Preschool minimal-pair discrimination (Creel, 2022)

Purpose: The primary aim was to assess whether children have difficulty distinguishing similar-sounding novel words. The secondary aim was to assess what task characteristics might hinder or facilitate perceptual discrimination.

Method: Three within-subjects experiments tested ninety-nine 3- to 5-year-old children total. Experiment 1 presented two cartoon characters each saying a novel word. Children were asked to report whether they said the same word or different words. Words were identical (e.g., deev/deev), were dissimilar (deev/vush), differed in onset consonant voicing (deev/teev), or differed in vowel tenseness (deev/div). Experiment 2 added accuracy feedback after each trial to remind children of task instructions. Experiment 3 interspersed many “same” trials containing a repeating standard word to assess the role of bottom-up stimulus support on difference detection.

Results: The d′ scores were highest for dissimilar words, next highest on different-vowel pairs, and lowest on different-consonant pairs. Performance was better with repeated standard stimuli (Experiment 3) than without (Experiment 1). Benefits for repeated task instructions (Experiment 2) were marginal. Exploratory analyses comparing these results to findings in a word-learning study using the same stimuli suggest an imperfect match to how easily children can learn similar-sounding words.

Conclusions: Overall, similar-sounding novel words are challenging for children to discriminate perceptually, although discrimination scores exceeded chance for all levels of similarity. Clinically speaking, same/different tests may be less sensitive to sound discrimination than change/no-change tests.

Supplemental Material S1. Age of acquisition in production consonants. 

Supplemental Material S2. Age of acquisition in production. 

Supplemental Material S3. Phonotactic probability of words.

Supplemental Material S4. All test lists for Experiments 1 and 2.

Supplemental Material S5. First four test lists for Experiment 3. 

Supplemental Material S6. Similarity effects when limiting analyses to first four or eight trials of each type (same, different-consonant, different-vowel, and dissimilar-word).

Creel, S. C. (2022). Preschoolers have difficulty discriminating novel minimal-pair words. Journal of Speech, Language, and Hearing Research. https://doi.org/10.1044/2022_JSLHR-22-00029