comparison of global positioning system (gps) tracking and parent-report diaries to characterize children's time–location patterns - best car alarm with gps tracking
Respondent errors, low resolution, and burden of study participants are known limitations of the diary schedule used in exposure studies, such as the National Human Exposure Assessment Survey (NHEXAS).
Latest developments in the global positioning system (GPS)
Technology has produced tracking devices that are portable enough, powerful and affordable to be used in the exposure assessment science.
In this study, a GPS with differential correction (dGPS)
The tracking device was compared to the NHEXAS diary timeline.
The study also explores how GPS can be used to evaluate and improve this diary schedule by determining which location categories and which respondents are most unlikely to record "correct" Time-Location responses.
A total of 31 children aged 3-5 were wearing a dGPS device for all their waking hours on the weekend, while their parents completed the NHEXAS diary schedule to record the child's time.
Compared to dGPS, parents used the NHEXAS timeline to misclassify the time position of the child, about 48% of the time.
The overall consistency between methods is small (κ=0. 33–0. 35).
Devices for differential GPS found that, on average, children spent 76% of 24-
During study at home.
This diary underestimates the child's time at home by 17%, while overestimating the time spent elsewhere, outside of home, elsewhere and in transit.
The diary data of the time spent outdoors and the time in transit were the lowest in response to the dGPS.
Diary of accommodation-at-
The consistency of mother at home and mother working in unskilled labor with dGPS was lower than that of other participants.
The ability of DGPS to track and collect continuous rather than Classification (ordinal)
Data is also shown.
The conclusion is that automated GPS tracking measurements can improve the quality and collection efficiency of time-location data in exposure assessment studies, although this is for small populations.
All human subjects protocol was approved by the University of Washington Human subjects review committee.
The study was conducted at the end of spring and early summer, as children in the northern climate may go out this season and often move between indoors and outdoors.
In the spring of 2003, a total of 35 families living in the city of Seattle were recruited from the early start-up program.
Sent a letter to the parents (
Spanish, Vietnamese and English)
Starting from four early Start centers, the researchers attended evening meetings at the center to explain the study to their parents.
Choose to start early in order to be able to reach the lowerincome (
150% poverty line and below)
And the minority population in the region, and access
Defining age groups (3–5 years old).
A total of 15 registered families speak only Spanish or mainly Spanish, and all materials and interviews are provided to these families in Spanish.
Register one child per family.
One of the parents of each family filled out the NHEXAS diary schedule for her/his children ().
The NHEXAS time location diary and the agreement to train the interviewer and manage the diary is obtained from one of the authors of the diary (
N Freeman, EOHSI, Rutgers University, Piscataway, New Jersey, USA;
Currently under the United States
Florida, Gainesville, Florida, United States of America).
The schedule and support materials are printed in Spanish and English.
The original diary has seven time location categories.
In this study, two categories of "internal work or school" and "external work or school" were removed because the subjects were participating in non-
The remaining five categories are: "At Home "(INHOME)
"In other places "(INOTHER)
"Outside Home "(OUTHOME)
"In other places "(OUTOTHER)
And "on the way "(TRANSIT).
In addition to the diary schedule, researchers at the University of Washington also worked on housing features, household demographics, and income to determine whether these factors affect responses to diaries or compliance with GPS agreements.
Portable GPS Portable
PAL data record form ()
The children were wearing vests.
The vest allows the antenna to be in the best position on the surface of the horizontal shoulder. GPS-
When you are a good friend, record the data at a time.
During the first home visit, the researchers spent about 45 minutes explaining the purpose of the study, how to fill out the diary, how and when to turn on the GPS-
PAL, and manage the questionnaire.
According to the NHEXAS guidelines, the researchers asked the parents to fill out a practice diary and correct the symbol error.
Practice diaries and notes are kept by parents as a guide to completing the diary on the day of study.
Parents are required to record the time position every 15 minutes.
Respondent parents are instructed to report the time location of the child in real time
Time, or close to the real time
As much time as possible
The respondent's parents were instructed to observe the child throughout the period, rather than relying on the child's description of the time and place.
Family participation on weekends is just to maximize the time for parental supervision.
The participation time is 24 hours. The GPS-
All the waking hours wear PAL.
The parents turned the GPS.
When the child wakes up in the morning, the friend puts on the vest.
The second family visit is usually the day after attending the event.
The researchers checked the diary with their parents and if the total time was not more than 24 hours, they made appropriate corrections.
The researchers asked parents to verify the time of GPS
PAL vest was put on while GPS
PAL vest removed and reported if the child left the house without GPSPAL vest on.
Left the house without wearing a vest (
Or forget to turn on GPS-PAL)
And participate in a non
Children are excluded from the analysis on weekends.
The completed diary is summarized according to the NHEXAS protocol.
For each topic, at a specific time of day, the total time reported for each location category is entered into the database. GPS-
PAL data is also classified into these categories after publication
Use differential correction measurements obtained from the National Earth survey continuous operation reference station to improve resolution (CORS)program.
Use GIS software to map the processed GPS path to a 1-meter-resolution Seattle City forward-shooting Image map (ArcGIS v. 8.
3, ESRI, Redlands, California, United States of America)
And record the location by time.
Records include zooming in the map multiple times to determine the outline of the house, and then creating a polygon around each house so that the points inside are distinguished from those outside.
In addition to the house the child entered, the same procedure was used in other buildings.
The dots inside the Polygans are coded as "at home" or "elsewhere ".
Similarly, a Polly gan was drawn in the yard of each child.
If there is no yard, a patio belonging to an apartment building such as a courtyard or shared outdoor space is considered the same as the yard.
These points are coded as "outside of home ".
In some cases, when the boundary of the building or courtyard cannot be seen clearly in the aerial image, the researchers went to the location and took ground measurements and photographs.
Measurements and photos are then used to help draw boundaries in GIS.
"On the way" is determined by the child's GPS path speed.
The elapsed time to move at more than the possible speed of walking (>5u2009m/s)
It is coded as "on the way ".
The rest of the points are encoded as "elsewhere ".
Ability to show GPS-
Record continuous data and keep the data set in its original format (
Record latitude and longitude about every 5 m/s).
This data is plotted as the average distance and direction per hour from the "tether" position.
In this case, the lanyard is an air monitoring station near the children's home in Seattle.
A key assumption in this study is GPS-
The PAL method is defined here as "the best practical standard ".
In this case, the "best practical standard" means GPS-
PAL provides good resolution and reception while providing children with fully portable and wearable devices.
Based on previous validation studies ()
, This can be defended if the child is wearing a vest and the parent is complying with the agreement.
Only the time between 0900 and 2100 is included in the analysis.
Some mornings shortly after 0900 and nights not long before 2100, there was no GPS data due to receiving, going to bed early or going to bed.
Therefore, if the child sleeps early or sleeps late, if the child lives in a building that prevents GPS reception, it is sometimes necessary to insert proxy GPS data in the evening and in the morning, or both
All children were verified (
Interview through parents)
Kids don't leave home without wearing a vest, GPS-
The PAL was opened before leaving home.
The missing time period is encoded as "at home" except that it is not the first record point far away "(>1u2009km)from the home.
If the subject's first login point on the day was> 1 kilometer away from home, the subject did not meet the criteria for inclusion in the study (
Although not excluded for this reason).
If the subject's GPS recorded more than 6 hours during the 12-hour period from 0900 to 2100 on the day of participation, and if the GPS device was used over the weekend, they would be included in the analysis.
4 children do not meet these inclusion criteria: one family attends on weekdays rather than weekends;
One family forgot to turn on GPS equipment;
The two children did not record a minimum of 6 Thanh on the GPS device.
As a result, a total of 31 subjects met the study criteria and were included in the analysis.
In order to compare the two methods, the measurement and significance test (; ).
The test of meaning is-
One-sided testing of Zero hypothesis = 0 and alternative hypothesis> 0. (
No meaning for negative values).
It is a coefficient in itself, not a statistic.
Based on two proportions: If the method is completely independent, the proportion of the expected consistency, and the proportion of the observed consistency between the two methods.
A misclassification matrix is constructed so that the five time position categories listed by each method are perpendicular to each other, creating a 25 square 5 × 5 grid (and ).
Again, create a 2x2 matrix each time
Location categories are separate compared to all other categories ().
The matrix allows the calculation of a consistent frequency for each possible combination reported by both methods.
Cells falling down the diagonal from the top left to the bottom right are consistent (
Both methods have the same rating).
These determine the proportion of observed consistency ().
Proportion of expected agreement ()
It is calculated under the zero assumption that there is no correlation between the two measurements.
The measurement formulas are: gender, age, first language, housing factors, family demographics, and income tests for reporting consistency of the two types of data through linear regression.
The first type of data is the consistency of sequential scores (
Or proportional measures).
This was tested in two ways: by using the report frequency correctly each time-
The position category is used as the dependent variable and the integer value is used as the dependent variable.
The second type of data is the total time (sum)
In each category.
This was tested using the Spielman grade related test.