Verbal mediation of visual memory on the Continuous Visual Memory Test
S. Patrick Whalen · The Mathematics Enthusiast · 1999
Nonverbal memory testing is an integral component o f clinical neuropsychological evaluation.The idea o f nonverbal memory is supported both by research on hemispheric asymmetry and by some factor analyses o f memory test batteries.The construct o f nonverbal memory has received inconsistent support from empirical research.One significant, potential confound o f nonverbal memory testing is verbal mediation or contamination o f visual stimuli.Investigation o f several nonverbal memory tests indicates that test items are susceptible to verbal contamination.The Continuous Visual Memory Test (CVMT, Trahan & Larrabee, 1988) is a relatively new and little researched test o f nonverbal memory.The CVMT utilizes a recognition format and uses complex geom etric figures as visual stimuli, and stimulus duration is brief.These characteristics are proposed as means to minimize confounds, including verbal contamination o f the test.The verbal characteristics o f the test items, however, have not been empirically examined.In the present study tw o experiments were performed.In the first, the verbal characteristics o f the CVMT w ere investigated.51 students from an introductory psychology class were administered the CVMT and asked to assign verbal labels to the items.Subject responses were recorded and several indices o f verbal loading o f items were calculated using those responses.Results indicate that the CVMT items are susceptible to verbal encoding, and that some classes o f visual stimuli are more prone to verbal labeling than others.In the second experiment the relationship between verbal load o f test items and subject performance was examined.51 volunteer students from an introductory psychology class participated.Each was administered the CVMT according to standard instructions.Verbal load o f the stimulus items was not consistently related to subjects' ability to correctly identify target items at delayed recognition.However, verbal load o f item classes did relate to subject performance using one index o f verbal contamination.In addition, the overall performance o f subjects in experiment 1 and experiment 2 was compared.Results indicate that there were no significant differences in performance at delayed recognition between the tw o groups.Implications for interpretation o f CVMT performance, future research, and test development were discussed.