My apologies for being delinquent in adding new things to this. I've been planning to release the trainer level one for a while, but I've been waiting in hopes that the data would improve. Alas, even with close to 500 users across the US and non-US sites, well over 5000 weeks of data and 10,000 pops, the numbers are still messy. So, I'm releasing it anyway, and trusting you all to take it with many grains of salt!
If you log in to the site
(http://bb-training.comuf.com/club/) and go to the analysis page, you'll see "Trainer level analysis" as an option.
I've allowed you to select the age range of players to include. There's a trade-off here, of course-- the smaller the range you set, the more consistent the data will be, but the less of it will be included. But, it's a useful experiment to play with it a bit to see how consistent the results are depending on the input data.
Part of the reason for doing this is that the results, especially for level 6 & 7 trainers, is really bizarre. Level 6 is almost always slightly *slower* than level 5. And level 7 almost always looks like a bigger effect than common sense suggests. I strongly suggest expanding the "raw data" tab and looking at the actual rates being used to generate the plot. Especially for level 7, you'll see that there's a very small number of data points, so you may want to mentally exclude those. It's also useful to see that even with what seem like large samples, the data can still be funny (for PG Pressure training, which is by far the most common, level 6 has produced pops at an almost 5% lower rate, despite having >300 weeks of data for each for just 18-20yo's).
So, take a look (as long as you've contributed at least 4 weeks of data) and draw your own conclusions!