Repeated job quits: stepping stones or learning about quality?

IZA Journal of European Labor Studies

Table 4 Variance decomposition - proportion explained by stepping stone theory

	Job satisfaction in old job:
	1	2	3	4	5	6	7	Average
A. Baseline
Multiple time job
	0.572**	0.828**	0.704**	0.672**	0.738**	0.704**	0.648**	0.695**
	(0.220)	(0.138)	(0.087)	(0.064)	(0.035)	(0.029)	(0.058)	(0.042)
N	91	187	316	355	765	1034	245
B. JS for new hires from unemployment
	1.018**	1.042**	0.823**	0.761**	0.787**	0.752**	0.697**	0.840**
	(0.260)	(0.126)	(0.076)	(0.053)	(0.029)	(0.024)	(0.049)	(0.044)
C. Ordinal variation measure
	0.713**	0.869**	0.767**	0.701**	0.746**	0.687**	0.591**	0.723**
	(0.153)	(0.104)	(0.067)	(0.051)	(0.0293)	(0.022)	(0.037)	(0.030)
D. JS in final year of new job
	0.790**	0.561**	0.858**	0.700**	0.773**	0.727**	0.498**	0.701**
	(0.191)	(0.145)	(0.072)	(0.061)	(0.030)	(0.024)	(0.076)	(0.039)
E. Conditioning on wage and hours worked
	0.465*	0.809**	0.667**	0.666**	0.698**	0.647**	0.616**	0.652**
	(0.248)	(0.130)	(0.087)	(0.063)	(0.037)	(0.033)	(0.066)	(0.045)

Note: The proportion explained by the stepping stone theory is computed as follows: s =(Var − Var(LM))/(Var(SS) − Var(LM)), where Var, Var(LM) and Var(S S) represent actual variance and expected variance according to the learning model and the stepping stone theory, respectively. Standard errors calculated using a non-parametric bootstrap with 5000 replications are in parentheses; a ** (*) indicates that the coefficient is different from zero at a 5% (10%) level of significance.
Source: BHPS.