There is a growing interest in exploring attribute non-attendance (ANA) in choice experiments. This study applied Random Response Share (RRS) approach for investigating inattention choice in choice experiments and assessed its out-of-sample predictive performance using 60 months of choice experiment data from 61,592 U.S households. Our results suggest that the RRS is not likely to be a dominant strategy to the conventional multinomial logit model in terms of the out-of-sample forecasting accuracy. However, the RRS could be a way to deal with attribute nonattendance when also considering the socio-economic characteristics of respondents.