Unfortunately I think the sample bias that you mentioned makes this useless as any sort of statistical study. The one useful result I'd take from this is "succeeding as a single founder is possible".
Exactly. And maybe one non statistical result as well: A significant number of very successful people _believed_ that it was possible, otherwise they wouldn't have tried to go it alone.
Five times as many must have believed they needed cofounders. Either way I still think the take home message can be that successful single founder companies aren't an insignificant bizarro event.
agreed: almost useless as a guide for prospective founders unless it takes into account the statistics for ventures that _failed_. For this reason I would be more interested in the stats for ycombinator funded companies that either succeeded or threw in the towel.
