevertheless, across all groups, more than 50% of participants reported experiencing awakenings caused by cold or heat during actual sleep, suggesting that current air conditioning modes are insufficient in fulfilling users’ thermal comfort requirements. Moreover, 68.7% of respondents expressed a preference for dynamic temperature variation during sleep, indicating a clear and widespread demand for adaptive thermal regulation strategies.

This study highlighted real-world air conditioning adjustment behaviors and the demand for dynamic temperature control, offering implications for the optimization of modes of the air conditioning systems.