The "timestamp" values in Tangerine data are automatically generated whenever the user transitions from one subtest (one screen) to the next. In your data file you will find a "timestamp" for each transition between subtests screens.
The values of each timestamp are provided in Epoch time, in milliseconds. These are the number of milliseconds that passed between January 1, 1970 and the time when the tablet user pressed the "Next" button to pass from one subtest to the next. On the tablet, the time is derived from whichever time was set on the device (in the tablet's own 'Settings').
The Epoch timestamps in your Tangerine data are calculated under the assumption that your user is on standard GMT (GMT + 0). As such, when converting your timestamps to human readable time, you must take into account the offset for the timezone in which the data was collected.
Please see the Tangerine user manual for further guidance or visit http://www.epochconverter.com/ to convert timestamps to human readable data.
Why are these timestamps useful?
These timestamps can provide you granular details into the duration of individual subtests screens (in aggregate or by individual users, to see if your users may be struggling with something), which can be helpful in finding ways to make your data collections more efficient.
If you are using the "Date / Time" subtest in Tangerine, the Epoc timestamps can also offer you a layer of verification. This is because, while the "Date/Time" subtest is visible to the user can be manually manipulated (i.e. the assessor can change the date/time on that particular subtest), the Epoc timestamps are not visible to the data collector and in most cases the data collector will not be aware that they exist. This can be helpful in investigating cases where you suspect a data collector may have attempted to fabricate data or otherwise falsify the date/time when the data was collected.