Precisely. If 2-person and 1-person startups are equally likely to succeed but 2-person startups just happen to be more likely then they will be over-represented in the data.
What is really necessary is a statistically significant sampling of startups from their inception and tracking the percentage of success and failure depending on various factors. There is simply not enough data in the study at hand to say that 1-person startups are less likely to succeed than 2-person startups, in fact it's quite possible that they are more likely to succeed, we just don't know.
Comments
Precisely. If 2-person and 1-person startups are equally likely to succeed but 2-person startups just happen to be more likely then they will be over-represented in the data.
What is really necessary is a statistically significant sampling of startups from their inception and tracking the percentage of success and failure depending on various factors. There is simply not enough data in the study at hand to say that 1-person startups are less likely to succeed than 2-person startups, in fact it's quite possible that they are more likely to succeed, we just don't know.