Handedness and earnings

Abstract

We examine whether handedness is related to performance in the labour market and, in particular, to earnings. We find a significant wage effect for left-handed men with high levels of education. This positive wage effect is strongest among those who have lower than average earnings relative to those of similar high education. This effect is not found among women.

An earlier version of this paper was presented at the January 2003 Meetings of the North American Econometric Society.

An earlier version of this paper was presented at the January 2003 Meetings of the North American Econometric Society.

Acknowledgments

We thank Susan Averett, Jeff Biddle, Ed Gamber, two anonymous referees, and the editor for their helpful comments.

Notes

An earlier version of this paper was presented at the January 2003 Meetings of the North American Econometric Society.

¹For a review of various theories regarding handedness, see Beaton (2003).

²We only became aware of this work in July 2006, after our analysis was completed.

³In unreported regressions, we investigated the inclusion of the “neither” and “ambidextrous” respondents in addition to those that identified themselves as right- or lleft-handed but obtained results little different from those reported below.

⁴For those who are not actually paid by the hour, the NLSY data set follows the usual practice of dividing earnings in the last paycheque by the number of hours worked in that period to obtain an estimate of the implicit hourly wage rate.

⁵We estimated similar regressions by replacing the dependent variable with hours of work per week, employment status, and the logarithm of annual earnings. The handedness variable was statistically insignificant in those regressions as well.

⁶We do not include the square of age, as most economic models do, because our men are relatively young in age and thus their wages have not yet started to exhibit much curvature.

⁷For this discussion, we took our estimates of the wage regression at the 25th percentile point (see footnote to Table 5 for the variables in the equation), predicted the 25th percentile point for each individual, and selected those individuals with 16 years of education whose actual wages were below that predicted value for the 25th percentile. We then tabulated the occupational distributions of left-handed and right-handed individuals within this group.

⁸The other two groups that had significantly large standard errors in the calculation of the chi-squared statistic were clerical workers and operators. Both are lower-skilled occupations, and both had fewer left-handed individuals than the right-handed distribution would predict.