The method for selecting the LPs and CDs could best be described as "trial and error"! It involved identifying "folk" albums that weren't predominantly singer-songwriter efforts. The criteria for inclusion was that the recording had to feature 3 or more songs not written by the performer, as I was researching the acquisition and adaptation of songs from documented or oral sources. So if it was Cyril Tawney performing Sally Free and Easy then the track was excluded from the research, but if was someone else's interpretation then that was included.

Analysing successive batches of records it became clear that certain artists and labels were more fruitful sources of detailed liner notes, so the later batches were biased towards these in order to get the most detail song comment. This probably led to a more traddy list, as well as the fact that 1970s recordings loomed rather large in the dataset. Needless to say, these skews in the sample were confessed to in the dissertaion.

Like I mentioned, it's a dubious yardstick to measure what was actually being sung in the clubs, and one of the overall impressions from recorded folk output was that performers were understandably seeking to record interesting and unusual songs. There were 3,263 song recordings included in the "weeded" sample- 2,041 different songs in all. 1,472 of these songs were only recorded once. Very few were recorded by more than 3 artists. It was all a bit before my time, but it would seem that in the earlier years of the revival there were a lot of people trying to be obscurer-than-thou!