Note that the article is only about securities fraud. From the abstract I assume they looked at the group that was reported to the SEC for fraud and compared t

Note that the article is only about securities fraud. From the abstract I assume they looked at the group that was reported to the SEC for fraud and compared to a random selection of 500 companies and then compared what percentage amongst those two groups were big corporations, growth pressures etc.. Regarding the control for undetected fraud, there could be various methods that one can employ (e.g. when one has some statistics about undetected fraud for example). From the formulation in the abstract I assume that when they apply the controls their findings don't change, so with or without the controls large corporations are more likely to commit fraud.