Skip to main content

Table 3 BMI coefficients in log hours of work for single respondents: OLS, IV, and FE estimates

From: Marriage markets as explanation for why heavier people work more hours

 

White

White

Black

Black

Hispanic

Hispanic

Single women

 Panel A: OLS estimates

  BMI

0.00451** (2.463)

0.00581*** (3.246)

0.00668*** (4.365)

0.00682*** (4.496)

−0.00120 (−0.359)

−0.000336 (−0.105)

 Panel B: IV estimates

  BMI

0.0104* (1.835)

0.0138** (2.416)

0.00308 (0.458)

0.00452 (0.690)

0.00936 (0.919)

0.0101 (1.050)

  First-stage F-stat

107.2

101.9

45.8

45.6

32.7

32.6

 Panel C: sibling FE estimates

  BMI

−0.00215 (−0.782)

−0.00172 (−0.628)

0.00504* (1.809)

0.00439 (1.591)

0.00507 (0.913)

0.00523 (0.948)

  Control for wage

No

Yes

No

Yes

No

Yes

  Observations

6469

6469

4671

4671

2020

2020

Single men

 Panel D: OLS estimates

  BMI

0.00345* (1.927)

0.00306* (1.777)

0.00603** (2.331)

0.00450* (1.723)

0.00173 (0.746)

0.00201 (0.935)

 Panel E: IV estimates

  BMI

0.0186** (2.540)

0.0185** (2.558)

0.00225 (0.293)

0.00170 (0.227)

0.00151 (0.164)

0.00283 (0.329)

  First-stage F-stat

63.8

63.6

85.5

84.9

42.0

41.9

 Panel F: sibling FE estimates

  BMI

0.000333 (0.118)

−0.000174 (−0.0634)

0.00564 (1.572)

0.00385 (1.095)

−0.00155 (−0.467)

−0.00156 (−0.481)

  Control for wage

No

Yes

No

Yes

No

Yes

  Observations

8648

8648

4777

4777

2997

2997

  1. Note 1: Control variables include work experience (quadratic), educational categories, dummies for whether the respondent believes in traditional gender roles, whether the respondent has any children, if the youngest child is below six, yearly age dummies, region of residence dummies, and year dummies. Note 2: For OLS and IV regressions, t-stats reported are based on standard errors clustered at the individual level. For the sibling FE model, the t-stats reported are based on standard errors clustered at the family level. *** p < 0.01, ** p < 0.05, and * p < 0.10