The significance of the finding "patents with higher reverse citation counts are less likely to survive review "?

Within s post titled Which Patents Get Instituted During Inter Partes Review ? , Michael Risch discusses a paper on SSRN titled Determinants of Patent Quality: Evidence from Inter Partes Review Proceedings, with a reference to an observation:


•Patents with more backward citations are more likely to be instituted (this is counterintuitive, but consistent with my own study of the patent litigation)



The abstract of the SSRN paper actually states:



patents with higher reverse citation counts are less likely to survive review


and from the body of the SSRN paper



Patents with more reverse citations and patents with more reverse citations added by the examiner are both more likely to be instituted, with an additional 10 reverse citations associated with a 0.15 percent increase in the chance of institution, and an additional 10 reverse citations added by the examiner associated with an impressive 1.8 percent increase in the chance of institution


AND


We next identified all prior art references that were cited during the patent’s prosecution (often referred to as “reverse citations”). In addition to determining the overall count of such citations, we also determined the number of reverse citations to foreign patents, as well as the number and type of reverse citations to “non-patent literature” (NPL) such as academic articles, books, and websites. Finally, for all patents issued in 2001 or thereafter, we were additionally able to determine whether reverse citations to patents and applications were disclosed by the applicant or, instead, were identified and cited by the examiner in an office action.185




As to terminology, a "reverse" citation is a reference in the patent to a document in existence BEFORE patent issue. A "forward" citation is a reference TO the issued patent by a document coming later in time.

As to expectations, the SSRN article wrote:


Moving next to data that proxies the scrutiny each application received from the USPTO, we again find a number of significant correlations with institution. First, as shown below in Table 10, we find a significant correlation between institution and various categories of “reverse citations.” While one might expect institution to be negatively correlated with counts of such citations—e.g., on the theory that more diligent applicants and examiners will tend to find and review more prior art224—we actually find the opposite. We observe that never-instituted patents cited fewer pieces of prior art overall, had fewer prior art citations added by the examiner, and cited to fewer pieces of nonpatent prior.



In the end



To the contrary, as our findings with respect to reverse citations also attest, it may be the case that more unique inventions have less prior art and thus face a speedier path to issuance

(...)

All in all, our findings suggest that reverse citations and the frequency of examination events are, at best, noisy proxies for quality.




The SSRN paper did not seem to address the issue that a relatively weak application might flood the examiner with many references in an attempt to "bury" the most relevant references. In some art areas, an applicant might include many prior art references to avoid any possibility of inequitable conduct for withholding references. This might be more likely when the prosecuting attorney is not familiar with the art in question, and has difficulty in assessing relevance.