Free-viewing laterality tasks: a multilevel meta-analysis.

OBJECTIVE

Chimeric free-viewing laterality tasks have been used extensively as measures of right-hemisphere functioning, with many variations in stimuli and samples typically showing an LVF bias. However, the questions remain concerning whether the LVF bias is significantly different from zero, and what factors might moderate this bias.

METHOD

The present meta-analysis answered these questions by retrieving a presumably exhaustive sample of studies published in English that involved free viewing of stimuli. The final analysis was based on 329 effect sizes drawn from 112 published studies. A hierarchical linear model (or multilevel) approach to meta-analysis was used to deal with the violation of the independence of effect-sizes assumption and to reflect better the hierarchical structure of the data.

RESULTS

A large and significant left visual-field (LVF) bias (estimated mean d = 1.024) was demonstrated across the entire set of retrieved effect sizes. It was also demonstrated that such tasks are a useful tool for discriminating between various clinical populations. Finally, the moderator analysis identified that emotion faces (estimated mean d = 1.052) and timed conditions (estimated mean d = 1.319) appear to promote large effects.

CONCLUSIONS

The present meta-analysis validated free-viewing laterality tasks as tools for neuropsychological assessment and for empirical research.