Project containing results from the precision analyses of question-answering data presented in Corps, R. E., Gambi, C., & Pickering, M. J. How do listeners time response articulation when answering questions? The role of speech rate. ABSTRACT During conversation, interlocutors often produce their utterances with little overlap or gap between their turns. But what mechanism underlies this striking ability to time articulation appropriately? In two verbal yes/no question-answering experiments, we investigated whether listeners use the speech rate of questions to time articulation of their answers. In Experiment 1, we orthogonally manipulated the speech rate of the context (e.g., Do you have a…) and final word (e.g., dog?) of questions using time-compression, so that each component was spoken at the natural rate or twice as a fast. Listeners responded earlier when the context was speeded rather than natural, suggesting they used the speaker’s context rate to time answer articulation. Additionally, listeners responded earlier when the speaker’s final syllable was speeded than natural, regardless of context rate, suggesting they adjusted the timing of articulation after listening to a single syllable produced at a different rate. We replicated this final word effect in Experiment 2, which also showed that our speech rate manipulation did not influence the timing of response preparation. Together, these findings suggest listeners use speech rate information to time articulation when answering questions. ANALYSIS Response precision was defined as how close participants responded to the end of the speaker's question (for analysis from final word offset) or how close participants responded to the onset of the speaker's final word (for analysis from final word onset). Response precision was calculated by taking the absolute value of response time. Before taking the absolute value, we first standardized response time to have a mean of zero, so that we could assume a half-normal distribution truncated at zero. Given that the distriburion of response precision is truncated at the lower boundary of zero, the distributional assumptions of lmer are not met. Therefore, we used Bayesian mixed effects models (BMM) as implemented in the *brms* package (version 2.1.0; Burkner, 2017). All models were fitted using a Weibull distribution (e.g., Pinder, Wiener, & Smith, 1978), as these models were a better fit to the data (assessed using LOO comparisons) than models fitted using a log-normal or a gamma distribution (note models with a normal distribution would not converge). We ran 4 chains per model, each for 1600 iterations, with a burn-in period of 800 and initial parameter values set to zero. All of the reported models diverged with no divergent transitions. The parameterization of the Weibull distribution implemented in brms is based on a scale and a shape parameter. The scale parameter quantifies the spread of the distribution, and is thus informative of the degree of precision in participants' responses. The shape parameter, however, is most often used to model failure or mortality rates, which is not relevant to response precision. Note that parameters were fitted on the log scale. In all instances, we fitted models using the maximal random effects structure (Barr, Levy, Scheepers, & Tily, 2013), except that correlations among random effects were fixed to zero to aid convergence. The full model structure for each experiment is detailed in the Data Analysis sections of the manuscript. The tables report coefficient estimates of effect size (*b*), estimate errors (*SE*), and the 95% credible interval (CrI; i.e., under the model assumptions, there is a 95% probability that the parameter estimate is contained in this interval) for each predictor. If zero lies outside the credible interval, then we conclude that there is sufficient evidence to suggest the estimate is different from zero.