A model with fixed relations between manifest and latent variables is presented for investigating choice reaction time data. The numbers for fixation originate from the polynomial function. Two options are considered: the component-based (1 latent variable for each component of the polynomial function) and composite-based options (1 latent variable for the weighted sum of components). The choice reaction time procedure yields reaction times for different numbers of stimuli that need to be monitored. The application of the model to such data enables the identification of the components of the polynomial function, which describe the effects of the different numbers of potential stimuli on reaction time. The investigation of the data stemming from a meta-analytic study revealed that the model with constant and quadratic components was most appropriate. Both the component-based and composite-based options led to this result. The composite-based option proved to be the more robust alternative.