AbstractMeasurements of the coefficient of luminous intensity R for prismatic retroreflectors contain errors due to misalignment of the photometric range and samples. A procedure for assessing these errors has been developed for the two‐orientation hexagon‐cube clear prismatic retrore‐flector that is included in the NBS Measurement Assurance Program (MAP) for retroreflection. This procedure involves experiments in which the geometrical setting of the range is deliberately varied so that the dependence of R on small variations in the geometry can be estimated. These experiments followed a fractional factorial design. Bounds on the uncertainties in the observation angle α, the two components of the entrance angle β1and β2, and the rotation angle ϵ can be used to bound the corresponding uncertainty in R due to geometrical effects. Results of this experiment are reported along with some repeated measurements of R for an internal NBS MAP. Similar coefficients have been reported previously for retroreflective bead sheeting, and their implication is