How atypical are the editions in Eric White's census of print runs?

In the introduction to his census of known fifteenth-century print runs, Eric White cautions against taking his results as representative for all incunables:

As the census will make immediately apparent, a large percentage of editions for which we know the print runs were produced to fulfill institutional functions.... Remunerative and relatively risk-free for printers, the original commissions for projects such as these tended to end up in surviving archives, and they tended to afford very large editions. It should be noted, therefore, that the print runs known from such institutional commissions do not represent a normative cross-section of fifteenth-century press production, but rather a selection of large scale projects carried out with institutional funding and pressure to produce. As a group they almost certainly reflect higher-than-average print runs.... Moreover, the majority of the recorded print runs reflect the output not of the ‘average’ printing shop, but rather that of a few exceptionally successful publishers who received commissions from well-funded institutions. It is worth remembering that a documented print run may not be a representative print run.

White's characterization of his sample is correct. Compared to all recorded incunables in the ISTC, folios are much more prevalent in the print run census, while quartos are underrepresented, and broadsides do not appear at all.

Comparison of format distribution 

For each format, the books are also substantially longer, with the average number of leaves 60-100% higher than for the ISTC as a whole. (NB: Averages can be a misleading way to describe the distribution of leaf counts, but they give a correct impression in this case.)

Comparison of average leaf count by format

White's suggestion that the sample of known print runs enjoyed a better survival rate than other incunables is also correct, with an average number of surviving copies 20-65% higher than what one finds for the ISTC as a whole. (NB: Averages can be even more misleading for describing survival rates.)

Comparison of average surviving copies by format

This doesn't mean that we should ignore White's census of print runs as an atypical sample, however. Rather, we can say that its sample differs from the body of known incunables in various ways, some of which have well-understood effects. For example, the size and format of editions in White's sample are larger on average than for the ISTC as a whole, and the included editions likely benefited from association with an institutional sponsor, all of which are associated with higher survival rates than other fifteenth-century printed books, so that we would expect the survival rate for White's sample to be higher than for the ISTC as a whole.