Abstract Four rodent control techniques—sustained baiting (SB) with coumachlor, pulsed baiting (PB) with brodifacoum, a lethal electrified barrier (LEB), and a nonlethal electrified barrier (NLEB)—were evaluated on the experimental farm of the International Rice Research Institute (IRRI) in the Philippines. A fifth treatment (no experimental rodent control) was established for reference. Mean tiller damage and rodent activity, respectively, near harvest were 1.0% and 1.5% in the LEB plots, 1.6% and 18.0% in the SB plots, 2.1% and 16.0% in the PB plots, and 4.1% and 32.5% in the NLEB plots. Highest mean tiller damage (9.3%) and rodent activity (56.0%) occurred in plots with no experimental rodent control. Both baiting methods (SB and PB) were less expensive than barrier methods (LEB and NLEB). During an 80-day crop protection period prevalent on the IRRI experimental farm, the total cost per hectare (US$) for protecting experimental plots from rodent damage was $26 for SB, $27 for PB, $268 for NLEB, and $1285 for LEB. Baiting methods were more cost-effective and are recommended for general rodent control on research farms that can tolerate ⩽2% rodent damage without losing experimental data. For small research plots demanding a greater degree of protection, an effective barrier system such as the LEB or a combination of the NLEB and LEB should be used.