There are 6 television programmes of different TV genres and their mean and mean ranks are put for each of them. But the reliability value of all the programmes together is 0.498 . So what exactly we infer from this 0.498 reliability value?
From the table in the provided paper, it seems that the "reliability" term is the "R" correlation factor of the data. Statistically, any "R" less than 0.8 (some may go lower) makes any correlation doubtful. Certainly, a value of 0.498 would mean that the data are randomly distributed and uncorrelated.
With regards to the research question, it just means that none of the TV surveys really agree with each other. The stated conclusion has nothing to do with either the initial intent nor the data collected.
From the documents itself it is mentioned that this is significant consistancy in overall watching all the six programmes, which he has taken for showing the reliability