There is a very simple solution on this "elite" vs. "non-elite" status of the Kenyans where everyone is right in one sense. It is that the best Kenyan participants were at least almost elite level athletes, but that they could not've cared less to give their best effort during the first test.
It should be stated at first that the PB times reported in the 2018 version of the experiment are mean values, and even if one assumes that there were only two people in the PB groups, the standard deviation dictates that there were runners in the study with PBs of 13:35 at 5K, 3:45 at 1500m, 1:47 at 800 m in addition to the half and full marathon times. If there were more people in the groups, the fastest were naturally even further faster.
As rekrunner mentioned, the 3000 m times of the Kenyans weren't even close to their capabilities, but the three fastest baseline "control" times are around 8:40, 8:50 and 8:55 (based on the chart in the page 96 of the PhD), which doesn't appear to be anything reminiscent of their talent. The simple solution is that the Kenyans weren't giving their best effort, far from it, which is highlighted by the fact that the rate perceived effort (RPE) is very low, only 14.6 and with the standard deviation of 1.5, there were some who estimated their 3K TT effort (elites run at 100 % of their Vo2Max) as low as 12-13 corresponding "light" or "somewhat hard" at the Borg scale.
The Scottish runners considered that they gave roughly similar effort after the weeks of rHuEPO injections (~18), but Kenyans gave more effort after rHuEPO injections, because the authors state that "nlike SCO whose RPE post time trial remained relatively constant throughout the study, RPE in KEN increased after rHuEpo administration compared to pre-injections levels (16.1 ± 1.1 vs. 14.6 ±1.5)". Because RPE is usually considered to correlate roughly with heart rate with the coefficient of 10, this indicates that the Kenyans felt that the post-rHuEPO time trial felt like they ran with 15 BPM higher heart rate.