In-barn field trials for new vaccines or antimicrobials are useful for gauging whether the product is worth a full-scale implementation. To accurate test the product, pigs (or pens of pigs) should be randomly assigned, sufficient numbers should be used, and preferably it should be a blind trial. When running a trial having control pigs is extremely important, as there can often be wide variation in performance between pigs or groups of pigs. To improve experimental units, more pens of smaller groups can be used, and if a pig is the unit spill-over effect and the effect of the pen must be monitored. Selection bias needs to be avoided by randomly assigning pigs to a pen, and pens to a trial. As well, pig days should be used to analyze data, and any mortalities recorded.