The issue of whether or not to correct errors in students' writing is controversial. Some scholars argue that error correction is helpful, while others argue that it is ineffective, perhaps even harmful. What is missing from the literature are studies about how error correction might affect the performance of specific types of students. This study, which included 140 undergraduate English majors from Taiwan, examined the relationship between Myers-Briggs Type Indicator (MBTI) personality types on the effectiveness of one kind of error correction of writing. To identify their personality types, the students completed Form G of the MBTI. To determine their opinions about error correction of writing, the students wrote--at the beginning and the end of the two-semester course--in-class essays about whether or not they wanted to receive error correction of their writing. The purpose of this repetition, using a pre-test/post-test format, was to assist the researchers in the comparison of the students' self-reported opinions about error correction over time. In addition, the researchers compared improvement in students' written grammatical accuracy by the variables of self-reported error-correction preferences and MBTI personality types. The students also completed the grammar section of the Michigan Test of English Language Proficiency (MTELP) at the beginning and the end of the course; its purpose was to double-check any gains in grammatical accuracy in the essays with the results of a standardized grammar assessment. Once the data were collected, the Kruskal-Wallis test for non-parametric data for chi-square results and the Siegel-Castellan formula for determining differences between ranked means were applied to look for patterns of differences among MBTI personality types about the effectiveness of error correction. All of the 16 MBTI personality types were found among the 140 undergraduate English majors included in this study. Although most of the students preferred receiving extensive error correction, and benefited from it at a statistically significant level, four MBTI personality types did not prefer it. What is more, for the four types, receiving extensive error correction against their expressed preferences did not help them to improve. These findings applied to both the essay and the MTELP results. Based upon the results of this study, the researchers recommend that teachers, when considering the correction of students' errors, should also consider carefully the self-reported error-correction preferences of students. (Contains 11 tables.)