Fish flight initiation distance (FID) is emerging as a useful metric of the response of fishes to fishing, with significant differences in FID demonstrated between fished and no-take marine reserves. However, studies investigating FID vary in methodology, and many of the potential confounding effects inherent to in-water estimation of FID have yet to be investigated. Here we examined relative effects of spear guns, dive gear, observer bias and protection status on FID estimates. Three observers estimated FID of parrot fishes in both a fished area and a no-take marine reserve, via both SCUBA and free-diving, and with and without a simulated spear gun (8 treatments). We found that FID was significantly influenced by protection status, increasing by 141 cm on average in the fished area compared to the no-take marine reserve, but not by dive type or spear gun presence. While there were some differences between observers' mean estimate of FID, there was no evidence of observer bias, nor were there any significant differences in the precision of FID estimates between observers. Overall, management status explained almost 60% of the variation in FID estimates, while observers accounted for only 4%.