This study considers the precision and accuracy of bipolar corneal electrodes compared with unipolar intravitreal methods in collecting electroretinographic (ERG) recordings from a small animal. Flash ERGs were obtained from 9 adult guinea pigs on three occasions. Corneal bipolar (Burian-Allen) electrodes were used to collect data on the first two occasions whereas unipolar intravitreal electrodes were used on the last. We identified the a-wave, b-wave, oscillatory potentials, PIII and PII responses. Intensity-response functions were fit using a Naka-Rushton relationship with a bootstrap estimating the 95% confidence limits. Discrepancy analysis was applied to determine the coefficient of agreement. We found significantly larger amplitudes with unipolar intravitreal electrodes (ANOVA; a-wave, p<0.002; b-wave, p<0.001; Oscillatory potentials (OPs), p<0.005) especially at high intensities. Implicit times showed little differences between electrodes for the a-wave, significantly faster (p<0.03) b-waves at some intensities, and significantly slower (p<0.005) OP implicit times across all intensities. The PIlI amplitude (log microV), sensitivity and timing were not significantly different (p>O.05) if expressed in logarithmic units but PII amplitude (log microV) was significantly smaller with corneal electrodes. We suggest that a conversion factor (x1.35) should be applied to data collected with bipolar corneal electrodes to estimate the amplitudes of the modelled parameters accurately. The corneal electrode gave a precision of +/-39 microV which yields a statistical power of 0.90 for a sample size of 7 subjects. We conclude that bipolar corneal electrodes provide smaller electroretinogram amplitudes due to their location and reduced span of the retinal generators.