Our friend Ben Alamar had a project to try to judge offensive linemen this year by looking at their performance on passing plays. I must admit to being reticent about some of the methodology; I'm not fully comfortable with the idea of judging linemen without knowing the line calls. However, Ben did a few things to try to make sure his results were accurate both scientifically and "football-wise." (Some of this information is from him, not the article linked.) He looked at only passing plays. He tossed out plays if the camera angle didn't allow him to see the whole line. He didn't blame any blockers for untouched rushers, and he didn't blame them for not holding a block more than 4.5 seconds. He also established averages for comparison purposes and used Cox Regression to make sure he had statistical significance. His results were big on Baltimore and New England, but not so big on last year's FO Offensive Lineman of the Year, Nick Mangold.
By the way, note what Jake Scott says at the end about the creation of individual stats for linemen becoming even more difficult with the rise of zone blocking.