Our study is only tracking GPS, but on the iPhone it is not working correctly. It is set up correctly, the person doesn’t have a ton of apps open and they have have accepted all notifications. The iPhone is in operation 55% of the time. Does anyone know how to fix this or has had this problem?
Hi @asusnacstudy,
I hope you’re doing well.
Could you please provide me with the user_id(s) of the participant(s) you are referring?
Kind regards,
Omid
AID 89008 and AID 89786
Hi @asusnacstudy,
I’ve reviewed the data for both user 89008 (iPhone) and user 89786 (Android).
For user 89008, the observed gap suggests that the app was terminated. Since the app cannot collect data once it’s terminated, please advise users to grant the ‘Allow Always’ permission for location services, keep the app open in the background, and avoid swiping the app away from recent apps.
For user 89786, the gap in data is attributed to battery optimization settings and app termination. Please advise Android users to disable battery optimization for the Avicenna app and to keep the app open in the background and avoid force-stopping it.
Please note that the app might be terminated by the operating system at any time, depending on the system behaviours.
Hello,
This is our study and our research assistant asked the question. We continue to have challenges with this, so we purchased the application and licenses. We are only getting about 50% of the data, on average.
Do you have any idea what the average is? A bit disappointed in all the limitations that are now coming up, even though we met with the team multiple times about this need.
Hi @aaron.guest
Sorry for the disappointment. I understand your team put a lot of effort in the study and not collecting the data quality that you expect can be very frustrating.
The thing is, the data quality can vary greatly by aspects of the study design like invectives in place, using provisioned-device or bring-your-own-device model, the demographic of participants, and so on.
The reason is, for the app to collect data, it should be running all the time on the participants’ phone. The less the participant interacts with the app, the higher the chances of the app getting terminated. But this varies a lot across different phones. For example, iOS has a consistent termination behaviour (which is its upside), but it aggressively terminates any app, you just have to swipe it off the screen, something many users do. Across Android, app termination is very fragmented. Google has a standard that is followed on Pixel phones, but every manufacturer has their own version as well. For example, Huawei is very aggressive in terminating apps (sometimes even navigation) to preserve battery, Samsung was pretty bad in older versions, and newer ones are more standard, and so on.
So there is no codebase that guarantees permanent data collection across all phone makes and models. The best solution is:
- Design the surveys in a way participants open the app multiple times a day. This way the app is active most of the time.
- Incentivize participants to ensure the app is running on the phone (our app tries to detect when its terminated and notify participants).
- Use wearables.
Depending on how the above is implemented in the study protocol, the data collection rate can be as high as 85% or as low as 30%.
Hope it provides more context.
Mohammad
1 Like
Thanks so much for the reply. We have looked into using wearables, but have been unable to find one consistently collecting data as most of them only collect active GPS data - as individuals have to activate the data collection device. We have implemented multiple check-in points throughout the day, to encourage participants to open the app. However, even when doing that, data collection still seems to stop shortly after it is opened, even if it is still running on the back of the phone.
Have you checked if Low Power Mode or background app refresh settings are limiting GPS tracking? Also, ensure location access is set to ‘Always’ in the app permissions to allow consistent tracking, even in the background. Hope this helps! 