The Zarit Burden Interview (ZBI) is the instrument most widely used in North America and Europe for assessing the burden experienced by family caregivers who look after the community-residing impaired elderly. The Japanese version of the ZBI (J-ZBI) is the most widely used in Japan for the same purpose. We used data from 691 caregivers in relation to their caregiver burden with the 22-item J-ZBI. Following a factor analysis, the 8-item short version of the J-ZBI, the J-ZBI_8, was proposed with the following two factors: Personal strain (5 items) and Role strain (3 items). Cronbach's alpha of the J-ZBI_8 was 0.89, indicating the high reliability of this instrument. Pearson's correlation coefficient between J-ZBI and J-ZBI_8 was 0.93, and the same coefficient between J-ZBI_8 and item 22 (a single global burden) was 0.68. These data indicated that J-ZBI_8 had a high concurrent validity. In addition, caregivers who declared that they did not encounter any difficulties in looking after the impaired elderly had a significantly lower J-ZBI_8 score (3.45; SD = 4.57) than those who claimed that they had such difficulties (9.31; SD = 7.19) by the t-test; this indicated that J-ZBI_8 had a high construct validity. These results indicated that the short version, the J-ZBI_8, had a high reliability and validity. Therefore, the J-ZBI_8 produced results comparable to those of the full version, i.e. the J-ZBI. The shorter yet no less reliable and valid 8 item version will thus lead to easier administration of the instrument for assessing family caregiver burden in clinical settings.