Two major confounders for this argument are the raising of national speed limits from the 1970s to the 1990s, during which time traffic accident and fatality rates went down significantly, and the post-COVID bump in traffic accident and fatality rates, during which time, for the most part, no speed limits went up.
Without a control, there can be no confidence in the claimed causal link. Moreover, "speed limit" vs "no speed limit" is not the same question as "lower speed limit" vs "higher speed limit". People often behave in non-intuitive ways, which is part of why studies are conducted in the first place.
That’s a pretty silly line of thought.. there’s absolutely no need for a “no speed limit” control when speed limits already are just chopping off the higher end of the bell curve of speed distributions.
And clearly there are other causes of accidents and other ways to reduce fatalities (antilock brakes and seatbelts for the period you mentioned) but it’s actually pretty straightforward to assess speed’s influence on accident rates and injuries as shown by the dozens of studies that have found a direct causal relationship. It’s not as if there’s some possibility of a discontinuity where no speed limits means everyone drives slower… especially if you know how speed limits are set in the first place.
The Autobahn has lower accident and fatality rates than U.S. interstates with similar levels of traffic and yet portions of it have no set speed limit at all.
The question is not of speed but speed-related policies and their effect in practice, not in deductively reasoned theory.
Without a control, there can be no confidence in the claimed causal link. Moreover, "speed limit" vs "no speed limit" is not the same question as "lower speed limit" vs "higher speed limit". People often behave in non-intuitive ways, which is part of why studies are conducted in the first place.