Come on people, population reports can absolutely provide useful information/trends
IF:
1) there is a sufficient sample size
2) avoid apples / oranges comparisons
There are always exceptions, however there is a wealth of information that can be gleaned from the 7,740 OJs that have been slabbed by SGC. An example of a good comparison would be to evaluate relative scarcity of HOFers within the set. Do not use the population report to provide an exact ranking, but instead provide a trend of most difficult and most common subjects. Likewise, as a Detroit collector I can look at population report for Stump Wiedman and Frank Scheibeck or Lady Baldwin and surmise that the 1888 Detroit OJs are more difficult than 1887 Detroits.
Keefe now has nearly 100 graded OJs (including his double player cards) and is the most common HOFer. I can say with high degree of confidence that he is more common than those with half as many graded examples (Welch, O'Rourke, Brouthers, Thompson, Clarkson, etc.). However, when looking at the four lowest HOF pops, I'd say there isn't a high enough sample size to definitively rank them (White, Robinson, Nichols, & McPhee).
An exception/risk would be if someone like John Drecker opted to send in his 50 or so Corcoran cards to all be graded. That would make the card appear far more common than it really is. I'm guilty of this to a lesser extent in that I only collect Detroits and have all my cards graded which artificially inflates the Detroit players population reports.
Here is an article I wrote back in 2005 when SGCs N172 population report was less than 1/3 of today's total yet the HOF rankings remain similar.
http://www.oldcardboard.com/eNews/20.../eNews17.htm#5