Drinking behavior in nursery pigs: determining the accuracy between an automatic water meter versus human observers

Assimilating accurate behavioral events over a long period can be labor-intensive and relatively expensive. If an automatic device could accurately record the duration and frequency for a given behavioral event, it would be a valuable alternative to the traditional use of human observers for behavioral studies. Therefore, the objective of this study was to determine the accuracy in the time spent at the waterer and the number of visits to the waterer by individually housed nursery pigs between human observers scoring video files using Observer software (OBS) and an automatic water meter Hobo (WM, control) affixed onto the waterline. Eleven PIC USA genotype gilts (22 +/- 2 d of age; 6.5 +/- 1.4 kg of BW) were housed individually in pens with ad libitum access to a corn-based starter ration and one nipple waterer. Behavior was collected on d 0 (day of weaning), 7, and 14 of the trial using 1 color camera positioned over 4 attached pens and a RECO-204 DVR at 1 frame per second. For the OBS method, 2 experienced observers recorded drinking behavior from the video files, which was defined as when the gilt placed her mouth over the nipple waterer. Data were analyzed using nonparametric methods and the general linear model and regression procedures in SAS. The experimental unit was the individual pen housing 1 gilt. The GLM model included the method of observation (WM vs. OBS) and time (24 h) as variables, and the gilt nested within method was used as the error term. Gilts consumed more water (P = 0.04) on d 14 than on d 0. The time of day affected (P < 0.001) the number of visits and the time spent at the waterer regardless of the method. However, the OBS method underestimated (P < 0.001) the number of visits to the waterer (3.48 +/- 0.33 visits/h for OBS vs. 4.94 +/- 0.33 for WM) and overestimated (P < 0.001) the time spent at the waterer (22.6 +/- 1.46 s/h for OBS vs. 13.9 +/- 1.43 for WM) compared with WM. The relationship between the 2 methods for prediction of time spent at the waterer and number of visits made by the gilts was weak (R(2) = 0.56 and 0.69, respectively). Collectively, these data indicate that the use of the traditional OBS method for quantifying drinking behavior in pigs can be misleading. Quantifying drinking behavior and perhaps other behavioral events via the OBS method must be more accurately validated.