Benefits of using GPS sensors in behavioural research

Currently it is still very common to track behaviour via self-report logs or observer logs. This can be expensive and inaccurate. For instance hiring observers for many days can increase costs dramatically and self-reports tend to have a bias of under-reporting and over-reporting certain events. GPS or other tracking technologies can improve or replace existing research tools to track movement.

Preparing a GPS experiment

Avoid obstacles

With GPS you track the movement of people on a global map. For the sensors to work they need a good signal from the GPS satellites. High buildings, trees or other obstacles can block these signals. Therefore you cannot track people within most buildings, between high buildings or within a dense forest, etc. If you plan an experiment you should try to avoid these obstacles. If the experiments takes place over several days and or kilometers this is a minor problem that can be ignored.

Accurate within 20 meters

The sensors have a certain deviation that depends on the quality of the signal but also on the quality of the tracker in use. A basic sensor will be within a range of 20 meter and a good sensor within 5 meter. To draw meaningful conclusions it is important that the participants walk at least several hundred meters so that behavioural changes can be observed within the data.

Test sensor quality and track

Like with all other research methods you should test the experiment before conducting it. This way you can determine weaknesses and remove them. You don’t want to be surprised by a sensor that is not working or an building that blocks you signal.

During an experiment

Cold start

Most sensors need around a minute to get good signals from the GPS satellites. Therefore it is important that the sensors are turned on outside, with as few obstacles as possible and to wait for about 1 minute until the sensors indicate that they receiving the signal.

Log sensor quality

If possible you should test frequently if the sensors are properly working and if necessary replace these sensors.

Sensor signal loss

Depending on the quality of the sensors it is possible that the sensors sometimes need a reset. For instance when the sensor lost the satellite signal and struggles to regain it. You could instruct the participants to check the connectivity and if needed to restart the sensor.

Sensor settings

Some sensors can be programmed to behave in a specific way. For instance with a turn on / off timer, tracking interval, etc.

Our advice would be:

  • Set the tracking interval at high as possible. For instance to save a coordinate every 5 or 10 seconds. The highest possible interval can be calculated by dividing the storage capacity with the duration of the experiment.
  • Disable user input and use the of / off timer. If your sensors offers this option it can be very useful to limit the actions your participants have to do.
  • Disable tracking auto pause. Many trackers stop the location logging if the user is standing still. For experiments this can be a drawback, because afterwards it is difficult to determine if the tracker paused or lost the signal.