The Boston Naming Test (Kaplan et al 1983) is standardized on different populations and it is the most widely used naming test worldwide in both research and clinical settings. For this reason, this test was used to investigate naming performance in Maltese-speaking adults in a bilingual context.
The first part of the research aimed to determine the demographic and word variables that affect naming performance in healthy Maltese people. It also aimed to propose an adaptation of the BNT to suit the linguistic and cultural characteristics of the Maltese population. The second part of the research, then, applied the findings obtained initially to establish whether the
proposed test could be used clinically in assessing Maltese people with naming impairments. It sought to establish whether the proposed modified test was able to differentiate between people with impaired and unimpaired naming, and whether it could contribute to making a differential
diagnosis.
In order to obtain a measure of name agreement the BNT was administered to 24 individuals prior to the main study. Naming performance data was then collected from 60 healthy people. The BNT pictures were presented via a laptop and naming accuracy and reaction times were recorded. Analysis of the variables which affect naming performance was carried out. Finally,
the BNT and additional tests of comprehension and repetition were administered on 18 people with aphasia to determine the suitability of the test for clinical use.
Only 38 items of the BNT had at least 70% name agreement. Therefore, a reduced test of 38 items was proposed and two alternative scoring methods were described: strict (accepting only dominant responses and their cognates) and lenient scoring (accepting all possible responses,including English equivalents).
The research on healthy Maltese adults showed that education was the strongest predictor of
naming accuracy. Age also remained significant in a stepwise regression analysis. The strongest predictor of naming latency was age, with evidence of a marked increase in reaction times after 66 years of age. Since older Maltese adults had fewer opportunities for education, cohort effects are likely.
Some of the controls obtained very low scores which overlapped with the scores obtained by the people with aphasia. This indicated that scores alone cannot be used to differentiate between impaired and unimpaired people. On the other hand, some types of errors were only produced
by people with impaired naming, and did not appear at all in error profiles of unimpaired
individuals. Although particular types of errors were associated with particular levels of breakdown in spoken word production, a degree of overlap between errors and levels ofbreakdown became apparent.
A reduced set of the BNT with lenient scoring was found to be the most suitable test for assessing Maltese people at risk of naming impairments. It is recommended that test standardization in the elderly should consider cohort effects and regular test updates should be carried out.
In order to differentiate between people with impaired and unimpaired naming it is necessary to look at error profiles, apart from the number of errors, as the presence of atypical errors is the best indicator of naming impairments. Further, it was concluded that, although error profiles
may prove to be valuable in differential diagnosis, they cannot be used on their own, as they do not provide sufficient information. A variety of additional tasks of semantic and phonological
integrity must also be utilized to make conclusive decisions regarding differential diagnosis.