In the last fifteen years the Dizziness Handicap Inventory (DHI) has gained wide acceptance as a useful measure of handicap resulting from dizziness and unsteadiness. The objective of this study was to calculate measurement error and test-retest reliability of the Dutch version of the DHI. The translation into Dutch was performed according to a double translation method. One hundred and six patients with balance problems (range 26-78 years), scheduled for vestibular rehabilitation, completed the DHI on two occasions on the same day. The test-retest reliability was excellent, with Intraclass Correlation Coefficients ranging from 0.94 to 0.99 for DHI sub-scores and DHI total score. Most weighted kappa values (kappaW) exceeded 0.80 indicating substantial item per item test-retest reliability. Over 80% agreement was noted for all items except for item 8 (74%). Item 8 asks whether the subject feels dizzy or unsteady while doing ambitious activities like sports, dancing and household activities. Consequently measurement errors were calculated, suggesting that, after an intervention, the pretreatment DHI total score should at least decrease with 12 points (lower bound 99% confidence interval for a true change) before the intervention could be said to be effective for an individual patient. Based on these data, the Dutch version of the DHI showed itself to be a highly reliable instrument to assess the self-perceived handicap imposed by vestibular system diseases.