My division by industry is a necessary refinement of your method, though.
The “recognition” metric is going to select for consumer-facing businesses and away from companies that sell to other companies. This will mean it’s deeply skewed between industries: we all know our airlines and utility suppliers, but even very big miners, engineering firms, or oil companies might be completely invisible to the public.

Unfortunately, though, “industries trend but stocks revert”—the big secular trends in the economy tend to move all the stocks in a particular industry in the same direction at once, over a long period, while news affecting an individual firm pulls it briefly away from that trend but usually dissipates quite rapidly.

This means that your results will depend radically on whether you perform your test over a period when the consumer is strong or weak. Your time period is indeed over a weak-consumer era. I suspect your comparator study was conducted during the consumer boom.

The solution is to split by industry and within each industry sector to look at whether public knowledge is systematically correlated with better performance relative to the industry.

Of course any retrospective study will have a survivor bias: names that dropped out of the FTSE (because they stopped making money) will be unfamiliar to your 2012 readership. So it’s very hard to use this sort of thing to compare with live data and make an effective investment strategy.

(Though even the slightest advantage, among the very liquid FTSE100 shares, could be cheaply and easily traded.)

There’s a somewhat related commentary article in last December’s Science magazine (Science, vol. 334, p.1503) entitled “Can Ignorance Promote Democracy”: When a group needs to reach a consensus decision, uninformed members can help to reduce the influence of a manipulative minority. I think there’s also a research paper attached to it. Let me know if you want a copy.

Cheers,

Matthew