Data reliability

I have been looking at Kobo data collection and was very positive about the efficiency, But I’m starting to be sceptical about the use of exported data. How one can be sure that data exported to excel, cleaned, prepared and analysed are based on data that were originally collected? There is a lot of interest in manipulation of data to present an “convenient” results. Is there any mean to verify by an “3th party” that the presented output is based on the “original” data collected?

Welcome to the community, @GJK!

So you mean you would like to know if the submissions were manipulated on the server by a user or not? Did I get you correctly?

Hi Kal_Lam,

Thanks for the quick question,

Both, “manipulation” on the server before extract and of course after (data transfer).

The data collected, frequently by external partners, are saved in Kobo. How fahr the data can be “manipulated” in Kobo I don’t know. But mostly the results are presented in excel after download from Kobo. The first verification of result is the data transfer. The external partner(s) collect the data electronically, and have no record nor real access to Kobo. The registered Kobo user has access and possible option to revise and change the data. The question is: how an 3th party can be sure that the presented extract from Kobo is the original one collected in the field.

@GJK, as of now we don’t have a feature that would keep logs of the changes made to the data from the server. We are soon coming up with a feature that should keep the logs of the actions made in the account.

Maybe you could then come up again if that feature does not fulfill the feature that you intend to have. We could then make improvements to the feature again.

But please feel free to reach us back when the feature is live. We would love to get your feedback on the same.

1 Like

Hi Kal-Lam,

Thanks for the explanation.

The Kobo is great tool and I was recommending it several times for data collection by GHG projects.

But, it comes to the point where the data verification is a key to confirmation/approval.

Without prof of transfer and of use of original data in the analysis the verification of the results becomes impossible.

This becomes a problem in cases where during verification of sampled data (users) different (questionable) results are found.

I will be happy to see how this will be resolved.

1 Like