The effect of inverting decades and units on the retention of two-digit numbers in working memory: a matter of the output mode

Abstract

Number-word structure affects transcoding and relates to working memory capacity (WMc). We tested the relationship between the structure of number-words (inverted "Units-Decades (UD) format" or not "Decades-Units (DU) format") and performance on the digit span task for recalling two-digit number-words presented verbally, using two modes of output. University students, native speakers of Arabic, who were also highly proficient in Hebrew, and native speakers of Hebrew, were asked to response by typing Arabic digits or recalling numbers verbally. Arabic and Hebrew speakers were equal in WMc for one-digit numbers in the two modes. However, when inversion was presented, performance was related to output and participants habits. In the typing response, both groups showed better performance for the DU format. In the native language (standard format), Arabic speakers (UD) remembered fewer numbers than Hebrew speakers (DU). Arabic speakers show inversion errors even in the noninverted format. For verbal responses, there were no differences between the groups in their standard format, and native speakers of Arabic weren't affected by the format, while Hebrew speakers still preferred the DU format. Experience with inverted and noninverted language affected WMc but is also related to the output demands.

Fig. 1

Percent recall in the correct order of 1- digit numbers for lists of different lengths. Error bars are standard errors

Fig. 2

WMc for both Arabic and Hebrew native speakers, using varying sequence lengths in the standard format for each language (UD for Arabic; DU for Hebrew). Response was by typing or verbal. Significant effects of Format are indicated by *. Error bars are standard errors

Fig. 3

WMc for both Arabic and Hebrew native speakers, using varying sequence lengths, when stimuli were presented in inverted (UD) or noninverted (DU) format, and when response was by typing or verbal. Significant effects of Format are indicated by *. Error bars are standard errors

Fig. 4

Mean of overall errors (inverted errors & other errors) for native speakers of Arabic and Hebrew, when stimuli were presented in inverted or noninverted format, and when response was by typing. Error bars are standard errors

Fig. 5

WMc for Native Speakers of Arabic in Arabic and Hebrew for sequences of different lengths, when stimuli were presented in the inverted and noninverted format and in the typing response mode. Error bars are standard errors