get_twin_data

get_twin_data(N_pairs=500, seed=42)

Generate twin study data for returns to education.

Inspired by Ashenfelter & Krueger (1994). Each twin pair shares an unobserved ability component. The true return to education is 0.08 log-points per year; naive OLS is biased upward because ability is correlated with both education and wages.

Parameters

Name	Type	Description	Default
N_pairs	int	Number of twin pairs. Total observations = 2 * N_pairs.	`500`
seed	int	Random seed for reproducibility.	`42`

Returns

Name	Type	Description
	pd.DataFrame	Columns: twin_pair_id, twin_id, ability, educ, age, experience, log_wage.

Examples

import pyfixest as pf

data = pf.get_twin_data()

# OLS is biased upward by ability, twin-pair FE recover the true 0.08
pf.etable(
    [
        pf.feols("log_wage ~ educ", data),
        pf.feols("log_wage ~ educ | twin_pair_id", data),
    ]
)

	log_wage
	(1)	(2)
coef
educ	0.096*** (0.007)	0.073*** (0.008)
Intercept	1.744*** (0.083)
fe
twin_pair_id	-	x
stats
Observations	1,000	1,000
R²	0.164	0.739
Significance levels: * p < 0.05, p < 0.01, * p < 0.001. Format of coefficient cell: Coefficient (Std. Error)