One potential solution is to make the Roleplaying award based upon a vote. This avoids the issue of GM favoritism. It doesn't eliminate favoritism overall, but does avoid GM favoritism, and can reduce the emotional impact of dissimilar awards.
One should, however, also remain aware of the relationship of reward cycles and player behavior: What is rewarded tends to be what gets done. If a player can't rise to the standard, then it can create resentment; in some cases, however, it simply results in acknowledgement of the defect. In order to maximize the positive reinforcement effect, however, the GM needs to be clear why each experience point award is given - if the system instructs awarding a karma for good roleplay, then when awarding Karma, award it saying something to the effect of, "Ok, Fred, Joe, George, 1 point for good roleplay tonight." Do not say, "Darryl, no RP award." Negative reinforcement works far less well than positive; a lack of negative with positive of others tends to encourage like behavior. (This is all well grounded in educational psychology theory. It also is used strongly in games like Burning Wheel.)
If you have a player who is incapable of rising to the standard, you can lower your standard, or you can work with the player to train them. You can also work with the deficient player on the needed skills if they are willing.
The one thing not to do is to provide alternate routes to the same total. Each category should stand alone, and either be met or not. Providing alternates simply makes experience awards meaningless for shaping behavior, and cheats those who are meeting the category.
If you want even advancement, just award a flat award every session, and ignore the guidelines. It's perfectly reasonable to drop a category as well.
In 33 years of GMing, I've only had two players (of over 50, total) complain about asymmetric awards. Of those two, one was a cheat and a liar, so I utterly discount his contribution. The other was emotionally unstable and constantly seeking unmerited reinforcement. It usually isn't going to be an issue as long as it's clear why the awards are dissimilar.