This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR mhealth and uhealth, is properly cited. The complete bibliographic information, a link to the original publication on http://mhealth.jmir.org/, as well as this copyright and license information must be included.
Air quality affects us all and is a rapidly growing concern in the 21st century. We spend the majority of our lives indoors and can be exposed to a number of pollutants smaller than 2.5 microns (particulate matter, PM2.5) resulting in detrimental health effects. Indoor air quality sensors have the potential to provide people with the information they need to understand their risk and take steps to reduce their exposure. One such sensor is the Speck sensor developed at the Community Robotics, Education and Technology Empowerment Lab at Carnegie Mellon University. This sensor provides users with continuous real-time and historical PM2.5 information, a Web-based platform where people can track their PM2.5 levels over time and learn about ways to reduce their exposure, and a venue (blog post) for the user community to exchange information. Little is known about how the use of such monitors affects people’s knowledge, attitudes, and behaviors with respect to indoor air pollution.
The aim of this study was to assess whether using the sensor changes what people know and do about indoor air pollution.
We conducted 2 studies. In the first study, we recruited 276 Pittsburgh residents online and through local branches of the Carnegie Library of Pittsburgh, where the Speck sensor was made available by the researchers in the library catalog. Participants completed a 10- to 15-min survey on air pollution knowledge (its health impact, sources, and mitigation options), perceptions of indoor air quality, confidence in mitigation, current behaviors toward air quality, and personal empowerment and creativity in the spring and summer of 2016. In our second study, we surveyed 26 Pittsburgh residents in summer 2016 who checked out the Speck sensor for 3 weeks on the same measures assessed in the first study, with additional questions about the perception and use of the sensor. Follow-up interviews were conducted with a subset of those who used the Speck sensor.
A series of paired t tests found participants were significantly more knowledgeable (t25=−2.61,
Providing people with low- or no-cost portable indoor air quality monitors, with a supporting Web-based platform that offers information about how to reduce risk, can help people better express perceptions and adopt behaviors commensurate with the risks they face. Thus, thoughtfully designed and deployed personal sensing devices can help empower people to take steps to reduce their risk.
Air quality affects all of us and is a rapidly growing concern in the 21st century [
Pittsburgh, in particular, has a long history of pollution stemming from coal mining and other industrial activities [
Although outdoor air pollution is widely accepted as a problem, indoor air quality can often be overlooked because the level of visible pollution indoors is relatively low. Indoor air pollution can be caused by outdoor contaminants seeping in through windows or poor air filtration systems, or generated from indoor sources such as smoking, cooking, and vacuuming. Many of these sources can produce PM2.5 inside our homes, schools, and offices, but because these particles are so small (about a 30th of the diameter of a human hair or less), they would be invisible to us, except in very high concentrations. We spend the majority of our lifetime in indoor spaces, so our level of exposure to these indoor pollutants can be very high. However, unlike outdoor air pollution, which is a significant challenge to mitigate and requires years to enact necessary air quality regulations, indoor air quality can be managed by anyone.
Although individuals are better able to control their air quality indoors, whether or not individuals or families take action to reduce their risk of exposure to pollution indoors is largely dependent on how they perceive this risk [
By contrast, studies in other domains, notably health, have often found that fear and worry can undermine individuals’ resolve to act, unless they see opportunities for effective action [
There are a growing number of personal sensor technologies, including those that detect ambient PM2.5 levels, available in the market [
Image of the Speck monitor home screen displaying air quality reading.
Little research has been conducted to evaluate the effect of these types of monitors on people’s knowledge, attitudes, and behaviors; however, one study investigating the use of air quality visualizations over a 4-week period with 14 participants did find changes in attitudes and an increase in prohealth behaviors [
Our research objective was to assess whether and how the use of a Speck sensor to monitor indoor air quality empowers people to reduce their risk of exposure to indoor pollutants. To that end, we conducted 2 studies. Our first exploratory study gathers baseline information about what Pittsburgh residents, recruited online or from their local library, generally know and do about indoor air pollution; how confident they are that those actions are effective; and how they would want to learn about it (eg, are indoor air quality sensors appropriate?). Our second study evaluates the effect of using a sensor to monitor indoor air pollution on what people know and do about indoor air pollution among those library patrons from our exploratory first study. These patrons were invited to use the monitor for a period of up to 3 weeks, with their views and behaviors being surveyed after using the monitor. To better explicate our findings, we also interviewed a select subset of those who checked out the monitor from their public library.
After a brief introduction to the study, eligible participants (18 years or older, and living in Pittsburgh) took a 10- to 15-min survey to assess their views and behaviors related to indoor air pollution, and basic demographics.
Knowledge of air pollution was assessed by asking, “How much do you know about air quality?” where 1=none and 5=everything.
The seriousness of perceived health consequences was assessed by asking, “Do you think air quality can cause or make worse the following issues?” Participants were encouraged to check all of the potential issues from a list of 8 items, which included asthma and other respiratory illnesses, heart disease, diabetes, lung cancer, stroke, epilepsy, allergic responses, and others they think may apply. For our analyses, we summed the number of perceived consequences where higher counts indicated greater severity of perceived health consequences (range of 0-8).
Perceived sources of indoor air pollution risks were assessed by asking, “What do you think are some of the sources of pollution inside your home?” Participants were encouraged to check all of the potential sources from a list of 11 items, which included cooking, vacuuming, smoking, microwave oven, gas heating, fireplace, open windows, insulation, pets, refrigerator, or other. For our analyses, we summed the number of perceived sources where a higher count indicates greater severity of perceived risk (range of 0-11).
Perceptions about the number of possible avenues to mitigate risk were assessed by asking, “What do you think are effective ways to reduce your exposure to indoor air pollution?” Participants were encouraged to check all effective ways from a list of 10 items, which included installing a range hood, opening windows, closing windows, installing an air purifier, changing air filters, cleaning the house, smoking outside instead of inside, installing an air quality monitor, cleaning air filters, and others that they think may apply. For our analyses, we summed the number of perceived mitigation strategies where a higher count indicates a greater number of perceived avenues for reducing risk (range of 0-10).
Perceptions of indoor air quality was assessed by asking participants, “On average, how would you rate the air quality in your home?” where 1=very poor and 5=very good.
Confidence in knowing what to do to mitigate risk was assessed by asking, “How confident are you that you will know what actions to take if you learned that your indoor air quality was poor?” where 1=not at all confident and 5=extremely confident.
Behaviors related to improve indoor air quality was assessed by asking participants, “In the past 3 months, have you made any changes in your home to improve the air quality?” where 1=yes, I have; 2=not yet, but I plan to; and 3=no, I have not and do not plan to. For our analyses, we recoded affirmative responses (“yes, I have or not yet, but I plan to”) as 1, and unenthusiastic responses (“no, I have not and don’t plant to”) as 0.
We used Rogers et al’s [
nine statements related to self-esteem and self-efficacy (eg, “I generally accomplish what I set out to do”)
seven statements related to power and powerlessness (eg, “I feel powerless most of the time”)
six statements related to community activism and autonomy (eg, “People have a right to make their own decisions, even if they are bad ones”)
four statements related to optimism and control over the future (eg, “People are limited only by what they think is possible”)
four statements related to righteous anger (eg, “Getting angry about something is often the first step toward changing it”)
We created an overall measure of empowerment by taking the average of all 27 items (Cronbach alpha=.86).
Previous research suggests creativity is inextricably linked to learning and experimentation [
Participants from the Pittsburgh area were recruited using Amazon’s Mechanical Turk, a Web-based survey platform [
Participants reported being on average 36.2 years old (SD 12.26), with 55.6% (149/268) being female, 78.0% (206/264) having at least a college degree, and 44.7% (118/264) with a household income of US $51,000 or greater per year. Most identified as Democrats (119/264, 45.1%), followed by Independents (70/264, 26.5%), Republicans (46/264, 17.4%), Other (14/264, 5.3%), or Prefer Not to Answer (15/264, 5.7%). Most households had at least one child under the age of 18 years living at home (230/267, 86.1%), and of those, 9.0% (24/267) had at least one child under the age of 5 years. Most households also had at least one adult over the age of 65 years living at home (212/264, 80.3%), suggesting that many households were multigenerational. About 21.3% (56/263) of our participants reported that they or someone in their household suffered from a respiratory illness. Overall, the average long-term outdoor PM2.5 levels experienced by our participants were good (mean 10.48, median 9.97, SD 1.85). Of note, the Environmental Protection Agency’s federal long-term (annual average) standard is 15 μg/m3 and short-term (24-hour average) standard is 35 μg/m3 [
Statistical analyses were conducted using Stata version 14 (Stata Corp, College Station, TX, USA). One-sample
Participants checking out the sensor completed the first survey following the same protocol described in study 1. Upon returning the sensor to the library, participants were asked whether they would like to take a 10- to 15-min follow-up survey. They answered the same set of questions as before, with the addition of a few questions regarding their opinions. No compensation was offered for the follow-up survey.
Participants were asked about their views on indoor air pollution, managing indoor air pollution and the Speck sensor, as well as basic demographic questions.
The variables for study 2 were exactly the same as for study 1, with Cronbach alpha for empowerment being .89 and for creativity being .72.
Of the 62 participants who checked out the sensor and completed study 1, 26 agreed to participate in study 2 (attrition rate of 58.1%). Those who agreed to participate in study 2 did not meaningfully differ from those who elected not to participate, based on demographics, baseline knowledge, perceived home air quality, and confidence in ability to mitigate risk. Please refer to
Of the 62 participants who checked out the sensor and completed study 1, 4 agreed to be interviewed. The interviews lasted approximately 1 hour, were audio-recorded, and were transcribed for later analysis.
Participants reported being on average 44.5 years old (SD 12.6), with 61% (14/23) being female, 87% (20/23) having at least a college degree, and 57% (13/23) with a household income of US $51,000 or greater per year. Most identified as Democrats (11/23, 48%), followed by Independents (9/23, 39%), Republicans (1/23, 4%), or other (2/23, 9%). Many households had at least one child under the age of 18 years living at home (9/22, 41%), and of those, all (9/9, 100%) had at least one child under the age of 5 years. Few households also had at least one adult over the age of 65 years living at home (1/23, 4%). About 17% (4/23) of our participants reported that they or someone in their household suffered from a respiratory illness. Overall, the average long-term outdoor PM2.5 levels experienced by our participants were good (mean 10.56, median 10.36, SD 1.17).
One-sample
In general, Pittsburgh residents reported knowing less than the average citizen (mean 2.62, SD 0.75) about indoor air quality (
On balance, most people thought that their indoor air quality is relatively good (mean 3.31, SD 0.72,
One-sample
Variables | Mean (SD) | N | ||
Knowledge | 2.62 (0.75) | 274 | −8.35 (273) | <.001 |
Air quality | 3.31 (0.72) | 274 | 7.15 (273) | <.001 |
Confidence | 2.42 (0.96) | 276 | −10.1 (275) | <.001 |
Percent of participants indicating possible health consequences, sources, and mitigation solutions related to indoor air pollution among the general public.
Survey prompts relating to indoor air quality knowledge | Participants who agreed, n (%) | ||
Asthma | 271 (100.0) | ||
Allergic responses | 263 (97.0) | ||
Lung cancer | 246 (90.8) | ||
Heart disease | 130 (48.0) | ||
Stroke | 68 (25.1) | ||
Epilepsy | 45 (16.6) | ||
Diabetes | 24 (8.9) | ||
Other | 19 (7.0) | ||
Pets | 163 (60.1) | ||
Cooking | 160 (59.0) | ||
Open windows | 147 (54.2) | ||
Gas heating | 140 (51.7) | ||
Vacuuming | 127 (46.9) | ||
Insulation | 111 (41.0) | ||
Fireplace | 73 (26.9) | ||
Refrigerator | 67 (24.7) | ||
Smoking | 66 (24.4) | ||
Microwave oven | 48 (17.7) | ||
Other | 39 (14.4) | ||
Installing air purifier | 239 (88.2) | ||
Changing air filter | 238 (87.8) | ||
Cleaning air filter | 231 (85.2) | ||
Cleaning the house | 217 (80.1) | ||
Installing air quality monitor | 201 (74.2) | ||
Smoking outside instead of inside | 151 (57.9) | ||
Installing range hood | 131 (48.3) | ||
Opening windows | 120 (44.3) | ||
Closing windows | 75 (27.7) | ||
Other | 14 (5.2) |
Among those who are not already interested (checked the sensor out of the library), most people report wanting to know whether their indoor air quality is good or bad (146/195, 74.9%), with those claiming that they would indeed take action being those expressing the most interest in knowing about it (beta=.76,
On balance, interviewed participants reported being interested in using the sensor because of health concerns (4/4, 100%), curiosity (3/4, 75%), and its free availability at the library (1/4, 25%). In general, survey participants viewed the sensor quite favorably. Participants thought that the sensor was more easy to use (mean 4.24, SD 0.97,
A paired
...like, running the vacuum and cooking, and you know, things like that.
After using the sensor, participants attributed indoor air pollution to biological (3/4, 75%), chemical (3/4, 75%), combustion (4/4, 100%), and dust/dander (3/4, 75%) sources and saw it as being much worse in the spring/summer (1/4, 25%) than at other times of the year.
Although we did not observe a significant difference in reported action, our findings suggest a trend toward taking or intending to take action to reduce indoor air pollution after using the sensor (McNemar χ21=2.7
...moved [the sensor] around and tested various behaviors to see if it had any impact [on PMreadings].
It was through this experimentation that participants discovered the impact of cooking (4/4, 100%), movement (2/4, 50%), and vacuuming (3/4, 75%) on indoor air pollution. They also used the sensor to monitor particulate levels in spaces such as their child’s room (Participant G) (3/4, 75%), where vulnerable people spend a lot of time, to make sure that air quality remained good.
We found that people felt more confident about knowing what to do to mitigate their risk after using the sensor (after: mean 2.62, SD 0.94; before: mean 2.31, SD 1.01;
We observed no difference in mean number of reported sources of indoor air pollution and ways to mitigate risk. However, participants we interviewed reported taking new measures they had not tried before to reduce their exposure to indoor air pollution. These included improved pet care and maintenance (1/4, 25%) to reduce dander, a new furnace (1/4, 25%), cleaning more frequently and thoroughly (1/4, 25%), opening windows when cooking (1/4, 25%), and running ventilation systems when necessary (3/4, 75%). Our participants also expressed more concern about the consequences of indoor air pollution after using the sensor than before (after: mean 4.64, SD 1.66; before: mean 3.88, SD 1.56;
I have asthma...how can I improve my own air quality to avoid having an asthma attack?
One-sample
Variables | Mean (SD) | N | ||
Easy | 4.24 (0.97) | 25 | 6.39 (24) | <.001 |
Accurate | 4.21 (0.72) | 24 | 8.21 (23) | <.001 |
Learn | 4.12 (1.01) | 25 | 5.53 (24) | <.001 |
Mitigation | 3.2 (1.15) | 25 | .87 (24) | .40 |
Recommendation | 3.8 (1.15) | 25 | 3.46 (24) | .01 |
Paired-sample
Variables | Before, mean (SD) | After, mean (SD) | N | ||
Knowledge | 2.38 (0.75) | 2.77 (0.71) | 26 | −2.61 (25) | .02 |
Air quality | 2.96 (0.77) | 3.65 (0.75) | 26 | −5.2 (25) | <.001 |
Confidence | 2.31 (1.01) | 2.62 (0.94) | 26 | −1.87 (25) | .07 |
Reported or intended mitigation action after using the sensor for those who did not take or intend to take previous action before using the sensor (no at baseline) and those who did (yes at baseline).
However, not everyone made a change since they found they did not really need to do anything because their indoor air quality was not bad. As a result, they would not adopt any new measures (2/4, 50%) and moreover, one participant said:
I had no idea what I would do if it said it was bad [laughs].
Participants also mentioned a number of barriers to reducing exposure, should the air quality be bad, such as lack of equipment (2/4, 50%), nearby polluters they have no control over (2/4, 50%), and pollution naturally being worse at certain times of year (2/4, 50%). For example:
We don’t have central air...if it’s hot, we need to have the window open.
The sensor was also used outside of the home to help participants learn about their indoor air pollution in other settings (2/4, 25%):
I took it into work so it’s there.
Indeed, one participant was able to use the output from the sensor to pressure building owners to make changes to improve the indoor air quality at work:
The office is right beside a nail salon and they were getting some really powerful smells and so they’re getting on the landlord about “something’s got to give, you know?” My folks can’t suffer like that, so I mean one of the things I’ve been—to be honest with you—one of the bargaining chips was, “well, listen we’re bringing this air monitor up so you’d better get your shit together...,” so they did, and we took the readings up there and they were generally pretty good.
Not only did participants bring the sensor to places outside of the home, they also talked to other people about the sensor, encouraging them to use it (3/4, 75%):
I did tell my parents who live near me that they should check it out and see what their quality looks like.
They also showed other people how to use it (1/4, 25%) and shared what they had learned about air pollution with others (1/4, 25%):
Some of the things we have learned just by seeing them...[I] would like to try to pass it on.
Percent of participants indicating possible health consequences, sources, and mitigation solutions related to indoor air pollution among the sensor users.
Survey prompts relating to indoor air quality knowledge | Participants who agreed, n (%) | |
26 (100) | ||
Lung cancer | 25 (96) | |
Heart disease | 17 (65) | |
Stroke | 10 (39) | |
Epilepsy | 8 (31) | |
Diabetes | 4 (15) | |
Other | 4 (15) | |
Asthma | 1 (4) | |
Cooking | 21 (81) | |
Vacuuming | 18 (69) | |
Open windows | 15 (58) | |
Pets | 14 (54) | |
Gas heating | 14 (54) | |
Refrigerator | 5 (19) | |
Microwave oven | 5 (19) | |
Other | 5 (19) | |
Smoking | 4 (15) | |
Insulation | 3 (12) | |
Fireplace | 2 (8) | |
Cleaning the house | 24 (92) | |
Changing air filter | 23 (89) | |
Installing air purifier | 22 (85) | |
Cleaning air filter | 20 (77) | |
Installing range hood | 18 (69) | |
Installing air quality monitor | 13 (50) | |
Opening windows | 12 (44) | |
Smoking outside instead of inside | 11 (42) | |
Closing windows | 8 (31) | |
Other | 2 (8) |
In general, most people see themselves as knowledgeable about indoor air pollution, the sources of the pollution, and ways to mitigate their risk should they learn that their indoor air quality is poor. Although people report that they believe they have fairly good indoor air quality, they are not completely certain and are generally open to learning about it through the use of a portable indoor air quality monitor. People are willing to pay for such a monitor providing them with information about indoor air quality; however, the amount they are willing to spend is considerably less than that of those currently available. Therefore, making these monitors freely available to the public at a place that is convenient for them, such as at their local public library, is a way to help people access needed tools for informed decision making about indoor air quality.
We found that after using the sensor people reported higher levels of knowledge about indoor air pollution, confidence in their ability to improve indoor air, and improved indoor air quality (possibly as a result of taking mitigation actions). Moreover, we found a significant increase in the number of perceived health impacts after using the sensor, suggesting enhanced perceptions of risk. We also found a positive trend in action-taking among those who already took action before using the sensor
Our findings also suggest that using the sensor was an interactive experience, where participants learned about the link between what they do in their home and what their exposure levels are. There is evidence that this type of experiential learning may be a more powerful way of helping people master new information and suggests a way to enhance motivation to make positive behavior changes [
Although our study has very strong external validity, it is not without its limitations. First, we did not recruit a representative sample of Pittsburgh residents to participate in either study 1 or study 2, and therefore we cannot generalize our findings. However, we were mostly interested in evaluating those individuals most likely to use an indoor air quality monitor when made freely available. Future studies could be conducted to more rigorously evaluate the effect of using such monitors through a randomized controlled trial, allowing for more generalizable findings and a more thorough examination of underlying predictive factors.
Second, we were not able to collect actual exposure data since it was logistically difficult to offload data in real time from every single Speck checked out from a library branch and because of data privacy concerns. However, in this study we were less interested in actual exposure level and more interested in how the information induced changes in perceptions and self-reported behavior. A future study could look at actual PM2.5 levels along with knowledge and behaviors as predictors or covariates to better understand the relationship between these factors and outcomes.
Third, we did not ask our participants in study 2 what actions they took to improve their indoor air quality, nor did we ask them or evaluate which features of the sensor they found to be most persuasive in pursuing the given actions. Due to the design of our study, responses would likely have been subject to recall bias; thus, we did not pursue these lines of questions. However, a future study could ask participants to keep a running log of their activities and changes in behavior (with rationale) related to engagement with the sensor.
Fourth, given our design, we do not know whether learning one time from using the sensor is enough to influence actions over the long term. A future longitudinal study could help determine whether this type of short-term learning can lead to long-term impacts.
Finally, only a small subset of individuals who checked out the sensor from the library agreed to participate in a phone interview for our study. One possible reason is that email-based recruitment from a small sample for a time-consuming activity—with interviews lasting approximately 1 hour in length—is challenging and usually does not yield large numbers. Nonetheless, the researchers believe that the information gathered from the interviews that were conducted does yield insights into people’s perceptions and behaviors. Future studies should collect this valuable qualitative data that allow for deeper understanding of people’s views and actions.
There is much to be hopeful about in these findings. Providing people with low- or no-cost portable indoor air quality monitors with a supporting Web-based platform that offers information about how to reduce risk can help people better express perceptions and adopt behaviors commensurate with the risks they face. Moreover, there appear to be other benefits from engaging in information about indoor air pollution through this experiential means, such as talking to others about the potential risks they may face and using the technology to make positive changes in indoor spaces other than the home. The emerging picture is that thoughtfully and well-designed personal sensor technologies can empower people to take control of the risks that they face and affect positive outcomes in their lives.
Comparison of the demographics of library patron participants who elected to participate in study 2 versus those who did not.
Comparison of knowledge, air quality, and confidence of library patron participants who elected to participate in study 2 versus those who did not.
Logistic regression evaluating whether opening or closing windows as a solution to poor indoor air quality predicts open windows as a source of pollution.
Logistic regression evaluating whether installing a range hood as a solution to poor indoor air quality predicts cooking as a source of pollution.
Logistic regression evaluating whether smoking outside as a solution to poor indoor air quality predicts smoking indoors as a source of pollution.
Logistic regression evaluating whether cleaning as a solution to poor indoor air quality predicts vacuuming as a source of pollution.
odds ratio
particulate matter smaller than 2.5 microns
This work was supported by a research grant from Metro21 at Carnegie Mellon University. The authors thank Toby Greenwalt, Mary Ann McHarg, Rita Botts, and Jody Bell from the Carnegie Public Library for their support with the study.
MT and MBD are consultants with Airviz, Inc, which manufactures Speck air quality monitors. They do not directly benefit from any Speck sales but have a very small share of nonqualified stock options with the company.