This paper reports on four studies that were all concerned with cognitive and non-cognitive correlates of proofreading (PR) ability. A new, five-minute PR test was devised and piloted. In the first pilot study (N = 191) it was correlated with a verbal reasoning test. In the second study (N = 103) PR scores were regressed onto measures of personality: approach to learning, general and fluid intelligence (verbal reasoning) as well as a test of general knowledge. All three cognitive ability measures were significant predictors of PR but when entered together, general knowledge alone remained significant and accounted for a third of the variance. The third study (N = 95) was similar, except it also included a measure of typical intellectual engagement. Fluid intelligence (verbal reasoning) and openness were the only significant predictors of PR. In the fourth study (N = 249) participants completed four tests: PR, fluid intelligence (verbal reasoning), vocabulary and spelling. PR was consistently correlated with verbal reasoning but more strongly correlated with measures of crystallised intelligence, namely general knowledge and vocabulary. Implications and limitations are considered. (Contains 9 tables.)