I noticed this, in looking at it a bit closer, the study was an update of an earlier systematic assessment of threats from 1998 to rare species in order to (1) prioritize research efforts and (2) best direct and inform policy mitigations (efforts/funding).
It seems their definition of ‘impact’ is based on a textual analysis of threats to ‘imperiled’, ‘threatened’, ‘endangered’, and ‘suspected extinct’ plant species (from NatureServe). They took these data, assessed and scored it using the IUCN threats taxonomy method scoring each threat for each species between pairs of researchers who rotated between pairs, finding ‘very high’ agreement between qualitative descriptions and the scores assigned to each threat for each species.
The IUCN method involved classifying threats under three hierarchical levels (L1-L3) based upon:
L1=“human intrusions and disturbance”, L2=“more specific (‘e.g. recreation’)”, and L3= “most detailed (‘e.g. off-road vehicular recreation’)”.
Their mean Cohen’s Kappa=0.98 for the threat analysis/quantification–this is a statistic that I am unfamiliar with, but which appears to validate convergence of quantitation of results by separate individuals on a scale from 0-1 with 0 being perfect disagreement, and 1 being perfect agreement between researchers on all scores).
So while there are limitations to the depth to which the study can characterize the threats, it does give you a numerical sense of what percentage of rare species may be affected by various classes of threats (e.g. [in order of %]: recreation, livestock, residential development, invasive species, roads/railroads, mining and quarrying, fire regime change, dams and water management/use, overharvest, etc.).
Long story short–off-road vehicles (ORVs) affected more rare plant species (19%) than hiking, bicycling, trail-riding, skiing, and recreational climbing combined (affected 13% of rare plants). Though to me, 13% vs. 19% is not significantly different, though obviously hiking is not as high-impact as ORVs on a per-person basis.
As I understand it the authors did not account for magnitude of impacts, only for potential for impact though I have to add that I need to look more into the IUCN method itself–though based on a cursory search, it does seem to stay relatively non-specific about the actual magnitude of potential threat impact itself.
Through their study, they found that based on their methodology, that threats such as missing species (e.g. pollinators), pathogens, and invasive species for example may be overstudied relative to the number of rare species they may potentially impact compared to outdoor recreation as a whole (affected 35% of these species). Interestingly, the categories which were most understudied in their results were (in order) 1) oil/gas drilling, 2) mining and quarrying, 3) landslides and volcanoes, and 4) tourism/recreation development (recreation was #7 for under-research).