Introduction
Around the world, the majority of people spend most of their time indoors (80–90%) (González-Martín et al., 2020). A secure and comfortable home or apartment is far from a healthy environment, with sources of household air pollution (Hoskins, 2010; Pérez-Padilla et al., 2010; Apte and Salvi, 2016; Mannan and Al-Ghamdi, 2021). The concentration of pollutants can be several times higher than outdoors (Simoni et al., 2003). In most cases, outside air is healthier than indoor air (Taştan and Gökozan, 2019). One efficient way to lower the concentration of pollutants at home is to open the windows (Haddad et al., 2021). Health agencies recommend opening the windows to allow fresh air in for a short period of time (at least 10 min each day is recommended).
Knowledge about window opening behavior at home remains scarce. Energy performance in buildings remains the main focus of interest (Fabi et al., 2012; Hou et al., 2017; Park and Choi, 2019). Most models seek to predict manual control of windows at home as a function of outdoor and indoor environmental variables (outdoor and indoor temperature, wind speed, level of PM2.5 and PM10, indoor CO2 level, relative indoor humidity, etc.) (Arethusa et al., 2014). Unlike many health behaviors, the social and psychological antecedents of window control have not been fully explored until now. The present research addresses a gap in the literature about the volitional and non-volitional antecedents of windows opening. We focus here on the willingness to open windows at home. Our aim is twofold. On the one hand, we will explore the social and psychological antecedents of the intention to open the windows at home. Within the framework of the Theory of Planned Behavior (TPB), we expect the intention to open windows to stem from motivational antecedents available in current social cognitive models of behavior prediction. On the other hand, we will assess the effect of the exposure to a recommendation to open the windows on intention to open the windows. The effectiveness of ventilation recommendations remains an unmapped area, as do the psychological antecedents of window opening.
Indoor air pollution is increasingly covered by the scientific literature as a main health issue (Zang and Smith, 2003; Kampa and Castanas, 2008; Pérez-Padilla et al., 2010; González-Martín et al., 2020; Sakhvidi et al., 2022), whether it stems from the work environment (Reijula and Sundman-Digert, 2004), the home and the daily exposure to pollutants (Missia et al., 2010). The health consequences of air pollution indoors cover a wide range of diseases, from immediate irritation of the eyes and respiratory mucosa (Bernstein et al., 2008) to long-term diseases such as lung cancer and cardiovascular problems (Berglund et al., 1992; Corlan et al., 2021). Most previous studies have focused on pollutants and their multivariate origins: the materials used for construction, biological sources, combustion of materials or occupants’ activities and behaviors (Park and Ikeda, 2006; Kampa and Castanas, 2008; Lin et al., 2017). Behaviors may largely contribute to worsening indoor air quality at home (Kim and Paulos, 2010). Whether it be from smoking habits (Repace and Lowrey, 1993; McAuley et al., 2012), the use of air fresheners/cleansers (Nazaroff and Weschler, 2004) or uses of personal care products (Pieri et al., 2013; Johnson et al., 2019). Indoor air behaviors have been identified as being responsible for a much higher level of fine particle pollution than building characteristics (Spilak et al., 2014). Behaviors related to indoor air can also have a beneficial effect (Park and Choi, 2019), such as the choice of furniture (Lee et al., 2006), or changing habits related to smoking behaviors (Thornley et al., 2013). Fresh air circulation with natural ventilation is an efficient way of maintaining good indoor quality (Saito et al., 2004; Kim et al., 2008; Bekö et al., 2011). Ventilation removes pollutant concentration from an interior space and replaces contaminated air with fresh air from outside (Awbi, 2017; Pan et al., 2019). An inverse relationship was also found between lung cancer and good household ventilation (Jin et al., 2014), and is also a low-cost way of improving air quality at home, which is not complex to implement for most people.
Environmental determinants are the only factors influencing window opening behavior studied by the literature (Fabi et al., 2012; Belafi et al., 2018), with temperature, CO2 concentrations, wind speed but also odors (Fabi et al., 2012). Fabi et al. (2012) carried out a review of the factors influencing occupant ventilation behaviors. Different categories were identified, the physical environment, contextual drivers, psychological drivers, and the physiological and social environment. Besides, the climate conditions have a decisive impact on natural ventilation (Escandón et al., 2019). In very hot climate, the ventilation is possible only in a precise moment of the day (night-time) because of extremely hot temperatures. This leads to an opening of windows motivated mainly by environmental factors.
The social dimension needs to be further explored. Fabi et al. (2012) are among the few to integrate a psychological dimension into the drivers of ventilation. However, no study has investigated the psychological determinants of ventilation as the main focus of interest. This exploration could also be used to clarify the effect of environmental cues on behavior when measured with psychological variables.
Environmental cues are susceptible to guide window opening behaviors, especially when they are related to comfort, occupants’ comfort plays an important part in window opening patterns (Mircea et al., 2014; Yao and Zhao, 2017). Odors seriously influence human comfort even if habituation from a prolonged exposure to a compound may decrease the feeling of discomfort (Zhu and Li, 2017). Subjective indoor air quality may also be inferred via cues that are detected by humans (bodily odor, kitchen odors, cleaning agents, etc.), even if many pollutants, such as carbon monoxide, have no detectable odor (Wolkoff, 2018). The absence of odors is not an indication that the place is healthy, and some pleasant odors (e.g., fragrances) may be associated with pollutants (Steinemann and Goodman, 2019; Steinemann et al., 2020; Goodman et al., 2021). Odors may contribute to the opening of windows as cues that are perceived by our sensory abilities as a means of detecting pathogens. Due to the individual’s odor sensitivity (Smeets et al., 2008), it is an interesting variable that may contribute change the willingness to ventilate at home.
Window opening is a repeated behavior that occurs in a predictable pattern of frequency, time of day and places. More window opening is observed in the early morning than at midday (Park and Choi, 2019). This predictable pattern shows potential for the calibration of a TPB application addressed to a behavior with temporal and situational specificities (Ajzen, 1991, 2002).
Considering the lack of research on the social and psychological antecedents of ventilation behavior, the guidelines of the Theory of Planned Behavior (TPB; Ajzen, 1991) will help to organize knowledge. The TPB has been extensively applied to health and environmental issues (Godin and Kok, 1996; De Leeuw et al., 2015), and provides strong predictions of health intention and behavior especially among behaviors that are repeated daily such as handwashing, physical activity and diet (McEachan et al., 2011; Liddelow et al., 2021). Closer to our purpose, Shi et al. (2017) were interested in inhabitants’ behaviors relating to the reduction of targeted pollutants at home. Controlling households’ behaviors relative to PM2.5 greatly helped to reduce pollution. The TPB model explained 56% of the variance of target behaviors. All three antecedents for behavioral intention were predictors for PM2.5 reduction intention. No studies have explored the application of the TPB to window opening. It is fruitful to apply this model to new health behaviors (Rich et al., 2015), especially when those can be explained in a relevant way by the model’s variables as described below.
Intention
In current life situations, ventilation may be stimulated by an external cue that prompts opening of the windows (Persily, 2015). Even though such cues trigger a behavior with a low cognitive cost, such behavior remains under the control of the will. Occupants may decide to initiate a new behavior or stop the behavior at any time. The first antecedent to opening the windows is therefore a behavioral intention.
Attitudes Toward Behavior
The evaluation (positive or negative) associated with any discriminable aspect of the world in which we live (Ajzen and Cote, 2008). Attitudes relative to indoor air and window opening has been little explored. In France, Marchand et al. (2018) did not find shared representations of indoor air as such representations are available for other environmental issues (e.g., flood). However, the interviews revealed positive attitudes toward ventilation as a means of protecting health.
Subjective Norm
The subjective norm is related to the perceived approval of the behavior by the social environment (Ajzen, 1991; Sniehotta, 2009). As people share homes with other occupants (family, roommates, etc.), they may seek others’ approval to decide to open the window. In some cases, opening may be highly approved by significant others for comfort or health motives (Shi et al., 2020). As an expectation, the subjective norm may result from the observation of the number of people who open the window each day (descriptive norm), or from explicit injunctions that inform about what people ought to do (injunctive norm; Cialdini et al., 2006).
Perceived Behavioral Control
The perceived behavioral control can be broadly defined as the perceived difficulty or ease of establishing the behavior (Ajzen and Driver, 1992). Compared to other health or pro-environmental behaviors (e.g., smoking cessation, regular physical exercise, effective waste management) that are investigated by the TPB model, the opening of windows is easier to learn, easier to perform and easier to apply in one’s own familiar home environment. Except for disabled people who live in inadequate home. So, the manual control of windows is a highly controllable behavior that has positive consequences in the short (fresh and healthy air) and long term (lifelong health), and some immediate negative consequences (exposure to uncomfortable cold air from outdoors, energy expenditure, etc.). Intentions are more likely to be translated into behavior when the targeted behaviors are easy to achieve (Sheeran et al., 2003; Sheeran and Webb, 2016). Perceived Behavior control may moderate the other constructs of the TPB model. Attitudes may lead to favorable intentions and behaviors to the extent that people have a high level of perceived control of the behavior (La Barbera and Ajzen, 2020). Such a moderation effect of perceived control is expected in relation to the opening of windows. Considering the interaction with the subjective norm, a high level of perceived behavioral control is expected to decrease the impact of the subjective norm. The weight of others’ approval may diminish when a behavior is well under control.
Habits
According to Orbell and Verplanken (2015), habits are defined as converging upon repetition, automaticity and cued by a stable environmental context. Habits are expected to be a predictor of behavior (De Bruijn, 2010; Chen and Chao, 2011; De Bruijn and Rhodes, 2011) and to moderate the relationship between intentions and behavior. In numerous models, the link between habits and intention remains strong and significant (Honkanen et al., 2005; Liddelow et al., 2021). A similar integration of habits in a TPB framed model was found within health-oriented behavior [habits significantly increased the explained variance (fruit consumption, physical exercising)] (De Bruijn, 2010; De Bruijn and Rhodes, 2011). The application of the habit construct to the repetitive feature of window opening behavior may be especially interesting. Verbruggen et al. (2019) defined self-reported habits about ventilation behavior as an action repeated daily before, after or during certain activities, in the everyday life. Also, the relationship between habits and intention is important for non-reflective behaviors (El-Khatib and Barki, 2012).
The effects of indoor air have led health organizations to communicate and establish evidence-based recommendations about indoor air pollution, in the United States with the Environmental Protection Agency (U.S. Environmental Protection Agency, 2003), in the United Kingdom (UK:100, 2019) and in France (INPES, 2009; ANSES, 2018). Detailed instructions include how to ventilate and the health advantages of ventilation. To our knowledge, no research has measured the impact of campaigns or recommendations about ventilation on window opening intentions at home. Considering the small effect size of health messages on behavior (Anker et al., 2016), the probability of a behavioral change following a single recommendation remains low. Nevertheless, other outcomes may benefit from recommendations given by a trustworthy source (e.g., a health agency): the transfer of knowledge, awareness of health importance and a greater willingness to open the windows.
Overview of Empirical Research and Hypothesis
The extended Theory of Planned Behavior (TPB; Ajzen, 1991, Ajzen and Driver, 1992) provides constructs that are relevant antecedents of ventilation behavior as a goal-directed and a health behavior (Ajzen, 2011). Expectations of ventilation benefits and costs (Attitudes), the subjective ease of opening the windows (Perceived behavioral control), and the perceived social pressure to open or close the windows at home may contribute to the intention to open the windows.
Beyond TPB variables, we will focus on two manipulated contextual variables that may shift intention to open the windows. Firstly, with an exposure to a ventilation health recommendation. Participants will receive information about the health benefit of performing the ventilation behavior. Secondly, with respect to the low level of engagement with indoor air quality, indoor air awareness should increase attention to indoor air as a sensory and physical reality. People who are aware of their own odor sensitivity will agree more with information about the hazards of indoor air pollution and therefore show more intention to open the windows to cope with the health risk (Smeets et al., 2008). In order to control the valence of odors, attention was focused either on positive or negative odors.
We hypothesized that all the predictors of the extended TPB model will play a role in explaining the behavioral intention to ventilate (H1a). We expect that Habits will also contribute (H1b). As suggested by La Barbera and Ajzen (2020), we expect that Perceived Behavioral Control may moderate the effect of Attitudes on intention (H1c).
Considering the health recommendation, we expect that it will increase the intention to open windows (H2a) and that the effect of the recommendation will be moderated by habits. Contrary to low habit occupants, high habit occupants will not change as a function of a recommendation (H2b). According to odor sensitivity, we expect that people who are aware of their own odor awareness (specially to negative odors) will report intention to open the windows when exposed to a recommendation (H3a) and this effect will be moderated by window opening habits (H3b).
Two studies were carried out in France. Study 1 was performed among students. Study 2 replicated the protocol among a larger convenience sample from the Web. The research was conducted in compliance with the Ethical Guidelines of the Psychology Department of the Paris Nanterre University as well as with the National Ethical Standards approved by the University.
Overall Discussion
Considering the gap in the literature about ventilation as a volitional and non-volitional behavior, our first aim was to identify the psychological determinants of willingness to open windows at home. In the two studies, the extended TPB model has provided significant predictors of intention to open the windows at home, which explained a large amount of variance with around sixty percent of explained variance similar to other health domains (McEachan et al., 2011). In the two convenience samples (students vs. the general population), Habits and Attitude toward behavior were the main predictors. The benefit of good indoor air quality was acknowledged by participants with greater habits compared with participants with low ones. Daily intention to opening the windows may stem both from deliberate intention and from less deliberate processes (habits) that are triggered by external cues (time, place, temperature, etc.) in a stable home environment. This result is consistent with studies showing that habits were found to explain differences in intention, indicating that forming intention does not necessarily have to be reasoned. When a strong habit is present, the expression of an intention might be guided by the salience of past behavior (Honkanen et al., 2005).
After controlling for the interaction with Perceived Behavior Control, social approval did not guide willingness to open the windows in our two samples. The TPB literature about health and pro-environmental behavior predicted a positive interaction between Attitudes and Perceived Behavioral Control, because in behavioral change studies Attitudes need a high level of Perceived Behavioral Control to be translated into behavior (La Barbera and Ajzen, 2020). In our two samples, we found a significant interaction between Perceived Behavioral Control and Attitude toward the behavior. Perceived Behavioral Control increased willingness to open the windows only among people with a low level of Attitude toward the behavior. The specificity of the target behavior must be stressed. The repeated manual control of the window is less challenging than difficult behaviors that have been observed in other health and pro-environmental issues (e.g., regular physical exercise, effective waste management). Positive expectations about the benefits of opening windows do not require a high level of perceived control to motivate window opening.
The second aim of our study was to measure the impact of a recommendation from a public health source on intention. Except for a small effect with our student sample, a single recommendation did not improve the intention to open the windows. Beyond a generic recommendation, effective messages need to be targeted at specific people and places (Petty et al., 2009). We also expected that the lack of awareness of indoor air quality might be improved by a focus about one’s own odor sensitivity. Our hypothesis was firmly rejected. In the first study, the persuasion effect of the ventilation recommendation decreased with odor awareness. In the second study, odor awareness decreased intention to open windows. As expected, the assessment of one’s own ability to detect odor decreased perceived vulnerability to poor indoor air quality.
Limitations
Student participants (Study 1) and Web-based volunteers (Study 2) may not allow a generalization to wider populations and wider areas. That is not to say that students are not a vulnerable population at home considering indoor air risks (Lai et al., 2018). A larger and representative sample is needed especially with regard to housing conditions and vulnerabilities associated with knowledge, low income, men and women repartition (sample for both Study 1 and Study 2 are constituted with a majority of women) and education levels. Specifically for Study 1, a larger sample size would have been required due to missing responses on the questionnaires used in the hierarchical regression, so these analyses were conducted on a smaller sample than what is recommended by the power analyses, regarding the effect of a recommendation. Nevertheless, the effect size of the prediction of intention by the TPB model is greater (Sheeran and Webb, 2016). Another indication that could be addressed for a larger sample, is taking account for participants local climate conditions. In the framework of our study, our French sample evolves in a temperate climate. The specific characteristics related to climate differences for people living in area with drastic climate differences were not mobilized into the framework of our study. The extension of this research on a large sample allowing the consideration of the climate of individuals on the behavioral determinants of ventilation could be interesting, since these characteristics have been identified to have an effect on behavior (Wei et al., 2017; Escandón et al., 2019).
Another point of improvement relates to the odor awareness manipulation. In Study 2, the strength of our manipulation of odor awareness was plausibly too weak compared to the first study. The focal message (INPES, 2009) was not targeted toward a specific population with a careful consideration of time, context or motivational background of behavior (Keller and Lehmann, 2008). Finally, we did not address the gap between intention and behavior (Sheeran and Webb, 2016). Nevertheless, the specificity of the behavior (to open the windows for 10 min), the global agreement with target behavior and the fact that people have direct experience with it (Glassman and Albarracin, 2006) may narrow the gap with behavior. Regarding the gap between intention and behavior, it seems difficult to provide general guidance on predictors of intention and its power to predict behavior, particularly in the context of meta-analyses concerned with an overall effect of intention on behavior (Armitage and Conner, 2001; Webb and Sheeran, 2006). The weight of determinants in the prediction of intention and behavior varies depending on the application domain (McEachan et al., 2011). Factors related to the generalized prediction of intention and behavior, the magnitude of whose effects vary considerably across behaviors (Rich et al., 2015). The psychosocial antecedents of the window opening behavior at home has not been explored before our research. Further research with a test of the relationship between intention and behavior would fulfill another main gap in the literature. Interestingly, health recommendations did not change intention beyond a student sample. To further illustrates the link between the intention and the behavior, quantitative measures of the behavior are available. Sensors provide an objective measure of window opening and closing (Guerra-Santin et al., 2016).