“Gender, Competitiveness and Career Choices”niederle/BNO.QJE.Appendix.pdf · Thomas Buser,...
31
Supplementary material to: “Gender, Competitiveness and Career Choices” Thomas Buser, Muriel Niederle and Hessel Oosterbeek This document provides supplementary material to our paper “Gender, Competitiveness and Career Choices”. Appendix I presents tables with additional results. This includes descriptive statistics of the control variables (A.I), a comparison of study tracks by gender in the national statistics and in our sample (A.II), the distribution of undergraduate major students by study track at the pre-university level (A.III), estimates of the relation between gender and track choice for different specifications without psychological variables (A.IV), ordered probit estimates of the relation between track choice and explanatory variables for the main specifications (A.V), ordered probit estimates of the relation between track choice and explanatory variables for the main specifications but with class fixed effects instead of school fixed effects (A.VI), ordered probit estimates of the relation between track choice and explanatory variables for the main specifications extended with controls for student age and name-group fixed effects (A.VII), ordered probit estimates of the relation between track choice and explanatory variables for the main specifications where observations are weighted to reproduce the average distribution of chosen tracks by gender in the Netherlands (A.VIII), ordered probit estimates of the relation between track choice and explanatory variables for the main specifications extended with controls for class-level variables (A.IX), ordered probit estimates of the relation between track choice and explanatory variables for the main specifications with alternative assignment of NT/NH and ES/CS students (A.X), ordered probit estimates of the relation between track choice and ex- planatory variables for the main specifications with NT/NH and ES/CS as separate tracks (A.XI), ordered probit estimates of the relation between track choice and explanatory variables for the main specifications where tracks are ranked according to students’ own rankings (A.XII), linear probabil- ity models of the relation between the choice of NT vs other tracks and explanatory variables for the main specifications (A.XIII), linear probability models of the relation between the choice of NT or NH vs ES or CS and explanatory variables for the main specifications (A.XIV), linear probability models of the relation between the choice of CS vs other tracks and explanatory variables for the main specifications (A.XV), linear probability models of the choice of self-ranked best track vs other tracks for the main specifications (A.XVI), and OLS estimates of the relation between track choice and explanatory variables for the main specifications. Appendix II presents a figure about the relation between tournament entry (conditional on performance) by gender and subsequent track choice. Appendix III analyzes the relation between gender and study track choice for four subsamples: i) competitive boys and non-competitive girls, ii) competitive boys and competitive girls, iii) non- competitive boys and non-competitive girls, and iv) non-competitive boys and competitive girls. 1
Transcript of “Gender, Competitiveness and Career Choices”niederle/BNO.QJE.Appendix.pdf · Thomas Buser,...
Supplementary material to:
“Gender, Competitiveness and Career Choices”
Thomas Buser, Muriel Niederle and Hessel Oosterbeek
This document provides supplementary material to our paper “Gender, Competitiveness and CareerChoices”.
Appendix I presents tables with additional results. This includes descriptive statistics of thecontrol variables (A.I), a comparison of study tracks by gender in the national statistics and in oursample (A.II), the distribution of undergraduate major students by study track at the pre-universitylevel (A.III), estimates of the relation between gender and track choice for different specificationswithout psychological variables (A.IV), ordered probit estimates of the relation between track choiceand explanatory variables for the main specifications (A.V), ordered probit estimates of the relationbetween track choice and explanatory variables for the main specifications but with class fixed effectsinstead of school fixed effects (A.VI), ordered probit estimates of the relation between track choiceand explanatory variables for the main specifications extended with controls for student age andname-group fixed effects (A.VII), ordered probit estimates of the relation between track choice andexplanatory variables for the main specifications where observations are weighted to reproduce theaverage distribution of chosen tracks by gender in the Netherlands (A.VIII), ordered probit estimatesof the relation between track choice and explanatory variables for the main specifications extendedwith controls for class-level variables (A.IX), ordered probit estimates of the relation between trackchoice and explanatory variables for the main specifications with alternative assignment of NT/NHand ES/CS students (A.X), ordered probit estimates of the relation between track choice and ex-planatory variables for the main specifications with NT/NH and ES/CS as separate tracks (A.XI),ordered probit estimates of the relation between track choice and explanatory variables for the mainspecifications where tracks are ranked according to students’ own rankings (A.XII), linear probabil-ity models of the relation between the choice of NT vs other tracks and explanatory variables forthe main specifications (A.XIII), linear probability models of the relation between the choice of NTor NH vs ES or CS and explanatory variables for the main specifications (A.XIV), linear probabilitymodels of the relation between the choice of CS vs other tracks and explanatory variables for themain specifications (A.XV), linear probability models of the choice of self-ranked best track vs othertracks for the main specifications (A.XVI), and OLS estimates of the relation between track choiceand explanatory variables for the main specifications.
Appendix II presents a figure about the relation between tournament entry (conditional onperformance) by gender and subsequent track choice.
Appendix III analyzes the relation between gender and study track choice for four subsamples:i) competitive boys and non-competitive girls, ii) competitive boys and competitive girls, iii) non-competitive boys and non-competitive girls, and iv) non-competitive boys and competitive girls.
1
Appendix IV contains a translation of the instructions for the in-class experiment, a translationof the questionnaire the was administered after the in-class experiment, and an example sheet of theaddition task used in the experiment.
2
Appendix I: Tables
Table A.I. Means and standard deviations of control variables
Mean Standard-dev. Name origin group PercentageFemale 0.51 0.50 English 22.9Math grade 6.63 1.07 Dutch modern 22.4GPA 6.89 0.62 Dutch premodern 16.3Math relative 0.37 0.26 Elite 8.3Math difficulty 3.80 2.80 Old-testament 5.3Math quartile 2.12 0.96 Traditional 4.1Guessed rank 2.35 0.94 Arabic 3.3Risk 6.23 1.90 French 3.0Lottery 3.22 1.32 Frisian 3.0Age 15.46 0.95 Scandinavian 2.8
Latin 2.2Modern 1.4Slavic 0.8Turkish 0.6Other 3.6
N=362. The variables in the left hand panel were collected through the questionnaire at the end of the experiment(with the exception of grades which were provided by the schools at the end of the school year). Grades run from1 to 10 with 6 the first passing grade. Math relative is the normalized rank in terms of math grades of a studentwithin his own class. Math difficulty is the answer to the question “How difficult do you think it would be for you topass mathematics track B” and goes from 0 - very easy to 10 - very hard. Math quartile is the answer to a questionasking the students to rank themselves on mathematical ability compared to other students in their year (and school)on a scale from 1 (in the best 25%) to 4 (in the worst 25%). Risk is the answer to the question “How do you seeyourself: Are you generally a person who is fully prepared to take risks or do you try to avoid taking risks?” and goesfrom 0 (“unwilling to take risks”) to 10 (“fully prepared to take risk”). Lottery is the outcome of a choice between acertain payoff and four 50/50 lotteries whereby 1 represents the safe option and 5 represents the riskiest and highestexpected reward lottery. The name origin groups in the right hand panel come from Bloothooft and Onland (2011),who show that in the Netherlands, first names are strongly predictive of social class, income and lifestyle, and developa classification of names into 14 socio-economic categories.
Table A.II. Study tracks by gender: national statistics and sample (percentages)
National statistics National statisticsBoys Girls Boys Girls
NT 43 23 NT 26 8NH 17 26 NT/NH 17 15ES 35 32 NH 17 26CS 5 18 ES 30 23
ES/CS 5 9CS 5 18
Our sample Our sampleBoys Girls Boys Girls
NT 40 17 NT 35 10NH 12 36 NT/NH 14 23ES 39 32 NH 8 21CS 8 15 ES 34 27
ES/CS 4 7CS 6 12
N 177 185 N 161 181Source of the national data: CBS (2012), the data are from 2012. In the left hand panel, we treat NT/NH studentsas NT and ES/CS students as ES in the national data. In our own sample, of the 22 boys who chose NT/NH, 15stated NT as their favorite track in the questionnaire, six NH and one CS. Of the 42 girls who chose NT/NH, 13 putNT and 29 put NH. Of the six boys that chose ES/CS all six put ES. Of the 12 girls that chose ES/CS, eight put ESand four put CS. We use this information to split them into the four tracks in the left hand panel.
Table A.III. Study tracks by tertiary education choices and gender (percentages): national averages
NT NH ES CSA Undergraduate major
Humanities 9 6 8 30Social Sciences 2 9 19 34Law 1 4 20 20Economics and Business 15 8 46 5Science and Engineering 64 18 2 0Health Care 7 48 1 1Other 2 7 4 9
B Going to university 81 72 69 60Source: Panel A: CBS (http://www.cbs.nl/nl-NL/menu/themas/onderwijs/publicaties/artikelen/archief/2007/2007-2193-wm.htm), the data are from 2006; Panel B: CBS (2010), the data are from 2009.
Table
A.IV
.G
ende
ran
dtr
ack
choi
ce(a
lter
nati
vesp
ecifi
cati
ons)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
NT
/NH
asN
Tan
dE
S/C
San
dE
SN
T/N
Han
dE
S/C
Sas
sepa
rate
trac
ksSt
uden
ts’o
wn
rank
ing
Nat
ure
vs.
Soci
ety
CS
vs.
rest
Fem
ale
-0.2
79⇤⇤
-0.3
52⇤⇤
⇤-0
.230
⇤-0
.445
⇤⇤⇤
-0.5
02⇤⇤
⇤-0
.391
⇤⇤⇤
-0.4
37⇤⇤
⇤-0
.533
⇤⇤⇤
-0.4
55⇤⇤
⇤0.
017
0.03
90.
072⇤
⇤0.
066⇤
(0.1
21)
(0.1
28)
(0.1
31)
(0.1
15)
(0.1
19)
(0.1
21)
(0.1
19)
(0.1
26)
(0.1
28)
(0.0
53)
(0.0
47)
(0.0
34)
(0.0
37)
Mat
hG
rade
0.28
00.
032
0.14
6-0
.082
-0.0
58-0
.232
-0.0
310.
053
(0.2
05)
(0.2
11)
(0.1
82)
(0.1
86)
(0.1
69)
(0.1
75)
(0.0
65)
(0.0
49)
GPA
0.30
0⇤⇤⇤
0.31
0⇤⇤⇤
0.21
6⇤⇤
0.21
3⇤⇤
0.22
3⇤⇤
0.21
5⇤⇤
0.12
3⇤⇤⇤
-0.0
46⇤
(0.1
09)
(0.1
07)
(0.1
00)
(0.0
98)
(0.0
94)
(0.0
92)
(0.0
33)
(0.0
28)
Rel
.M
ath
Gr.
-0.0
140.
011
-0.1
47-0
.127
-0.2
51⇤
-0.2
38-0
.030
0.03
9
(0.1
64)
(0.1
63)
(0.1
49)
(0.1
48)
(0.1
45)
(0.1
48)
(0.0
55)
(0.0
41)
Mat
hD
ifficu
lty
-0.2
44⇤⇤
⇤-0
.219
⇤⇤-0
.219
⇤⇤-0
.088
⇤⇤⇤
0.02
9
(0.0
94)
(0.0
86)
(0.0
90)
(0.0
33)
(0.0
23)
Mat
hQ
uart
ile-0
.334
⇤⇤⇤
-0.3
26⇤⇤
⇤-0
.138
⇤-0
.129
⇤⇤⇤
0.04
4⇤⇤
(0.0
81)
(0.0
76)
(0.0
75)
(0.0
29)
(0.0
22)
Cut
1-1
.604
⇤⇤⇤
3.47
6⇤⇤
0.84
6-1
.450
⇤⇤⇤
1.65
8-0
.917
-1.4
62⇤⇤
⇤0.
175
-1.6
22
Cut
2-0
.391
⇤⇤4.
792⇤
⇤⇤2.
278
-1.1
74⇤⇤
⇤1.
944
-0.6
16-0
.529
⇤⇤⇤
1.17
0-0
.596
Cut
3-0
.002
5.23
9⇤⇤⇤
2.77
0⇤-0
.236
2.97
2⇤⇤
0.49
90.
302⇤
⇤2.
065
0.32
9
Cut
40.
144
3.40
3⇤⇤⇤
0.97
1
Cut
50.
710⇤
⇤⇤4.
021⇤
⇤⇤1.
638
F/(
Cm
ax-C
1)-0
.174
⇤⇤-0
.200
⇤⇤⇤
-0.1
12⇤⇤
-0.2
06⇤⇤
⇤-0
.212
⇤⇤⇤
-0.1
53⇤⇤
⇤-0
.248
⇤⇤⇤
-0.2
82⇤⇤
⇤-0
.233
⇤⇤⇤
Obs
erva
tion
s34
234
234
234
234
234
235
435
435
436
236
236
236
2N
ote:
Dep
ende
ntva
riab
le:
Trac
kch
oice
.C
oeffi
cien
tsin
colu
mns
(1)
to(9
)ar
efr
omor
dere
dpr
obit
regr
essi
ons;
coeffi
cien
tsin
colu
mns
(10)
to(1
3)ar
efr
omO
LSre
gres
sion
s.A
llre
gres
sion
sco
ntro
lfo
rsc
hool
fixed
effec
ts.
Rob
ust
stan
dard
erro
rsin
pare
nthe
ses;
p-va
lues
for
F/(
Cm
ax-C
1)ar
ebo
otst
rapp
ed;
*p<
0.10
,**
p<
0.05
,**
*p<
0.01
.
Table
A.V
.Tr
ack
choi
ce:
orde
red
prob
itre
gres
sion
(ful
ltab
le)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
Fem
ale
-0.3
33⇤⇤
⇤-0
.250
⇤⇤-0
.462
⇤⇤⇤
-0.3
99⇤⇤
⇤-0
.205
⇤-0
.140
-0.3
37⇤⇤
⇤-0
.272
⇤⇤-0
.355
⇤⇤⇤
-0.2
99⇤⇤
-0.2
83⇤⇤
-0.2
30⇤
-0.3
01⇤⇤
-0.2
56⇤
(0.1
18)
(0.1
22)
(0.1
28)
(0.1
31)
(0.1
22)
(0.1
27)
(0.1
29)
(0.1
34)
(0.1
30)
(0.1
34)
(0.1
32)
(0.1
34)
(0.1
32)
(0.1
34)
Ent
ry0.
373⇤
⇤⇤0.
318⇤
⇤0.
306⇤
⇤0.
333⇤
⇤0.
425⇤
⇤⇤0.
336⇤
⇤0.
414⇤
⇤⇤
(0.1
31)
(0.1
33)
(0.1
32)
(0.1
35)
(0.1
44)
(0.1
43)
(0.1
52)
Mat
hgr
ade
0.16
30.
111
-0.0
94-0
.150
-0.0
78-0
.129
-0.1
13-0
.170
-0.0
97-0
.148
(0.1
85)
(0.1
88)
(0.1
93)
(0.1
96)
(0.1
93)
(0.1
95)
(0.1
96)
(0.1
98)
(0.1
96)
(0.1
97)
GPA
0.24
9⇤⇤
0.27
7⇤⇤⇤
0.25
0⇤⇤⇤
0.27
9⇤⇤⇤
0.24
8⇤⇤⇤
0.28
2⇤⇤⇤
0.25
2⇤⇤⇤
0.27
3⇤⇤⇤
0.25
1⇤⇤⇤
0.27
7⇤⇤⇤
(0.0
97)
(0.0
96)
(0.0
95)
(0.0
95)
(0.0
95)
(0.0
95)
(0.0
95)
(0.0
95)
(0.0
95)
(0.0
95)
Mat
hre
lati
ve-0
.171
-0.1
87-0
.168
-0.1
87-0
.164
-0.1
83-0
.174
-0.1
89-0
.171
-0.1
85
(0.1
54)
(0.1
56)
(0.1
56)
(0.1
57)
(0.1
55)
(0.1
56)
(0.1
59)
(0.1
60)
(0.1
58)
(0.1
58)
Mat
hdi
fficu
lty
-0.3
42⇤⇤
⇤-0
.329
⇤⇤⇤
-0.2
25⇤⇤
-0.2
18⇤⇤
-0.2
28⇤⇤
-0.2
24⇤⇤
-0.2
44⇤⇤
⇤-0
.242
⇤⇤⇤
-0.2
46⇤⇤
⇤-0
.246
⇤⇤⇤
(0.0
81)
(0.0
81)
(0.0
89)
(0.0
89)
(0.0
90)
(0.0
90)
(0.0
93)
(0.0
92)
(0.0
94)
(0.0
94)
Mat
hqu
arti
le-0
.357
⇤⇤⇤
-0.3
62⇤⇤
⇤-0
.329
⇤⇤⇤
-0.3
36⇤⇤
⇤-0
.334
⇤⇤⇤
-0.3
50⇤⇤
⇤-0
.338
⇤⇤⇤
-0.3
45⇤⇤
⇤-0
.343
⇤⇤⇤
-0.3
57⇤⇤
⇤
(0.0
77)
(0.0
78)
(0.0
76)
(0.0
76)
(0.0
76)
(0.0
77)
(0.0
77)
(0.0
77)
(0.0
77)
(0.0
78)
Gue
ssed
rank
0.06
00.
143⇤
0.06
10.
131
(0.0
79)
(0.0
83)
(0.0
81)
(0.0
86)
Ris
k-0
.059
-0.1
03-0
.049
-0.0
92
(0.0
68)
(0.0
69)
(0.0
68)
(0.0
69)
Lot
tery
0.18
1⇤⇤
0.16
9⇤⇤
0.18
1⇤⇤
0.16
5⇤⇤
(0.0
73)
(0.0
74)
(0.0
74)
(0.0
74)
Cut
1-1
.180
⇤⇤⇤
-1.1
66⇤⇤
⇤2.
007
1.96
8-2
.637
⇤⇤⇤
-2.6
15⇤⇤
⇤-0
.645
-0.7
05-0
.375
-0.0
72-0
.583
-0.8
91-0
.253
-0.2
60
Cut
2-0
.049
-0.0
203.
261⇤
⇤3.
235⇤
⇤-1
.321
⇤⇤⇤
-1.2
88⇤⇤
⇤0.
712
0.66
60.
983
1.30
30.
795
0.49
81.
124
1.13
1
Cut
30.
625⇤
⇤0.
661⇤
⇤4.
029⇤
⇤⇤4.
008⇤
⇤⇤-0
.515
⇤-0
.476
1.54
71.
508
1.81
92.
151
1.63
71.
348
1.96
71.
985
Fem
ale/
(C3-
C1)
-0.1
84⇤⇤
⇤-0
.137
⇤⇤-0
.229
⇤⇤⇤
-0.1
96⇤⇤
-0.0
97⇤
-0.0
65-0
.154
⇤⇤⇤
-0.1
23⇤⇤
-0.1
62⇤⇤
⇤-0
.134
⇤⇤-0
.128
⇤⇤-0
.103
⇤⇤-0
.136
⇤⇤-0
.114
⇤⇤
Diff
.25
.9%
14.3
%32
.3%
20.0
%17
.1%
19.4
%16
.1%
Boo
tstr
app-
valu
e0.
003
0.01
00.
013
0.00
90.
005
0.01
40.
012
Obs
erva
tion
s36
236
236
236
236
236
236
236
236
236
236
236
236
236
2N
ote:
Coe
ffici
ents
are
from
orde
red
prob
itre
gres
sion
s,w
here
NT
>N
H>
ES>
CS.
All
spec
ifica
tion
sin
clud
eco
ntro
lsfo
rpe
rfor
man
cein
Rou
nds
1an
d2
ofth
eex
peri
men
t,th
e
chan
ceof
win
ning
the
Rou
nd2
tour
nam
ent,
scho
olfix
edeff
ects
and
test
vers
ion
fixed
effec
ts.
Rob
ust
stan
dard
erro
rsin
pare
nthe
ses;
p-va
lues
for
F/(
C3-
C1)
and
Diff
.ar
e
boot
stra
pped
;*p<
0.10
,**
p<
0.05
,**
*p<
0.01.T
heim
pact
ofco
nfide
nce
(com
pari
ngco
lum
ns(7
)an
d(9
))an
dri
skat
titu
des
(com
pari
ngco
lum
ns(7
)an
d(1
1))
onth
e
gend
erga
p(F
emal
e/(C
3-C
1))
and
the
asso
ciat
edp-
valu
esar
e5.
3%(i
ncre
asin
g)(p
=0.
76)
and
16.0
%(p
=0.
02),
resp
ecti
vely
.
Table
A.V
I.
Trac
kch
oice
:or
dere
dpr
obit
regr
essi
on(c
lass
fixed
effec
ts)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
Fem
ale
-0.4
36⇤⇤
⇤-0
.343
⇤⇤⇤
-0.5
58⇤⇤
⇤-0
.487
⇤⇤⇤
-0.3
05⇤⇤
-0.2
39⇤
-0.4
39⇤⇤
⇤-0
.375
⇤⇤⇤
-0.4
66⇤⇤
⇤-0
.409
⇤⇤⇤
-0.3
92⇤⇤
⇤-0
.339
⇤⇤-0
.417
⇤⇤⇤
-0.3
70⇤⇤
⇤
(0.1
20)
(0.1
24)
(0.1
28)
(0.1
31)
(0.1
23)
(0.1
28)
(0.1
30)
(0.1
33)
(0.1
30)
(0.1
33)
(0.1
33)
(0.1
35)
(0.1
33)
(0.1
34)
Ent
ry0.
458⇤
⇤⇤0.
376⇤
⇤⇤0.
340⇤
⇤0.
355⇤
⇤0.
456⇤
⇤⇤0.
378⇤
⇤0.
461⇤
⇤⇤
(0.1
42)
(0.1
44)
(0.1
42)
(0.1
46)
(0.1
52)
(0.1
56)
(0.1
62)
Mat
hgr
ade
0.26
90.
212
0.01
5-0
.036
0.03
9-0
.006
-0.0
26-0
.067
-0.0
05-0
.039
(0.2
14)
(0.2
19)
(0.2
20)
(0.2
26)
(0.2
22)
(0.2
25)
(0.2
23)
(0.2
26)
(0.2
25)
(0.2
26)
GPA
0.25
7⇤⇤
0.28
3⇤⇤⇤
0.27
2⇤⇤⇤
0.29
6⇤⇤⇤
0.27
2⇤⇤⇤
0.30
2⇤⇤⇤
0.27
2⇤⇤⇤
0.29
0⇤⇤⇤
0.27
3⇤⇤⇤
0.29
6⇤⇤⇤
(0.1
04)
(0.1
04)
(0.1
01)
(0.1
01)
(0.1
01)
(0.1
01)
(0.1
01)
(0.1
01)
(0.1
01)
(0.1
01)
Mat
hre
lati
ve-0
.099
-0.1
17-0
.089
-0.1
09-0
.080
-0.0
99-0
.113
-0.1
16-0
.109
-0.1
09
(0.1
79)
(0.1
81)
(0.1
84)
(0.1
86)
(0.1
83)
(0.1
85)
(0.1
87)
(0.1
88)
(0.1
87)
(0.1
87)
Mat
hdi
fficu
lty
-0.3
77⇤⇤
⇤-0
.359
⇤⇤⇤
-0.2
42⇤⇤
⇤-0
.231
⇤⇤-0
.248
⇤⇤⇤
-0.2
38⇤⇤
⇤-0
.261
⇤⇤⇤
-0.2
56⇤⇤
⇤-0
.264
⇤⇤⇤
-0.2
60⇤⇤
⇤
(0.0
82)
(0.0
82)
(0.0
91)
(0.0
91)
(0.0
92)
(0.0
92)
(0.0
95)
(0.0
95)
(0.0
96)
(0.0
97)
Mat
hqu
arti
le-0
.324
⇤⇤⇤
-0.3
24⇤⇤
⇤-0
.286
⇤⇤⇤
-0.2
90⇤⇤
⇤-0
.294
⇤⇤⇤
-0.3
05⇤⇤
⇤-0
.297
⇤⇤⇤
-0.3
02⇤⇤
⇤-0
.304
⇤⇤⇤
-0.3
15⇤⇤
⇤
(0.0
79)
(0.0
79)
(0.0
78)
(0.0
78)
(0.0
78)
(0.0
79)
(0.0
78)
(0.0
78)
(0.0
79)
(0.0
79)
Gue
ssed
rank
0.08
90.
169⇤
⇤0.
089
0.15
5⇤
(0.0
83)
(0.0
85)
(0.0
85)
(0.0
87)
Ris
k-0
.060
-0.1
12-0
.044
-0.0
95
(0.0
69)
(0.0
71)
(0.0
70)
(0.0
72)
Lot
tery
0.15
6⇤⇤
0.14
2⇤0.
155⇤
⇤0.
136⇤
(0.0
75)
(0.0
75)
(0.0
75)
(0.0
75)
Cut
1-1
.545
⇤⇤⇤
-1.4
43⇤⇤
⇤2.
462
2.43
2-3
.034
⇤⇤⇤
-2.9
33⇤⇤
⇤0.
010
-0.0
010.
424
0.78
4-0
.123
-0.3
230.
343
0.45
2
Cut
2-0
.365
-0.2
433.
773⇤
⇤3.
760⇤
⇤-1
.677
⇤⇤⇤
-1.5
62⇤⇤
⇤1.
411
1.41
61.
825
2.20
41.
294
1.10
71.
760
1.88
5
Cut
30.
353
0.48
74.
602⇤
⇤⇤4.
597⇤
⇤⇤-0
.830
⇤-0
.708
2.29
72.
310
2.71
43.
107⇤
2.18
42.
008
2.65
42.
793
Fem
ale/
(C3-
C1)
-0.2
30⇤⇤
⇤-0
.178
⇤⇤⇤
-0.2
61⇤⇤
⇤-0
.225
⇤⇤⇤
-0.1
38⇤⇤
⇤-0
.108
⇤⇤-0
.192
⇤⇤⇤
-0.1
62⇤⇤
⇤-0
.204
⇤⇤⇤
-0.1
76⇤⇤
⇤-0
.170
⇤⇤⇤
-0.1
45⇤⇤
⇤-0
.180
⇤⇤⇤
-0.1
58⇤⇤
⇤
Diff
.22
.7%
13.6
%22
.1%
15.4
%13
.6%
14.4
%12
.3%
Boo
tstr
app-
valu
e0.
001
0.00
60.
014
0.01
30.
004
0.01
40.
013
Obs
erva
tion
s36
236
236
236
236
236
236
236
236
236
236
236
236
236
2N
ote:
Coe
ffici
ents
are
from
orde
red
prob
itre
gres
sion
s,w
here
NT
>N
H>
ES>
CS.
All
spec
ifica
tion
sin
clud
eco
ntro
lsfo
rpe
rfor
man
cein
Rou
nds
1an
d2
ofth
eex
peri
men
t,th
e
chan
ceof
win
ning
the
Rou
nd2
tour
nam
ent,
clas
sfix
edeff
ects
and
test
vers
ion
fixed
effec
ts.
Rob
ust
stan
dard
erro
rsin
pare
nthe
ses;
p-va
lues
for
Fem
ale/
(C3-
C1)
and
Diff
.ar
e
boot
stra
pped
;*p<
0.10
,**
p<
0.05
,**
*p<
0.01.T
heim
pact
ofco
nfide
nce
(com
pari
ngco
lum
ns(7
)an
d(9
))an
dri
skat
titu
des
(com
pari
ngco
lum
ns(7
)an
d(1
1))
onth
e
gend
erga
p(F
emal
e/(C
3-C
1))
and
the
asso
ciat
edp-
valu
esar
e6.
1%(i
ncre
asin
g)(p
=0.
84)
and
11.5
%(p
=0.
05),
resp
ecti
vely
.
Table
A.V
II.
Trac
kch
oice
:or
dere
dpr
obit
regr
essi
on(a
geco
ntro
land
nam
e-gr
oup
fixed
effec
ts)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
Fem
ale
-0.3
21⇤⇤
-0.2
54⇤
-0.4
80⇤⇤
⇤-0
.429
⇤⇤⇤
-0.2
00-0
.147
-0.3
50⇤⇤
⇤-0
.295
⇤⇤-0
.364
⇤⇤⇤
-0.3
17⇤⇤
-0.2
83⇤⇤
-0.2
35⇤
-0.2
92⇤⇤
-0.2
52⇤
(0.1
25)
(0.1
31)
(0.1
33)
(0.1
37)
(0.1
28)
(0.1
35)
(0.1
35)
(0.1
40)
(0.1
37)
(0.1
41)
(0.1
39)
(0.1
42)
(0.1
41)
(0.1
43)
Ent
ry0.
322⇤
⇤0.
270⇤
0.26
2⇤0.
291⇤
⇤0.
364⇤
⇤0.
319⇤
⇤0.
371⇤
⇤
(0.1
39)
(0.1
41)
(0.1
37)
(0.1
41)
(0.1
48)
(0.1
50)
(0.1
58)
Ara
bic
0.77
3⇤⇤
0.75
7⇤⇤
1.18
5⇤⇤⇤
1.14
5⇤⇤⇤
0.81
5⇤⇤⇤
0.78
9⇤⇤⇤
1.01
1⇤⇤⇤
0.96
2⇤⇤⇤
1.00
7⇤⇤⇤
0.94
2⇤⇤⇤
0.99
9⇤⇤⇤
0.96
2⇤⇤⇤
0.99
5⇤⇤⇤
0.94
3⇤⇤⇤
(0.3
11)
(0.3
14)
(0.2
88)
(0.2
88)
(0.3
03)
(0.2
94)
(0.3
04)
(0.2
99)
(0.3
04)
(0.2
96)
(0.3
08)
(0.3
08)
(0.3
06)
(0.3
04)
Elit
e-0
.403
⇤-0
.442
⇤-0
.530
⇤⇤-0
.564
⇤⇤-0
.559
⇤⇤-0
.590
⇤⇤-0
.604
⇤⇤-0
.642
⇤⇤-0
.590
⇤⇤-0
.616
⇤⇤-0
.701
⇤⇤⇤
-0.7
46⇤⇤
⇤-0
.690
⇤⇤⇤
-0.7
21⇤⇤
⇤
(0.2
39)
(0.2
36)
(0.2
55)
(0.2
53)
(0.2
66)
(0.2
65)
(0.2
65)
(0.2
63)
(0.2
69)
(0.2
66)
(0.2
59)
(0.2
59)
(0.2
64)
(0.2
62)
Eng
lish
0.16
10.
123
0.25
00.
222
0.14
60.
118
0.22
30.
194
0.22
60.
195
0.15
00.
123
0.15
20.
125
(0.1
83)
(0.1
83)
(0.1
80)
(0.1
81)
(0.1
90)
(0.1
92)
(0.1
88)
(0.1
90)
(0.1
89)
(0.1
90)
(0.1
91)
(0.1
93)
(0.1
92)
(0.1
92)
Fren
ch-0
.189
-0.1
39-0
.037
-0.0
04-0
.246
-0.2
04-0
.194
-0.1
59-0
.199
-0.1
61-0
.281
-0.2
36-0
.284
-0.2
38(0
.361
)(0
.363
)(0
.402
)(0
.395
)(0
.426
)(0
.425
)(0
.436
)(0
.428
)(0
.442
)(0
.443
)(0
.419
)(0
.415
)(0
.423
)(0
.427
)Sc
andi
navi
an0.
260
0.20
90.
091
0.05
0-0
.290
-0.3
27-0
.262
-0.3
07-0
.256
-0.3
03-0
.263
-0.3
08-0
.259
-0.3
05(0
.391
)(0
.384
)(0
.420
)(0
.421
)(0
.331
)(0
.325
)(0
.387
)(0
.384
)(0
.386
)(0
.380
)(0
.393
)(0
.385
)(0
.394
)(0
.383
)Fr
isia
n0.
067
0.00
1-0
.002
-0.0
610.
204
0.15
00.
114
0.05
20.
122
0.05
70.
106
0.05
60.
109
0.05
9(0
.340
)(0
.345
)(0
.389
)(0
.402
)(0
.311
)(0
.320
)(0
.351
)(0
.366
)(0
.352
)(0
.367
)(0
.359
)(0
.371
)(0
.360
)(0
.371
)Lat
in-0
.575
-0.6
47⇤
-0.4
77-0
.535
-0.6
07⇤
-0.6
65⇤
-0.5
48-0
.612
⇤-0
.524
-0.5
67-0
.608
⇤-0
.673
⇤-0
.591
-0.6
36(0
.376
)(0
.384
)(0
.347
)(0
.355
)(0
.340
)(0
.356
)(0
.337
)(0
.350
)(0
.346
)(0
.370
)(0
.362
)(0
.371
)(0
.371
)(0
.388
)M
oder
n-0
.137
-0.1
02-0
.247
-0.2
32-0
.329
-0.3
01-0
.440
-0.4
27-0
.427
-0.3
90-0
.503
-0.4
89-0
.493
-0.4
57(0
.448
)(0
.431
)(0
.459
)(0
.441
)(0
.428
)(0
.416
)(0
.411
)(0
.392
)(0
.420
)(0
.406
)(0
.349
)(0
.336
)(0
.356
)(0
.351
)O
ther
0.35
30.
296
0.30
30.
256
0.37
30.
326
0.34
80.
298
0.35
00.
290
0.30
70.
276
0.30
50.
267
(0.3
16)
(0.3
27)
(0.3
58)
(0.3
61)
(0.3
13)
(0.3
18)
(0.3
55)
(0.3
56)
(0.3
52)
(0.3
48)
(0.3
69)
(0.3
68)
(0.3
67)
(0.3
59)
Old
-tes
tam
ent
0.61
0⇤0.
596⇤
0.63
8⇤0.
632⇤
0.55
9⇤0.
554
0.58
5⇤0.
578⇤
0.59
6⇤0.
605⇤
0.57
1⇤0.
561⇤
0.58
0⇤0.
584⇤
(0.3
51)
(0.3
51)
(0.3
49)
(0.3
53)
(0.3
39)
(0.3
41)
(0.3
36)
(0.3
39)
(0.3
38)
(0.3
39)
(0.3
33)
(0.3
30)
(0.3
35)
(0.3
32)
Slav
ic0.
526
0.32
60.
591
0.41
70.
360
0.20
00.
452
0.26
60.
483
0.29
70.
298
0.11
80.
318
0.14
6(0
.484
)(0
.515
)(0
.366
)(0
.393
)(0
.421
)(0
.440
)(0
.362
)(0
.383
)(0
.371
)(0
.393
)(0
.350
)(0
.364
)(0
.360
)(0
.375
)Tra
diti
onal
0.27
50.
226
0.21
90.
179
0.21
40.
176
0.23
80.
198
0.25
00.
218
0.12
40.
064
0.13
50.
086
(0.3
32)
(0.3
32)
(0.3
40)
(0.3
38)
(0.3
21)
(0.3
19)
(0.3
32)
(0.3
30)
(0.3
33)
(0.3
29)
(0.3
48)
(0.3
49)
(0.3
51)
(0.3
48)
Tur
kish
1.57
4⇤⇤⇤
1.46
4⇤⇤⇤
1.80
8⇤⇤⇤
1.68
9⇤⇤⇤
2.09
9⇤⇤⇤
1.97
3⇤⇤⇤
2.02
4⇤⇤⇤
1.88
2⇤⇤⇤
2.05
5⇤⇤⇤
1.91
6⇤⇤⇤
2.16
5⇤⇤⇤
2.05
6⇤⇤⇤
2.18
1⇤⇤⇤
2.07
8⇤⇤⇤
(0.3
87)
(0.4
34)
(0.3
01)
(0.3
30)
(0.3
20)
(0.3
36)
(0.3
21)
(0.3
38)
(0.3
29)
(0.3
34)
(0.3
52)
(0.3
54)
(0.3
58)
(0.3
56)
Pre
-mod
ern
0.46
7⇤⇤
0.41
2⇤⇤
0.45
0⇤⇤
0.39
9⇤⇤
0.46
0⇤⇤
0.41
7⇤⇤
0.40
9⇤⇤
0.35
5⇤0.
420⇤
⇤0.
368⇤
0.36
6⇤0.
310
0.37
4⇤0.
322
(0.1
97)
(0.1
99)
(0.2
00)
(0.2
03)
(0.1
99)
(0.2
01)
(0.1
99)
(0.2
02)
(0.2
03)
(0.2
04)
(0.2
01)
(0.2
03)
(0.2
04)
(0.2
05)
Age
-0.1
64⇤⇤
-0.1
58⇤⇤
-0.1
67⇤⇤
-0.1
61⇤⇤
-0.0
81-0
.076
-0.0
93-0
.086
-0.0
91-0
.077
-0.1
07-0
.098
-0.1
05-0
.091
(0.0
65)
(0.0
66)
(0.0
66)
(0.0
66)
(0.0
69)
(0.0
69)
(0.0
69)
(0.0
69)
(0.0
69)
(0.0
70)
(0.0
69)
(0.0
69)
(0.0
69)
(0.0
70)
Mat
hgr
ade
0.15
70.
111
-0.1
26-0
.177
-0.1
14-0
.160
-0.1
41-0
.198
-0.1
33-0
.183
(0.1
98)
(0.2
01)
(0.2
08)
(0.2
11)
(0.2
09)
(0.2
11)
(0.2
09)
(0.2
11)
(0.2
10)
(0.2
10)
GPA
0.23
3⇤⇤
0.25
8⇤⇤
0.25
0⇤⇤
0.27
7⇤⇤⇤
0.24
8⇤⇤
0.27
9⇤⇤⇤
0.24
5⇤⇤
0.26
8⇤⇤⇤
0.24
5⇤⇤
0.27
0⇤⇤⇤
(0.1
02)
(0.1
02)
(0.1
00)
(0.1
01)
(0.1
01)
(0.1
01)
(0.1
01)
(0.1
01)
(0.1
01)
(0.1
01)
Mat
hre
lati
ve-0
.236
-0.2
50-0
.240
-0.2
58-0
.236
-0.2
52-0
.239
-0.2
54-0
.237
-0.2
49(0
.163
)(0
.164
)(0
.164
)(0
.165
)(0
.163
)(0
.163
)(0
.166
)(0
.166
)(0
.165
)(0
.165
)M
ath
diffi
culty
-0.3
77⇤⇤
⇤-0
.367
⇤⇤⇤
-0.2
35⇤⇤
-0.2
30⇤⇤
-0.2
38⇤⇤
-0.2
36⇤⇤
-0.2
70⇤⇤
⇤-0
.271
⇤⇤⇤
-0.2
71⇤⇤
⇤-0
.274
⇤⇤⇤
(0.0
85)
(0.0
85)
(0.0
96)
(0.0
95)
(0.0
98)
(0.0
97)
(0.1
01)
(0.1
00)
(0.1
02)
(0.1
01)
Mat
hqu
arti
le-0
.356
⇤⇤⇤
-0.3
60⇤⇤
⇤-0
.330
⇤⇤⇤
-0.3
37⇤⇤
⇤-0
.334
⇤⇤⇤
-0.3
49⇤⇤
⇤-0
.336
⇤⇤⇤
-0.3
43⇤⇤
⇤-0
.338
⇤⇤⇤
-0.3
52⇤⇤
⇤
(0.0
79)
(0.0
80)
(0.0
79)
(0.0
80)
(0.0
79)
(0.0
80)
(0.0
80)
(0.0
81)
(0.0
81)
(0.0
81)
Gue
ssed
rank
0.04
40.
113
0.03
10.
090
(0.0
83)
(0.0
87)
(0.0
87)
(0.0
91)
Ris
k-0
.106
-0.1
49⇤⇤
-0.1
00-0
.139
⇤
(0.0
73)
(0.0
75)
(0.0
74)
(0.0
75)
Lot
tery
0.21
4⇤⇤⇤
0.20
4⇤⇤⇤
0.21
3⇤⇤⇤
0.20
0⇤⇤⇤
(0.0
75)
(0.0
75)
(0.0
75)
(0.0
75)
Cut
1-6
.100
⇤⇤⇤
-5.9
32⇤⇤
⇤-3
.358
-3.1
99-5
.096
⇤⇤-4
.940
⇤⇤-3
.774
-3.5
93-3
.483
-2.8
04-4
.286
⇤-4
.332
⇤-4
.050
-3.6
59C
ut2
-4.9
09⇤⇤
-4.7
31⇤⇤
-2.0
22-1
.854
-3.7
09⇤
-3.5
44⇤
-2.3
35-2
.143
-2.0
44-1
.352
-2.8
21-2
.857
-2.5
85-2
.183
Cut
3-4
.193
⇤⇤-4
.009
⇤⇤-1
.195
-1.0
23-2
.845
-2.6
76-1
.435
-1.2
38-1
.144
-0.4
43-1
.913
-1.9
41-1
.676
-1.2
63Fe
mal
e/(C
3-C
1)-0
.176
⇤⇤⇤
-0.1
37⇤⇤
-0.2
27⇤⇤
⇤-0
.202
⇤⇤⇤
-0.1
01⇤
-0.0
75-0
.161
⇤⇤⇤
-0.1
36⇤⇤
-0.1
71⇤⇤
⇤-0
.147
⇤⇤-0
.134
⇤⇤-0
.114
⇤-0
.145
⇤⇤-0
.126
⇤⇤
Diff
.22
.1%
11.3
%25
.8%
15.8
%14
.0%
14.8
%12
.9%
Boo
tstr
app-
valu
e0.
014
0.03
50.
036
0.02
50.
013
0.02
60.
024
Obs
erva
tion
s35
835
835
835
835
835
835
835
835
835
835
835
835
835
8N
ote:
Coe
ffici
ents
are
from
orde
red
prob
itre
gres
sion
s,w
here
NT
>N
H>
ES>
CS.
All
spec
ifica
tion
sin
clud
eco
ntro
lsfo
rpe
rfor
man
cein
Rou
nds
1an
d2
ofth
eex
peri
men
t,th
e
chan
ceof
win
ning
the
Rou
nd2
tour
nam
ent,
scho
olfix
edeff
ects
and
test
vers
ion
fixed
effec
ts.
The
base
cate
gory
for
the
nam
e-gr
oup
dum
mie
sis
”Dut
chm
oder
n”.
Rob
ust
stan
dard
erro
rsin
pare
nthe
ses;
p-va
lues
for
Fem
ale/
(C3-
C1)
and
Diff
.ar
ebo
otst
rapp
ed;*
p<
0.10
,**p<
0.05
,***
p<
0.01.T
heim
pact
ofco
nfide
nce
(com
pari
ngco
lum
ns
(7)
and
(9))
and
risk
atti
tude
s(c
ompa
ring
colu
mns
(7)
and
(11)
)on
the
gend
erga
p(F
emal
e/(C
3-C
1))
and
the
asso
ciat
edp-
valu
esar
e6.
5%(i
ncre
asin
g)(p
=0.
70)
and
16.5
%
(p=
0.02
),re
spec
tive
ly.
Table
A.V
III.
Trac
kch
oice
:or
dere
dpr
obit
regr
essi
on(w
eigh
ted
regr
essi
ons)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
Fem
ale
-0.4
90⇤⇤
⇤-0
.410
⇤⇤⇤
-0.6
14⇤⇤
⇤-0
.554
⇤⇤⇤
-0.3
51⇤⇤
⇤-0
.294
⇤⇤-0
.481
⇤⇤⇤
-0.4
23⇤⇤
⇤-0
.505
⇤⇤⇤
-0.4
52⇤⇤
⇤-0
.423
⇤⇤⇤
-0.3
74⇤⇤
⇤-0
.446
⇤⇤⇤
-0.4
03⇤⇤
⇤
(0.1
21)
(0.1
24)
(0.1
28)
(0.1
30)
(0.1
24)
(0.1
29)
(0.1
31)
(0.1
35)
(0.1
30)
(0.1
34)
(0.1
33)
(0.1
35)
(0.1
32)
(0.1
33)
Ent
ry0.
358⇤
⇤⇤0.
298⇤
⇤0.
269⇤
⇤0.
298⇤
⇤0.
404⇤
⇤⇤0.
310⇤
⇤0.
407⇤
⇤
(0.1
35)
(0.1
38)
(0.1
36)
(0.1
40)
(0.1
52)
(0.1
50)
(0.1
60)
Mat
hgr
ade
0.13
00.
078
-0.1
43-0
.195
-0.1
26-0
.179
-0.1
68-0
.223
-0.1
49-0
.206
(0.1
96)
(0.2
00)
(0.2
04)
(0.2
07)
(0.2
03)
(0.2
05)
(0.2
07)
(0.2
10)
(0.2
07)
(0.2
08)
GPA
0.24
2⇤⇤
0.27
0⇤⇤⇤
0.24
4⇤⇤
0.27
2⇤⇤⇤
0.24
3⇤⇤
0.27
9⇤⇤⇤
0.24
6⇤⇤
0.26
9⇤⇤⇤
0.24
6⇤⇤
0.27
5⇤⇤⇤
(0.0
99)
(0.0
99)
(0.1
00)
(0.1
00)
(0.0
99)
(0.0
99)
(0.0
99)
(0.0
99)
(0.0
99)
(0.0
98)
Mat
hre
lati
ve-0
.221
-0.2
36-0
.230
-0.2
47-0
.227
-0.2
47-0
.237
-0.2
50-0
.235
-0.2
50
(0.1
60)
(0.1
62)
(0.1
62)
(0.1
64)
(0.1
61)
(0.1
62)
(0.1
66)
(0.1
67)
(0.1
65)
(0.1
65)
Mat
hdi
fficu
lty
-0.3
55⇤⇤
⇤-0
.343
⇤⇤⇤
-0.2
34⇤⇤
-0.2
28⇤⇤
-0.2
38⇤⇤
-0.2
35⇤⇤
-0.2
58⇤⇤
⇤-0
.257
⇤⇤⇤
-0.2
61⇤⇤
⇤-0
.263
⇤⇤⇤
(0.0
86)
(0.0
87)
(0.0
93)
(0.0
93)
(0.0
94)
(0.0
93)
(0.0
96)
(0.0
95)
(0.0
96)
(0.0
96)
Mat
hqu
arti
le-0
.353
⇤⇤⇤
-0.3
54⇤⇤
⇤-0
.325
⇤⇤⇤
-0.3
29⇤⇤
⇤-0
.331
⇤⇤⇤
-0.3
42⇤⇤
⇤-0
.333
⇤⇤⇤
-0.3
37⇤⇤
⇤-0
.339
⇤⇤⇤
-0.3
50⇤⇤
⇤
(0.0
81)
(0.0
82)
(0.0
79)
(0.0
80)
(0.0
79)
(0.0
81)
(0.0
80)
(0.0
80)
(0.0
80)
(0.0
81)
Gue
ssed
rank
0.07
40.
157⇤
0.08
00.
151⇤
(0.0
80)
(0.0
87)
(0.0
84)
(0.0
90)
Ris
k-0
.067
-0.1
11-0
.055
-0.1
03
(0.0
69)
(0.0
71)
(0.0
70)
(0.0
71)
Lot
tery
0.18
4⇤⇤
0.17
5⇤⇤
0.18
5⇤⇤
0.17
3⇤⇤
(0.0
75)
(0.0
75)
(0.0
75)
(0.0
75)
Cut
1-1
.340
⇤⇤⇤
-1.3
25⇤⇤
⇤1.
468
1.44
0-2
.814
⇤⇤⇤
-2.7
89⇤⇤
⇤-1
.292
-1.3
31-0
.964
-0.6
49-1
.280
-1.5
51-0
.869
-0.8
56
Cut
2-0
.227
-0.1
992.
708⇤
2.69
0⇤-1
.516
⇤⇤⇤
-1.4
83⇤⇤
⇤0.
051
0.02
20.
380
0.70
90.
083
-0.1
800.
495
0.52
0
Cut
30.
371
0.40
63.
394⇤
⇤3.
381⇤
⇤-0
.795
⇤⇤⇤
-0.7
57⇤⇤
0.80
00.
777
1.13
11.
469
0.83
90.
582
1.25
31.
288
Fem
ale/
(C3-
C1)
-0.2
86⇤⇤
⇤-0
.237
⇤⇤⇤
-0.3
19⇤⇤
⇤-0
.286
⇤⇤⇤
-0.1
74⇤⇤
⇤-0
.145
⇤⇤-0
.230
⇤⇤⇤
-0.2
01⇤⇤
⇤-0
.241
⇤⇤⇤
-0.2
14⇤⇤
⇤-0
.199
⇤⇤⇤
-0.1
75⇤⇤
⇤-0
.210
⇤⇤⇤
-0.1
88⇤⇤
⇤
Diff
.17
.4%
10.4
%16
.8%
12.7
%11
.4%
12.2
%10
.7%
Boo
tstr
app-
valu
e0.
004
0.01
70.
027
0.02
00.
008
0.02
30.
014
Obs
erva
tion
s36
236
236
236
236
236
236
236
236
236
236
236
236
236
2N
ote:
Coe
ffici
ents
are
from
orde
red
prob
itre
gres
sion
s,w
here
NT
>N
H>
ES>
CS.
All
spec
ifica
tion
sin
clud
eco
ntro
lsfo
rpe
rfor
man
cein
Rou
nds
1an
d2
ofth
eex
peri
men
t,th
e
chan
ceof
win
ning
the
Rou
nd2
tour
nam
ent,
scho
olfix
edeff
ects
and
test
vers
ion
fixed
effec
ts.
Rob
ust
stan
dard
erro
rsin
pare
nthe
ses;
p-va
lues
for
Fem
ale/
(C3-
C1)
and
Diff
.ar
e
boot
stra
pped
;*p<
0.10
,**p<
0.05
,***
p<
0.01.E
ach
obse
rvat
ion
isw
eigh
ted
acco
rdin
gto
gend
eran
dtr
ack
soas
tore
prod
uce
the
aver
age
dist
ribu
tion
ofch
osen
trac
ksby
gend
erin
the
Net
herl
ands
.T
heim
pact
ofco
nfide
nce
(com
pari
ngco
lum
ns(7
)an
d(9
))an
dri
skat
titu
des
(com
pari
ngco
lum
ns(7
)an
d(1
1))
onth
ege
nder
gap
(Fem
ale/
(C3-
C1)
)
and
the
asso
ciat
edp-
valu
esar
e4.
9%(i
ncre
asin
g)(p
=0.
80)
and
13.2
%(p
=0.
02),
resp
ecti
vely
.
Table
A.IX
.Tr
ack
choi
ce:
orde
red
prob
itre
gres
sion
(cla
ss-le
velc
ontr
ols)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
Fem
ale
-0.4
31⇤⇤
⇤-0
.342
⇤⇤⇤
-0.5
46⇤⇤
⇤-0
.477
⇤⇤⇤
-0.2
98⇤⇤
-0.2
28⇤
-0.4
16⇤⇤
⇤-0
.348
⇤⇤⇤
-0.4
28⇤⇤
⇤-0
.372
⇤⇤⇤
-0.3
62⇤⇤
⇤-0
.308
⇤⇤-0
.374
⇤⇤⇤
-0.3
30⇤⇤
(0.1
21)
(0.1
24)
(0.1
28)
(0.1
31)
(0.1
24)
(0.1
29)
(0.1
30)
(0.1
34)
(0.1
31)
(0.1
34)
(0.1
33)
(0.1
34)
(0.1
33)
(0.1
34)
Per
cent
fem
ale
2.47
6⇤⇤⇤
2.72
9⇤⇤⇤
2.34
9⇤⇤⇤
2.54
6⇤⇤⇤
2.17
2⇤⇤⇤
2.37
6⇤⇤⇤
2.03
9⇤⇤⇤
2.23
1⇤⇤⇤
2.02
0⇤⇤⇤
2.21
3⇤⇤⇤
2.02
1⇤⇤⇤
2.20
4⇤⇤⇤
2.00
2⇤⇤⇤
2.18
5⇤⇤⇤
(0.6
10)
(0.6
17)
(0.6
54)
(0.6
55)
(0.6
23)
(0.6
29)
(0.6
52)
(0.6
53)
(0.6
51)
(0.6
50)
(0.6
49)
(0.6
51)
(0.6
49)
(0.6
49)
Mea
nab
ility
0.10
20.
105
0.11
80.
126
0.10
10.
105
0.12
60.
134
0.12
30.
125
0.11
00.
123
0.10
60.
114
(0.0
92)
(0.0
92)
(0.0
90)
(0.0
91)
(0.0
92)
(0.0
93)
(0.0
93)
(0.0
94)
(0.0
93)
(0.0
94)
(0.0
94)
(0.0
95)
(0.0
94)
(0.0
94)
Ent
ry0.
447⇤
⇤⇤0.
384⇤
⇤⇤0.
369⇤
⇤⇤0.
388⇤
⇤⇤0.
474⇤
⇤⇤0.
396⇤
⇤⇤0.
468⇤
⇤⇤
(0.1
33)
(0.1
37)
(0.1
34)
(0.1
39)
(0.1
47)
(0.1
47)
(0.1
56)
Mat
hgr
ade
0.16
00.
098
-0.0
97-0
.161
-0.0
86-0
.140
-0.1
15-0
.182
-0.1
03-0
.160
(0.1
86)
(0.1
90)
(0.1
95)
(0.2
00)
(0.1
96)
(0.1
99)
(0.1
98)
(0.2
00)
(0.1
98)
(0.2
00)
GPA
0.21
7⇤⇤
0.24
8⇤⇤
0.22
5⇤⇤
0.25
5⇤⇤⇤
0.22
3⇤⇤
0.25
8⇤⇤⇤
0.22
7⇤⇤
0.25
0⇤⇤⇤
0.22
6⇤⇤
0.25
3⇤⇤⇤
(0.0
97)
(0.0
97)
(0.0
95)
(0.0
95)
(0.0
95)
(0.0
95)
(0.0
95)
(0.0
95)
(0.0
95)
(0.0
95)
Mat
hre
lati
ve-0
.199
-0.2
20-0
.195
-0.2
19-0
.192
-0.2
13-0
.200
-0.2
20-0
.197
-0.2
14
(0.1
54)
(0.1
55)
(0.1
57)
(0.1
59)
(0.1
57)
(0.1
58)
(0.1
60)
(0.1
61)
(0.1
60)
(0.1
59)
Mat
hdi
fficu
lty
-0.3
58⇤⇤
⇤-0
.344
⇤⇤⇤
-0.2
37⇤⇤
⇤-0
.230
⇤⇤-0
.239
⇤⇤⇤
-0.2
36⇤⇤
-0.2
55⇤⇤
⇤-0
.255
⇤⇤⇤
-0.2
57⇤⇤
⇤-0
.258
⇤⇤⇤
(0.0
82)
(0.0
83)
(0.0
91)
(0.0
90)
(0.0
92)
(0.0
92)
(0.0
95)
(0.0
94)
(0.0
96)
(0.0
96)
Mat
hqu
arti
le-0
.329
⇤⇤⇤
-0.3
31⇤⇤
⇤-0
.303
⇤⇤⇤
-0.3
09⇤⇤
⇤-0
.307
⇤⇤⇤
-0.3
21⇤⇤
⇤-0
.312
⇤⇤⇤
-0.3
18⇤⇤
⇤-0
.315
⇤⇤⇤
-0.3
29⇤⇤
⇤
(0.0
79)
(0.0
79)
(0.0
77)
(0.0
77)
(0.0
77)
(0.0
78)
(0.0
78)
(0.0
78)
(0.0
78)
(0.0
79)
Gue
ssed
rank
0.04
20.
133
0.04
40.
121
(0.0
81)
(0.0
85)
(0.0
83)
(0.0
87)
Ris
k-0
.054
-0.1
06-0
.047
-0.0
95
(0.0
69)
(0.0
71)
(0.0
69)
(0.0
70)
Lot
tery
0.17
5⇤⇤
0.16
0⇤⇤
0.17
5⇤⇤
0.15
6⇤⇤
(0.0
73)
(0.0
74)
(0.0
74)
(0.0
74)
Cut
10.
534
0.67
73.
345⇤
⇤3.
405⇤
⇤-1
.031
-0.8
890.
721
0.76
40.
884
1.29
90.
691
0.46
60.
901
1.00
0
Cut
21.
697⇤
⇤1.
859⇤
⇤4.
632⇤
⇤⇤4.
709⇤
⇤⇤0.
314
0.47
12.
106
2.16
72.
269
2.70
5⇤2.
095
1.88
42.
305
2.42
0
Cut
32.
389⇤
⇤⇤2.
563⇤
⇤⇤5.
421⇤
⇤⇤5.
507⇤
⇤⇤1.
138
1.30
42.
959⇤
⇤3.
030⇤
⇤3.
122⇤
⇤3.
573⇤
⇤2.
954⇤
2.75
5⇤3.
164⇤
⇤3.
295⇤
⇤
Fem
ale/
(C3-
C1)
-0.2
32⇤⇤
⇤-0
.181
⇤⇤⇤
-0.2
63⇤⇤
⇤-0
.227
⇤⇤⇤
-0.1
37⇤⇤
-0.1
04⇤⇤
-0.1
86⇤⇤
⇤-0
.154
⇤⇤⇤
-0.1
91⇤⇤
⇤-0
.164
⇤⇤⇤
-0.1
60⇤⇤
⇤-0
.135
⇤⇤-0
.165
⇤⇤⇤
-0.1
44⇤⇤
⇤
Diff
.22
.0%
13.6
%24
.3%
17.3
%14
.5%
15.9
%12
.9%
Boo
tstr
app-
valu
e0.
001
0.00
30.
004
0.00
40.
005
0.00
80.
015
Obs
erva
tion
s36
236
236
236
236
236
236
236
236
236
236
236
236
236
2N
ote:
Coe
ffici
ents
are
from
orde
red
prob
itre
gres
sion
s,w
here
NT
>N
H>
ES>
CS.
Per
cent
fem
ale
isth
epe
rcen
tage
ofgi
rls
inth
est
uden
t’s
clas
san
dm
ean
abili
tyis
mea
nG
PAin
the
stud
ent’
scl
ass.
All
spec
ifica
tion
sin
clud
eco
ntro
lsfo
rpe
rfor
man
cein
Rou
nds
1an
d2
ofth
eex
peri
men
t,th
ech
ance
ofw
inni
ngth
eR
ound
2to
urna
men
t,sc
hool
fixed
effec
ts
and
test
vers
ion
fixed
effec
ts.
Rob
ust
stan
dard
erro
rsin
pare
nthe
ses;
p-va
lues
for
Fem
ale/
(C3-
C1)
and
Diff
.ar
ebo
otst
rapp
ed;
*p<
0.10
,**
p<
0.05
,**
*p<
0.01.T
he
impa
ctof
confi
denc
e(c
ompa
ring
colu
mns
(7)
and
(9))
and
risk
atti
tude
s(c
ompa
ring
colu
mns
(7)
and
(11)
)on
the
gend
erga
p(F
emal
e/(C
3-C
1))
and
the
asso
ciat
edp-
valu
esar
e
2.9%
(inc
reas
ing)
(p=
0.68
)an
d13
.8%
(p=
0.02
),re
spec
tive
ly.
Table
A.X
.Tr
ack
choi
ce:
orde
red
prob
itre
gres
sion
(NT
/NH
asN
Tan
dE
S/C
Sas
ES)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
Fem
ale
-0.2
68⇤⇤
-0.1
98-0
.359
⇤⇤⇤
-0.3
04⇤⇤
-0.1
21-0
.073
-0.2
40⇤
-0.1
90-0
.251
⇤-0
.206
-0.1
96-0
.152
-0.2
03-0
.166
(0.1
24)
(0.1
29)
(0.1
33)
(0.1
36)
(0.1
29)
(0.1
35)
(0.1
34)
(0.1
39)
(0.1
35)
(0.1
38)
(0.1
38)
(0.1
40)
(0.1
38)
(0.1
39)
Ent
ry0.
313⇤
⇤0.
272⇤
0.23
00.
260⇤
0.32
6⇤⇤
0.28
9⇤0.
343⇤
⇤
(0.1
40)
(0.1
40)
(0.1
41)
(0.1
44)
(0.1
56)
(0.1
55)
(0.1
67)
Mat
hgr
ade
0.27
00.
226
0.02
0-0
.021
0.02
9-0
.007
-0.0
01-0
.051
0.00
6-0
.038
(0.2
05)
(0.2
08)
(0.2
13)
(0.2
16)
(0.2
13)
(0.2
13)
(0.2
16)
(0.2
18)
(0.2
15)
(0.2
15)
GPA
0.30
4⇤⇤⇤
0.32
5⇤⇤⇤
0.31
9⇤⇤⇤
0.33
9⇤⇤⇤
0.31
7⇤⇤⇤
0.33
8⇤⇤⇤
0.31
7⇤⇤⇤
0.33
5⇤⇤⇤
0.31
6⇤⇤⇤
0.33
4⇤⇤⇤
(0.1
09)
(0.1
07)
(0.1
05)
(0.1
05)
(0.1
05)
(0.1
05)
(0.1
05)
(0.1
05)
(0.1
05)
(0.1
05)
Mat
hre
lati
ve-0
.023
-0.0
40-0
.002
-0.0
20-0
.001
-0.0
210.
007
-0.0
120.
008
-0.0
13
(0.1
65)
(0.1
67)
(0.1
67)
(0.1
68)
(0.1
66)
(0.1
66)
(0.1
70)
(0.1
71)
(0.1
69)
(0.1
69)
Mat
hdi
fficu
lty
-0.3
27⇤⇤
⇤-0
.318
⇤⇤⇤
-0.2
31⇤⇤
-0.2
25⇤⇤
-0.2
32⇤⇤
-0.2
27⇤⇤
-0.2
63⇤⇤
⇤-0
.262
⇤⇤⇤
-0.2
63⇤⇤
⇤-0
.261
⇤⇤⇤
(0.0
85)
(0.0
85)
(0.0
94)
(0.0
93)
(0.0
95)
(0.0
94)
(0.0
98)
(0.0
97)
(0.0
98)
(0.0
98)
Mat
hqu
arti
le-0
.373
⇤⇤⇤
-0.3
73⇤⇤
⇤-0
.348
⇤⇤⇤
-0.3
50⇤⇤
⇤-0
.351
⇤⇤⇤
-0.3
60⇤⇤
⇤-0
.358
⇤⇤⇤
-0.3
60⇤⇤
⇤-0
.360
⇤⇤⇤
-0.3
69⇤⇤
⇤
(0.0
84)
(0.0
85)
(0.0
83)
(0.0
84)
(0.0
83)
(0.0
84)
(0.0
84)
(0.0
84)
(0.0
83)
(0.0
85)
Gue
ssed
rank
0.03
70.
103
0.02
60.
087
(0.0
84)
(0.0
90)
(0.0
89)
(0.0
95)
Ris
k-0
.120
-0.1
58⇤⇤
-0.1
16-0
.152
⇤⇤
(0.0
74)
(0.0
77)
(0.0
75)
(0.0
77)
Lot
tery
0.19
9⇤⇤
0.18
8⇤⇤
0.19
9⇤⇤
0.18
5⇤⇤
(0.0
79)
(0.0
79)
(0.0
79)
(0.0
79)
Cut
1-1
.441
⇤⇤⇤
-1.4
18⇤⇤
⇤3.
407⇤
⇤3.
361⇤
⇤-2
.819
⇤⇤⇤
-2.7
91⇤⇤
⇤0.
936
0.88
71.
087
1.29
60.
795
0.51
60.
921
0.88
7
Cut
2-0
.208
-0.1
694.
743⇤
⇤⇤4.
710⇤
⇤⇤-1
.405
⇤⇤⇤
-1.3
66⇤⇤
⇤2.
389
2.35
32.
540
2.76
5⇤2.
264
1.99
72.
390
2.37
0
Cut
30.
189
0.23
05.
197⇤
⇤⇤5.
166⇤
⇤⇤-0
.929
⇤⇤⇤
-0.8
89⇤⇤
⇤2.
887⇤
2.85
2⇤3.
039⇤
3.26
7⇤⇤
2.77
1⇤2.
506
2.89
7⇤2.
880⇤
Fem
ale/
(C3-
C1)
-0.1
64⇤⇤
-0.1
20⇤
-0.2
01⇤⇤
⇤-0
.169
⇤⇤-0
.064
-0.0
39-0
.123
⇤⇤-0
.097
⇤-0
.129
⇤⇤-0
.105
⇤-0
.099
⇤-0
.076
-0.1
03⇤
-0.0
83
Diff
.26
.9%
16.0
%39
.6%
21.5
%18
.5%
23.2
%19
.0%
Boo
tstr
app-
valu
e0.
013
0.02
70.
055
0.03
90.
023
0.03
60.
031
Obs
erva
tion
s34
234
234
234
234
234
234
234
234
234
234
234
234
234
2N
ote:
Coe
ffici
ents
are
from
orde
red
prob
itre
gres
sion
s,w
here
NT
>N
H>
ES>
CS.
All
spec
ifica
tion
sin
clud
eco
ntro
lsfo
rpe
rfor
man
cein
Rou
nds
1an
d2
ofth
eex
peri
men
t,th
e
chan
ceof
win
ning
the
Rou
nd2
tour
nam
ent,
scho
olfix
edeff
ects
and
test
vers
ion
fixed
effec
ts.
Rob
ust
stan
dard
erro
rsin
pare
nthe
ses;
p-va
lues
for
Fem
ale/
(C3-
C1)
and
Diff
.ar
e
boot
stra
pped
;*p<
0.10,**
p<
0.05,**
*p<
0.01.T
heim
pact
ofco
nfide
nce
(com
pari
ngco
lum
ns(7
)an
d(9
))an
dri
skat
titu
des
(com
pari
ngco
lum
ns(7
)an
d(1
1))
onth
e
gend
erga
p(F
emal
e/(C
3-C
1))
and
the
asso
ciat
edp-
valu
esar
e4.
6%(i
ncre
asin
g)(p
=0.
66)
and
18.3
%(p
=0.
08),
resp
ecti
vely
.
Table
A.X
I.
Trac
kch
oice
:or
dere
dpr
obit
regr
essi
on(N
T/N
Han
dE
S/C
Sas
sepa
rate
trac
k)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
Fem
ale
-0.4
43⇤⇤
⇤-0
.373
⇤⇤⇤
-0.5
13⇤⇤
⇤-0
.459
⇤⇤⇤
-0.3
20⇤⇤
⇤-0
.268
⇤⇤-0
.405
⇤⇤⇤
-0.3
53⇤⇤
⇤-0
.421
⇤⇤⇤
-0.3
75⇤⇤
⇤-0
.361
⇤⇤⇤
-0.3
15⇤⇤
-0.3
73⇤⇤
⇤-0
.336
⇤⇤⇤
(0.1
18)
(0.1
21)
(0.1
23)
(0.1
26)
(0.1
19)
(0.1
25)
(0.1
24)
(0.1
28)
(0.1
25)
(0.1
28)
(0.1
26)
(0.1
28)
(0.1
26)
(0.1
28)
Ent
ry0.
325⇤
⇤0.
277⇤
⇤0.
250⇤
0.27
2⇤⇤
0.34
9⇤⇤
0.29
5⇤⇤
0.36
0⇤⇤
(0.1
31)
(0.1
31)
(0.1
31)
(0.1
34)
(0.1
43)
(0.1
41)
(0.1
51)
Mat
hgr
ade
0.13
50.
089
-0.0
97-0
.141
-0.0
85-0
.126
-0.1
12-0
.162
-0.1
01-0
.147
(0.1
82)
(0.1
85)
(0.1
89)
(0.1
91)
(0.1
89)
(0.1
90)
(0.1
90)
(0.1
92)
(0.1
90)
(0.1
90)
GPA
0.21
6⇤⇤
0.23
7⇤⇤
0.21
7⇤⇤
0.23
8⇤⇤
0.21
5⇤⇤
0.24
0⇤⇤
0.21
3⇤⇤
0.23
0⇤⇤
0.21
2⇤⇤
0.23
2⇤⇤
(0.1
00)
(0.0
99)
(0.0
98)
(0.0
97)
(0.0
98)
(0.0
97)
(0.0
98)
(0.0
97)
(0.0
98)
(0.0
98)
Mat
hre
lati
ve-0
.156
-0.1
73-0
.143
-0.1
61-0
.141
-0.1
62-0
.138
-0.1
54-0
.137
-0.1
55
(0.1
51)
(0.1
53)
(0.1
51)
(0.1
53)
(0.1
50)
(0.1
51)
(0.1
53)
(0.1
54)
(0.1
52)
(0.1
53)
Mat
hdi
fficu
lty
-0.2
91⇤⇤
⇤-0
.281
⇤⇤⇤
-0.2
06⇤⇤
-0.2
01⇤⇤
-0.2
08⇤⇤
-0.2
04⇤⇤
-0.2
31⇤⇤
⇤-0
.230
⇤⇤⇤
-0.2
31⇤⇤
⇤-0
.231
⇤⇤⇤
(0.0
77)
(0.0
77)
(0.0
85)
(0.0
85)
(0.0
86)
(0.0
86)
(0.0
89)
(0.0
88)
(0.0
90)
(0.0
89)
Mat
hqu
arti
le-0
.356
⇤⇤⇤
-0.3
57⇤⇤
⇤-0
.334
⇤⇤⇤
-0.3
37⇤⇤
⇤-0
.338
⇤⇤⇤
-0.3
48⇤⇤
⇤-0
.342
⇤⇤⇤
-0.3
46⇤⇤
⇤-0
.346
⇤⇤⇤
-0.3
55⇤⇤
⇤
(0.0
79)
(0.0
79)
(0.0
78)
(0.0
79)
(0.0
78)
(0.0
79)
(0.0
78)
(0.0
79)
(0.0
78)
(0.0
79)
Gue
ssed
rank
0.05
20.
121
0.04
50.
109
(0.0
78)
(0.0
83)
(0.0
82)
(0.0
87)
Ris
k-0
.090
-0.1
29⇤
-0.0
84-0
.123
⇤
(0.0
67)
(0.0
68)
(0.0
67)
(0.0
69)
Lot
tery
0.18
1⇤⇤
0.17
1⇤⇤
0.18
0⇤⇤
0.16
7⇤⇤
(0.0
73)
(0.0
73)
(0.0
73)
(0.0
74)
Cut
1-1
.292
⇤⇤⇤
-1.2
70⇤⇤
⇤1.
561
1.50
0-2
.583
⇤⇤⇤
-2.5
53⇤⇤
⇤-0
.912
-0.9
68-0
.689
-0.4
61-0
.995
-1.2
75-0
.768
-0.7
94
Cut
2-1
.016
⇤⇤⇤
-0.9
88⇤⇤
⇤1.
847
1.79
1-2
.284
⇤⇤⇤
-2.2
51⇤⇤
⇤-0
.610
-0.6
61-0
.388
-0.1
54-0
.688
-0.9
63-0
.461
-0.4
82
Cut
3-0
.065
-0.0
232.
888⇤
⇤2.
842⇤
⇤-1
.188
⇤⇤⇤
-1.1
44⇤⇤
⇤0.
515
0.47
50.
737
0.98
60.
457
0.19
30.
683
0.67
7
Cut
40.
324
0.36
73.
326⇤
⇤⇤3.
282⇤
⇤-0
.727
⇤⇤-0
.683
⇤⇤0.
992
0.95
41.
215
1.46
60.
941
0.68
01.
168
1.16
5
Cut
50.
898⇤
⇤⇤0.
945⇤
⇤⇤3.
953⇤
⇤⇤3.
910⇤
⇤⇤-0
.065
-0.0
201.
667
1.62
91.
891
2.14
41.
621
1.36
11.
849
1.84
9
Fem
ale/
(C5-
C1)
-0.2
02⇤⇤
⇤-0
.168
⇤⇤⇤
-0.2
14⇤⇤
⇤-0
.190
⇤⇤⇤
-0.1
27⇤⇤
⇤-0
.106
⇤⇤⇤
-0.1
57⇤⇤
⇤-0
.136
⇤⇤⇤
-0.1
63⇤⇤
⇤-0
.144
⇤⇤⇤
-0.1
38⇤⇤
⇤-0
.120
⇤⇤⇤
-0.1
43⇤⇤
⇤-0
.127
⇤⇤⇤
Diff
.16
.8%
11.3
%16
.7%
13.6
%11
.8%
13.2
%11
.0%
Boo
tstr
app-
valu
e0.
007
0.01
80.
030
0.02
20.
009
0.02
00.
017
Obs
erva
tion
s34
234
234
234
234
234
234
234
234
234
234
234
234
234
2N
ote:
Coe
ffici
ents
are
from
orde
red
prob
itre
gres
sion
s,w
here
NT
>N
T/N
H>
NH
>E
S>E
S/C
S>C
S.A
llsp
ecifi
cati
ons
incl
ude
cont
rols
for
perf
orm
ance
inR
ound
s1
and
2of
the
expe
rim
ent,
the
chan
ceof
win
ning
the
Rou
nd2
tour
nam
ent,
scho
olfix
edeff
ect
and
test
vers
ion
fixed
effec
ts.
Rob
ust
stan
dard
erro
rsin
pare
nthe
ses;
p-va
lues
for
Fem
ale/
(C3-
C1)
and
Diff
.ar
ebo
otst
rapp
ed;*p<
0.10,**
p<
0.05,**
*p<
0.01.T
heim
pact
ofco
nfide
nce
(com
pari
ngco
lum
ns(7
)an
d(9
))an
dri
skat
titu
des
(com
pari
ngco
lum
ns(7
)an
d
(11)
)on
the
gend
erga
p(F
emal
e/(C
3-C
1))
and
the
asso
ciat
edp-
valu
esar
e4.
0%(i
ncre
asin
g)(p
=0.
74)
and
10.9
%(p
=0.
04),
resp
ecti
vely
.
Table
A.X
II.
Trac
kch
oice
:or
dere
dpr
obit
regr
essi
on(s
tude
nts’
own
rank
ing)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
Fem
ale
-0.4
15⇤⇤
⇤-0
.344
⇤⇤⇤
-0.5
15⇤⇤
⇤-0
.457
⇤⇤⇤
-0.3
36⇤⇤
⇤-0
.279
⇤⇤-0
.434
⇤⇤⇤
-0.3
75⇤⇤
⇤-0
.437
⇤⇤⇤
-0.3
91⇤⇤
⇤-0
.406
⇤⇤⇤
-0.3
63⇤⇤
⇤-0
.415
⇤⇤⇤
-0.3
80⇤⇤
⇤
(0.1
20)
(0.1
27)
(0.1
28)
(0.1
34)
(0.1
24)
(0.1
29)
(0.1
31)
(0.1
36)
(0.1
34)
(0.1
37)
(0.1
34)
(0.1
37)
(0.1
36)
(0.1
38)
Ent
ry0.
316⇤
⇤0.
290⇤
⇤0.
260⇤
0.29
5⇤⇤
0.34
6⇤⇤
0.26
1⇤0.
310⇤
⇤
(0.1
39)
(0.1
41)
(0.1
39)
(0.1
42)
(0.1
49)
(0.1
50)
(0.1
58)
Mat
hgr
ade
-0.1
50-0
.195
-0.3
35⇤
-0.3
81⇤⇤
-0.3
33⇤
-0.3
70⇤⇤
-0.3
45⇤
-0.3
82⇤⇤
-0.3
38⇤
-0.3
71⇤⇤
(0.1
71)
(0.1
73)
(0.1
77)
(0.1
80)
(0.1
78)
(0.1
80)
(0.1
78)
(0.1
80)
(0.1
80)
(0.1
81)
GPA
0.22
6⇤⇤
0.24
9⇤⇤⇤
0.22
2⇤⇤
0.24
5⇤⇤⇤
0.22
2⇤⇤
0.24
7⇤⇤⇤
0.23
4⇤⇤
0.24
8⇤⇤⇤
0.23
3⇤⇤
0.25
0⇤⇤⇤
(0.0
94)
(0.0
93)
(0.0
92)
(0.0
92)
(0.0
92)
(0.0
92)
(0.0
93)
(0.0
92)
(0.0
93)
(0.0
93)
Mat
hre
lati
ve-0
.306
⇤⇤-0
.318
⇤⇤-0
.301
⇤⇤-0
.314
⇤⇤-0
.301
⇤⇤-0
.312
⇤⇤-0
.313
⇤⇤-0
.319
⇤⇤-0
.312
⇤⇤-0
.317
⇤⇤
(0.1
49)
(0.1
50)
(0.1
51)
(0.1
52)
(0.1
51)
(0.1
51)
(0.1
52)
(0.1
52)
(0.1
52)
(0.1
52)
Mat
hdi
fficu
lty
-0.2
61⇤⇤
⇤-0
.248
⇤⇤⇤
-0.2
04⇤⇤
-0.1
97⇤⇤
-0.2
04⇤⇤
-0.1
99⇤⇤
-0.1
99⇤⇤
-0.1
97⇤⇤
-0.1
99⇤⇤
-0.1
99⇤⇤
(0.0
81)
(0.0
81)
(0.0
94)
(0.0
92)
(0.0
95)
(0.0
93)
(0.0
93)
(0.0
92)
(0.0
94)
(0.0
93)
Mat
hqu
arti
le-0
.182
⇤⇤-0
.186
⇤⇤-0
.164
⇤⇤-0
.172
⇤⇤-0
.165
⇤⇤-0
.179
⇤⇤-0
.167
⇤⇤-0
.173
⇤⇤-0
.169
⇤⇤-0
.180
⇤⇤
(0.0
76)
(0.0
76)
(0.0
77)
(0.0
77)
(0.0
77)
(0.0
78)
(0.0
77)
(0.0
77)
(0.0
77)
(0.0
77)
Gue
ssed
rank
0.01
00.
078
0.02
90.
081
(0.0
83)
(0.0
89)
(0.0
84)
(0.0
89)
Ris
k0.
050
0.01
60.
055
0.02
3
(0.0
67)
(0.0
69)
(0.0
67)
(0.0
69)
Lot
tery
0.05
80.
047
0.05
80.
044
(0.0
66)
(0.0
67)
(0.0
66)
(0.0
67)
Cut
1-0
.817
⇤⇤⇤
-0.7
94⇤⇤
⇤0.
069
0.04
7-1
.642
⇤⇤⇤
-1.6
12⇤⇤
⇤-1
.753
-1.7
87-1
.708
-1.4
45-1
.422
-1.6
25-1
.271
-1.2
46
Cut
20.
146
0.17
61.
086
1.07
1-0
.613
⇤⇤-0
.577
⇤⇤-0
.702
-0.7
28-0
.657
-0.3
84-0
.367
-0.5
66-0
.216
-0.1
84
Cut
31.
002⇤
⇤⇤1.
040⇤
⇤⇤1.
999
1.99
00.
310
0.35
10.
241
0.22
20.
285
0.56
60.
578
0.38
50.
729
0.76
7
Fem
ale/
(C3-
C1)
-0.2
28⇤⇤
⇤-0
.188
⇤⇤⇤
-0.2
67⇤⇤
⇤-0
.235
⇤⇤⇤
-0.1
72⇤⇤
⇤-0
.142
⇤⇤-0
.218
⇤⇤⇤
-0.1
86⇤⇤
⇤-0
.219
⇤⇤⇤
-0.1
94⇤⇤
⇤-0
.203
⇤⇤⇤
-0.1
81⇤⇤
⇤-0
.208
⇤⇤⇤
-0.1
89⇤⇤
⇤
Diff
.17
.8%
11.9
%17
.4%
14.3
%11
.5%
11.0
%9.
0%
Boo
tstr
app-
valu
e0.
013
0.02
20.
033
0.02
00.
013
0.04
80.
034
Obs
erva
tion
s35
435
435
435
435
435
435
435
435
435
435
435
435
435
4N
ote:
Coe
ffici
ents
are
from
orde
red
prob
itre
gres
sion
s.A
llsp
ecifi
cation
sin
clud
eco
ntro
lsfo
rpe
rfor
man
cein
Rou
nds
1an
d2
ofth
eex
peri
men
t,th
ech
ance
ofw
inni
ngth
eR
ound
2to
urna
men
t,sc
hool
fixed
effec
tsan
dte
stve
rsio
nfix
edeff
ects
.R
obus
tst
anda
rder
rors
inpa
rent
hese
s;p-
valu
esfo
rFe
mal
e/(C
3-C
1)an
dD
iff.
are
boot
stra
pped
;*p<
0.10,**
p<
0.05,**
*p<
0.01.T
heim
pact
ofco
nfide
nce
(com
pari
ngco
lum
ns(7
)an
d(9
))an
dri
skat
titu
des
(com
pari
ngco
lum
ns(7
)an
d(1
1))
onth
ege
nder
gap
(Fem
ale/
(C3-
C1)
)
and
the
asso
ciat
edp-
valu
esar
e0.
0%(p
=0.
55)
and
6.4%
(p=
0.12
),re
spec
tive
ly.
Table
A.X
III.
Bin
ary
regr
essi
on:
NT
vsre
st
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
Fem
ale
-0.2
25⇤⇤
⇤-0
.197
⇤⇤⇤
-0.2
38⇤⇤
⇤-0
.221
⇤⇤⇤
-0.1
72⇤⇤
⇤-0
.153
⇤⇤⇤
-0.1
95⇤⇤
⇤-0
.178
⇤⇤⇤
-0.2
02⇤⇤
⇤-0
.187
⇤⇤⇤
-0.1
89⇤⇤
⇤-0
.174
⇤⇤⇤
-0.1
96⇤⇤
⇤-0
.183
⇤⇤⇤
(0.0
46)
(0.0
48)
(0.0
45)
(0.0
46)
(0.0
43)
(0.0
44)
(0.0
43)
(0.0
44)
(0.0
43)
(0.0
44)
(0.0
45)
(0.0
45)
(0.0
45)
(0.0
45)
Ent
ry0.
122⇤
⇤0.
080
0.08
5⇤0.
078
0.10
7⇤⇤
0.08
70.
112⇤
⇤
(0.0
55)
(0.0
51)
(0.0
50)
(0.0
50)
(0.0
54)
(0.0
53)
(0.0
56)
Mat
hgr
ade
0.07
80.
065
0.00
7-0
.006
0.01
30.
001
0.00
5-0
.008
0.01
1-0
.002
(0.0
64)
(0.0
65)
(0.0
65)
(0.0
65)
(0.0
65)
(0.0
65)
(0.0
65)
(0.0
66)
(0.0
66)
(0.0
66)
GPA
0.02
80.
035
0.02
40.
031
0.02
30.
031
0.02
40.
029
0.02
30.
030
(0.0
32)
(0.0
32)
(0.0
31)
(0.0
32)
(0.0
31)
(0.0
32)
(0.0
31)
(0.0
32)
(0.0
32)
(0.0
32)
Mat
hre
lati
ve-0
.064
-0.0
68-0
.057
-0.0
60-0
.055
-0.0
59-0
.057
-0.0
59-0
.056
-0.0
58
(0.0
55)
(0.0
55)
(0.0
54)
(0.0
54)
(0.0
54)
(0.0
54)
(0.0
54)
(0.0
54)
(0.0
54)
(0.0
54)
Mat
hdi
fficu
lty
-0.1
09⇤⇤
⇤-0
.105
⇤⇤⇤
-0.0
68⇤⇤
-0.0
67⇤⇤
-0.0
69⇤⇤
-0.0
68⇤⇤
-0.0
71⇤⇤
-0.0
70⇤⇤
-0.0
71⇤⇤
-0.0
71⇤⇤
(0.0
30)
(0.0
30)
(0.0
32)
(0.0
32)
(0.0
32)
(0.0
32)
(0.0
33)
(0.0
32)
(0.0
33)
(0.0
33)
Mat
hqu
arti
le-0
.098
⇤⇤⇤
-0.0
98⇤⇤
⇤-0
.086
⇤⇤⇤
-0.0
87⇤⇤
⇤-0
.088
⇤⇤⇤
-0.0
91⇤⇤
⇤-0
.087
⇤⇤⇤
-0.0
88⇤⇤
⇤-0
.089
⇤⇤⇤
-0.0
91⇤⇤
⇤
(0.0
27)
(0.0
27)
(0.0
26)
(0.0
26)
(0.0
26)
(0.0
26)
(0.0
26)
(0.0
26)
(0.0
26)
(0.0
26)
Gue
ssed
rank
0.02
30.
044
0.02
30.
042
(0.0
26)
(0.0
28)
(0.0
27)
(0.0
29)
Ris
k-0
.009
-0.0
21-0
.005
-0.0
17
(0.0
24)
(0.0
24)
(0.0
25)
(0.0
25)
Lot
tery
0.02
00.
017
0.02
00.
015
(0.0
24)
(0.0
24)
(0.0
24)
(0.0
24)
Boo
tstr
app-
valu
e0.
014
0.06
10.
048
0.06
00.
025
0.05
20.
027
Diff
.12
.6%
7.1%
10.9
%8.
4%7.
6%7.
7%6.
7%
Obs
erva
tion
s36
236
236
236
236
236
236
236
236
236
236
236
236
236
2N
ote:
Coe
ffici
ents
are
from
OLS
regr
essi
ons.
All
spec
ifica
tion
sin
clud
eco
ntro
lsfo
rpe
rfor
man
cein
Rou
nds
1an
d2
ofth
eex
peri
men
t,th
ech
ance
ofw
inni
ngth
eR
ound
2
tour
nam
ent,
scho
olfix
edeff
ects
and
test
vers
ion
fixed
effec
ts.
Rob
ust
stan
dard
erro
rsin
pare
nthe
ses;
p-va
lues
for
Diff
.ar
ebo
otst
rapp
ed;*
p<
0.10,**
p<
0.05,**
*p<
0.01.
Table
A.X
IV
.B
inar
yre
gres
sion
:N
atur
evs
Soci
ety
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
Fem
ale
0.01
40.
043
-0.0
29-0
.010
0.08
5⇤0.
101⇤
⇤0.
031
0.04
90.
024
0.04
00.
044
0.06
10.
039
0.05
4
(0.0
53)
(0.0
55)
(0.0
50)
(0.0
51)
(0.0
48)
(0.0
50)
(0.0
48)
(0.0
50)
(0.0
49)
(0.0
50)
(0.0
49)
(0.0
50)
(0.0
50)
(0.0
50)
Ent
ry0.
123⇤
⇤0.
088
0.07
60.
086
0.11
5⇤⇤
0.10
5⇤0.
129⇤
⇤
(0.0
61)
(0.0
56)
(0.0
54)
(0.0
54)
(0.0
56)
(0.0
57)
(0.0
60)
Mat
hgr
ade
0.07
00.
055
-0.0
27-0
.041
-0.0
21-0
.034
-0.0
31-0
.048
-0.0
27-0
.042
(0.0
67)
(0.0
68)
(0.0
66)
(0.0
67)
(0.0
66)
(0.0
67)
(0.0
65)
(0.0
66)
(0.0
66)
(0.0
66)
GPA
0.12
6⇤⇤⇤
0.13
3⇤⇤⇤
0.12
1⇤⇤⇤
0.12
8⇤⇤⇤
0.12
0⇤⇤⇤
0.12
8⇤⇤⇤
0.11
7⇤⇤⇤
0.12
3⇤⇤⇤
0.11
6⇤⇤⇤
0.12
4⇤⇤⇤
(0.0
35)
(0.0
35)
(0.0
32)
(0.0
32)
(0.0
32)
(0.0
32)
(0.0
32)
(0.0
32)
(0.0
32)
(0.0
32)
Mat
hre
lati
ve-0
.042
-0.0
46-0
.033
-0.0
37-0
.031
-0.0
35-0
.030
-0.0
33-0
.030
-0.0
32
(0.0
59)
(0.0
59)
(0.0
55)
(0.0
56)
(0.0
55)
(0.0
55)
(0.0
56)
(0.0
56)
(0.0
56)
(0.0
55)
Mat
hdi
fficu
lty
-0.1
28⇤⇤
⇤-0
.124
⇤⇤⇤
-0.0
82⇤⇤
-0.0
81⇤⇤
-0.0
83⇤⇤
-0.0
82⇤⇤
-0.0
91⇤⇤
⇤-0
.091
⇤⇤⇤
-0.0
91⇤⇤
⇤-0
.091
⇤⇤⇤
(0.0
30)
(0.0
30)
(0.0
33)
(0.0
33)
(0.0
33)
(0.0
33)
(0.0
34)
(0.0
34)
(0.0
34)
(0.0
34)
Mat
hqu
arti
le-0
.144
⇤⇤⇤
-0.1
44⇤⇤
⇤-0
.128
⇤⇤⇤
-0.1
29⇤⇤
⇤-0
.130
⇤⇤⇤
-0.1
33⇤⇤
⇤-0
.129
⇤⇤⇤
-0.1
30⇤⇤
⇤-0
.130
⇤⇤⇤
-0.1
33⇤⇤
⇤
(0.0
30)
(0.0
30)
(0.0
29)
(0.0
29)
(0.0
29)
(0.0
29)
(0.0
29)
(0.0
29)
(0.0
29)
(0.0
29)
Gue
ssed
rank
0.02
20.
045
0.01
70.
038
(0.0
29)
(0.0
30)
(0.0
30)
(0.0
31)
Ris
k-0
.043
⇤-0
.057
⇤⇤-0
.040
-0.0
54⇤⇤
(0.0
24)
(0.0
25)
(0.0
25)
(0.0
25)
Lot
tery
0.05
7⇤⇤
0.05
3⇤⇤
0.05
6⇤⇤
0.05
1⇤
(0.0
27)
(0.0
27)
(0.0
27)
(0.0
27)
Boo
tstr
app-
valu
e0.
022
0.05
60.
080
0.05
20.
022
0.03
50.
022
Diff
.20
4.3%
64.3
%19
.9%
57.9
%69
.9%
40.2
%38
.8%
Obs
erva
tion
s36
236
236
236
236
236
236
236
236
236
236
236
236
236
2N
ote:
Coe
ffici
ents
are
from
OLS
regr
essi
ons.
All
spec
ifica
tion
sin
clud
eco
ntro
lsfo
rpe
rfor
man
cein
Rou
nds
1an
d2
ofth
eex
peri
men
t,th
ech
ance
ofw
inni
ngth
eR
ound
2
tour
nam
ent,
scho
olfix
edeff
ects
and
test
vers
ion
fixed
effec
ts.
Rob
ust
stan
dard
erro
rsin
pare
nthe
ses;
p-va
lues
for
Diff
.ar
ebo
otst
rapp
ed;*
p<
0.10,**
p<
0.05,**
*p<
0.01.
Table
A.X
V.
Bin
ary
regr
essi
on:
CS
vsre
st
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
Fem
ale
0.07
1⇤⇤
0.05
20.
088⇤
⇤0.
072⇤
0.04
80.
033
0.06
6⇤0.
051
0.06
6⇤0.
054
0.04
70.
037
0.04
80.
040
(0.0
36)
(0.0
38)
(0.0
38)
(0.0
39)
(0.0
35)
(0.0
37)
(0.0
38)
(0.0
39)
(0.0
38)
(0.0
39)
(0.0
38)
(0.0
38)
(0.0
38)
(0.0
38)
Ent
ry-0
.080
⇤⇤-0
.077
⇤⇤-0
.066
⇤-0
.076
⇤⇤-0
.087
⇤⇤-0
.058
-0.0
68⇤
(0.0
35)
(0.0
36)
(0.0
35)
(0.0
35)
(0.0
39)
(0.0
37)
(0.0
41)
Mat
hgr
ade
0.03
00.
043
0.06
50.
078
0.06
50.
075
0.07
00.
079
0.06
90.
076
(0.0
48)
(0.0
49)
(0.0
48)
(0.0
49)
(0.0
49)
(0.0
49)
(0.0
49)
(0.0
49)
(0.0
49)
(0.0
49)
GPA
-0.0
51⇤
-0.0
58⇤⇤
-0.0
49⇤
-0.0
56⇤⇤
-0.0
49⇤
-0.0
56⇤⇤
-0.0
52⇤
-0.0
55⇤⇤
-0.0
52⇤
-0.0
56⇤⇤
(0.0
28)
(0.0
28)
(0.0
27)
(0.0
27)
(0.0
27)
(0.0
27)
(0.0
27)
(0.0
27)
(0.0
27)
(0.0
27)
Mat
hre
lati
ve0.
049
0.05
30.
046
0.05
00.
046
0.04
90.
049
0.05
10.
049
0.05
0
(0.0
43)
(0.0
43)
(0.0
42)
(0.0
42)
(0.0
42)
(0.0
42)
(0.0
43)
(0.0
43)
(0.0
43)
(0.0
42)
Mat
hdi
fficu
lty
0.03
8⇤⇤
0.03
5⇤0.
030
0.02
90.
030
0.02
90.
032
0.03
20.
033
0.03
3
(0.0
19)
(0.0
19)
(0.0
23)
(0.0
23)
(0.0
23)
(0.0
23)
(0.0
24)
(0.0
24)
(0.0
24)
(0.0
24)
Mat
hqu
arti
le0.
050⇤
⇤0.
050⇤
⇤0.
047⇤
⇤0.
048⇤
⇤0.
047⇤
⇤0.
050⇤
⇤0.
048⇤
⇤0.
049⇤
⇤0.
049⇤
⇤0.
050⇤
⇤
(0.0
22)
(0.0
22)
(0.0
22)
(0.0
22)
(0.0
22)
(0.0
22)
(0.0
22)
(0.0
22)
(0.0
22)
(0.0
22)
Gue
ssed
rank
0.00
0-0
.017
-0.0
05-0
.016
(0.0
25)
(0.0
26)
(0.0
25)
(0.0
26)
Ris
k-0
.003
0.00
5-0
.004
0.00
3
(0.0
20)
(0.0
20)
(0.0
19)
(0.0
20)
Lot
tery
-0.0
50⇤⇤
-0.0
48⇤⇤
-0.0
50⇤⇤
-0.0
48⇤⇤
(0.0
20)
(0.0
20)
(0.0
20)
(0.0
20)
Boo
tstr
app-
valu
e0.
012
0.01
80.
028
0.01
60.
016
0.06
00.
058
Diff
.26
.3%
18.2
%30
.4%
23.8
%18
.9%
20.8
%16
.6%
Obs
erva
tion
s36
236
236
236
236
236
236
236
236
236
236
236
236
236
2N
ote:
Coe
ffici
ents
are
from
OLS
regr
essi
ons.
All
spec
ifica
tion
sin
clud
eco
ntro
lsfo
rpe
rfor
man
cein
Rou
nds
1an
d2
ofth
eex
peri
men
t,th
ech
ance
ofw
inni
ngth
eR
ound
2
tour
nam
ent,
scho
olfix
edeff
ects
and
test
vers
ion
fixed
effec
ts.
Rob
ust
stan
dard
erro
rsin
pare
nthe
ses;
p-va
lues
for
Diff
.ar
ebo
otst
rapp
ed;*
p<
0.10,**
p<
0.05,**
*p<
0.01.
Table
A.X
VI.
Bin
ary
regr
essi
on:
Self-
rank
edbe
stvs
rest
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
Fem
ale
-0.2
05⇤⇤
⇤-0
.174
⇤⇤⇤
-0.2
21⇤⇤
⇤-0
.196
⇤⇤⇤
-0.1
71⇤⇤
⇤-0
.147
⇤⇤⇤
-0.1
91⇤⇤
⇤-0
.168
⇤⇤⇤
-0.1
88⇤⇤
⇤-0
.170
⇤⇤⇤
-0.1
78⇤⇤
⇤-0
.161
⇤⇤⇤
-0.1
76⇤⇤
⇤-0
.163
⇤⇤⇤
(0.0
48)
(0.0
50)
(0.0
48)
(0.0
50)
(0.0
47)
(0.0
49)
(0.0
48)
(0.0
50)
(0.0
49)
(0.0
50)
(0.0
49)
(0.0
50)
(0.0
50)
(0.0
51)
Ent
ry0.
134⇤
⇤0.
116⇤
⇤0.
106⇤
0.11
3⇤⇤
0.12
2⇤⇤
0.10
3⇤0.
111⇤
(0.0
57)
(0.0
56)
(0.0
55)
(0.0
56)
(0.0
60)
(0.0
59)
(0.0
62)
Mat
hgr
ade
-0.0
34-0
.054
-0.0
88-0
.107
-0.0
91-0
.105
-0.0
91-0
.107
-0.0
92-0
.105
(0.0
71)
(0.0
70)
(0.0
72)
(0.0
72)
(0.0
73)
(0.0
73)
(0.0
72)
(0.0
72)
(0.0
73)
(0.0
73)
GPA
0.04
30.
052
0.04
10.
050
0.04
10.
050
0.04
40.
050
0.04
40.
050
(0.0
36)
(0.0
36)
(0.0
36)
(0.0
36)
(0.0
36)
(0.0
36)
(0.0
36)
(0.0
36)
(0.0
36)
(0.0
36)
Mat
hre
lati
ve-0
.108
⇤-0
.113
⇤-0
.100
⇤-0
.105
⇤-0
.100
⇤-0
.104
⇤-0
.103
⇤-0
.105
⇤-0
.103
⇤-0
.105
⇤
(0.0
60)
(0.0
60)
(0.0
60)
(0.0
60)
(0.0
60)
(0.0
60)
(0.0
60)
(0.0
60)
(0.0
61)
(0.0
60)
Mat
hdi
fficu
lty
-0.1
01⇤⇤
⇤-0
.096
⇤⇤⇤
-0.0
80⇤⇤
-0.0
78⇤⇤
-0.0
79⇤⇤
-0.0
78⇤⇤
-0.0
79⇤⇤
-0.0
79⇤⇤
-0.0
79⇤⇤
-0.0
79⇤⇤
(0.0
31)
(0.0
30)
(0.0
34)
(0.0
33)
(0.0
34)
(0.0
33)
(0.0
34)
(0.0
33)
(0.0
34)
(0.0
33)
Mat
hqu
arti
le-0
.037
-0.0
38-0
.032
-0.0
33-0
.031
-0.0
35-0
.033
-0.0
34-0
.032
-0.0
35
(0.0
28)
(0.0
27)
(0.0
28)
(0.0
28)
(0.0
29)
(0.0
29)
(0.0
29)
(0.0
28)
(0.0
29)
(0.0
29)
Gue
ssed
rank
-0.0
100.
014
-0.0
050.
014
(0.0
30)
(0.0
32)
(0.0
31)
(0.0
32)
Ris
k0.
013
-0.0
010.
012
0.00
0
(0.0
26)
(0.0
27)
(0.0
27)
(0.0
27)
Lot
tery
0.02
80.
025
0.02
80.
024
(0.0
25)
(0.0
25)
(0.0
25)
(0.0
25)
Boo
tstr
app-
valu
e0.
011
0.01
90.
031
0.02
30.
020
0.04
50.
044
Diff
.15
.2%
11.1
%13
.8%
12.3
%9.
3%9.
7%7.
4%
Obs
erva
tion
s36
236
236
236
236
236
236
236
236
236
236
236
236
236
2N
ote:
Coe
ffici
ents
are
from
OLS
regr
essi
ons.
All
spec
ifica
tion
sin
clud
eco
ntro
lsfo
rpe
rfor
man
cein
Rou
nds
1an
d2
ofth
eex
peri
men
t,th
ech
ance
ofw
inni
ngth
eR
ound
2
tour
nam
ent,
scho
olfix
edeff
ects
and
test
vers
ion
fixed
effec
ts.
Rob
ust
stan
dard
erro
rsin
pare
nthe
ses;
p-va
lues
for
Diff
.ar
ebo
otst
rapp
ed;*
p<
0.10,**
p<
0.05,**
*p<
0.01.
Table
A.X
VII.
Trac
kch
oice
:O
LSre
gres
sion
s
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
Fem
ale
-0.2
83⇤⇤
⇤-0
.207
⇤-0
.355
⇤⇤⇤
-0.3
03⇤⇤
⇤-0
.136
-0.0
85-0
.230
⇤⇤-0
.180
⇤-0
.245
⇤⇤-0
.200
⇤⇤-0
.192
⇤-0
.150
-0.2
05⇤⇤
-0.1
69⇤
(0.1
08)
(0.1
12)
(0.1
04)
(0.1
06)
(0.0
96)
(0.1
01)
(0.0
99)
(0.1
02)
(0.0
99)
(0.1
02)
(0.1
00)
(0.1
02)
(0.1
01)
(0.1
01)
Ent
ry0.
325⇤
⇤⇤0.
245⇤
⇤0.
226⇤
⇤0.
240⇤
⇤0.
310⇤
⇤⇤0.
250⇤
⇤0.
309⇤
⇤⇤
(0.1
19)
(0.1
10)
(0.1
04)
(0.1
04)
(0.1
11)
(0.1
10)
(0.1
17)
Mat
hgr
ade
0.11
80.
077
-0.0
85-0
.125
-0.0
74-0
.109
-0.0
96-0
.136
-0.0
85-0
.121
(0.1
39)
(0.1
41)
(0.1
36)
(0.1
38)
(0.1
37)
(0.1
37)
(0.1
37)
(0.1
38)
(0.1
37)
(0.1
37)
GPA
0.20
6⇤⇤⇤
0.22
5⇤⇤⇤
0.19
5⇤⇤⇤
0.21
4⇤⇤⇤
0.19
3⇤⇤⇤
0.21
5⇤⇤⇤
0.19
2⇤⇤⇤
0.20
7⇤⇤⇤
0.19
1⇤⇤⇤
0.20
9⇤⇤⇤
(0.0
75)
(0.0
75)
(0.0
70)
(0.0
70)
(0.0
70)
(0.0
70)
(0.0
70)
(0.0
69)
(0.0
70)
(0.0
70)
Mat
hre
lati
ve-0
.156
-0.1
67-0
.136
-0.1
47-0
.133
-0.1
43-0
.136
-0.1
43-0
.134
-0.1
40(0
.121
)(0
.121
)(0
.114
)(0
.114
)(0
.114
)(0
.113
)(0
.116
)(0
.115
)(0
.115
)(0
.114
)M
ath
diffi
culty
-0.2
75⇤⇤
⇤-0
.264
⇤⇤⇤
-0.1
81⇤⇤
-0.1
76⇤⇤
-0.1
83⇤⇤
-0.1
79⇤⇤
-0.1
94⇤⇤
⇤-0
.193
⇤⇤⇤
-0.1
95⇤⇤
⇤-0
.195
⇤⇤⇤
(0.0
64)
(0.0
64)
(0.0
70)
(0.0
69)
(0.0
71)
(0.0
70)
(0.0
72)
(0.0
71)
(0.0
73)
(0.0
72)
Mat
hqu
arti
le-0
.292
⇤⇤⇤
-0.2
93⇤⇤
⇤-0
.262
⇤⇤⇤
-0.2
65⇤⇤
⇤-0
.266
⇤⇤⇤
-0.2
74⇤⇤
⇤-0
.264
⇤⇤⇤
-0.2
67⇤⇤
⇤-0
.267
⇤⇤⇤
-0.2
75⇤⇤
⇤
(0.0
61)
(0.0
61)
(0.0
59)
(0.0
60)
(0.0
60)
(0.0
60)
(0.0
60)
(0.0
60)
(0.0
60)
(0.0
60)
Gue
ssed
rank
0.04
50.
106⇤
0.04
50.
097
(0.0
59)
(0.0
62)
(0.0
61)
(0.0
64)
Ris
k-0
.049
-0.0
83-0
.041
-0.0
75(0
.051
)(0
.052
)(0
.051
)(0
.052
)Lo
tter
y0.
127⇤
⇤0.
118⇤
⇤0.
127⇤
⇤0.
114⇤
⇤
(0.0
54)
(0.0
54)
(0.0
55)
(0.0
54)
Diff
.26
.9%
14.5
%37
.1%
21.7
%18
.2%
21.8
%17
.7%
Boo
tstr
app-
valu
e0.
003
0.01
30.
017
0.01
10.
005
0.01
40.
012
Obs
erva
tion
s36
236
236
236
236
236
236
236
236
236
236
236
236
236
2N
ote:
Coe
ffici
ents
are
from
OLS
regr
essi
ons,
whe
reN
T(=
4)>
NH
(=3)
>E
S(=
2)>
CS(
=1)
.A
llsp
ecifi
cati
onsin
clud
eco
ntro
lsfo
rpe
rfor
man
cein
Rou
nds1
and
2of
the
expe
rim
ent,
the
chan
ceof
win
ning
the
Rou
nd2
tour
nam
ent,
scho
olfix
edeff
ects
and
test
vers
ion
fixed
effec
ts.
Rob
ust
stan
dard
erro
rsin
pare
nthe
ses;
p-va
lues
for
Diff
.ar
ebo
otst
rapp
ed;*
p<
0.10
,**p<
0.05
,***
p<
0.01.
Appendix II: Figure
Figure A.I: Tournament entry by gender and subsequent track choice (conditional on performance)
Note: In the Figure, competitiveness is measured as the residual from a regression of tournament entry on the measuresof performance in the experiment, school and test version fixed effects.
Appendix III: Gender Effects by Subsample
To investigate the impact of competitiveness on gender differences in track choice in an intuitiveway, we compare the impact of gender on educational choices for different subpopulations split bygender and tournament entry. If the gender gap in track choice is unrelated to competitiveness, theimpact of gender on track choice should, for example, be the same for the subsample made up ofcompetitive boys (Comp Boys) and non-competitive girls (N-comp Girls) as it is for the subsamplemade up of non-competitive boys and competitive girls.
This idea is explored in Table A.XVIII which reports coefficients of regressions of track choice on afemale dummy (and controls for performance in the experiment, school fixed effects, test version fixedeffects, objective and subjective ability, risk attitudes, and guessed rank) for the various subsamples.The top part reports ordered probit estimations that rank tracks by prestige: NT>NH>ES>CS.The table shows that the gender gap in track choice, which is significant for the whole sample, variesstrongly with competitiveness as measured by tournament entry. The gender gap in track choiceincreases with the competitiveness of boys and decreases with the competitiveness of girls. Whenwe consider competitive boys and non-competitive girls, gender bridges about 34 percent of the gapbetween choosing the most and the least prestigious track (see column (2) of the upper half of TableA.XVIII). When, on the other hand, we consider non-competitive boys and competitive girls, thereis no gender difference in track choices. Furthermore, the change in the gender dummy betweenthe group of competitive boys and non-competitive girls and the other way round is significant atp=0.02.
Table A.XVIII. Gender effects by subsample
(1) (2) NOrdered probit Female/(C3-C1)
(1) Comp B & n-comp. G -0.84⇤⇤⇤ -0.34⇤⇤⇤ 230(0.23)
(2) Comp. B & comp. G -0.52⇤⇤ -0.20⇤⇤ 129(0.23)
(3) N-comp. B & n-comp. G -0.16 -0.07 233(0.16)
(4) N-comp. B & comp. G -0.00 -0.00 132(0.20)
(5) Whole sample -0.30⇤⇤⇤ -0.14⇤⇤⇤ 262(0.13)
P-value (1) vs (4) 0.02 0.02
Binary OLS (1) (2) (3) (4)NT vs Rest N vs S Rest vs CS Best vs Rest
(1) Comp B & n-comp. G -0.33⇤⇤⇤ -0.08 -0.15⇤⇤⇤ -0.24⇤⇤⇤
(0.08) (0.09) (0.06) (0.08)(2) Comp. B & comp. G -0.25⇤⇤⇤ 0.03 -0.08 -0.17⇤
(0.08) (0.09) (0.05) (0.10)(3) N-comp. B & n-comp. G -0.14⇤⇤⇤ 0.06 -0.02 -0.15⇤⇤
(0.06) (0.06) (0.05) (0.06)(4) N-comp. B & comp. G -0.12 0.14 -0.01 -0.06
(0.09) (0.09) (0.07) (0.10)(5) Whole sample -0.20⇤⇤⇤ 0.04 0.05 -0.18⇤⇤⇤
(0.04) (0.05) (0.04) (0.05)P-value (1) vs (4) 0.05 0.05 0.06 0.09Note: Coefficients are from regressions of track choice on a female dummy and controls for performance in Rounds 1and 2 of the experiment, the chance of winning the Round 2 tournament, school fixed effects, test version fixed effects,objective and subjective ability, risk attitudes, and guessed rank; robust standard errors in parentheses; * p < 0.10
, ** p < 0.05 , *** p < 0.01 ; p-values are bootstrapped. For the ordered probit regression, F/(C3-C1) representsthe relative size of the female coefficient to the distance between the cuts provided by the ordered probit regression(between on the one hand choosing the least and the second to least prestigious track and on the other hand betweenchoosing the second and the most prestigious track). For the OLS regressions, the value is 1 for the left variable (e.g.NT in column (1)) and 0 otherwise.
The probit models in the lower part of Table A.XVIII give a more detailed view on this re-sult. We first assess the probability of choosing the most prestigious NT track compared to anyother study track. When we consider only competitive boys and non-competitive girls, girls are 33percentage points less likely to choose NT. When instead we consider competitive girls and non-competitive boys, girls are only 12 percentage points less likely to choose the NT track, a differencethat is not significant. Furthermore, the gender difference in choices is significantly smaller when weconsider competitive girls and non-competitive boys than when we consider competitive boys andnon-competitive girls. The results are qualitatively similar when we either consider choices between
the Nature and the Society tracks (the top two versus the bottom two in terms of prestige), orwhen we consider the option to choose CS, the least prestigious track, compared to any other track.Finally, we consider the student specific ordering of study tracks. Specifically, for each student weask whether they pick the track that they deem to be the one chosen by the best students, or anothertrack. In all cases the results are very similar. Gender differences are reduced when we reduce thecompetitiveness of boys and increase the competitiveness of girls.
Appendix IV: Instructions, Questionnaire and Example Sheet
WELCOME
In the experiment today you will be asked to complete three different tasks. None of these will take
more than 3 minutes. At the end of the experiment, we will randomly select one of the tasks and pay
you based on your performance in that task. Once you have completed the three tasks we determine
which task counts for payment by rolling a die. The method we use to determine your earnings
varies across tasks. Before each task we will describe in detail how your payment is determined.
Task 1 – Piece Rate
For Task 1 you will be asked to calculate the sum of four randomly chosen two-digit numbers. You
will be given 3 minutes to calculate the correct sum of a series of these problems. You cannot use a
calculator to determine these sums, however you are welcome to make use of the provided scratch
paper.
Example:
23 81 15 47
If Task 1 is the one randomly selected for payment, then you get 25 cents per problem you solve
correctly in the 3 minutes. Your payment does not decrease if you provide an incorrect answer to a
problem. We refer to this payment as the piece rate payment.
The problem sheets are in the envelopes in front of you. We will tell you when you can open the
envelopes and start working. You will then have exactly 3 minutes. At the end of the three minutes
we will say “Time’s up”. You then have to immediately stop writing and stand up. If you don’t stand
up or keep writing, you will not get paid.
Please do not talk with one another for the duration of the experiment. If you have any questions,
please raise your hand.
ARE THERE ANY QUESTIONS BEFORE WE BEGIN?
Task 2 - Tournament
As in Task 1 you will be given 3 minutes to calculate the correct sum of a series of four 2-digit
numbers. However for this task your payment depends on your performance relative to that of a
group of other participants. Each group consists of four people, the three other members of your
group are randomly selected members of your class. You will not know who is in your group.
If Task 2 is the one randomly selected for payment, the individual in the group who correctly solves
the largest number of problems will receive €1 per correct problem. The other participants receive
no payment. We refer to this as the tournament payment. You will not be informed of how you did
in the tournament until later. If there are ties the winner will be randomly determined.
Please do not talk with one another. If you have any questions, please raise your hand.
ARE THERE ANY QUESTIONS BEFORE WE BEGIN?
Task 3 – Choice
As in the previous two tasks you will be given 3 minutes to calculate the correct sum of a series of
four 2-digit numbers. However you will now get to choose how you want to be payed: piece rate or
tournament.
If Task 3 is the one randomly selected for payment, then your earnings for this task are determined
as follows. If you choose the piece rate you receive 25 cents per problem you solve correctly. If you
choose the tournament your performance will be compared to the performance of the other three
participants of your group in Task 2. Task 2 is the one you just completed. If you correctly solve
more problems than they did in Task 2, then you receive four times the payment from the piece rate,
which is €1 per correct problem. You will receive no earnings for this task if you choose the
tournament and do not solve more problems correctly now, than the others in your group did in Task
2.
You will not be informed of how you did in the tournament until later. If there are ties the winner
will be randomly determined.
Please do not talk with one another. If you have any questions, please raise your hand.
Please indicate below which payment scheme you choose: piece rate or tournament.
ARE THERE ANY QUESTIONS BEFORE WE BEGIN?
Make your choice:
Piece rate
Tournament
Name: ____________________________________
Name: ____________________________________
Birth date:___________________________________
School:____________________________________
Gender:____________________________________
This queson is about your performance in Task 2 (the tournament task). What do you think was
your rank within the group in terms of sums solved correctly. Please choose a number from 1
(meaning that you were the best in your group of four) to 4 (meaning that you were the 4th in your
group of four). If your guess is correct, you receive €1.
1 2 3 4
__________________________________________________________________________
In this part, you can earn money with your choices. You will have to choose between lo)eries. Each
lo)ery gives you a high amount of money with a 50% probability and a lower amount with a 50%
probability. The roll of a die will then determine whether you get the high or the low payo.. We will
roll the die in front of your eyes at the end of the experiment.
Please pick one of the following �ve op�ons:
€ 2 for certain
€3.50 with a 50%
chance
€1.50 with a 50%
chance
€4 with a 50%
chance
€1 with a 50%
chance
€5 with a 50%
chance
€0.50 with a 50%
chance
€6 with a 50%
chance
€0 with a 50%
chance
How do you see yourself: Are you generally a person who is fully prepared to take risks or do you try
to avoid taking risks?
Please �ck a box on the scale, where the value 0 means: ‘unwilling to take risks’ and the value 10
means: ‘fully prepared to take risk’.
0 1 2 3 4 5 6 7 8 9 1
0
_________________________________________________________________________
Which track do you think you will pick this summer?
ES
NT
NH
CS
__________________________________________________________________________
Do you plan to study at a university in the future?
Yes
No
__________________________________________________________________________
If yes, which topic will you most likely pick?
__________________________________________________________________________
How di<cult is it for you to get a passing grade in mathemacs? Please answer on a scale from 0 to
10 where 0 means very easy and 10 means very di<cult.
0 1 2 3 4 5 6 7 8 9 1
0
_________________________________________________________________________
We would like to know from you which track you think the smartest students in your class will pick.
Please order the order the four tracks from 1 to 4 whereby you assign 1 to the track which you think
the smartest students will choose and 4 to the track the least smart students will choose.
ES
NT
NH
CS
__________________________________________________________________________
With which pro=le do you think you would earn most in ten year's me? Rank the pro=les from 1 to 4
where 1 means that you would earn most if you chose that pro=le and 4 that you would earn least if
you chose that pro=le
ES
NT
NH
CS
__________________________________________________________________________
What was your Cito score?
__________________________________________________________________________
Do you think your mathemacs ability is:
... in the top 25% of your school? Yes No
...in the top 50% of your school? Yes No
....in the top 75% of your school? Yes No
__________________________________________________________________________
Do you agree or disagree with the following proposions:
1 Agree strongly
2 Agree
3 Neither agree nor disagree
4 Disagree
5 Disagree strongly
Boys are be)er at maths than girls
1 2 3 4 5
Girls are be)er at languages than boys
1 2 3 4 5
Boys are be)er at sciences than girls
1 2 3 4 5
A pre-school child is likely to su.er if his or her mother works
1 2 3 4 5
A working mother can establish just as warm and secure a relaonship with her children as a mother
who does not work
1 2 3 4 5
In general, fathers are as well suited to look aDer their children as mothers
1 2 3 4 5
Both the husband and wife should contribute to household income
1 2 3 4 5
Having a job is the best way for a woman to be an independent person
1 2 3 4 5
Version 1
Name: ____________________________________
Round 1 Answer:
67 87 29 24
75 59 32 32
59 24 95 11
19 10 29 74
80 12 70 56
31 17 21 23
38 91 26 92
21 88 99 46
96 99 76 86
74 56 94 37
71 33 34 28
23 68 84 66
35 60 97 15
Version 1
19 25 40 30
23 24 41 52
28 94 87 92
83 73 23 85
17 33 49 79
43 50 34 15
44 60 90 90
47 97 57 62
29 35 77 36
42 65 60 45
54 12 80 10
89 88 67 84
89 49 98 99
64 52 20 27
21 55 58 13
68 83 64 13
