I scraped running log data of some online logs a while back and just looked into this question of how does average MPW correlate with 5k race time.
In my database have about 8000+ 5k race results where I also have the prior 28 days of training for each athlete. The distribution of 5k times are as follows:
14:00-14:30 - 13 observations / race results
14:30-15:00 - 108 observations
15:00-15:30 - 288 observations
15:30-16:00 - 512 observations
16:00-16:30 - 619 observations
I also have the results for slower times but dont feel like writing it all out. For each of the above race time groups, I calculated the average mpw the athlete ran during the previous 28 days leading up to the race.
For 14:00-14:30 times, people ran on average 63 miles per week in the previous 28 days. In other words, in the 28 days leading up to their 5k races athletes who ran between 14:00 and 14:30 for the 5k, ran approximately 252 miles on average. Note the average here is calculated for a very small sample of 13 race results/athletes. Also this is for the 28 days leading up to the race, which quite possibly is the athletes MPW during racing season/tapering and not base building. I'd have to dig into each athlete more to confirm this.
For 14:30 -15:00 guys, they ran on average 65.1 miles per week.
For 15:00-15:30 guys, they ran on average 65.9 miles per week.
For 15:30-16:00 guys, they ran on average 59.0 miles per week.
For 16:00-16:30 guys, they ran on average 53.4 miles per week.
This trend of lower miles per week continues as the times decrease. So what can we conclude after this real rough and preliminary analysis?
1) Increasing MPW does on average lower athletes 5k times until they're running in the 15:30 range.
2) Athletes who can run between 14:00 and 15:30 on average tend to run about the same MPW. Or in other words increasing MPW does not appear to lower 5k times after 15:30 and does not explain the difference between someone running 14:30 and someone running 15:30.