Data from a previously reported study were analyzed to determine the effect of the Δa-increment on the error in calculating the fatigue crack growth rate from recorded a versus N data. The data had been recorded from 68 replicate fatigue crack growth tests on center crack panels of 2024-T3 aluminum alloy using a Δa-increment of 0.2 mm. Two different differentiation methods were considered, the secant method and the quadratic seven-point incremental polynomial method. The influence of the Δa-increment was determined by evaluating da/dN using all the data, every second, every fourth, every eighth, and every sixteenth data point. The minimum error for calculating da/dN was found to be for Δa between 0.8 and 1.6 mm. An evaluation of the type of distribution for da/dN showed that of the five distributions considered no single distribution adequately represented all da/dN regardless of the Δa-increment.