**Project contents** This projects contains materials, scripts and supplementary material for the paper "Improving the robustness of infant lexical processing speed measures" by Julia Egger, Caroline Rowland & Christina Bergmann. 1.Materials: This folder contains all materials used for the study which we are able to share freely. Please note that both the audiofiles and the questionnaires are in Dutch, given that we tested a Dutch speaking population. 2.Data and Analysis (November 2019): In this folder, the scripts used for the analysis as well as the data are available. - *Preprocessing_Study_Fixation.R* is the revised preprocessing script and therefore, slightly different from our preregistered script. We decided to filter the fixations assigned by the eyetracker such that we only included fixations that last for at least 100ms in our analysis, in line with our verbal preregistration. Furthermore, we refinded the image areas and changed variable names. - *Analysis_Paper_Fixation.Rmd* uses the the file *rawdata_Fixation.csv* (created by *Preprocessing_Study_Fixation.R*) to conduct all the analyses and create all figures and tables. - *QuestionnaireDate.csv* contains the information necessary for the analyses as extracted from the questionnaires. - *CDIwords.csv* shows which words used in the study were known or unknown by the individual participants, according to their parents. This file is needed for the exploratory analysis on the effect of word knowledge on the reaction times. - *RTTrialData.csv* contains an overview of the reaction times per participant, per condition, per trial, including the label they heard at the specific trials. This file was created to facilitate (external) follow-up analyses. 3.Supplementary Material (November 2019): 1. Additional Graphs: This folder contains further graphs that were not added to the manuscript. *FixationsOnScreenPlot.jpg* shows the locations of all the fixations made by our participants. It confirms that most fixations fall onto the image areas. *RTPlot_Numeric.jpg* shows the numeric mean reaction times for each participant between conditions. *RTPlot_Numeric_ExcludedWords.jpg*, *RTPlot_Ranked_ExcludedWords.jpg* & *Violinplot_ExcludedWords.jpg* were created to inspect the effects of excluding words individually based on word knowledge. ---------- **Abstract** Visual reaction times to target pictures after naming events are an informative measurement in language acquisition research, because gaze shifts measured in looking-while-listening paradigms are an indicator of infants’ lexical speed of processing. This measure is very useful as it can be applied from a young age onwards and has been linked to later language development. However, to obtain valid reaction times, the infant is required to switch the fixation of their eyes from a distractor to a target object. This means that usually at least half the trials have to be discarded - those where the participant is already fixating the target at the onset of the target word - so that no reaction time can be measured. With few trials, reliability suffers, which is especially problematic when studying individual differences. In order to solve this problem, we developed a gaze-triggered looking-while-listening paradigm. The trials do not differ from the original paradigm apart from the fact that the target object is chosen depending on the infant’s eye fixation just before naming. The object the infant is looking at becomes the distractor and the label of the other object is used as the target, requiring a fixation switch, and thus providing a reaction time. We tested our paradigm with forty-four 18-month-old infants. We also included the original paradigm to compare the methods within participants. The results show that the gaze-triggered paradigm results in a higher percentage of reaction time trials, allowing for a less varied speed of processing measure. However, a ranked correlation between the conditions confirms that the manipulated paradigm measures the same concept.