For many overfished marine stocks, recreational fishing continues even though recovery plans are implemented and commercial landings regulated. In such cases, unbiased and precise estimates of recreational harvest are important for successful management. Harvest estimation often relies on interviewed-based surveys where fishers are asked to recall harvest within a given timeframe. However, the importance of whether fishers are requested to provide figures in weight or number is unresolved. Therefore, a recall survey aiming at estimating recreational harvest was designed, such that respondents could report harvest using either weight or numbers. It was found that: (1) a preference for reporting in numbers dominated; (2) reported mean individual weight of fish caught, differed between units preferences; and (3) when an estimate of total harvest in weight are calculated, these difference could result in a substantial bias through the conversion from numbers to weight. Based upon these results it is recommended that recreational harvest should be requested in numbers and not weight.