Programme Performance Assessment in Results-based Management

Session 3 - Collecting Data by Method: Records

Records data like logs, expenditures, website hits, conference service usage require little collection, since they are series that are routinely registered. There is a cost, however, in extracting the data for use in evaluations. The issue here is first determining when to extract data. Some points are obvious. Data should be extracted at the outset of a period, since they will define the baseline. Obviously, data will have to be extracted at the point of the self-evaluation. If, however, trends are to be mapped, you may want to specify more frequent intervals. It is possible, depending on the structure of the registry, to extract the data only once, at the end, and classify it by time. However, if this is likely to be excessively cumbersome, it may be better to have periodic extractions.

The data can be extracted by the owner of the registry, or by the self-evaluation team. This will depend on whether the database has restricted access (as is likely to be the case with accounts) or whether it is open (as in the case of manually entered logs). If it is to be extracted by the owner of the database, this should be agreed in advance so that the appropriate staff time can be set aside.

If the data are in electronic form, it may be possible to set up an automatic process for extraction. This may only require a subroutine in the underlying database programme or a regular extraction that might be permitted by the programme. This would be clearly preferable to manual extraction.