Experimental data comparing hypothetical and real dichotomous choice respon
ses for two different goods were used to estimate a statistical bias functi
on to calibrate the hypothetical yes responses. The probability that a hypo
thetical yes response would be a real yes response was estimated as a funct
ion of the individual's self-assessed certainty of the hypothetical yes res
ponse (assessed on a 0-10 scale) and a variable representing the price leve
l. Without calibration the hypothetical yes responses significantly exceede
d the proportion of real yes responses, but after calibration the null hypo
thesis of no difference between hypothetical and real responses could not b
e rejected in any of the experiments.