Reading Difficulties in Adult Deaf Readers of French: Phonological Codes, Not Guilty!

Abstract

Deaf people often achieve low levels of reading skills. The hypothesis that the use of phonological codes is associated with good reading skills in deaf readers is not yet fully supported in the literature. We investigated skilled and less skilled adult deaf readers' use of orthographic and phonological codes in reading. Experiment 1 used a masked priming paradigm to investigate automatic use of these codes during visual word processing. Experiment 2 used a serial recall task to determine whether orthographic and phonological codes are used to maintain words in memory. Skilled hearing, skilled deaf, and less skilled deaf readers used orthographic codes during word recognition and recall, but only skilled hearing readers relied on phonological codes during these tasks. It is important to note that skilled and less skilled deaf readers performed similarly in both tasks, indicating that reading difficulties in deaf adults may not be linked to the activation of phonological codes during reading.

ACKNOWLEDGMENTS

The authors wish to extend their gratitude to all the deaf and hearing participants in this study, to Marc-André Bernier for invaluable help in participant recruitment and data collection, and to Dr. Debra Titone for the use of her lab equipment. We thank Dr. Keith Rayner, Dr. Tim Slattery, Dr, Rauno Parrila, Maria Dimitropoulou, and anonymous reviewers for comments on earlier versions of the article, and Bernhard Angele for extensive help with lme analyses. This research was supported by SSHRC (752-2002-2323 and 831-2002-0100) and FQRSC (83796) Doctoral Fellowships, and by the Centre for Language, Mind and Brain Bridge Funding Award to N. N. Bélanger, along with a SSHRC grant (410-200401775) to R. I. Mayberry.

Notes

¹Phonological overlap was calculated as the number of phonemes shared between a prime and a target. Orthographic overlap was calculated as the number of letters shared between a prime and a target. The letters did not have to be in the exact same position within the prime as they were within the target, but they had to respect the relative position of letters within the target.

²O stands for orthographic and P stands for phonological. The + sign indicates that the prime/target pairs share a high percentage of orthographic or phonological information, and the – sign indicates that the prime/target pairs share a lower percentage of either type of information.

³Thirty-one participants were tested, but 2 participants were removed from the analyses in Experiment 1, and 1 participant was removed from the analyses in Experiment 2 because these participants did not understand the tasks.

⁴To perform the successive difference contrasts, conditions were analyzed in the following order: 1 = Unrelated, 2 = O+P+, 3 = O–P+, and 4 = O–P–. The first contrast compared Condition 2 versus Condition 1 to get a measure of overall priming. Because of the way the analyses were set up, for the overall priming contrast (O+P+ minus Unrelated) the regression coefficient estimates should be negative (but would indicate facilitation because it is expected that the O+P+ condition should be responded to faster than Unrelated condition). The second contrast compared Condition 3 versus Condition 2 to get a measure of orthographic priming. Finally, the third contrast compared Condition 4 versus Condition 3 to get a measure of phonological priming. These two contrasts should yield positive regression coefficient estimates, which would be indicative of orthographic and phonological facilitation effects.

⁵Net phonological priming effects were calculated as the difference in reaction times between the O–P+ and the O–P– conditions for the 60 ms prime duration data, where phonological priming is most likely to occur.

⁶To perform the successive difference contrasts, lists were analyzed in the following order: 1 = O+P+ control, 2 = O+P+, 3 = O–P+, and 4 = O–P+ control. The first contrast compared Condition 2 versus Condition 1 to get a measure of combined orthographic and phonological encoding. Better recall of items in the O+P+ list relative to the O+P+ control list will result in a negative regression coefficient estimate; similarly for the measure of orthographic encoding (Condition 3 vs. Condition 2). Because of the way the analyses were set up (Condition 4 vs. Condition 3), if an inhibitory effect of phonologically similar words is found (fewer items would be recalled in the O–P+ list than in the O–P+ control list), the regression coefficient estimate will result in a positive value.