OBJECTIVES: Researchers and clinicians consider thinking to be important in the development and maintenance of tinnitus distress, and altering thoughts or thinking style is an object of many forms of psychological therapy for tinnitus. Those working with people with tinnitus require a reliable, psychometrically robust means of measuring both positive and negative thinking related to it. The Tinnitus Cognitions Questionnaire (TCQ) was designed as such a measure and its authors showed it to be reliable, with good psychometric properties. However, no research teams have yet carried out independent validation. This study aimed to use the TCQ to investigate thinking amongst members of the general population with both bothersome and nonbothersome tinnitus and also to verify its factor structure. DESIGN: Three hundred forty-two members of the public with tinnitus completed the TCQ online or on paper. They also rated their tinnitus on a scale as "not a problem," "a small problem," "a moderate problem," "a big problem," or a "very big problem." The authors tested the original factor structure of the TCQ using confirmatory factor analysis and then calculated the mean scores for each item, comparing mean total scores across "problem categories" for the full questionnaire and for the positive and negative subscales. RESULTS: The original two-factor structure of the TCQ was a good fit to the data when the correlation between positive and negative factors was fixed at zero (root mean square error of approximation = 0.064, 90% confidence interval = 0.058 to 0.070). Items pertaining to wishing the tinnitus would go away and despairing that it would ever get better had the highest mean scores. The mean total score for the "no problem" group (M = 31.17, SD = 16.03) was not significantly different from the mean total score for the "small problem" group (M = 34.00, SD = 12.44, p = 0.99). Differences between mean scores for all other groups were statistically significant. For the negative subscale, differences were statistically significant between all problem categories. For the positive subscale, the differences between mean scores were only statistically significant for the "no problem" group (M = 28.40, SD = 17.11) compared with the "moderate problem" group (M = 18.55, SD = 8.64, p = 0.02) and for the "moderate problem" group compared with the "very big problem" group (M = 26.79, SD = 11.66, p = 0.002). Positive and negative factors were uncorrelated (ρ = -0.03.) CONCLUSIONS:: The TCQ is a valid measure of positive and negative thinking in tinnitus, and the authors recommend its use in research and therapeutic settings. Negative thinking appears to be associated with more problematic tinnitus, but positive thinking is not associated with unproblematic tinnitus, suggesting that reducing negative thinking may be more important than teaching positive thinking in therapy.