The influence of evaluative right/wrong feedback on phonological and semantic processes in word learning

Abstract

Feedback is typically incorporated in word learning paradigms, in both research studies and commercial language learning apps. While the common-sense view is that feedback is helpful during learning, relatively little empirical evidence exists about the role of feedback in spoken vocabulary learning. Some work has suggested that long-term word learning is not enhanced by the presence of feedback, and that words are best learned implicitly. It is also plausible that feedback might have differential effects when learners focus on learning semantic facts, or when they focus on learning a new phonological sequence of sounds. In this study, we assess how providing evaluative (right/wrong) feedback on a spoken response influences two different components of vocabulary learning, the learning of a new phonological form, and the learning of a semantic property of the phonological form. We find that receiving evaluative feedback improves retention of phonological forms, but not of semantic facts.

Keywords: feedback; foreign language learning; phonological; semantic; vocabulary; word learning.

Figure 1.

Schematic of experimental design. Participants are randomly assigned to either the phonological or semantic condition. (a) During training, participants complete 11 alternating blocks of reproduction and cued recall. In each block, they encounter 12 items from List 1, and 12 items from List 2. Items in one of the lists are designated as those receiving feedback. (b) In a reproduction trial, participants hear a fact, like ‘Tavepu has a striped body’. They then have to reproduce either the phonological form (Tavepu) or the fact (has a striped body). No feedback is given. (c) In a cued-recall trial, participants encounter a visual target, and are cued to give their response (the phonological form or the semantic fact, based on the training condition). An experimenter marks each spoken response as correct/incorrect. However, only items in one of the lists receives evaluative feedback (here, items in List 1); for items in the other list, participants see a screen acknowledging their response. (d) Tests are completed one week after training. All 24 items are tested using cued-recall tests of phonology and semantics. Here, the first set tests knowledge of the phonological forms associated with each referent and the second set of tests assesses memory for the semantic facts associated with each target. Participants also complete a free-recall test. No feedback is provided during tests.

Figure 2.

Cued-recall performance in the final test conducted a week after learning, separated by training condition and cued-recall condition. The figure shows accuracy in the (a) phonological and the (b) semantic cued-recall conditions. For each cued-recall condition, performance is grouped by training condition, with performance following phonological training depicted in red and that of participants who received semantic training shown in blue. Box plots show the median in black, the edges of the boxes represent the interquartile range and the whiskers denote the upper and lower extremes of the distribution (25th or 75th percentile ±1.5 of the interquartile range). Black dots indicate outliers beyond these extremes.

Figure 3.

Individual differences in cued-recall accuracy in the phonological and semantic training conditions. Specifically, panel (a) shows the effect of feedback on cued recall of phonology for individuals in the phonological training condition. Panel (b) shows the effect of feedback on cued recall of semantics for individuals in the semantic training condition. The influence of evaluative feedback is significant for (a), but not (b). The black dots represent the performance of individual participants; the grey lines depict the change in performance across feedback conditions for each individual. Points are vertically jittered in order to show data from all the participants. The edges of the boxes on the box plots depict the first and third quartile; the whiskers show the upper and lower extremes. The median is shown within the box.

Figure 4.

Mean recall accuracy in the two training conditions, (a) phonological and (b) semantic. Red bars show performance for items that received evaluative feedback, and blue bars show performance for the items when no feedback was received. Performance is grouped by block, only blocks 2, 4, 6, 8 and 10 are shown, as blocks 1, 3, 5, 7, 9 and 11 are blocks where participants had to repeat the phonological form or semantic fact. Error bars denote ±1 standard error of the mean.

Figure 5.

Box plot showing the distribution of different error types in the last block of training. Red boxes depict performance in the phonological training condition and the blue boxes show performance in the semantic training condition. See the legend to figure 2 for details.