Scan sampling, observing animals at set intervals is often used by behaviorists to reduce labor inputs which are necessary with continuous sampling. The interval used when scan sampling sows housed outdoors on pasture should take into consideration the desired behavior to be collected. The objective of this study was to validate scan sampling intervals for sow behaviors and postures when sows were housed outdoors on pasture. Validation for the scanning interval was conducted on 23 sows. Sows were scanned every minute between 0730 and 1130 and then again between 1430 and 1830 by one trained observer on one day in May 2001 at the Sustainable Pork Farm. Six behavioral categories were measured; standing, feeding, head down (rooting and grazing), inactive (lying and sitting), drinking and walking. Scan samples could be taken out to as long as 60 minutes or as little as 15 minutes and still obtain accurate data for most behaviors. Validation of scanning intervals based on correlations for walking (0.81 vs. 0.34) and head down (0.98 vs. 0.85) were less (P < 0.01) accurate at 15 min scan than when a 1-min interval was used. When LSMeans were compared, standing, head and walking became less accurate (P < 0.05) at a 30 min scan compared to the 1 min scan control. For inactive behaviors there was no differences from 1 min scans until a 60 min scan (P = 0.03). In conclusion, when the period of scanning was shorter, a more precise prediction of behavior and postures was obtained. However, a balance must be struck between practical feasibility and accuracy when conducting 24 hours of behavioral observations in the field.