A comparison is presented between the sensitivity to measurement error of the crack compliance and layer removal methods of residual stress measurement when applied to glass fiber reinforced plastic (GFRP) pipes. This is done by adding random scatter to the exact strain distribution associated with a known stress distribution. This defines strain data that simulate experimental measurements. These data are used to determine the corresponding residual stress distributions. The error in the residual stress distribution when scatter is included can thereby be determined. It is shown that the layer removal and crack compliance methods are equally suitable for the measurement of axial and circumferential stresses in a pipe wound at only ±55 deg. The layer removal method, however, is shown to have significantly lower sensitivity to measurement error when the axial residual stresses in layered GFRP pipes are considered.