678 pitchers appeared in a game last year. 613 of them were 6'+ Median K/9 was 7.3, median WAR was 0.2. For the 65 guys under 6', median K/9 was 7.4, and median WAR was 0.2. Which does look like an ever so slight advantage to the short guys.
It's worth noting that we're not really looking at good sample sizes for either group though, given that the pool of small guys is 1/3 the size of tall for the hitters, and 1/10 the size for pitchers, so let's look at roughly equal size groups around a dividing line.
And you can't come up with bigger sample sizes by changing the parameters of the study. When you change the parameters from players under 6' to players under 6'3", you are essentially changing the study. The numbers then become irrelevant.
You absolutely can make more valid sample sizes this way, you take the median height of the population and compare those above it to those below it. MLB pitchers are already substantially taller than the population at large, if we're just talking about the ability of tall people in general to be better pitchers, the fact that there are so many fewer short MLB pitchers is substantial evidence on its face. By splitting the MLB sub-population around its mean, we can try to determine if height within that sub-population is advantageous.
FWIW, as I stated somewhere upthread, I think it's likely that height within the MLB population is less predictive than height vs the population at large. That is if you gave me 500 random american males who were under 6', and 500 who were over 6' I think it is a no-brainer that you can build a better pitching staff out of the tall guys more often. If you give me the same distribution of guys who've pitched in MLB, I'd still take the tall guys, but I'd be less confident about the outcome.