Attention, Attention Impaired, Decision Making, Risk Behavior, Risk-Taking, Social Preferences
Conditions
Keywords
Attention, Air pollution, Black carbon, Particulate matter, Risk behavior, Social preferences, Decision making, Air purifier, Air cleaner
Brief summary
Previous observational studies have reported an association between higher air pollution exposure and lower attention in children. With this project, the investigators aim to confirm this association in adolescents using an experimental design. In addition, the study will assess the relationship between air pollution exposure and individual preferences with respect to risk, time and social considerations. High school students in 3rd grade (ESO, 14-15 years of age) in different high schools in the Barcelona province (Spain) will be invited to participate. For each class in each high school, participating students will be randomly split into two equal-sized groups. Each group will be assigned to a different classroom where they will complete several activities during two hours, including an attention test (Flanker task) and a reduced version of the Global Preferences Survey. One of the classrooms will have an air purifier that will clean the air. The other classroom will have the same device but without the filters, so it will only re-circulate the air without cleaning it. Students will be masked to intervention allocation. The investigators hypothesize that students assigned to the clean air classroom will have better scores in the attention test, and that decision-making will also present differences in the two classrooms.
Interventions
OTHERPurifying the air with a Pure Airbox device (Zonair 3D)
Purifying the air of the classroom where the experiment is conducted using a Pure Airbox device (Zonair 3D)
OTHERUsing a sham air purifier (same device without filters)
Use the air purifier (Pure Airbox, Zonair 3D) without filters in the classroom where the experiment is conducted
Sponsors
Recercaixa
Centre de Recerca per a l'Educació Científica i Matemàtica (CRECIM)
Institute of Environmental Assessment and Water Research (IDAEA-CSIC)
Economics and Bussiness Department, Universtitat Pompeu Fabra
Barcelona Institute for Global Health
Study design
Allocation
RANDOMIZED
Intervention model
PARALLEL
Primary purpose
PREVENTION
Masking
SINGLE (Subject)
Eligibility
Sex/Gender
ALL
Age
13 Years to 16 Years
Healthy volunteers
Yes
Inclusion criteria
* Students in the 3rd ESO course in participating high schools with signed informed consent
Exclusion criteria
* None
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Trust Score | Obtained from a test administered approximately 100 minutes after entering the classroom | Answer to the question I assume that people have only the best intentions, which is ranked from 0 to 10. Higher values imply more trust. |
| Response Speed Consistency Throughout the Attention Network Task-Flanker Task (Post ANT) | Obtained from a test administered approximately 90 minutes after entering the classroom | This is a measure related to attentiveness that is calculated as hit reaction time standard error for correct responses (HRT-SE). A higher HRT-SE indicates highly variable reactions during the ANT test that is related to inattentiveness. A lower HRT-SE indicates indicates consistent reaction times and thus, a better attention performance. |
| Combined Risk Taking Score | Obtained from a test administered approximately 100 minutes after entering the classroom | The Combined risk taking score measures how willing the person is to take risks and is calculated as 0.4729985 × Risk preference score + 0.5270015 × Willingness to take risks, as defined in Falk et al (2018). The Combined risk taking score is a weighted average of two z-scores (risk preference score and willingness to take risks score), and therefore it can be interpreted as a z-score. The central value of Combined risk taking score is 0 and represents the population mean. The higher the values of the Combined risk taking score, the more willing is the person to take risks. The Risk preference score is the final node in the risk tree (see the online appendix of Falk et al (2018)), which is a value between 1 and 32, ranked by the level of risk aversion, which was then transformed to a z-score. The Willingness to take risks refers to the answer to the question on one's willingness to take risks, with values from 0 to 10), which was then transformed to a z-score. |
| Combined Patience Score | Obtained from a test administered approximately 100 minutes after entering the classroom | The Combined patience score measures the patience of the person and is calculated as 0.7115185 × Time preference score + 0.2884815 × Self assessment of patience, as defined in Falk et al (2018). The Combined patience score is a weighted average of two z-scores (time preference score and self assessment of patience score), and therefore it can be interpreted as a z-score. The central value of Combined patience score is 0 and represents the population mean. The higher the values of the Combined patience score the more patience the person has. The Time preference score was obtained from the final node in the time tree (see the online appendix of Falk et al (2018)), which is a value between 1 and 32, ranked by the level of patience, which was then transformed to a z-score. The Self assessment of patience score refers to the answer to the question on the self-assessment of patience, with values from 0 to 10, which was then transformed to a z-score. |
| Positive Reciprocity Score | Obtained from a test administered approximately 100 minutes after entering the classroom | The Positive reciprocity score measures positive reciprocity. A person with high reciprocity agrees with the statement When someone does me a favor, I am willing to return it. The score is calculated as 0.4847038 × Willingness to return favor + 0.5152962 × Size of gift, as defined in Falk et al (2018). The Positive reciprocity score is a weighted average of two z-scores (willingness to return a favor score and size of gift score), and therefore it can be interpreted as a z-score, with the central value of 0 representing the population mean. Higher values of the score imply more positive reciprocity. Willingness to return favor refers to the answer to the question on the willingness to return a favor, ranked from 0 to 10 and then transformed to a z-score. Size of gift refers to the answer to the question on the reported size of the gift that would return to a stranger, from a list of 7 possible values (from 0 to 30€, by 5€ steps), then transformed to a z-score. |
| Altruism Score | Obtained from a test administered approximately 100 minutes after entering the classroom | The Altruism score measures the altruism of a person and is calculated as 0.6350048 × Willingness to give to good causes + 0.3649952 × Hypothetical donation, as defined in Falk et al (2018). The Altruism score is a weighted average of two z-scores (willingness to give to good causes score and hypothetical donation score), and therefore it can be interpreted as a z-score, with the central value of 0 representing the population mean. Higher values of the score imply more altruism. Willingness to give to good causes refers to the answer to the question on the willingness to give to good causes, ranked from 0 to 10 and then transformed to a z-score. Hypothetical donation refers to the answer to the question on the size of the hypothetical donation one would be willing to make, transformed to a z-score. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Impulsivity (From ANT) | Obtained from a test administered approximately 90 minutes after entering the classroom | Number of incorrect responses (responses made in the opposite direction to the direction of the target arrow) Lower score indicates a better attention performance. Total range score: 0-38. |
| Selective Attention (From ANT) | Obtained from a test administered approximately 90 minutes after entering the classroom | Number of omission errors (failure to respond) Lower score indicates a better attention performance Total range score: 0-21. Higher values represent a worse outcome |
| Alerting Score (From ANT) | Obtained from a test administered approximately 90 minutes after entering the classroom | Alerting represents the ability to maintain a state of high vigilance to incoming stimuli It is calculated by subtracting the median reaction time for double cue from median reaction time for the no cue condition (calculations performed after removing the incongruent trials). A lower score indicates a better attention performance |
| Orienting Score (From ANT) | Obtained from a test administered approximately 90 minutes after entering the classroom | Orienting is the ability to select information and it is calculated by subtracting the median reaction time for spatial cue from the reaction time for central cue (calculations performed after removing the incongruent trials).A lower score indicates a better attention performance |
| Conflict Score (Executive Attention) (From ANT) | Obtained from a test administered approximately 90 minutes after entering the classroom | Conflict generally refers to the ability to monitor and resolve conflict among responses and in this study it is calculated as the median reaction time for each flanker condition (across cue conditions) and subtracting the congruent from the incongruent reaction times. A lower score indicates a better attention performance |
| Self Assessment of How Good They Are in Math | Obtained from a test administered approximately 100 minutes after entering the classroom | Answer to the question on math skills, ranging from 0 to 10. The higher the score, the better the person feels s/he is doing in math. |
Countries
Spain
Participant flow
Recruitment details
To recruit participants (students), the objectives and design of the study were presented during teacher training sessions opened to all science teachers from all high schools of Barcelona metropolitan area. Teachers interested in participating signed up in an online form. Teachers from 33 high schools reached out to the research team to participate.
Pre-assignment details
Based on 33 participating high schools, 2670 students were estimated eligible. As the total number of eligible students was not retrieved at the onset of the study, this estimation was based on the maximum number of students per class (n = 30), and thus it was likely to be overestimated. Of these, 547 individuals were excluded prior to randomization (parents or legal guardians who did not sign the consent form or students who declined to participate).
Participants by arm
| Arm | Count |
|---|---|
| Purified Air Purifying the air with a Pure Airbox device (Zonair 3D). Use of air purifier (Pure Airbox, Zonair 3D) in the classroom 30 minutes before the participants enter the room and during the 2 hours of the experiment.
Purifying the air with a Pure Airbox device (Zonair 3D): Purifying the air of the classroom where the experiment is conducted using a Pure Airbox device (Zonair 3D) | 999 |
| Normal Air Using a sham air purifier (same device without filters). Use of the same air purifier but without filters, so that it only recirculates the air without purifying it. Used for the same time period than the other arm.
Using a sham air purifier (same device without filters): Use the air purifier (Pure Airbox, Zonair 3D) without filters in the classroom where the experiment is conducted | 994 |
| Total | 1,993 |
Withdrawals & dropouts
| Period | Reason | FG000 | FG001 |
|---|---|---|---|
| Overall Study | Appointment to the doctor | 1 | 1 |
Baseline characteristics
| Characteristic | Purified Air | Normal Air | Total |
|---|---|---|---|
| Age, Continuous | 14.7 years | 14.7 years | 14.7 years |
| Alerting | 34.5 ms | 35.5 ms | 34.5 ms |
| Conflict | 78 ms | 76.5 ms | 77 ms |
| HRT-SE, hit reaction time standard error | 164.1 ms | 158.6 ms | 161.7 ms |
| Impulsivity | 2 total score | 2 total score | 2 total score |
| Orienting | 14 ms | 14.8 ms | 14 ms |
| PMA-R, Primary Mental Aptitudes - Reasoning | 17 score | 16 score | 16 score |
| Race and Ethnicity Not Collected | — | — | 0 Participants |
| Selective attention | 0 total score | 0 total score | 0 total score |
| Sex/Gender, Customized Sex Female | 514 Participants | 483 Participants | 997 Participants |
| Sex/Gender, Customized Sex Male | 481 Participants | 507 Participants | 988 Participants |
| Sex/Gender, Customized Sex Other | 4 Participants | 4 Participants | 8 Participants |
Adverse events
| Event type | EG000 affected / at risk | EG001 affected / at risk |
|---|---|---|
| deaths Total, all-cause mortality | 0 / 0 | 0 / 0 |
| other Total, other adverse events | 0 / 0 | 0 / 0 |
| serious Total, serious adverse events | 0 / 0 | 0 / 0 |
Outcome results
Primary
Altruism Score
The Altruism score measures the altruism of a person and is calculated as 0.6350048 × Willingness to give to good causes + 0.3649952 × Hypothetical donation, as defined in Falk et al (2018). The Altruism score is a weighted average of two z-scores (willingness to give to good causes score and hypothetical donation score), and therefore it can be interpreted as a z-score, with the central value of 0 representing the population mean. Higher values of the score imply more altruism. Willingness to give to good causes refers to the answer to the question on the willingness to give to good causes, ranked from 0 to 10 and then transformed to a z-score. Hypothetical donation refers to the answer to the question on the size of the hypothetical donation one would be willing to make, transformed to a z-score.
Time frame: Obtained from a test administered approximately 100 minutes after entering the classroom
| Arm | Measure | Value (MEDIAN) |
|---|---|---|
| Purified Air | Altruism Score | 0.1118 Z-score |
| Normal Air | Altruism Score | 0.0597 Z-score |
Comparison: The hypothesis was that air filtration could result in higher Altruism behavior, relative to the classroom without air filter (normal air), in the Altruism measures among adolescents in high schools.p-value: 0.5895% CI: [-0.0595, 0.1035]Regression, Linear
Primary
Combined Patience Score
The Combined patience score measures the patience of the person and is calculated as 0.7115185 × Time preference score + 0.2884815 × Self assessment of patience, as defined in Falk et al (2018). The Combined patience score is a weighted average of two z-scores (time preference score and self assessment of patience score), and therefore it can be interpreted as a z-score. The central value of Combined patience score is 0 and represents the population mean. The higher the values of the Combined patience score the more patience the person has. The Time preference score was obtained from the final node in the time tree (see the online appendix of Falk et al (2018)), which is a value between 1 and 32, ranked by the level of patience, which was then transformed to a z-score. The Self assessment of patience score refers to the answer to the question on the self-assessment of patience, with values from 0 to 10, which was then transformed to a z-score.
Time frame: Obtained from a test administered approximately 100 minutes after entering the classroom
| Arm | Measure | Value (MEDIAN) |
|---|---|---|
| Purified Air | Combined Patience Score | -0.1453 Z-score |
| Normal Air | Combined Patience Score | -0.0732 Z-score |
Comparison: The hypothesis was that air filtration could result in higher level of patience, relative to the classroom without air filter (normal air), in the combined patience measures among adolescents in high schools.p-value: 0.595% CI: [-0.117, 0.0584]Regression, Linear
Primary
Combined Risk Taking Score
The Combined risk taking score measures how willing the person is to take risks and is calculated as 0.4729985 × Risk preference score + 0.5270015 × Willingness to take risks, as defined in Falk et al (2018). The Combined risk taking score is a weighted average of two z-scores (risk preference score and willingness to take risks score), and therefore it can be interpreted as a z-score. The central value of Combined risk taking score is 0 and represents the population mean. The higher the values of the Combined risk taking score, the more willing is the person to take risks. The Risk preference score is the final node in the risk tree (see the online appendix of Falk et al (2018)), which is a value between 1 and 32, ranked by the level of risk aversion, which was then transformed to a z-score. The Willingness to take risks refers to the answer to the question on one's willingness to take risks, with values from 0 to 10), which was then transformed to a z-score.
Time frame: Obtained from a test administered approximately 100 minutes after entering the classroom
| Arm | Measure | Value (MEDIAN) |
|---|---|---|
| Purified Air | Combined Risk Taking Score | -0.0210 Z-score |
| Normal Air | Combined Risk Taking Score | -0.0121 Z-score |
Comparison: The hypothesis was that air filtration could result in higher riskloving behavior, relative to the classroom without air filter (normal air), in the combined risk taking measures among adolescents in high schools.p-value: 0.7295% CI: [-0.0607, 0.0865]Regression, Linear
Primary
Positive Reciprocity Score
The Positive reciprocity score measures positive reciprocity. A person with high reciprocity agrees with the statement When someone does me a favor, I am willing to return it. The score is calculated as 0.4847038 × Willingness to return favor + 0.5152962 × Size of gift, as defined in Falk et al (2018). The Positive reciprocity score is a weighted average of two z-scores (willingness to return a favor score and size of gift score), and therefore it can be interpreted as a z-score, with the central value of 0 representing the population mean. Higher values of the score imply more positive reciprocity. Willingness to return favor refers to the answer to the question on the willingness to return a favor, ranked from 0 to 10 and then transformed to a z-score. Size of gift refers to the answer to the question on the reported size of the gift that would return to a stranger, from a list of 7 possible values (from 0 to 30€, by 5€ steps), then transformed to a z-score.
Time frame: Obtained from a test administered approximately 100 minutes after entering the classroom
| Arm | Measure | Value (MEDIAN) |
|---|---|---|
| Purified Air | Positive Reciprocity Score | -0.0186 Z-score |
| Normal Air | Positive Reciprocity Score | -0.0359 Z-score |
Comparison: The hypothesis was that air filtration could result in higher level of positive reciprocity behavior, relative to the classroom without air filter (normal air), in the positive reciprocity measures among adolescents in high schoolsp-value: 0.1395% CI: [-0.0184, 0.1394]Regression, Linear
Primary
Response Speed Consistency Throughout the Attention Network Task-Flanker Task (Post ANT)
This is a measure related to attentiveness that is calculated as hit reaction time standard error for correct responses (HRT-SE). A higher HRT-SE indicates highly variable reactions during the ANT test that is related to inattentiveness. A lower HRT-SE indicates indicates consistent reaction times and thus, a better attention performance.
Time frame: Obtained from a test administered approximately 90 minutes after entering the classroom
| Arm | Measure | Value (MEDIAN) |
|---|---|---|
| Purified Air | Response Speed Consistency Throughout the Attention Network Task-Flanker Task (Post ANT) | 161.3 ms |
| Normal Air | Response Speed Consistency Throughout the Attention Network Task-Flanker Task (Post ANT) | 160.2 ms |
Comparison: The hypothesis was that air filtration could result in better performance, relative to the classroom without air filter (normal air), in attention measures among adolescents in high schools.p-value: 0.5295% CI: [-2.81, 5.56]Conditional linear regression
Primary
Trust Score
Answer to the question I assume that people have only the best intentions, which is ranked from 0 to 10. Higher values imply more trust.
Time frame: Obtained from a test administered approximately 100 minutes after entering the classroom
| Arm | Measure | Value (MEDIAN) |
|---|---|---|
| Purified Air | Trust Score | 5 score on a scale |
| Normal Air | Trust Score | 5 score on a scale |
Comparison: The hypothesis was that air filtration could result in higher trust, relative to the classroom without air filter (normal air), in the subjective trust measure among adolescents in high schools.p-value: 0.8995% CI: [-0.1786, 0.2049]Ordered logistic regression
Secondary
Alerting Score (From ANT)
Alerting represents the ability to maintain a state of high vigilance to incoming stimuli It is calculated by subtracting the median reaction time for double cue from median reaction time for the no cue condition (calculations performed after removing the incongruent trials). A lower score indicates a better attention performance
Time frame: Obtained from a test administered approximately 90 minutes after entering the classroom
| Arm | Measure | Value (MEDIAN) |
|---|---|---|
| Purified Air | Alerting Score (From ANT) | 42 ms |
| Normal Air | Alerting Score (From ANT) | 44.5 ms |
p-value: 0.295% CI: [-8.51, 1.74]Conditional linear regression
Secondary
Conflict Score (Executive Attention) (From ANT)
Conflict generally refers to the ability to monitor and resolve conflict among responses and in this study it is calculated as the median reaction time for each flanker condition (across cue conditions) and subtracting the congruent from the incongruent reaction times. A lower score indicates a better attention performance
Time frame: Obtained from a test administered approximately 90 minutes after entering the classroom
| Arm | Measure | Value (MEDIAN) |
|---|---|---|
| Purified Air | Conflict Score (Executive Attention) (From ANT) | 75 ms |
| Normal Air | Conflict Score (Executive Attention) (From ANT) | 75 ms |
p-value: 0.4695% CI: [-4.94, 2.22]Conditional linear regression
Secondary
Impulsivity (From ANT)
Number of incorrect responses (responses made in the opposite direction to the direction of the target arrow) Lower score indicates a better attention performance. Total range score: 0-38.
Time frame: Obtained from a test administered approximately 90 minutes after entering the classroom
| Arm | Measure | Value (MEDIAN) |
|---|---|---|
| Purified Air | Impulsivity (From ANT) | 2 total score |
| Normal Air | Impulsivity (From ANT) | 2 total score |
p-value: 0.595% CI: [-0.49, 0.24]Conditional linear regression
Secondary
Orienting Score (From ANT)
Orienting is the ability to select information and it is calculated by subtracting the median reaction time for spatial cue from the reaction time for central cue (calculations performed after removing the incongruent trials).A lower score indicates a better attention performance
Time frame: Obtained from a test administered approximately 90 minutes after entering the classroom
| Arm | Measure | Value (MEDIAN) |
|---|---|---|
| Purified Air | Orienting Score (From ANT) | 12.5 ms |
| Normal Air | Orienting Score (From ANT) | 13.5 ms |
p-value: 0.9295% CI: [-5.1, 4.6]Conditional linear regression
Secondary
Selective Attention (From ANT)
Number of omission errors (failure to respond) Lower score indicates a better attention performance Total range score: 0-21. Higher values represent a worse outcome
Time frame: Obtained from a test administered approximately 90 minutes after entering the classroom
| Arm | Measure | Value (MEDIAN) |
|---|---|---|
| Purified Air | Selective Attention (From ANT) | 0 total score |
| Normal Air | Selective Attention (From ANT) | 0 total score |
p-value: 0.4795% CI: [-0.2, 0.1]Conditional linear regression
Secondary
Self Assessment of How Good They Are in Math
Answer to the question on math skills, ranging from 0 to 10. The higher the score, the better the person feels s/he is doing in math.
Time frame: Obtained from a test administered approximately 100 minutes after entering the classroom
| Arm | Measure | Value (MEDIAN) |
|---|---|---|
| Purified Air | Self Assessment of How Good They Are in Math | 6 score on a scale |
| Normal Air | Self Assessment of How Good They Are in Math | 6 score on a scale |
Comparison: The hypothesis was that air filtration could result in higher assessment of math skills, relative to the classroom without air filter (normal air), in the self assessed math skills measure among adolescents in high schools.p-value: 0.23295% CI: [-0.382, 0.0928]Ordered logistic regression