A meta-analysis was performed on 10 studies that directly compared an informant questionnaire with a brief cognitive test at screening for dementia. The effectiveness (effect size) of the screening tests used in each study was measured using the standardized difference between the means of the demented and control samples. The informant questionnaires were found to have a weighted mean effectiveness of 1.74 [95% confidence interval (CI): 1.39-2.09], whereas for the cognitive tests, the mean was 1.48 (95% CI: 1.23-1.73). These means translate into sensitivities and specificites of 0.86 and 0.80 for informant questionnaires, compared with 0.79 and 0.80 for brief cognitive tests. The effectiveness of the informant questionnaires was found to differ significantly from study to study, but the reason for this variation is not clear. It was concluded that informant questionnaires are as effective as brief cognitive tests at screening for dementia and deserve to be used more extensively.