Development, psychometric properties, and pilot norms of the first Arabic indigenous memory test: The Verbal Memory Arabic Test (VMAT)

ABSTRACT

Background

Tests of verbal memory (list learning) are particularly useful for clinicians and researchers globally, yet there are no psychometrically robust tests that are built indigenously for Arabic-speaking populations, which comprise more than 370 million.

Objective

To develop a verbal memory Arabic test using a systematic procedure of item selection and then provide evidence of validity and reliability in an Arabic-speaking sample in Lebanon.

Method

In study 1, we conducted a word prototypicality study (n = 77), and identified 932 words across 7 semantic categories. Following quantitative analyzes and qualitative judgments by an expert panel, we selected a sufficient number of words and categories, and constructed and piloted the items, instructions, and protocol for the Verbal Memory Arabic Test (VMAT). In study 2, we administered the VMAT on a community sample (n = 199; screened for depression and cognitive impairment) and patients with Multiple Sclerosis (n = 16).

Results

Scores decreased with age as expected, they discriminated well between healthy and clinical populations (matched on age, sex, and years of education), and showed acceptable consistency within items and across time. Conclusions: The VMAT is the first Arabic test developed indigenously. It can be used in clinical and research settings with Arabic-speaking populations to assess verbal learning.

KEYWORDS:

• Verbal Memory Arabic Test
• Arabic norms
• assessment
• learning and memory
• normative data
• test development

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. “Mushroom” was not perceived as a vegetable by participants, and was replaced by “radish”; “stapler” was unfamiliar to participants because it has different pronunciations, and was replaced by “stamp”; “pigeon: was perceived as having a sexual connotation, so it was replaced by another type of bird (bulbul); “Nai” in recognition was considered unfamiliar and was replaced by “bell” (both words are distractors as an unrelated category).

2. The raw d′ score reflects the absolute difference in SD units between the examinee’s hit rate (signal) and false-positive rate (noise). Hit rates were computed using the following formula: [total # of hits + 0.5]/[total # of targets + 1], and false-alarm rates were computed using the following formula [total # of false alarms + 0.5]/[total # of distracters + 1].

3. For semantic clustering, EXP was = $\frac{\left[\left(\mathrm{r}-1\right)\left(\mathrm{m}-1\right)\right]}{\mathrm{N}\mathrm{L}-1}$ = $\frac{\left[\left(\mathrm{r}-1\right)\left(3-1\right)\right]}{15-1}$ = (r – 1)/7, and for serial clustering EXP was = (r – 1)/NL = (r – 1)/15.

4. In the short delay cued recall trial residuals were not equal across the regression line, while in the long delay cued recall scores, and recognition discriminability, assumptions of homoscedasticity and normality were violated.

5. Since age and sex, predicted only one score, we made a judgment to not include gender-based norms.

Additional information

Funding

This work was supported by Novartis, MENACTRIMS funded by SANOFI-AVENTIS GROUPE (DMCC Branch) (NUR.HD1.05), representative office of Sanofi - Aventis Groupe SA.