Lexical production research has relied extensively on in- person naming experiments such as Picture Word Interference (PWI). While the interference effects observed in the standard paradigm have been well-attested, PWI experiments have tra- ditionally been underpowered, owing to limitations associated with large-scale data collection. In this study, we validate a scalable, typing time-based PWI paradigm that can be de- ployed on the internet, and enables large-scale replications of the interference effects observed in the spoken modality. We also propose an automated response coding process that in- corporates production errors such as incorrect responses and restarts, and can be leveraged to illuminate aspects of response conflict and correction processes in incremental production.