![Sample our Behavioral Sciences journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/110801/TOC-Subject-Sample-2021-Behavioral-Sciences.jpg"})
Abstract
Performance in numerical classification tasks involving either parity or magnitude judgements is quicker when small numbers are mapped onto a left-sided response and large numbers onto a right-sided response than for the opposite mapping (i.e., the spatial–numerical association of response codes or SNARC effect). Recent research by Gevers et al. [Gevers, W., Santens, S., Dhooge, E., Chen, Q., Van den Bossche, L., Fias, W., & Verguts, T. (2010). Verbal-spatial and visuospatial coding of number–space interactions. Journal of Experimental Psychology: General, 139, 180–190] suggests that this effect also arises for vocal “left” and “right” responding, indicating that verbal–spatial coding has a role to play in determining it. Another presumably verbal-based, spatial–numerical mapping phenomenon is the linguistic markedness association of response codes (MARC) effect whereby responding in parity tasks is quicker when odd numbers are mapped onto left-sided responses and even numbers onto right-sided responses. A recent account of both the SNARC and MARC effects is based on the polarity correspondence principle [Proctor, R. W., & Cho, Y. S. (2006). Polarity correspondence: A general principle for performance of speeded binary classification tasks. Psychological Bulletin, 132, 416–442]. This account assumes that stimulus and response alternatives are coded along any number of dimensions in terms of – and + polarities with quicker responding when the polarity codes for the stimulus and the response correspond. In the present study, even–odd parity judgements were made using either “left” and “right” or “bad” and “good” vocal responses. Results indicated that a SNARC effect was indeed present for the former type of vocal responding, providing further evidence for the sufficiency of the verbal–spatial coding account for this effect. However, the decided lack of an analogous SNARC-like effect in the results for the latter type of vocal responding provides an important constraint on the presumed generality of the polarity correspondence account. On the other hand, the presence of robust MARC effects for “bad” and “good” but not “left” and “right” vocal responses is consistent with the view that such effects are due to conceptual associations between semantic codes for odd–even and bad–good (but not necessarily left–right).
Notes
1The use of such predictors also allowed for a test of the interaction between numerical magnitude and parity status in each of the four conditions. None of the t-tests involving this term was significant.
2One important qualification of this statement with respect to the size of the MARC effect must be made, however, because one of the two participants in this group that could not be part of the manual response task analyses actually had a very large positive parity status regression coefficient in the verbal response task. When these two participants are omitted from the verbal response task analyses, the size of MARC effect for this task (M = −61.16, SD = 103.72) becomes quite comparable to that obtained for the reduced set of 17 participants in the manual response task analyses and is now also significantly different from zero, t(16) = −2.432, p < .05. On the other hand, the size of the SNARC effect is still substantially smaller for the verbal response task than for the manual response task even when only this reduced set of 17 participants is considered (M = −4.13, SD = 7.11).
3Interestingly, although the correlation between the individual regression coefficients indexing the size of the MARC effects across the verbal and manual response tasks was indeed moderately large and positive for the “left” and “right” verbal response group (r = .568, p < .05, N = 17), for the “bad” and “good” verbal response group, this same correlation was now also moderately large but negative (r = −.505, p < .05, N = 19). Moreover, with respect to the correlations between the individual regression coefficients indexing the size of the SNARC effects across the verbal and manual response tasks, these were actually very low for both the “bad” and “good” verbal response group (r = −.094, p > .50, N = 19) and, perhaps surprisingly, also the “left” and “right” verbal response group (r = .042, p > .50, N = 17).
4None of the t-tests involving the term for the interaction between numerical magnitude and parity status was significant for Experiment 2.
5It also seemed relevant to compare the SNARC and MARC effects in the left–right vocal responses provided by the eight participants in Experiment 2 who performed this vocal response task first to those same effects in the left–right vocal responses provided by the eight participants in Experiment 2 who performed this vocal response task second (i.e., after previously having responded with bad–good vocal responses). Although not significantly so, the SNARC effect turned out to be smaller (−6.72 vs. −10.65) and the MARC effect larger (−106.99 vs. −34.68) for the eight participants who performed the left–right response vocal task first.
With respect to comparing the SNARC- and MARC-like effects in the bad–good vocal responses provided by the eight participants in Experiment 2 who performed this vocal response task first to those same effects in the bad–good vocal responses provided by the eight participants in Experiment 2 who performed this vocal response task second (i.e., after previously having responded with left–right vocal responses), although not significantly different in each case, the SNARC-like effects were 5.29 and −7.49, respectively, whereas the MARC-like effects were −63.26 and −121.60, respectively.
6Note that such findings could also be regarded as being rather inconsistent with the linguistic markedness account for the MARC effect given that, according to such a view, MARC effects would have been expected to be especially robust for verbal responding.