The focus in this thread on the single issue of who did or didn't compose what we have come to call 'traditional songs' obscures the width and depth of this book. It is much more than a single issue work and has many implications. It does not set out to rubbish the achievements of earlier researchers and credits the scholars whose work built the foundations of folk song study. Roud does regret instances amongst the pioneers' work where he can produce evidence that they brought a preconceived agenda to their work or where he can point to assumptions made on insufficent evidence, eg, Sharp writing Some Conclusions.... a relatively short time after hearing a traditional song sung for the first time.
The attitude towards only using 'evidence' in historical research is now the norm amongst modern academics not just in folk song studies.
The research involved in Roud's book is meticulous and has been made over four decades of deep interest in the subject. It is part of a very wide interest and knowledge in a whole range of popular musics. I know that he was surprised recently to be asked to deliver a series of lectures on the history of rock music to London-based college of an American university. He told me that he was amazed how little knowledge the students had in an aspect of a subject that they were majoring in.
Throughout the book he bemoans the lack of solid information currently available and calls for more and deeper research on the subject. I defy anyone to provide anything in the book to challenge the rigorous discipline on his fact-based approach. The widespread praise and admiration expressed in the reviews that I and others have posted or linked to here is fully justified.