Subjective and Objective Measurement of Indoor Environmental Quality and Occupant Comfort in a Multinational Graduate Student Office

Abstract

In an air-conditioned multinational graduate students’ office in Japan during the winter season, we examined indoor environmental conditions, occupants’ perceptions, and their acceptance levels over five consecutive days. Indoor air quality (IAQ) acceptance peaked on the third day, coinciding with the most favourable thermal sensation vote, which was “neutral” at a geometric mean indoor temperature of 25.1 °C. Aural comfort received the lowest acceptance due to ongoing construction work, but did not significantly impact overall IEQ acceptance, thus suggesting that unacceptable aspects of indoor environmental quality (IEQ) can be offset by acceptable aspects. IAQ and thermal comfort compensated for its effects, offering insights into occupants’ environmental tolerance. IAQ sensation votes and visual comfort votes exhibit a strong relationship with overall comfort, as indicated by their respective R² values. However, variations in overall comfort are primarily explained by IAQ, which has the highest R² value of 0.50, suggesting that IAQ accounts for 50% of the changes in overall occupant comfort. Non-Japanese participants had lower IEQ acceptance and a significantly higher number of complaints than Japanese participants more so in visual comfort where acceptable luminance levels were higher in Japan than other participants’ countries of origin. Thermal comfort was mutually highly accepted by both groups. Nose and eye irritation were significantly experienced by the international participants due to low RH levels but experiencing loss of concentration and lethargy was comparable in both groups (p > 0.05, t-test). We recommend global coherence in indoor environmental quality standards as is the case with drinking water standards for public health protection and seamless transitions in new indoor environments.

1. Introduction

Preventing exposure to highly contagious airborne pathogens, limiting the circulation of harmful chemical compounds, and enhancing occupant comfort in indoor environments remain critical in a world where people are increasingly spending prolonged periods indoors [1,2,3]. Indoor Carbon IV Oxide (CO₂) concentration has been widely used [4,5,6,7,8,9,10,11,12,13,14,15,16,17] as a primary reference for indoor ventilation efficiency and airborne transmission risk in shared spaces. Exhaled CO₂ serves as a reliable indicator since, in the absence of other sources, human occupants are the primary contributors of bio-effluents, and indoor CO₂ is a byproduct of metabolic processes [18,19,20]. Consequently, various studies have utilized CO₂ to assess infection risk, physiological responses, and occupant comfort in settings such as classrooms, buses, hospitals, trains, and offices [4,6,7,12,16,20,21,22,23,24,25,26,27,28,29,30,31].

While CO₂ concentration reflects ventilation levels and, by extension, indoor air quality (IAQ), other factors such as thermal comfort, aural comfort, and visual comfort significantly influence the overall indoor environmental status and user experience in buildings. To enhance occupant well-being and productivity, indoor environmental quality (IEQ) should be continuously studied and improved [32,33,34]. Nagano and Horikoshi [35] emphasized the importance of integrating these different attributes to evaluate overall IEQ and clarify relationships for targeted interventions aimed at improving occupant comfort. Korsavi et al. [36,37] found that sensation votes have a greater impact on overall comfort than physical parameters when assessing occupant experiences.

In this study, we examined ethnically diverse occupant responses and heart rate variability to various indoor environmental factors, including thermal comfort, IAQ, sound pressure levels, and horizontal illumination levels, in an air-conditioned office during the winter season. Simultaneously, we measured indoor and outdoor environmental conditions using sensors. This dual approach provides both physical and perceptual data that can inform building design and operational strategies for creating people- and planet-centered built environments.

The data can also support the development of machine learning models for predicting occupant satisfaction. Furthermore, it can help prioritize key IEQ attributes in building design and management, guiding the implementation of strategies that enhance user comfort and performance.

1.1. Literature Review and Research Question Formulation

Perceptual quality assessment of the indoor environment has been extensively explored in numerous studies focusing on classrooms, hospitals and offices in various parts of the world [3,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70]. Other studies have conducted assessments of the four core IEQ variables, visual comfort, thermal comfort, aural comfort and indoor air quality in climate chambers, and findings generalized to shared enclosed spaces in the real world [67]. Some studies [33,41,42,43,44,45,46,47,48,63,64,65,67,68] have conducted separate or bivariate [35] or collective [32,36,37,49,50,51,52,53,54,55,56,57,58,59,60,61,62,64,66,70] analysis of the influence of the four IAQ basic components on occupant comfort or IEQ acceptance and/or learning performance.

Occupant comfort and IEQ acceptance are subjective and are influenced by various environmental and physiological parameters [68]. While these studies have provided invaluable insights into occupant comfort and well-being, to the best of our understanding, none have examined tolerance and IEQ acceptance collectively in a real-world air-conditioned space with continuously varying levels of occupancy and associated environmental data with subjective responses (both physiological–psychological perceptions) in an ethnically diverse indoor population.

Our study examines how people perceive, accept, and tolerate an indoor environment that is conditioned differently from the standards of their country of origin. The study participants were primarily under 35 years old from seven different nationalities. Therefore, in this study, we first describe the indoor environment conditions and analyze what variables influence occupants’ IEQ acceptance in such a setting and the relationship between the variables. Occupant comfort and acceptance of IEQ were analyzed based on responses to four key indoor environmental variables: thermal comfort, horizontal illumination level, IAQ, and sound pressure level in decibels A [dBA]. We also compare the perceptions, acceptance and tolerance levels of the natives to the same indoor conditions and overall occupant acceptance/comfort and examine whether occupants from different nationalities have acclimatized to the local environment and have any significant differences in subjective evaluation and acceptance of IEQ compared to the native occupants.

1.2. Research Gaps

The world is increasingly embracing social integration as many people migrate to new places for study or work and/or residency due to various reasons including conflicts and climate change [71,72]. To foster synergy in workspaces besides promoting social cohesion and harmony, it is important to create conducive environments that enhance worker’s productivity and well-being [73,74,75]. People adapt differently to new environments, and this can influence their health as well as social–economic outputs. Numerous studies have examined social integration of newcomers’ well-being into various communities globally [72,76,77].

Additionally, study comparisons between objective and subjective analyses reveal that objective assessment underestimates IEQ with respect to perceptual evaluation of IEQ [69]. Therefore, considering the complexity of IEQ dynamics, we carried out this study in a real office scenario and included monitoring of the subject’s heart rate variability (HRV) to obtain subjective physiological responses alongside psychological responses to reduce as much as possible the difference between subjective and objective analyses of IEQ. The heart rate variability parameter can provide an important link for the physiological–psychological response of occupants to the indoor environment. It can highlight any psychological and physiological discordancy. A greater understanding of underlying human neurological and biological mechanisms of perception to changes in indoor environmental quality is needed and is important to explore [70].

Therefore, this study builds upon aforesaid studies combining sensor-based IEQ data, heart rate variability and repeated subjective assessments of a shared indoor workplace environment with a varying diverse occupancy.

1.3. Policy Implication

Different countries have different acceptable limits or standards for the four key components of IEQ (

Table 1. Acceptable IEQ standards in participants’ country of origin.

Country	Indoor Air Quality Acceptable Levels	Thermal Comfort Acceptable Levels	Visual Comfort Acceptable Levels	Aural Comfort Acceptable Levels
China	CO2 levels < 1000 ppm relative humidity (RH) 30–60% in winter	20 °C to 26 °C	≥300–500 lux.	<85 dBA
Egypt	700~1000 ppm above outdoor levels	22 °C to 27 °C	>500 lux	45–65 decibels (dBA)
Japan	<1000 ppm (RH) 40–70%	18–28 °C	750 to 1000 lux	55 decibels (dB)
Kenya	<1000 ppm *	-	≥100 lux	45 decibels (dB)
Philippines	<1000 ppm * 40–60% RH	23–26 °C *	300–500 lux	90 dBA
South Korea	<1000 ppm (RH) 40–60%	<20 °C	≥500 lux	<90 dBA
Vietnam	<1000 ppm * RH-40–60% *	20 °C and 26 °C *	300–500 lux	70 dBA

* Policy documents refer to WHO guidelines or national and international professional bodies.

). Our study examines whether these differences affect occupant tolerance in new indoor environments and possible policy recommendations to support seamless transitions especially since humans, more so adults, adapt to new environments differently [78,79]. Table 1 presents a summary of the IEQ standards for countries of origin for study participants. China, Japan, Philippines, South Korea and Vietnam have IAQ standards [80,81,82,83,84]. Where there is a lack of national IAQ guidelines or standards on some of the IEQ components, some of the countries such as Kenya [85,86], Vietnam and Philippines refer to WHO guidelines in their policy documents or national and international professional bodies such as the American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE). All the student participants were between 21 and 33 years old at the time of the study and faculty members were below 51 years.

Studies examining the subjective response and acceptance of measured indoor environment conditions can demonstrate coherence regarding the prioritization of these variables in building design and development of country-specific green building rating tools while considering geographical differences, climatic changes and social integration.

In recognition of the importance of indoor environmental quality on occupant well-being, the built environment worldwide is increasingly adopting green building standards. Green buildings prioritize occupant well-being and environmental consciousness. IEQ is accorded various points/weightings in different green building rating tools [67]. According to Wei et al. [87], the average apportionment of credits attributed to the four core aspects of IEQ among different green building rating tools was 27%, 22%, 17% and 34% for thermal comfort, acoustic comfort, visual comfort and IAQ, respectively [67,87]. The Leadership in Energy and Environmental Design (LEED) rating tool gives more credits to IAQ (47%) and visual comfort (35%) [67,87]. The International Finance Corporation Excellence in Design for Greater Efficiencies (IFC EDGE) tool widely promoted in developing countries focuses solely on energy, water and material efficiency [88], not IEQ. Some green building rating tools such as LEED, Building Research Establishment Environmental Assessment Method (BREEAM) and Green Star Australia have been adopted widely within and beyond their countries of origin. Green Star Australia rating gives IEQ a weighting of 12%. It has been adapted in South Africa as Green Star South Africa as well as in Kenya and Nigeria. Green Star South Africa gives IEQ a category weighting of (15%) [89]. Our study findings can contribute to the ongoing discourse for the selection of parameters that can characterize IEQ as countries improve existing green building tools/schemes and others begin to adapt and customize established green building rating tools from other countries. Furthermore, in a world where climate analogue analysis [90] is gaining tract in serving as a complement to climate impact projections, our study can stimulate policy dialogue in that regard but for indoor environments.

2. Materials and Methods

The study design involved the collection and analysis of objective data about environmental conditions, including indoor and outdoor CO₂ concentration, temperature, relative humidity, luminance and sound levels using sensors (Appendix A). Occupants’ subjective data on physiological (heart rate variability) and psychological (perceptual) data were collected using a Fitbit Sense 2 watch (Appendix A) and questionnaire (Appendix B), respectively.

2.1. Objective Measurement: Study Set-Up Measures of CO₂, Temperature and Relative Humidity

To provide a detailed understanding of the prevailing environmental conditions, indoor and outdoor CO₂ concentration (ppm), temperature (°C), and relative humidity (%) were measured simultaneously using sensors installed at 10 locations, as illustrated in Figure 1a,b. Nine sensors were placed indoors, while one (sensor 22-07) was positioned outdoors. Sensor 22-10 was mounted on the ceiling in the center of the room (Figure 1).

The study utilized the CO₂, Temperature, and Humidity Data Logger RTR-576-S, paired with a Wireless Base Station RTR500BC (USB communication-type, RTR500B series parent unit) (Appendix A). CO₂ concentrations were recorded every second for 24 h over 5 consecutive days in January 2023. A sound level meter (SL1373SD) was positioned at a height of 1.2 m in the center of the room, next to sensor 22-04 (RTR-576-S), and recorded indoor sound pressure levels every two seconds throughout the study period (Appendix A). A lux meter AP-881E (Appendix A) was used to measure indoor and outdoor lighting levels at each occupant’s desk.

Before all the measurements, the IEQ instruments were calibrated according to the manufacturer’s instructions. Their accuracies and measurement ranges are detailed in

Table A1. Accuracy of physical environment monitoring instruments.

Parameter	Instrument	Measuring Range	Accuracy
CO2 concentration [ppm]	RTR-576-S CO2, Temperature, and humidity sensor	0~9999 ppm	±50 ppm
Air Temperature		25~70 °C	±0.3 °C
Relative humidity		0~99%RH (at −20~70 °C)	±2.5%RH (at 15 °C~35 °C, 30~80%RH)
Heart rate	Fitbit wristwatch Sense 2	30–220 bpm	-
Lux meter	AP-881E	0.1 Lux–300,000 Lux	20,000 LUX/2000 FC: ±4% rdg ± 0.5% f. s >20,000 LUX/2000 FC: ±5% rdg ± 10 dgt
Sound level	SL1373SD	30–130 dB	±1.4 dB (at 1 kHz and 23 ± 5 °C)

of Appendix A.

As shown in Figure 1a, eight sensors (22-01, 22-02, 22-03, 22-04, 22-06, 22-08, and 22-09) were placed on desks at a height of 1.2 m above the floor to measure CO₂ concentration, temperature, and relative humidity within the vicinity of seated occupants. The sensors were positioned 68 cm away from the occupants to minimize direct exhalation effects. Sensor 22-05 was placed in an area containing a bookshelf and a microwave, positioned 1.6 m above the floor to measure CO₂ concentration, temperature, and relative humidity at standing height, as occupants primarily used this area while standing (Figure 1a). These sensor placements [22,30] ensured that the measurements captured were representative environmental conditions throughout the office, providing a comprehensive overview of air quality within the study area.

During the study period, occupancy patterns varied, as participants entered and exited the graduate students’ office based on their individual schedules [38]. Consequently, ∆CO₂ concentrations at different locations were assessed using sensor records, as shown in Figure 1a,b. Indoor occupancy density, personal and environmental adaptive behaviors were observed using a high-resolution video monitoring camera.

Statistical significance of the mean differences of daily readings of indoor conditions was analyzed through Analysis of Variance (ANOVA). In addition, a Pearson correlation coefficient was computed to determine the relationship between occupancy levels and indoor CO₂ concentrations. To analyze differences between nationality groups, a two-sample t-test was carried out [84].

2.2. Subjective Measurement Assessment of Occupant Comfort

A self-reported questionnaire was administered to the occupants via the SurveyMonkey platform to assess their comfort levels. Twenty participants from different nationalities were physically present in the student office during the study period. At the time of the study, all students in the office were aged between 21 and 33 years old and faculty members aged below 51 years. In total, 70% of the participants present indoors were male and 30% were females. Participants’ countries of origin were Vietnam, South Korea, the Philippines, Kenya, Japan, Egypt, and China. With Japan being the host country of the learning institution, non-Japanese occupants in the study were categorized as “international participants”. For purposes of analyzing occupants’ comfort and tolerance to the environment, one student of mixed Korean and Japanese descent but born and raised in Japan was categorized as Japanese. Overall, 40% of the participants were categorized as ‘international participants’ and 60% were thus categorized as ‘Japanese participants’. Two new international students joined the graduate student’s lab/office 3 months before the study, while other participants (Japanese and international) had been using the same graduate student office for at least 6 months [38]. As per the video monitoring data, the participants spent between 8 and 58 h per week in the graduate students’ office. Eighty-two percent (82%) spent more than 27 h per week in the office [38].

The questionnaire (

Table A2. Study’s subjective evaluation questions.

Sub-Section			Questions
General Information	Age		Are you a smoker?    Yes    NO
	During a regular week, how many hours do you spend in this students’ office/lab?
	<16 h		16–40 h		40–60 h	>60 h		I don’t know
Thermal Comfort	1. How do you judge the thermal environment of the lab?
	Very Acceptable		Slightly Acceptable		Just Acceptable	Just Unacceptable	Slightly Unacceptable	Totally Unacceptable
	2. Generally, how do you perceive the thermal environment in the lab?
	Hot	Warm	Slightly warm		Neutral	Slightly cool	Cool	Cold
	3. In general, I would like the thermal environment in the lab to be
	Warmer				No Change		Cooler
	4. Are you experiencing any thermal discomfort in your sitting space in the lab?    Yes    NO
	5. If yes, what is the thermal condition in your sitting space in the lab?
	Hot		Warm		Slightly warm	Slightly Cool	Cool	Cold
Occupant Comfort	6. When you are in this lab space, how do you generally feel?
	Very Comfortable		Just Comfortable		Slightly Comfortable	Slightly Uncomfortable	Not very Comfortable	Not at all Comfortable
	7. Have you suffered any of these symptoms when staying in this lab today? Please tick all that apply
	Lethargy			Loss of concentration		Nausea and dizziness	Headache	Hoarseness
	Skin rash			Wheezing		Itching	Eye irritation	Nose irritation
Air Quality	8. How do you judge the air quality of the lab?
	Very acceptable		Acceptable		Slightly acceptable	Slightly unacceptable	Not acceptable	Not at all acceptable
Visual Comfort	9. During daytime, are you satisfied with the level of natural lighting in the lab?
	Very Satisfied		Satisfied	Slightly Satisfied	Neutral	Slightly Dissatisfied	Dissatisfied	Very Dissatisfied
	10. Are you satisfied with the overall lighting in your sitting space in the lab? (Please consider both electric and natural lighting)
	Very Satisfied		Satisfied	Slightly Satisfied	Neutral	Slightly Dissatisfied	Dissatisfied	Very Dissatisfied
	11. If you are dissatisfied with the lighting in the space, which of the following contribute to your dissatisfaction? Please tick all that apply
	Too bright		Too dark		Not enough daylight	Too much daylight	Not enough electric lighting	Too much electric lighting
	Electric lighting flickers		Electric lighting in an undesirable colour		No task lighting	Reflection	Glare	Other (please specify)
Aural Comfort	12. On a scale of 1 to 7, where 7 is very satisfied and 1 is very dissatisfied, how satisfied are you with the noise level in the room?
	13. If you are dissatisfied with the noise level in the lab. Which of the following contributes to this problem? (Please select all that apply
	Loud laughter		People talking in the lab		People talking outside the lab	Equipment noise		Lighting noise
	Mechanical noises (e.g., heating, cooling, ventilation systems)				Outdoor traffic noise	Other outdoor noises		Other (please specify)

of Appendix B) was administered in English language and was pretested among half of the participants to validate the language and survey design. Because participants entered the graduate students’ office according to their own schedules, they completed the questionnaire on the days they were present, resulting in an overall response rate of 87.5%. Of these responses received, 48.4% were from international participants while 51.6% were from Japanese participants. The questionnaire consisted of 16 questions divided into three sections: participants’ biodata, indoor environmental conditions, and occupant comfort and close contact behavior [38].

The indoor environment section of the questionnaire included a subjective assessment (Table A2 of Appendix B) of thermal comfort, noise levels, indoor air quality (IAQ), and illumination. It incorporated relevant IEQ questions from previous studies [39,40,41]. The survey utilized a semantic differential scale and visual analog scales to assess thermal comfort, IAQ, and audio and visual comfort [39,40,41]. Thermal comfort was evaluated using a semantic differential scale with the following categories: “cold, cool, slightly cool, neutral, slightly warm, warm, hot”. Acceptability of thermal, audio, and visual conditions was assessed on a scale ranging from “very satisfied” to “very dissatisfied” (Table A2 of Appendix B).

A separate Likert scale was used to assess self-reported experiences of building-related illnesses, including loss of concentration, nausea, dizziness, headache, hoarseness, skin rash, eye irritation, nose irritation, wheezing, and itching (Table A2 of Appendix B). Confirmation questions regarding occupant comfort and IEQ improvements were included to verify response consistency. Additionally, participants were provided with a Fitbit Sense 2 smartwatch equipped with a multi-path optical heart rate sensor (Appendix A) and were required to wear it while in the graduate student laboratory to monitor their heart rates during the study period.

We did not analyze occupant comfort differences between males and females. Future field surveys with a more balanced sample distribution could help examine gender-based differences in comfort perception besides nationality.

2.3. Geometric Mean

We calculated the geometric mean (GM) (Equation (1)) [60,91] instead of arithmetic mean to quantify certain indoor parameters (temperature, CO₂, RH) since unlike the arithmetic mean, the geometric mean accounts for compounding and serial correlations.

$$GM=({\prod }_{i=1}^n{x}_i{)}^{\frac{1}{n}}=(X_1.X_2.X_3\dots X_n{)}^{\frac{1}{n}}$$

(1)

whereby X_i is the recorded reading (observation) for each parameter and n is the number of observations.

2.4. Variable Relationship with Overall Acceptance of IEQ

In our field survey, we measured these coefficients, which can be used to model the relationship between comfort acceptance and sensation (thermal comfort, visual comfort, indoor air, and aural comfort) using a binary logistic regression model (LR).

Equation (2) [60,92] shows the logistic regression equation with X_i as the input variable, Y as the target variable, α as the intercept term, and β_i as the coefficient for X_i.

$$Y=\frac{\exp (\alpha +{\sum }_{i=1}^4{\beta }_1{X}_i)}{1+\mathrm{e}\mathrm{x}\mathrm{p} (\alpha +{\sum }_{i=1}^4{\beta }_1{X}_i)}$$

(2)

During data analysis, the comfort votes were further classified into two categories: ‘acceptable’ and ‘unacceptable’. The occupant comfort ‘feel’ was also classified as ‘comfortable’ and ‘uncomfortable’.

In addition to examining the relationship between the individual variables with the overall IEQ acceptance vote, we also examined intervariable relationships using Yule’s Q value (Equation (3)) [60] through a 2 × 2 matrix.

$$Q=\frac{X_{11}{Y}_{22}-X_{21}{Y}_{12}}{X_{11}{Y}_{22}+X_{21}{Y}_{12}}$$

(3)

2.5. Heart Rate Variability

To measure heart rate variability of study participants while in the indoor environment under study, we employed nonlinear analysis of the RR time series through Poincaré plots [93]. We utilized the standard deviation (SD1, SD2) and SD1/SD2 ratio as the Poincaré plot parameters to analyze physiological changes while in the indoor environment. This method provides a non-invasive assessment of beat-to-beat behavior of the heart. Analyzing the beat-to-beat variability can provide deductions on physiological response to stress. The RR interval from time series RR = {RR1, RR2, …, RRn, RRn + 1} downloaded from the Fitbit Sense 2 watch for the duration the occupant was in the student’s office was plotted against the next RR interval [94,95,96,97].

The SD1/SD2 ratio assesses the balance between sympathetic and parasympathetic activities. SD1 (short-term variability) is an index of parasympathetic (vagal) activity. On the Poincaré plot, it represents standard deviation of the distance of each point from the line of identity (y = x), reflecting beat-to-beat variation. SD2 (long-term variability) is an index of the sympathetic activity on the Poincaré plot; it represents the distance of each point from the line of identity (y = x + mean RR interval) indicative of overall fluctuation. A higher SD1/SD2 ratio suggests a greater influence of short-term variability, potentially indicating a more balanced autonomic nervous system [94,95,96,97].

These parameters can be defined by following equations (Equations (3)–(8)) [94,95,96,97,98,99,100]:

$$x=\left\{x1,x2,x3,\dots ,xn\right\}=\{RR1,RR2,RR3,\dots ,RRn\}$$

(4)

$$y=\left\{y1,y2,y3,\dots ,yn\right\}=\{RR1,RR2,RR3,\dots .,RRn+1\}$$

(5)

$$d_1=\frac{x-y}{\sqrt{2}};d_2=\frac{x+y}{\sqrt{2}}$$

(6)

$$SD1=\sqrt{var\left(d_1\right)};SD2=\sqrt{var\left(d_2\right)};$$

(7)

$$Ratio=\frac{SD1}{SD2}$$

(8)

3. Results and Discussion

3.1. CO₂ Concentration

Figure 2 illustrates the daily high- and low-occupancy profiles alongside indoor CO₂ variations. The indoor CO₂ concentration fluctuated proportionally to the number of occupants present. Sudden drops in CO₂ levels (Figure 2a at 2:00 p.m., Figure 2c at 6:00 p.m., and Figure 2e at 1:00 p.m.) occurred when participants opened doors to increase ventilation due to discomfort. Door operation was the occupants’ environmental adaptive behavior observed in this study to improve ventilation. The observed door-opening CO₂ concentration (ppm) was between 996 ppm and 1766 ppm. The researchers observed that occupants would check the readings of CO₂ when they felt uncomfortable and opened the doors. But as the week (study period) progressed, the CO₂ door-opening concentration increased suggesting that initial placement of sensors aroused awareness and consciousness on IEQ but with time, occupants may have adjusted and became complacent. This needs to be investigated further, especially where sensors are placed to monitor IEQ but are not connected to the ventilation system and IEQ improvement interventions are solely implemented by occupants. One of the interventions in managing COVID-19 spread and tuberculosis spread is placement of CO₂ sensors in public buildings in many countries around the world, including the UK and South Africa [6]. From the results of our study, it was concluded that the success of this public health intervention (functionally) if not linked to a smart ventilation system is dependent on human adaptive behavior.

Our previous study [38] established a significant positive correlation between the number of people indoors and the number of close contacts. While this study could not directly link indoor CO₂ concentration to the number of close contacts, Pearson correlation analysis revealed a strong significant correlation between co-presence duration (i.e., the number of occupants present simultaneously and their duration of shared indoor time) and CO₂ concentration (p < 0.05). This study assumes that the metabolic rate of each occupant remained relatively constant during their respective indoor periods.

No significant differences in indoor CO₂ concentration were observed across the various indoor sensors (Figure 1a and Figure 3) at a given time. Since the room was mechanically ventilated using a mixing ventilation system (Figure 1a), indoor CO₂ was well-distributed at any given time (Figure 3), regardless of the number of occupants or their positions in the room. The standard deviation of CO₂ concentrations across different locations in the students’ office could be attributed to sensor errors. This aligns with the findings of Rackes et al. [30], who concluded that multiple sensors are not essential in mechanically ventilated spaces due to uniform air mixing.

Through one-way ANOVA, we analyzed the statistical significance of mean differences between daily readings. The mean differences were statistically significant (p < 0.05). Our analysis determined that indoor CO₂ concentration was highly sensitive to occupancy at a given ventilation rate. Even slight fluctuations in the number of occupants were reflected in CO₂ levels more rapidly than in relative humidity or temperature variations (Figure 3). Thus, indoor CO₂ concentration serves as a useful—though somewhat ambiguous—proxy for the quantity of exhaled aerosols, infection risk in indoor environments, and, by extension, ventilation efficiency and indoor air quality [22]. Figure 2 data can be useful for modelling in build-up methods that use the increase in CO₂ to determine ventilation rate, occupant comfort and/or infection spread.

On day 5, six-hour data (18:00–23:59) for sensors 22-07, 22-08, 22-09 and 22-10 were not stored due to human error in reconnecting the sensors to the router after data transfer. Hence, there is a data gap on day 5 (Figure 3). However, since the air is uniformly mixed, the data transmitted from sensors 22-01 to 22-06 can be assumed to be representative of the indoor conditions at the time except for the temperature at the ceiling level.

Regarding measurement locations, we monitored CO₂ levels at a single outdoor point, whereas other studies have measured outdoor CO₂ concentrations at multiple locations or relied on empirical constants. As a result, we cannot rule out potential inaccuracies in outdoor CO₂ measurements due to methodological differences.

3.2. Indoor Air Quality (IAQ) Acceptance

The survey results revealed that 82.3% of the occupants found the IAQ acceptable. IAQ acceptance was significantly higher among Japanese participants (93.3%) than international participants (70%) (Figure 4f). Day 1 had the lowest acceptance rate (Figure 4a) and the highest number of complaints (Figure 4b). Loss of concentration was the most frequently reported issue among occupants. Occupant well-being has been linked to productivity, where even mild discomfort can significantly reduce work performance [49,101]. Based on the number of complaints recorded during the study period, the field survey results suggest that Days 1 and 5 were the least productive for all occupants.

We also examined the relationship between the four aforementioned variables and self-reported loss of concentration and feelings of lethargy. IAQ had the strongest correlation (0.268) with lethargy and loss of concentration (p < 0.05). Similar findings have been reported in various studies [102,103].

When participants were asked to report the frequency of experiencing building illness-related symptoms for the previous 3~6 months they had worked in the students’ office, overall, Japanese participants reported fewer complaints associated with IAQ (Figure 4c) compared with their international colleagues. However, Japanese participants were more likely to experience lethargy (15%) and loss of concentration (19%) than international participants. Eye and nose irritation were a shared complaint, although international participants reported eye irritation to be a daily experience, but nose and eye irritation were less probable among the Japanese participants.

In the assessment of the same symptoms during the study period (Figure 4e), experiencing loss of concentration and lethargy was comparable in both groups (p > 0.05, t-test), with no significant difference. Nose and eye irritation were significantly experienced in the international participant group.

During the study period, the indoor CO₂ levels were largely within 1000 ppm above the corresponding outdoor CO₂ levels (

Table 2. Descriptive statistics of room occupancy, indoor and outdoor CO₂ concentration.

CO2 Concentration [ppm]
Outdoor					Indoor				Occupancy
Day	Max	Min	Median	Geometric Mean	Max	Min	Median	Geometric Mean	Occupants	Co-Presence Duration [h]
Day 1	814	483	514	519.23	1400.7	428.4	595	689.66	12	14
Day 2	556	440	453	461.65	1206	473.7	768	748.17	14	16
Day 3	574	435	454	459.85	1247.2	487	694	757.33	14	20
Day 4	614	444	488	482.92	1645.4	410.1	616	684.01	11	14
Day 5	626	452	476	472.91	1671.7	624.6	892	973.95	16	21

) as required by standards. This partially explains the high IAQ acceptance levels despite the proportion of complaints registered. Day 4 and Day 5 had indoor CO₂ levels at >1000 ppm above outdoor CO₂ levels (Table 2). The air quality acceptance level declined on Day 4 and 5 (Figure 4a). However, our study did not measure particulate matter and odor levels to ascertain IAQ or their contribution to the acceptance levels by occupants.

3.3. Indoor Environment Conditions: Relative Humidity

While CO₂ concentration and temperature levels increased with indoor occupancy, relative humidity tended to decrease as occupancy increased (Figure 3). Because the study was conducted during winter, a typically dry season, both outdoor and indoor relative humidity levels remained below 60% throughout the study period (

Table 3. Daily indoor and outdoor measures of relative humidity (%).

Relative Humidity [%]
Outdoor Conditions					Indoor Conditions
Day	Max	Min	Median	Geometric Mean	Max	Min	Median	Geometric Mean
Day 1	57.7	46.6	55.2	54.72	35.2	12.4	29.9	29.1
Day 2	56.5	23.6	30.2	35.33	33.2	10.6	22.4	22.9
Day 3	35.2	23.8	28.9	28.8	21.5	7.5	16	16.1
Day 4	47.8	34.8	39.6	40.08	26.9	5.3	20.7	20.6
Day 5	52.2	24.5	42.25	38.44	31.7	12.6	22.9	22.4

). This falls below the recommended indoor range of 40–60% for occupant health [42]. During the study, 9–29% of participants reported experiencing eye irritation, while 11–27% reported nose irritation (Figure 4e). These symptoms have been linked to low indoor relative humidity levels [42,43]. Japanese participants reported fewer eye and nose irritation complaints (Figure 4e) during the study period, suggesting that they have adapted better or are used to this environment condition compared to their international colleagues.

3.4. Thermal Comfort

Indoor temperature was notably high on Day 1 (Figure 3) because the air-conditioning system had been turned off during the weekend. However, since electronic equipment such as computers and refrigerators remained operational, they contributed significantly to heat generation, in addition to the metabolic heat produced by occupants. There was no significant difference in temperature readings from the sensors positioned at the occupants’ breathing zone level. However, the sensor on the ceiling (22-10) recorded significantly higher temperatures (Figure 3) than the other sensors due to the buoyancy-driven rise of warm air. The measured ranges of indoor and outdoor temperatures over the five-day period are detailed in

Table 4. Descriptive statistics of room occupancy, indoor and outdoor temperature readings.

Temperature [°C]
Outdoor Conditions					Indoor Conditions
Day	Max	Min	Median	Geometric Mean	Max	Min	Median	Geometric Mean	TocfEN	Tdiff
Day 1	15.4	11.4	12.4	12.32	39	19	24.2	23.5	23.5	0
Day 2	12.8	9.1	10.4	10.86	33.5	20.2	24.3	24.2	23.5	0.7
Day 3	10.5	9	9.8	9.82	33.7	21	24.3	25.1	23.5	1.6
Day 4	10.1	7.6	9.4	9.08	39	14.9	21.7	21.6	23.5	−1.9
Day 5	10.7	9.1	9.8	9.8	35.1	19.3	25.2	25.3	23.5	1.8

.

The survey results indicated that 90.3% of study participants regardless of their gender or nationality were satisfied with thermal comfort. For the thermal sensation votes (TSV), a 7-point semantic differential scale from −3 to +3 was used to assess thermal comfort in the office and at occupants’ sitting spaces. The scale ranged from cold (−3), cool (−2), slightly cool (−1), neutral (0), slightly warm (+1), warm (+2), to hot (+3). For the office, 1.61% of occupants voted for cold (−3), 1.61% for cool (−2), 11.29% for slightly cool (−1), 29.03% for neutral (0), 29.03% for slightly warm (+1), 25.81% for warm (+2), and 1.61% for hot (+3). Because this study was conducted during winter, with the room’s conditioning system set to heat the space, thermal sensation votes were skewed towards the warm side, with 56.45% of responses indicating TSV > 0 compared to 14.51% indicating TSV < 0.

For thermal comfort in individual sitting spaces, a dichotomous assessment revealed that 29% of occupants experienced thermal discomfort at some point during their indoor presence. Specifically, 3.57% of occupants reported feeling cold, 7.14% cool, 35.71% slightly cool, 25.00% slightly warm, 10.71% warm, and 17.86% hot. Based on survey responses, 14.52% of occupants preferred a warmer room, while 20.97% wanted it to be cooler. Table 4 and Figure 5 present a summary of daily indoor thermal conditions and occupant acceptance.

Since the room was air-conditioned, temperature variations due to human occupancy were minimal. Given that this was an air-conditioned office during winter, the acceptance levels were deemed reasonable and could be used to model the temperature associated with a neutral thermal sensation while balancing occupant comfort and energy efficiency. As shown in Figure 5a,b and Table 4, the geometric mean of indoor temperature at which overall comfort was achieved—where occupants expressed the highest neutral vote and the least desire for temperature change—was 25.1 °C. This empirical finding shows that for sedentary work with light clothing, occupants preferred a temperature close to 25.6 °C, as predicted by Fanger’s equation. We observed this while noting that despite the objective measurements, we did not rule out subject differences informed by various sensitivities, tolerance levels and traits not examined in this field survey.

Thus, we further calculated occupant comfort temperature (Equation (9)) [104] based on prevailing outdoor conditions as per EN 15251 [105], which adopts the exponentially weighted running mean temperature (T_rm) that considers the significance of temperatures based on their distance in the past.

$$T_{rm}=\frac{T_{od\text{-}1}+\alpha T_{od\text{-}2}+{\alpha }^2{T}_{od\text{-}3}+{\alpha }^3{T}_{od\text{-}4}+\dots }{1+\alpha +{\alpha }^2+{\alpha }^3+\dots }$$

(9)

where T_od-1 is the daily average outdoor temperature for the day before (Day 4 of the study), T_od-2 is Day 3, T_od-3 is Day 2, T_od-4 is Day 1 and α is a constant = 0.8. EN 15251 [105], ISO 7730 [106] and ASHRAE 55 [107] use thermo-physiological and adaptive models to calculate comfort temperature, since occupants interact with their environment and are in a dynamic equilibrium with it [104]. During heating season, the EN occupant comfort temperature (T_ocfEN) is calculated by (Equation (10)) [104] yielding an occupant comfort temperature of 23.5 °C.

$$T_{ocfEN}=0.09T_{rm}+22.6 °\mathrm{C}$$

(10)

The T_diff temperature between calculated EN comfort temperature (T_ocfEN) and observed comfort temperature was 1.6. In this study, T_diff = 0 corresponds to 23.5 °C. This is a similar value to Day 1 mean indoor temperature (Table 4). However, in the study period, the neutral vote was highest on Day 3 at (54.5%) (Figure 5a), where 18% of the occupants voted wanted ‘warmer’ and 82% of the occupants did not want the indoor temperatures changed (Figure 5b).

Acceptance of the thermal sensation votes was more skewed towards the warm side for Japanese participants with 68.8% of responses indicating TSV > 0 compared to 3.125% indicating TSV < 0 (Figure 6a). There were significant differences in the TSV vote among the two groups. However, the native participants were more satisfied with the thermal conditions than the international participants (Figure 6b,c) However, acceptance vote on thermal conditions had no significant difference (Figure 6b,c) (p > 0.05, t-test).

Managing and optimizing occupant thermal comfort involves various factors beyond ambient air temperature. However, our study did not measure the mean radiant temperature (MRT) or air velocity, nor did we calculate operative temperature. Additionally, we did not account for the BMI of each occupant; future studies could incorporate these parameters for a more comprehensive analysis. Our study was conducted in an air-conditioned room under controlled conditions, where certain variables, such as temperature, were regulated. Because the study took place only during the winter season, future research could explore seasonal variations and compare different indoor space types with varying ventilation systems.

3.5. Aural Comfort

Sound levels ranged from 39.6 dBA to 91.7 dBA (Figure 7a and

Table 5. Daily measurement of sound pressure level.

Sound Pressure Level	Day 1	Day 2	Day 3	Day 4	Day 5
Max [dBA]	79.5	90.1	88	91.7	91.7
Min [dBA]	39.6	41	41	39.7	40.4
Median [dBA]	43.5	44.1	44.8	43.6	44.5
Mode [dBA]	41.9	43.5	43.8	40.9	43.5
Average	45.73	47.1	48.3	46.8	48.6

), primarily due to major renovation work taking place right outside the students’ office. As a result, these values were more comparable to those recorded at construction sites rather than in offices and classrooms [44,45,46,47,48]. The daily median sound levels ranged between 43.5 and 44.8 dBA (Table 5), exceeding the WHO’s recommended indoor daytime limit of <35 dBA [108] but within the limits of the country of origin of the participants (Table 1). The survey results indicated overall that only 37.1% found the indoor noise levels acceptable. As the sound level deteriorated with the progression of construction works, the aural comfort steeped (Figure 7a,b). Japanese participants seemed to have better tolerance levels and higher acceptance levels compared to the international counterparts (Figure 7c). The high sound level was not continuous but in spurts; there were longer periods of relative silence and low indoor sound level interrupted by episodes of loud noise from the construction works. Day 4 had the highest recorded sound level and correspondingly low acceptance votes (Figure 7a,b).

3.6. Visual Comfort

Outdoor light levels ranged from 2621 to 3404 lux, with a geometric mean (GM) of 2950.68 lux. Indoor horizontal illumination levels at occupants’ desks ranged from 166 to 566.5 lux, with a GM of 316.09 lux, which is below the Japanese Industrial Standards (JIS Z 9110, 2010) recommendation of >750 lux for office lighting [109] but within the limits of acceptable standards of many countries of origin for some participants (Table 1). Deficiencies in natural lighting are typically compensated for by artificial lighting. However, due to scaffolding work at the time of the study, natural light in the room was further reduced, contributing to visual discomfort and various complaints were registered daily by the participants (Figure 8c). The survey results indicated that 74.2% of occupants found the horizontal illuminance in the room acceptable (Figure 8a). There was no correlation observed between the acceptance of overall room lighting and acceptance lighting at individual desks (Figure 8a) (p > 0.05).

Japanese participants mostly reported insufficient natural lighting and inadequate brightness from artificial lighting (Figure 8c). Adjustments are necessary to improve visual comfort.

Overall, international participants registered more complaints on visual comfort than Japanese participants (Figure 8c); this can be partly explained by the background differences in horizontal luminance standards (Table 1). Considering the higher standards on horizontal luminance in Japan, complaints from Japanese participants were on insufficiency of light while international participants complaints were associated with high brightness levels and reflection (Figure 8c).

3.7. Heart Rate Variability (HRV)

We utilized the SD1, SD2 and SD1/SD2 ratios from the HRV Poincaré plot parameters to analyze physiological responses of the occupants while in the indoor environment. The SD1/SD2 ratio analyzes measures of parasympathetic contributions to heart rate variability (HRV) in scenarios of disease or acute mental stress [94,95,96,97]. There were discrepancies in the hypothesized directional relationship between SD1/SD2 because a higher SD1/SD2 ratio may imply a reduction in SD2 or increase in SD1, or both [98]. However, it is still a proposed measure for characterizing sympathetic influences on HRV during periods of psychological stress. In this study, the SD1/SD2 ratio ranged from 0.06 to 0.2. Normal SD1/SD2 ratios for healthy people range from 0.1 to 0.4 [110].

There was a statistically significant mean difference between Day 4 and Day 5 for the first 3 days (p < 0.05). However, there was no statistically significant difference between Day 4 and Day 5. The slightly higher SD1/SD2 ratios on Day 4 and Day 5 indicate the possibility of being subjected to stress although variations were minimal. A higher SD1/SD2 ratio indicates decreased parasympathetic activity or an increased sympathetic influence on heart rate regulation [94,95,96,97,98,99,100]. In the indoor environment analysis, Day 4 and Day 5 had the highest aural discomfort votes and corresponding sensor measurements showed high sound level values. Day 4 also had a lower IAQ acceptance vote, visual comfort and thermal comfort than other days. Correspondingly, the HRV of occupants in day 4 had a slight surge (Figure 9a).

We analyzed the HRV among the two groups of international participants (Figure 9c) and Japanese participants (Figure 9b). The mean differences were not statistically significant (p > 0.05). On Day 4, participants from both groups had a slight increase in their SD1/SD2 ratio, yet in the subjective analysis, Japanese participants indicated more acceptance of the IEQ conditions than the international participants. This is indicative of two possibilities: international participants have not adapted to the environment and/or Japanese participants, due to their reserved culture and indirectness [111,112] in conveying disappointing feedback, did not fully disclose their discomfort or dissatisfaction with any indoor conditions. This is indicative of the potential interference of culture in subjective responses and the need for inclusion of physiological assessment.

3.8. Overall IEQ Acceptance

The field survey was conducted to examine the relationships between four major IEQ factors, thermal comfort, visual comfort, aural comfort, and IAQ acceptance, in relation to overall IEQ acceptance. Each day, occupants voted on their perceived comfort under the given indoor conditions (Figure 10). An occupant’s overall sense of comfort is a combination of multiple interrelated physical parameters [50,51].

Table 6. Coefficients in the regression analysis between comfort votes and the four main attributes of IEQ.

	Thermal	Visual	Aural	IAQ
Multiple R	0.254639	0.538805	0.087989	0.70409
R Square	0.064841	0.290311	0.007742	0.495743
ADJUSTED	−0.16895	0.112889	−0.24032	0.327657
Intercept	−0.15427	−0.32609	−0.02579	0.909747
p-value	0.018519	0.004649	0.000501	0.826733

summarizes the regression coefficients determined from the field surveys. Because not all factors contribute equally to IEQ acceptance and consequently, occupant comfort, it is essential to identify the most influential factors in order to enhance comfort through building design [36,37]. To achieve this, we conducted a regression analysis (Table 6) to evaluate the contribution of each variable to overall IEQ acceptance and comfort satisfaction.

Korsavi et al. [36] suggested that a better perception of IAQ can compensate for higher temperatures. This was also observed on Day 4, when thermal, aural, and visual comfort were poor, yet IAQ received higher acceptance votes and the lowest number of complaints.

Despite the increased aural discomfort reported throughout the week, overall indoor comfort remained high, particularly on Day 4 (Figure 7 and Figure 10), when aural comfort was at its lowest. This observation aligns with the findings reported by Humphreys (2005) [52], who found that dissatisfaction with a single element of the indoor environment does not necessarily affect overall comfort perception. This finding contrasts with the conclusions of Nagano and Horikoshi (2005) [35] and Tahsildoost and Zomorodian (2018) [53], who suggested that extreme discomfort in a single attribute—such as excessive noise or heat—can significantly contribute to overall discomfort.

According to Humphreys (2005) [52], unacceptable aspects of the indoor environment can be compensated for by acceptable aspects of IEQ. In this study, we observed that favorable IAQ and thermal conditions helped mitigate deficiencies in other variables. As shown in Table 6, IAQ and visual comfort votes exhibit a strong correlation with overall comfort, as indicated by their R² values [36]. However, IAQ had the most significant impact on overall comfort, with an R² value of 0.5, meaning it accounted for 50% of variations in overall comfort. Visual comfort contributed 29% (R² = 0.29), temperature accounted for 6.4% (R² = 0.064), and aural comfort had the lowest influence, contributing just 0.8% (R² = 0.008). These findings align with previous studies demonstrating that satisfaction with indoor air quality is a key determinant of overall satisfaction with indoor environments [36,37,52,53,55].

In terms of inter-variable relationships, Yule’s Q value indicated a weak positive relationship of 0.21 between thermal comfort and IAQ. Korsavi et al. [37] found that among schoolchildren, IAQ acceptance was highest when participants reported a “cool” thermal sensation. Similarly, in our study, IAQ acceptance peaked on Day 3 (Figure 4a), coinciding with the most favorable thermal sensation vote of “neutral” (Figure 5a). Overall, in the study period, IEQ acceptance was the highest among Japanese participants. In total, only 20% of international participants and 6.2% of Japanese participants voted IEQ to be unacceptable.

4. Conclusions

Through subjective and objective measurements, we investigated overall occupant comfort (IEQ acceptance) in an ethnically diverse graduate student’s office during the winter season. While this study provides valuable insights, we acknowledge the number of participants may limit the generalizability of the study’s findings to the larger population. In addition, a more prolonged study period may capture and describe the dynamics on a grander scale. Future work could address these limitations by using a larger, more diverse sample and an extended study timeframe, as well as incorporating seasonal variation.

Considering the limitations mentioned earlier, the results give a general overview of IEQ variation with indoor occupancy and variables that contribute both significantly and insignificantly to the acceptance of indoor environment quality and affect performance in an ethnically diverse population. Among the four variables examined, thermal comfort and IAQ received the highest acceptance votes, while visual and aural comfort had the lowest acceptance rates. Regression analysis indicated that IAQ accounted for nearly 50% of overall occupant comfort perception and IEQ acceptance. The daily number of complaints recorded in this study corresponded closely with indoor air quality perception and visual comfort. International participants had a significantly higher number of complaints than native (Japanese) participants more so in visual comfort where acceptable luminance levels are higher in Japan than other countries represented in the study. Prioritizing IAQ and thermal comfort while enhancing natural lighting—without excessive reliance on artificial lighting—can improve occupant comfort, reduce complaints, and positively impact productivity.

Thermal comfort was to a significant extent mutually accepted by both groups. The geometric mean of recorded indoor temperatures was within the range of acceptable/recommended limits by countries of origin for the participants. The observed geometric mean of the indoor temperature at which overall comfort can be obtained with a higher neutral vote and least desire to change temperature was 25.1 °C, which was close to the 25.6 °C predicted by Fanger’s equation and corresponded to the 23.5 °C occupant comfort calculated temperature based on prevailing outdoor conditions as per the EN15251, ISO7730 and ASHRAE 55 thermo-physiological and adaptive models.

Loss of concentration and lethargy was comparable in both groups (p > 0.05, t-test), with no significant difference. International participants reported eye irritation to be a daily experience, but nose and eye irritation were less probable among the Japanese participants. Nose and eye irritation were significantly experienced in the international participant group. These symptoms have been linked to low indoor relative humidity levels. Japanese participants reported fewer eye and nose irritation complaints during the study period, suggesting that they have adapted/coping better to this environment condition than their international colleagues.

We recommend global coherence in indoor environmental quality standards/guidelines as is the case with drinking water standards for public health protection. In this way, there can be seamless transitions as the world increasingly integrates socially and can ensure occupant comfort for both native and non-native populations. This can enhance and sustain productivity in the shared indoor environments.

From our HRV analysis, we recommend inclusion of physiological assessment in subjective evaluation of perceptions to the environment.

Although aural comfort was the least accepted due to ongoing construction and demolition activities, our findings suggest that this factor did not significantly affect overall IEQ acceptance, as IAQ and thermal comfort compensated for it. This observation provides insights into occupants’ environmental tolerance. Intra-variable analysis using Yule’s Q indicated a positive relationship among the four IEQ factors.

In recognition of our study limitations and applicability to generalize our study findings, our results demonstrate and support the need for further research of this type aimed towards informing future policy decisions and other measures by building designers and facility managers on occupants’ comfort and well-being in new environments, especially where adult occupants are not from the same country of origin.