Skip to main content

Table 5 Main results Oaxaca-Blinder decomposition by Field of Education

From: Same degree but different outcomes: an analysis of labour market outcomes for native and international PhD students in Australia

 

Pooled Natives-Foreigners

Natives ESB-NESB

ESB Natives-Foreigners

ESB Nat.-NESB Foreigners

STEM

Hum’s

Me&H

STEM

Hum’s

Me&H

STEM

Hum’s

Me&H

STEM

Hum’s

Me&H

\({\Delta }_{t}, {\Delta }_{t}\left(\tau \right)\)

 − .037***

(.012)

.029

(.028)

.091**

(.044)

 − .022*

(.012)

 − .009

(.019)

.066**

(.026)

 − .012

(.027)

.063

(.051)

 − .381

(1.59)

 − .046***

(.014)

.038

(.048)

.108

(.095)

N

9,860

6,380

2,846

8,454

5,890

2,600

6,700

5,213

2,129

7,448

5,393

2,249

Decomposition:

            

Explained (E)

 − .067***

(.010)

 − .021

(.033)

 − .011

(.079)

 − .041***

(.008)

 − .010

(.012)

 − .022

(.014)

 − .073***

(.025)

 − .097

(.060)

.148

(2.90)

 − .075***

(.015)

 − .080**

(.033)

 − .047

(.202)

Unexplained (U)

.023**

(.012)

.062**

(.028)

.075*

(.044)

.036***

(.012)

.017

(.020)

.072***

(.026)

.067***

(.025)

.080

(.051)

 − .392

(1.59)

.046***

(.016)

.099**

(.049)

.128

(.095)

Interaction^

Yes

Yes

Yes

 − .016**

(.007)

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Contribution to E^

        

Gender

 − .007***

(.002)

Yes

Yes

 − .005***

(.001)

.004*

(.002)

Yes

Yes

Yes

Yes

 − .010***

(.003)

Yes

Yes

Age

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

NESB

Yes

Yes

Yes

 − 

 − 

 − 

 − 

 − 

 − 

 − 

 − 

 − 

Go8

Yes

Yes

Yes

Yes

Yes

 − .012*

(.006)

Yes

Yes

Yes

Yes

Yes

Yes

Work part-time

.023***

(.003)

.021**

(.009)

.025***

(.009)

.007***

(.002)

.005**

(.002)

.006**

(.003)

.017***

(.005)

Yes

Yes

.027***

(.003)

.028***

(.010)

.029**

(.013)

  1. Mean wage gap decompositions using Oaxaca-Blinder method (Stata command: oaxaca) following the two-step method proposed by Firpo, Fortin, and Lemieux (2011). The covariates used in the model are human capital controls (gender, age, age square, if disable, if Aboriginal, if English is main language spoken at home, if graduated from Go8 university, if worked in last year of study, mode of attendance, share of foreign students in same field of education and university), institutional and labour market controls (lagged average wage and lagged unemployment rate by field of study and year), and dummy variables for the survey year and the geographical location of the employer. Adjustment is made for selection into emigration. Standard errors are bootstrapped (50 draws) and clustered by university. The signs *, **, and *** indicate p-values of < .1, < .05, and < .01, respectively
  2. ^ To ease reading of the details of the Table, only estimates that are different from zero at a 10% of lower level of statistical significance are reported. Estimates that are statistically identical to zero are reported as ‘Yes’