Figure 3.

(Color online) Percent correct is plotted as a function of signal level for each listener group and stimulus condition. Results for non-native speakers of English are shown in the top panel, and those for native speakers of English are shown in the bottom panel. Symbol shape represents the stimulus condition, as defined in the legend, and symbol size reflects the number of observations associated with each point. Solid lines indicate logit function fits. Dotted lines indicate 50% correct in the non-native listeners' data and the 80.7% correct in the native listeners' data; defining threshold according to these criteria results in matched SNR at baseline and very similar masking release across groups in the modulated masker conditions.