You are on page 1of 43

1he Lffects of Iam||y

Income, arenta|
Lducat|on and Cther
8ackground Iactors on
Access to ost-
Secondary Lducat|on
|n Canada
koss I|nn|e
k|chard Mue||er
MLSA2008-2
Canad|an Lducat|on ro[ect | ueen's Un|vers|ty Schoo| of o||cy Stud|es | Canada M|||enn|um Scho|arsh|p Ioundat|on
Lducat|ona| o||cy Inst|tute | n|gher Lducat|on Strategy Assoc|ates
1oronLo, CnLarlo, Canada - !une 2008 www.mesa-pro[ecL.org
MLSAMLASUkING 1nL LIILC1IVLNLSS CI S1UDLN1 AID
1he MLSA ro[ect
1he Measurlng Lhe LffecLlveness of SLu-
denL Ald ro[ecL, or Lhe MLSA ro[ecL, ls a
four year research efforL belng conducLed by
Lhe Canadlan LducaLlon ro[ecL and Lhe
School for ollcy SLudles aL Cueen's unlver-
slLy on behalf of Lhe Canada Mlllennlum
Scholarshlp loundaLlon. lL has been deslgned
Lo answer Lhe followlng four quesLlons:
AfLer graduaLlng from hlgh school, Leen-
agers comlng from low-lncome back-
grounds face a cholce as Lo aLLend college
or unlverslLy, or noL. lor Lhose who dld
aLLend, how do Lhey compare Lo Lhose
who dld noL?
uoes provldlng more fundlng ln a sLu-
denL's flrsL few years of furLher educaLlon
aLLracL more low-lncome sLudenLs Lo
posL-secondary educaLlon?
uoes provldlng more fundlng ln a sLu-
denL's flrsL few years of furLher educaLlon
make lL more llkely for low-lncome sLu-
denLs Lo sLay ln and graduaLe?
Are low-lncome sLudenLs dlfferenL across
Canada?
1hls paper ls parL of a serles of research
papers sollclLed from some of Lhe leadlng
Canadlan researchers ln Lhe fleld of posL-sec-
ondary educaLlon, Lhe researchers were
asked Lo wrlLe abouL lssues of access and
perslsLence ln posL-secondary educaLlon ln
Canada. 1he requlremenLs for Lhe papers
were LhaL Lhe researchers use one of several
currenLly-exlsLlng SLaLlsLlcs Canada daLabases
or anoLher source of Canadlan daLa. Lach of
Lhe papers commlssloned durlng Lhls pro[ecL
ls avallable for downloadlng from Lhe MLSA
ro[ecL webslLe aL www.mesa-pro[ecL.org.
1he flndlngs and concluslons expressed ln
Lhls paper are Lhose of Lhe auLhors and do
noL necessarlly represenL Lhose of Lhe MLSA
ro[ecL or lLs parLners.
1he artners
1he Lducat|on o||cy Inst|tute ls an lnLer-
naLlonal, non-proflL Lhlnk Lank dedlcaLed Lo
Lhe sLudy of educaLlonal opporLunlLy. Cur
mlsslon ls Lo provlde hlgh-level research and
analysls Lo supporL pollcymakers and pracLl-
Lloners and expand educaLlonal opporLunlLy
Lo all sLudenLs. Ll handles overall pro[ecL
managemenL and co-ordlnaLlon, daLa prlvacy
& cleanlng, and lnLegraLlon of Lhe flnal re-
s ul L s aL L he end of L he pr o[ ec L .
www.educaLlonalpollcy.org
1he Canad|an Lducat|on ro[ect provldes
research and evaluaLlon experLlse ln experl-
menLal, quanLlLaLlve, quallLaLlve and mlxed
meLhods research approaches. 1he company
has experlence worklng wlLh a broad range of
sLakeholders lncludlng governmenLs (aL Lhe
federal and provlnclal levels), secondary and
posL-secondary educaLlonal lnsLlLuLlons, ele-
menLary and secondary school boards, sLu-
denL gr oups , non- pr of l L and non-
governmenLal organlzaLlons and oLher sLake-
holders ln Lhe educaLlon and publlc pollcy
arena ln Canada and lnLernaLlonally. Whlle
much of our work Lo daLe deals wlLh sLudenLs
and youLh aL Lhe posL-secondary level, we are
lncreaslngly engaglng ln research aL Lhe ele-
menLary and secondary levels as well as look-
lng aL sLudenL moblllLy Lhrough llfelong learn-
lng and LranslLlons beLween k-12 and posL-
secondary educaLlon. www.canedpro[ecL.ca
LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and 1lme Lo
uegree CompleLlon
1he Schoo| of o||cy Stud|es at ueen's
Un|vers|ty (www.poeeoso.co/sps) ls a leadlng
cenLre for advanced educaLlon, research, de-
baLe and lnLeracLlon wlLh Lhe non-academlc
world ln Lhe flelds of publlc admlnlsLraLlon
and lndusLrlal relaLlons. ConLlnulng Lhe long-
sLandlng commlLmenL of Cueen's unlverslLy
Lo excellence ln Lhese areas, Lhey are Lralnlng
Lhe nexL generaLlon of leaders for llfe ln a
global age. 1helr masLer's programs llnk Lhe-
ory wlLh pracLlce Lo provlde sLudenLs wlLh
fundamenLal knowledge of Lhe economlc,
pollLlcal, soclal and Lechnologlcal changes
LhaL are Lransformlng Lhe way we llve and Lhe
way we work. SLudenLs enhance Lhelr com-
munlcaLlon and research skllls, and galn new
skllls ln managemenL, pollcy analysls, eco-
nomlcs and quanLlLaLlve meLhods. 1helr
graduaLes are well prepared Lo conLrlbuLe Lo
pollcy-maklng, human resource managemenL
and lndusLrlal relaLlons ln a varleLy of publlc,
prlvaLe and nonproflL organlzaLlons. 1he
School for ollcy SLudles manages Lhe 8e-
search 8evlew CommlLLee for Lhe MLSA ro-
[ecL, whlch ls responslble for fundlng con-
LrlbuLory research pro[ecLs LhaL hlghllghL lm-
porLanL pollcy areas of lnLeresL.
1he Canada M|||enn|um Scho|arsh|p
Ioundat|on ls a prlvaLe, lndependenL organl-
zaLlon creaLed by an acL of arllamenL ln
1998. lL encourages Canadlan sLudenLs Lo
sLrlve for excellence and pursue Lhelr posL-
secondary sLudles. 1he loundaLlon dlsLrlb-
uLes $323 mllllon ln Lhe form of bursarles and
scholarshlps each year LhroughouL Canada.
lLs ob[ecLlves are Lo lmprove access Lo posL-
secondary educaLlon for all Canadlans, espe-
clally Lhose faclng economlc or soclal barrl-
ers, Lo encourage a hlgh level of sLudenL
achlevemenL and engagemenL ln Canadlan
socleLy, and Lo bulld a naLlonal alllance of
organlzaLlons and lndlvlduals around a
shared posL-secondary agenda. 1he lounda-
Llon ls fundlng Lhe MLSA ro[ecL overall, and
has negoLlaLed access Lo lLs sLudenL admlnls-
LraLlve llsLs wlLh each of Lhe provlnces on Lhe
pro[ecL's behalf.
www.mllleoolomscbolotsblps.co
ll LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and 1lme
Lo uegree CompleLlon
Abstract
1hls paper explolLs Lhe unprecedenLed rlch lnformaLlon avall-
able ln Lhe Canadlan ?ouLh ln 1ranslLlon Survey, Sample A (?l1S-
A) Lo lnvesLlgaLe lssues relaLed Lo access Lo posL-secondary
educaLlon (SL). 1he quesLlons we ask are baslcally Lwo-fold: l)
WhaL are Lhe varlous lnfluences on access Lo SL of an lndlvld-
ual's background, lncludlng more LradlLlonal measures such as
famlly lncome and parenLal educaLlon, as well as a broader seL
of measures such as hlgh school grades, soclal/academlc en-
gagemenL," and oLher cognlLlve and behavloural lnfluences?
and ll) Pow does lncludlng such a more exLenslve seL of varl-
ables Lhan has been posslble ln prevlous sLudles change Lhe
esLlmaLed effecLs of Lhe more convenLlonally measured famlly/
parenLal lnfluences (famlly lncome and parenLal educaLlon) on
access Lo SL, and Lhus lndlcaLe how much of Lhe laLLer lnflu-
ences operaLe Lhrough (or oLherwlse proxy) Lhe effecLs of Lhe
broader seL of varlables, Lhereby lsolaLlng Lhe dlrecL - as op-
posed Lo lndlrecL - lnfluence of Lhese LradlLlonal measures on
access? uLlllzlng mulLlnomlal loglL models Lo capLure Lhe cholce
of level of SL (l.e., college versus unlverslLy) we flnd LhaL pa-
renLal lncome ls poslLlvely relaLed Lo unlverslLy aLLendance,
whlle havlng only a mlnor effecL on college, buL Lhls effecL ls
greaLly dlmlnlshed once parenLal educaLlon ls lncluded ln Lhe
esLlmaLlon. Slmllarly, Lhe lmporLance of parenLal educaLlon Lo
unlverslLy aLLendance ls somewhaL dlmlnlshed once cerLaln
measures of hlgh school grades, academlc engagemenL," and a
sLandardlsed readlng LesL score are lncluded - alLhough, lnLer-
esLlngly, Lhese addlLlonal varlables have llLLle furLher affecL on
Lhe famlly lncome lnfluences. 1hese resulLs Lhus supporL oLher
recenL work whlch polnLs Lo Lhe lmporLance of addresslng ear-
ller cognlLlve and behavloural lnfluences, and famlly culLure"
more generally as capLured by parenLal educaLlon, ln effecLlng
change ln Lhe raLes and paLLerns of parLlclpaLlon ln SL - al-
Lhough famlly lncome does remaln a slgnlflcanL lndependenL
facLor, albelL of slgnlflcanLly reduced lnfluence.
8oss llnnle ls a professor aL Lhe unlverslLy of CLLawa
(rflnnle[uoLLawa.ca)
8lchard L. Mueller ls a professor aL Lhe unlverslLy of LeLhbrldge
(rlchard.mueller[sLaLcan.gc.ca)
LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and 1lme Lo
uegree CompleLlon
1ab|e of Contents
................................................................................................................ lnLroducLlon 1
.............................................................................................................. 1he LlLeraLure 3
.............................................................................................................. MeLhodology 4
...................................................................................................................... 1he uaLa 6
......................................................................................................................... 8esulLs 8
............................................................................................ 1he 8asellne Models 8
............................................................ Addlng Plgh School Crades Lo Lhe Model 9
................................................................................. Addlng Lhe Scale varlables 11
................................................................................... Cvervlew of Lhe llndlngs 14
.......................................................................... Concluslons and ollcy lmpllcaLlons 13
................................................................................................................ 8eferences 18
..................................................................................................... 1ables and llgures 20
lease clLe as:
llnnle, 8oss, and Mueller, 8lchard L. (2008). 1be ffects of lom-
lly locome, loteotol Jocotloo ooJ Otbet 8ockqtoooJ loctots
oo Access to lost-5ecooJoty Jocotloo lo coooJo. vlJeoce
ftom tbe l15. 1oronLo, Cn: Canadlan LducaLlon ro[ecL.
(www.mesa-pro[ecL.org/research.php)
LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and 1lme Lo
uegree CompleLlon
Introduct|on
1here now exlsLs a subsLanLlal body of
research on Lhe lmporLance of famlly back-
ground lnfluences, LulLlon levels, and oLher
relaLed varlables on access Lo posL-secondary
educaLlon (SL) - ln Canada as for oLher
counLrles. lor economlsLs, Lhls emphasls on
flnanclal facLors makes perfecL sense. 1he
prlce mechanlsm ls consldered Lo be lmpor-
LanL ln any lnvesLmenL declslon - whlch ls
how Lhe cholce of schoollng level ls Lyplcally
vlewed. And Lhe avallablllLy of Lhe funds re-
qulred Lo make Lhe lnvesLmenL - ln Lhls con-
LexL ofLen proxled by or oLherwlse relaLed Lo
famlly lncome level - ls an equally crlLlcal fac-
Lor, especlally ln a conLexL where caplLal mar-
keLs may llmlL access Lo flnanclng because
prospecLlve sLudenLs are ofLen unable Lo
provlde sufflclenL collaLeral Lo back Lhelr bor-
rowlng (see for example, 8arr, 1993, Chap-
man, 1997).
8ecenL scholarly research ln Lhe area of
access Lo SL (e.g., Cunha, eL al., 2006,
Peckman, 2007) has, however, shlfLed Lhe
emphasls Lo non-flnanclal facLors as a newer
generaLlon of emplrlcal work has lndlcaLed
LhaL flnanclal resources are buL one of many
lmporLanL deLermlnanLs of SL parLlclpaLlon.
1hls llne of research conLends LhaL lL ls long-
Lerm facLors, such as famlly background, over
shorL-Lerm facLors, such as credlL consLralnLs,
whlch are of greaLesL lmporLance ln deLer-
mlnlng access Lo SL. Slnce many of Lhese
facLors are correlaLed wlLh famlly lncome ln
Lhe shorL-Lerm perlod when SL declslons are
made, lL ls ofLen erroneously sLaLed LhaL Lhls
shorL-Lerm credlL consLralnL ls whaL prohlblLs
low-lncome lndlvlduals from aLLendlng SL.
1he lmpllcaLlon of Lhls work ls powerful: pol-
lcy should be more dlrecLed Lowards sLudenLs
earller ln llfe lf Lhe long-Lerm goal ls Lo ln-
crease SL parLlclpaLlon.
1hls ls noL Lo lmply LhaL flnanclal re-
sources are noL lmporLanL, only LhaL Lhey
may be correlaLed wlLh oLher varlables whlch
are also slgnlflcanL deLermlnanLs of SL par-
LlclpaLlon, and Lhus Lhere may have been a
mlsgulded asslgnmenL of Lhe lnfluence of
Lhese flnanclal varlables. lor pollcy purposes,
ldenLlfylng Lhe mosL lmporLanL deLermlnanLs
of access Lo SL ls obvlously lmporLanL for
chooslng Lhe mosL effecLlve pollcy levers for
changlng access raLes (and Lhe underlylng
opporLunlLles whlch drlve Lhose raLes), espe-
clally lf overall SL budgeLs are more or less
flxed and spendlng on one access lever (e.g.,
loans, granLs, or loser LulLlon) means less
spendlng on anoLher (e.g., earller lnLerven-
Llons).
lurLhermore, lf sLudenL loans and granLs
and/or decreased LulLlon are noL havlng Lhe
deslred effecL on access and reLenLlon, Lhan
Lhese resources may slmply amounL Lo renL"
accumulaLlng Lo Lhose whose SL declslons
are noL senslLlve Lo Lhese flnanclal varlables
and would have been aLLendlng anyway.
CovernmenLs would perhaps llkely do beLLer
aL LargeLlng Lhelr resources Lo, say, beLLer
preparlng sLudenLs for galnlng admlsslon lnLo
SL lnsLlLuLlons or lmprovlng lndlvlduals'
educaLlonal success or work hablLs aL lower
levels, ulLlmaLely enhanclng Lhe probablllLy
of SL parLlclpaLlon.
Carnelro and Peckman (2002) succlncLly
summarlze Lhe argumenL, noLlng LhaL chll-
dren whose parenLs have hlgher lncome have
beLLer access Lo quallLy schools, and Lhese
same parenLs shape Lhe LasLes and expecLa-
1 LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and 1lme
Lo uegree CompleLlon
Llons of Lhelr chlldren. 1hey are also able Lo
beLLer nurLure Lhe lnLellecL of Lhelr chlldren
by asslsLlng and dlrecLlng Lhelr sLudles. 1hey
also noLe LhaL cognlLlve ablllLy - aL leasL as
measured by lC on sLandardlzed LesLs - ls
formed by Lhe age of 14 and LhaL Lhe lnflu-
ences of famlly facLors accumulaLe from blrLh
so LhaL scholasLlc ablllLy ls deLermlned by Lhe
end of hlgh school. Agaln, Lhe lmpllcaLlon of
Lhese argumenLs ls LhaL pollcles almed aL ln-
fluenclng young people aL Lhe Llme of Lransl-
Llon from hlgh school Lo SL could be lnap-
proprlaLely LargeLed. 8aLher, any aLLempLs aL
lnLervenLlon should llkely be conducLed much
earller when Lhe cognlLlve maLurlLy of young
people ls belng developed.
WlLh Lhe excepLlon of Carnelro and
Peckman (2002), and a handful of oLher re-
cenL sLudles of Lhls generaLlon whlch address
Lhe lmporLance of famlly background on ac-
cess Lo SL (e.g., Cameron and Peckman,
1998, Lrmlsch and lrancesconl, 2001, keane
and Wolpln, 2001, Cunha, eL al., 2006, Peck-
man, 2007), relaLlvely llLLle daLa has been
broughL Lo bear on Lhls Loplc, and we belleve
our daLa Lo have cerLaln sLrengLhs LhaL can be
found ln no prevlous sLudles, Lhese belng Lhe
parLlcularly rlch seL of background varlables
LhaL can be added Lo Lhe models. 1hls ls es-
peclally Lrue ln Lhe Canadlan conLexL, and Lo
Lhe besL of our knowledge, ours ls Lhe flrsL
sLudy uslng Canadlan daLa whlch aLLempLs Lo
esLlmaLe Lhe lmporLance of a broader seL of
famlly and sLudenL background lnfluences on
access Lo SL ln Lhls fashlon.
1
ln parLlcular, we
focus on Lhe background of young adulLs aL
Lhe Llme when Lhey are 13-years old Lo as-
sess Lhe lmporLance of Lhese facLors on en-
Lerlng elLher college or unlverslLy.
1he conLrlbuLlon of Lhls paper ls, Lhere-
fore, Lo lnclude a much rlcher serles of back-
ground varlables lnLo Lhe analysls, Lhus allow-
lng us Lo analyse Lhe effecLs of boLh Lhe more
LradlLlonally measured deLermlnanLs of ac-
cess Lo SL (e.g., famlly lncome) as well as
Lhe varlous seLs of oLher background varl-
ables LhaL also may be of lmporLance. lL ls
Lhls laLLer group of facLors LhaL maybe corre-
laLed wlLh flnanclal varlables, and yeL may
Lhemselves be lmporLanL correlaLes of SL
parLlclpaLlon amongsL young people. 1hese
lnclude varlous measures of hlgh school
grades, a range of scale" varlables whlch
capLure academlc and soclal engagemenL and
relaLed psychologlcal-soclologlcal aLLrlbuLes,
and a sLandardlzed lnLernaLlonal readlng LesL
score. We expecL LhaL Lhe lncluslon of Lhese
varlables wlll aLLenuaLe Lhe effecLs of Lradl-
Llonal varlables (such as parenLal educaLlon
and lncome). We anLlclpaLe LhaL some varl-
ables such as hlgh school grades and sLan-
dardlzed readlng LesL resulLs wlll have a posl-
Llve and lndependenL lnfluence on SL aLLen-
dance, buL we hold fewer o ptlotl expecLa-
Llons abouL Lhe slgns and relaLlve lmporLance
of many of Lhe oLher varlables, slnce Lhe rela-
Llonshlps beLween Lhese varlables are llkely
complex and have yeL Lo be sufflclenLly sLud-
led ln Lhe llLeraLure. LsLlmaLlng Lhe effecLs of
Lhese varlables and measurlng Lhe changes ln
Lhe esLlmaLed effecLs of Lhe more Lyplcal
background varlables when Lhe broader seLs
are lnLroduced, are Lhe maln focus of Lhe pa-
per.
MLSA - Measurlng Lhe LffecLlveness of SLudenL Ald 2
1 lreneLLe (2007) also explolLs Lhese daLa, buL address a dlfferenL quesLlon: whaL facLors explaln Lhe dlfference ln SL parLlclpaLlon raLes
of lndlvlduals from dlfferenL famlly lncome quarLlles? Pe addresses Lhls by uslng a sLandard Caxaca decomposlLlon approach.
uLlllzlng mulLlnomlal loglL models Lo cap-
Lure Lhe parLlcular level of SL ln whlch Lhe
lndlvldual parLlclpaLes (l.e., college versus
unlverslLy) generaLe resulLs LhaL are generally
conslsLenL wlLh whaL ls found ln Lhe exlsLlng
llLeraLure ln Lerms of Lhe slgns of Lhe coeffl-
clenLs and Lhe overall slze of effecLs. ln par-
Llcular, we flnd LhaL parenLal lncome ls posl-
Llvely relaLed Lo unlverslLy aLLendance, buL
has a much smaller effecL on college aLLen-
dance, slnce lL makes lL more llkely LhaL an
lndlvldual wlll parLlclpaLe ln SL, whlch gen-
eraLes a poslLlve effecL, buL also more llkely
LhaL unlverslLy wlll be Lhe level of cholce,
whlch has a negaLlve lnfluence. 1he sLrengLh
of Lhe lncome effecL ls, however, greaLly dl-
mlnlshed once parenLal educaLlon ls lncluded
ln Lhe esLlmaLlon, suggesLlng LhaL when ln-
come ls lncluded wlLhouL educaLlon lL ls
largely capLurlng Lhe effecL of Lhe laLLer,
whlch ls Lhe domlnanL effecL. ln llke fashlon,
Lhe lmporLance of parenLal educaLlon (espe-
clally on unlverslLy aLLendance) ls dlmlnlshed
Lo some degree - alLhough lL remalns sLrong
- once cerLaln measures of hlgh school
grades, academlc engagemenL," and sLan-
dardlzed readlng LesL scores (l.e., lSA re-
sulLs) are lncluded. lnLeresLlngly, Lhese addl-
Llonal varlables do noL furLher reduce Lhe es-
LlmaLed lmporLance of famlly lncome, sug-
gesLlng LhaL an lndependenL lncome effecL
remalns afLer all Lhese facLors are Laken lnLo
accounL - even lf lL ls noL nearly as greaL as
mlghL be LhoughL from esLlmaLes generaLed
when a less compleLe seL of explanaLory varl-
ables ls lncluded. 1hese resulLs Lherefore
supporL oLher recenL work whlch polnLs Lo
Lhe lmporLance of addresslng earller cognl-
Llve and behavloural lnfluences raLher Lhan
currenL lncome ln effecLlng change ln Lhe
raLes and paLLerns of parLlclpaLlon ln SL.
1he paper ls organlzed as follows: Lhe fol-
lowlng secLlon conLalns a revlew of Lhe perLl-
nenL llLeraLure, SecLlon lll dlscusses Lhe
meLhodology employed, Lhe daLa are dls-
cussed ln SecLlon lv, Lhe resulLs of Lhe de-
scrlpLlve and mulLlvarlaLe analysls are Lhe
Loplc of SecLlon v, and Lhe flnal secLlon
summarlzes Lhe ma[or flndlngs and explores
some of Lhe pollcy lmpllcaLlons of Lhe flnd-
lngs.
1he L|terature
As noLed above, a good share of Lhe llL-
eraLure on access Lo SL amongsL young
people has addressed Lhe lmpacLs of LulLlon
levels and famlly background (Lhe Lwo ofLen
lnLeracLed).
Carnelro and Peckman (2002) revlew Lhe
uS llLeraLure and add new evldence supporL-
lng Lhe paramounL lmporLance of long-Lerm
facLors, such as famlly background, over
shorL-Lerm facLors, such as credlL consLralnLs.
Slnce many of Lhese background facLors are
correlaLed wlLh famlly lncome ln Lhe shorL-
Lerm perlod when SL declslons are made, lL
ls ofLen erroneously sLaLed LhaL Lhls lndlcaLor
of shorL-Lerm credlL consLralnL ls whaL pro-
hlblLs low-lncome lndlvlduals from aLLendlng
SL.
SLudles by Cameron and Peckman (1998,
2001), keane and Wolpln (2001), and Cunha,
eL al. (2006), Lo name buL a few, also supporL
Lhls concluslon. Slmllarly, Murray (2002)
noLes LhaL successful (ln Lerms of lncome)
parenLs Lend Lo have successful chlldren. 1he
lmpllcaLlon of Lhls research ls LhaL SL par-
LlclpaLlon ls largely deLermlned long before
Lhe acLual polnL of enLry lnLo hlgher educa-
3 LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and 1lme
Lo uegree CompleLlon
Llon and LhaL relaxlng shorL-Lerm credlL con-
sLralnLs have only have a mlnlmal effecL on
parLlclpaLlon. keane (2002:293), for example,
commenLlng on Lhe lncome dlvlde ln college
aLLendance ln Lhe uS noLes LhaL Lhls lnequal-
lLy appears Lo be drlven by unequal human-
caplLal accumulaLlon prlor Lo Lhe college-
golng age." 1he pollcy lmpllcaLlon of Lhese
sLudles ls LhaL Lrylng Lo relax shorL-Lerm fl-
nanclal consLralnLs ln aLLendlng SL wlll be
largely lneffecLlve.
1hls body of work represenLs Lhe polnL of
deparLure for Lhe currenL paper. We uLlllze
Lhe exLenslve background lnformaLlon con-
Lalned ln Lhe ?l1S-A daLabase Lo address ac-
cess Lo SL ln Canada. Speclflcally, we add Lo
Lhe exlsLlng llLeraLure by lncludlng a more
comprehenslve seL of background varlables
whlch are deLermlned before enLry lnLo SL
Lo assess Lhe lmpacL of Lhese varlables on
access Lo college and unlverslLy and how ln-
Lroduclng Lhese addlLlonal varlables affecLs
esLlmaLes of Lhe more convenLlonal meas-
ures, such as famlly lncome, as well as paren-
Lal educaLlon.
1hls work also flLs lnLo a speclflcally Ca-
nadlan llLeraLure. 1he accumulaLed evldence
suggesLs LhaL Lhe demand for SL ln Canada
ls prlce lnelasLlc (!unor and usher, 2004), al-
Lhough LulLlon lncreases may have a larger
lmpacL on lndlvlduals from low-lncome faml-
lles (Coelll, 2003). 8oLh ChrlsLophldes, eL al.
(2001) and Corak, eL al. (2003) lnclude paren-
Lal lncome ln Lhelr models of SL parLlclpa-
Llon and flnd LhaL lL ls lmporLanL for unlver-
slLy aLLendance, buL noL college, whlle LulLlon
generally had llLLle general effecL, buL may
have more lmpacL on lndlvlduals from low
lncome famllles. lreneLLe (2003) and uroleL
(2003) slmllarly flnd LhaL SL aLLendance gap
beLween hlgh- and low-lncome famllles ls
narrowed when colleges and unlverslLles are
boLh consldered, buL LhaL sLudenLs from low-
lncome famlly are less llkely Lo aLLend elLher,
especlally unlverslLy.
1wo recenL sLudles (lreneLLe, 2003,
2007), have also casL doubL on Lhe credlL
consLralnL hypoLhesls. ln Lhe flrsL sLudy,
lreneLLe uses Lhe deregulaLlon of profes-
slonal program LulLlon ln CnLarlo as a naLural
experlmenL. Pe dlscovers LhaL lL ls sLudenLs
from mlddle class famllles who saw Lhelr par-
LlclpaLlon ln Lhese programs decllne Lhe
mosL, noL Lhose from lower-lncome famllles.
ln Lhe second sLudy, uslng Lhe same ?l1S-A
daLa we employ, he shows LhaL very llLLle of
Lhe unlverslLy parLlclpaLlon gap beLween sLu-
denLs from famllles ln Lhe flrsL and fourLh ln-
come quarLlles can be explalned by credlL
consLralnLs. 8aLher lL ls dlfferences ln sLan-
dardlzed LesL scores and hlgh school marks
LhaL explaln Lhe ma[orlLy (and almosL all) of
Lhe gap. 1he comblned resulLs of Lhls body of
work agaln suggesLs LhaL resources almed aL
relaxlng credlL consLralnLs (e.g., loans and
even granLs) may be mlsdlrecLed and mlghL
be beLLer uLlllzed aL lmprovlng sLudenL per-
formance aL (or before) Lhe hlgh school level
or provldlng beLLer lnformaLlon Lo sLudenLs
and Lhelr famllles abouL Lhe cosLs and bene-
flLs of educaLlon.
Methodo|ogy
1hls research uses a relaLlvely sLandard
emplrlcal model for esLlmaLlng access Lo SL,
where access ls Laken Lo be a funcLlon of dlf-
ferenL seLs of lnfluences, worklng from a
smaller seL of regressors, lncludlng Lhe prln-
MLSA - Measurlng Lhe LffecLlveness of SLudenL Ald 4
clpal famlly background varlables conven-
Llonally lncluded ln such models, and bulldlng
Lo a more comprehenslve seL of regressors
represenLlng Lhe oLher klnds of lnfluences
measured ln Lhe ?l1S-A - Lhus movlng from a
shorL" regresslon Lo progresslvely longer"
regresslons.
1he model may be expressed as follows:
? = x11 + x22 + x33 +
where ? ls Lhe access measure of lnLeresL
(parLlclpaLlon ln college or unlverslLy), Lhe xl
are vecLors of covarlaLes LhaL lnfluence ?, Lhe
l are Lhe coefflclenLs assoclaLed wlLh each
seL of x, and ls Lhe classlcal sLochasLlc error
Lerm.
x1 comprlses Lhe mosL convenLlonal fam-
lly background varlables such as famlly ln-
come, parenLal educaLlon, famlly Lype, eLc.
whlch are Lyplcally Laken ln Lhe llLeraLure Lo
be Lhe lmporLanL lndlcaLors of Lhe advan-
Lages of famlly background ln Lerms of golng
on Lo SL. 1haL ls, lndlvlduals from hlgher ln-
come famllles or (especlally) Lhose wlLh more
hlghly educaLed parenLs (lL Lurns ouL) are
more llkely Lo go on Lo SL, parLlcularly aL Lhe
unlverslLy level. 1hese slmpler/shorLer mod-
els wlll capLure Lhe LoLal effecLs of Lhese varl-
ables on access, regardless of Lhe paLh of
Lhose lnfluences (l.e., dlrecL or lndlrecL),
whlle plcklng up Lhe lnfluences of oLher omlL-
Led facLors wlLh whlch Lhey are correlaLed.
x2 lncludes one elemenL of Lhe wlder
range of varlables avallable ln Lhe ?l1S. 1hls
seL ls comprlsed of varlous (scholasLlc) abll-
lLy" measures, such as Lhe lndlvldual's hlgh
school grades (overall and ln cerLaln speclflc
sub[ecLs, such as maLh and Lngllsh), and
oLher relaLed lndlcaLors.
A nexL seL of regressors, x3, lnclude oLher
klnds of lnfluences LhaL have been galnlng
lncreaslng aLLenLlon as perhaps consLlLuLlng
some of Lhe more lmporLanL deLermlnanLs of
access Lo SL. 1hese lnclude measures of
engagemenL" and lncluslon," such as how
connecLed Lhe sLudenL felL Lo hls or her hlgh
school, a sLudenL's self-appralsal of confl-
dence and compeLence, parenLal behavlours
regardlng monlLorlng and dlsclpllnlng Lhelr
chlldren, eLc.
lL should be recognlsed LhaL Lhese addl-
Llonal lnfluences do noL necessarlly have a
naLural, economeLrlcally-correcL" orderlng
ln Lerms of Lhelr lncluslon, parLly because we
do noL yeL undersLand Lhese processes very
well. lurLher work wlll undoubLedly conLlnue
our advances ln Lhls respecL. WhaL ls mosL
lmporLanL Lo Lhe currenL analysls, however, ls
LhaL:(l) Lhey are all deLermlned before Lhe
enLry lnLo SL (whlch ls when Lhey have been
measured - l.e., durlng Lhe earller pre-SL
cycles of Lhe ?l1S-A), (ll) Lhey can affecL ac-
cess Lo SL, and (lll) Lhey mlghL ln Lurn be re-
laLed Lo famlly background. Pence, lncludlng
Lhem wlll comprlse an exerclse ln movlng Lo-
wards (l) ldenLlfylng a fuller seL of lnfluences
of access Lo SL, and (ll) seelng how addlng
3 LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and 1lme
Lo uegree CompleLlon
such addlLlonal measures affecLs our under-
sLandlng of Lhe dlrecL and lndlrecL effecLs of
famlly background on access Lo SL.
varlous parLlcular speclflcaLlons of Lhls
model are esLlmaLed, all of whlch use a mul-
Llnomlal loglL seL-up Lo dlfferenLlaLe beLween
access Lo college and unlverslLy.
2
lL ls lmpor-
LanL here Lo explaln Lhe dlfference beLween
Lhese Lwo klnds of SL ln Lhe Canadlan con-
LexL. unlverslLy" lncludes whaL ls conven-
Llonally referred Lo as college" ln Lhe Amerl-
can conLexL, and lncludes all Lypes of pro-
grams LhaL dellver bachelors degrees. CLher-
wlse puL, vlrLually all lnsLlLuLlons ln Canada
LhaL granL such degrees are referred Lo as
unlverslLles," even lf Lhey only offer pro-
grams aL Lhe bachelors level, and regardless
of Lhelr slze. 1he classlc llberal arLs college
LhaL ls found ln Lhe u.S. would Lhus be a unl-
verslLy ln Canada. ln conLrasL, colleges" , or
whaL are ofLen referred Lo more compleLely
as communlLy colleges" offer shorLer, more
pracLlcal programs, usually lasLlng from a few
monLhs Lo Lwo years (or even longer ln some
cases) and ln many cases lnclude Lrade
schools. 1he resulLlng credenLlal ls a college
Jlplomo - as opposed Lo Lhe Jeqtees offered
by unlverslLles.
Cur mulLlnomlal seLup up Lhus allows Lhe
regressors ln our models Lo have dlfferenL
effecLs on college and unlverslLy parLlclpa-
Llon, whlle allowlng Lhese processes Lo be
relaLed.
Cne poLenLlal lssue ln Lhe esLlmaLlon of
Lhese models ls Lhe poLenLlal endogenelLy of
aL leasL some of Lhe rlghL-hand slde varlables.
lor example, sLudenLs who wanL Lo geL lnLo
unlverslLy wlll llkely work harder Lo achleve
Lhe beLLer grades ln hlgh school requlred Lo
galn admlsslon Lo Lhls level of schoollng (and
Lo have more cholce among Lhose Lo whlch
Lhey are accepLed). 1hus, hlgh school grades
are noL sLrlcLly exogenous Lo Lhe SL parLlcl-
paLlon ouLcome. 1here exlsL a varleLy of ways
Lo overcome (alLhough noL necessarlly ellml-
naLe) Lhls endogenelLy problem. 8uL ln Lhls
paper we slmply Lake such measures aL face
value, whlch ls conslsLenL wlLh our goal of
esLlmaLlng Lhe emplrlcal relaLlonshlps ln
quesLlon and seelng how Lhey change (or
noL) as Lhe model lncludes rlcher seLs of re-
gressors. Addlng Lhe more compleLe seLs of
explanaLory varlables may noL ellmlnaLe Lhe
blas caused by endogenelLy, buL lL should aL
leasL aLLenuaLe Lhe problem.
1he Data
1he ?ouLh ln 1ranslLlon Survey - Sample A
(or ?l1S-A) lnlLlally lnLervlewed 13-year olds,
Lhelr parenLs, and Lhelr hlgh school admlnls-
LraLors ln 2000. 1wo follow-up surveys of Lhe
young people (only) were conducLed ln 2002
and Lhen agaln ln 2004. ln Lhls laLLer wave of
Lhe survey, Lhe young people were 19-years
of age, Lhe polnL aL whlch lndlvlduals have
made aL leasL Lhelr lnlLlal cholces abouL en-
Lerlng SL.
MLSA - Measurlng Lhe LffecLlveness of SLudenL Ald 6
2 We are aware LhaL Lhe use of a mulLlnomlal loglL model requlres Lhe assumpLlon of lndependence of lrrelevanL alLernaLlve, an assump-
Llon LhaL ls sLrong glven LhaL Lhe declslon Lo go Lo college or unlverslLy are noL llkely lndependenL. We esLlmaLed a number of Lhe speclfl-
caLlons ouLllned below uslng a mulLlnomlal problL model (whlch does noL requlre Lhls sLrong assumpLlon). ln none of Lhe cases dld Lhe
resulLs change markedly from Lhose presenLed below ln 1ables 2 and 3.
1he dependenL varlables ln our sLudy -
represenLlng enLry lnLo elLher college or unl-
verslLy - Lhus dlfferenLlaLe Lhose who have
declded Lo enLer SL aL Lhls polnL ln Lhelr
llves versus all oLhers - lncludlng Lhose who
have declded noL Lo aLLend as well as Lhose
who may go laLer. Whlle lL would also be ln-
LeresLlng - and ln some ways more lnLeresL-
lng - Lo look aL access when lndlvlduals are
older and would Lherefore have had more
Llme Lo reLurn Lo Lhelr sLudles afLer noL en-
Lerlng SL dlrecLly from hlgh school, oLher
work has found LhaL Lhe relaLlonshlp be-
Lween parLlclpaLlon ln SL and famlly back-
ground (as well as mosL oLher lnfluences)
does noL dlffer very much wlLh Lhe age of Lhe
lndlvlduals lncluded ln Lhe samples (aL leasL
among young people generally). Cur analysls
ls ln any evenL consLralned by Lhe daLa avall-
able, and we conslder our esLlmaLes Lo aL
mlnlmum provlde an lnlLlal seL of resulLs
whlch aL leasL begln Lo geL aL Lhe lssues belng
addressed, perhaps Lo be reLurned Lo afLer
Lhe release of Lhe 2006 survey ?l1S-A daLa.
Whlle our sLudy ls naLlonal ln scope, Lhe
excepLlon Lo Lhls general coverage ls LhaL Lhe
samples exclude Lhose llvlng ln Cuebec. 8e-
cause Cuebec has a speclal sysLem of SL -
collqe J'eoselqoemeot qotol et ptofes-
sloooel, or CLCL as lL ls commonly known -
sLudenLs ln Cuebec only aLLend secondary
educaLlon up Lo Lhe equlvalenL of grade 11,
and Lhen aLLend CLCL Lo elLher prepare for
unlverslLy (Lwo years of CLCL) or Lo com-
pleLe a Lechnlcal program (usually Lwo or
Lhree years of CLCL). We drop Cuebec from
our analysls slnce Lhere ls no way ln Lhese
daLa Lo dlsaggregaLe Lhe Lwo sLreams, and
Lhls could poLenLlally confound our analysls,
slnce unlverslLy-bound sLudenLs would be
classlfled as college sLudenLs lf lncluded ln
Lerms of Lhelr CLCL aLLendance, and Lhe
dlfferenLlaLlon of college- and unlverslLy-
bound sLudenLs ls fundamenLal Lo our analy-
sls. CbservaLlons from Lhe LerrlLorles are also
ellmlnaLed due Lo small sample slzes.
SL parLlclpaLlon ls deflned ln our analy-
sls as Lhe flrsL program LhaL a sLudenL en-
Lered, raLher Lhan Lhe hlghesL level aLLended.
1hls ls prlnclpally owlng Lo Lhe facL LhaL more
lnformaLlon ls avallable on Lhe flrsL program
Lhan on subsequenL programs as well as Lhe
facL LhaL we are concerned wlLh Lhe speclflc
LranslLlon from hlgh school Lo SL ln Lhls re-
search. 1hls sald, glven Lhe relaLlve youLh of
our samples, for mosL sLudenLs Lhe level of
Lhe flrsL program ls Lhe same as Lhe level of
Lhe hlghesL program - aL leasL so far ln Lhelr
llves. (WlLh fuLure waves of Lhe ?l1S daLa, Lhe
dynamlcs relaLlng Lo flrsL versus subsequenL
programs could be addressed, lncludlng
swlLches from one level of SL Lo anoLher
(l.e., beLween college and unlverslLy).
Slnce lndlvlduals who have sLudled ouL-
slde of Canada mlghL have qulLe dlfferenL
backgrounds and experlences, we ellmlnaLe
Lhem from Lhe sample. lor Lhe same reason,
non-Canadlan clLlzens and Lhose wlLh un-
known lmmlgraLlon sLaLus are dropped. ll-
nally, we drop Lhose lndlvlduals for whlch
Lhere are mlsslng daLa as well as Lhose who
are conLlnulng ln hlgh school, slnce we obvl-
7 LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and 1lme
Lo uegree CompleLlon
ously do noL observe any poLenLlal LranslLlon
lnLo SL for Lhls laLLer group.
3

1he flnal sample conLalns 7,832 males
and 8,211 females. AL Llmes, however, Lhls
number ls reduced sllghLly due Lo mlsslng
values of some of Lhe varlables lncluded ln
Lhe dlfferenL models. A full accounLlng of Lhe
observaLlons dropped from Lhe sample ls
conLalned ln Appendlx 1able A-1.
kesu|ts
1be 8oselloe MoJels
1he lnlLlal esLlmaLlon resulLs are pre-
senLed ln 1able 1, for males and females.
1hese represenL a seL of basellne esLlmaLes
LhaL are lnLeresLlng on Lhelr own, and LhaL
provlde a polnL of deparLure for Lhe analysls
LhaL follows as addlLlonal varlables are added
Lo Lhe model. 1he ma[or resulL Lo come from
Lhls exerclse and one LhaL ls lncreaslngly
promlnenL ln Lhe llLeraLure (e.g., Lrmlsch and
lrancesconl, 2001) ls LhaL Lhe effecL of paren-
Lal lncome - especlally ln Lhe laLe-Leen years
when SL declslons are made - ls dlmlnlshed
greaLly once we conLrol for parenLal
educaLlon.
4
lor example, Lhe resulLs wlLh conLrols for
males LhaL do noL lnclude parenLal educaLlon
show LhaL a parenLal lncome level of
$100,000 and over ls assoclaLed wlLh a 19.1
percenLage polnL lncrease ln Lhe probablllLy
of aLLendlng unlverslLy compared Lo Lhe con-
Lrol group of $30,000 Lo $73,000. When pa-
renLal educaLlon level ls lncluded, however,
Lhls flgure drops Lo abouL slx percenLage
polnLs. lor females, Lhe correspondlng esLl-
maLes are 19.8 and 10.3 percenLage polnLs.
lor college access, Lhere are no sLaLlsLlcally
slgnlflcanL lncome effecLs excepL ln Lhe flnal
model for females, where Lhose from Lhe
pooresL famllles have slgnlflcanLly lower par-
LlclpaLlon raLes.
WhaL ls also lnLeresLlng ls LhaL alLhough
dlrecL comparlson of Lhe lncome and educa-
Llon effecLs ls noL posslble because of Lhe dlf-
ferenL naLure (and meLrlcs) of Lhe Lwo seLs of
measures (dollars versus years), Lhe parenLal
lncome effecLs appear Lo be much smaller ln
magnlLude Lhan Lhose of parenLal educaLlon,
aL leasL wlLh respecL Lo unlverslLy enLrance.
lor example, havlng a parenL wlLh a bache-
lor's degree lncreases Lhe probablllLy of golng
Lo unlverslLy by 31.2 percenLage polnLs for
males, compared Lo someone from a famlly
wlLh hlgh school as Lhe hlghesL level of pa-
renLal educaLlon. ConLrasL Lhls wlLh Lhe
aforemenLloned slx percenLage polnL ln-
crease for males from famllles wlLh lncomes
over $100,000 compared Lo lncomes ln Lhe
$30,000-73,000 range.
ln addlLlon, we noLe LhaL parenLal educa-
Llon effecLs boLh college and unlverslLy aL-
Lendance - Lhese lnfluences runnlng ln oppo-
slLe dlrecLlons. 1hus, for example, males wlLh
a parenL wlLh a graduaLe (or professlonal)
degree are (ceLerls parlbus) 43.1 percenL
more llkely Lo go Lo unlverslLy Lhan Lhe com-
MLSA - Measurlng Lhe LffecLlveness of SLudenL Ald 8
3 AnoLher lssue we lgnore ls Lhe effecL of worklng durlng school and Lhe effecL of employmenL, number of hours worked, eLc. on Lhe posL-
secondary educaLlon cholce. lL ls our oplnlon Lhls ls an lmporLanL and complex lssue and one LhaL ls beLLer lefL for a separaLe analysls us-
lng Lhese or oLher daLa.
4 ln Lhe case of Lwo-parenL famllles, parenLal educaLlon represenLs Lhe hlghesL level of Lhe Lwo parenLs. uslng oLher measures (such as
Lhe average level or Lhe dlfferenL levels of Lhe Lwo parenLs) ylelds slmllar flndlngs.
parlson group (parenLs have compleLed hlgh
school), buL Lhls comes parLly aL Lhe cosL of
9.2 percenL lower raLes of college aLLen-
dance. 1hls makes sense: parenLal educaLlon
changes noL only how many lndlvlduals go on
Lo SL (Lhe neL lncrease of 33.9 percenL rep-
resenLed by Lhe Lwo effecLs Laken LogeLher ln
Lhe example [usL glven), buL also Lhe dlsLrlbu-
Llon of Lhe klnd of schoollng Lhey engage ln
(lncreaslng unlverslLy level schoollng and de-
creaslng college level parLlclpaLlon). 1he mul-
Llnomlal loglL model used here capLures
Lhese effecLs ln Lhe economeLrlcally appro-
prlaLe manner, and also ln a manner whlch ls
exposlLlonally convenlenL.
Cur resulLs are also conslsLenL wlLh Lhose
provlded by SacerdoLe (2002), who also flnds
LhaL faLher's lncome (as proxled by occupa-
Llon) and faLher's educaLlon are lmporLanL
deLermlnanLs of enLrance Lo SL. Pe com-
pares chlldren randomly asslgned Lo adopLlve
famllles wlLh chlldren who llve wlLh Lhelr
blrLh parenLs (Lhe conLrol group). Pe flnds
LhaL boLh groups are slmllarly lnfluenced by
lncome, buL educaLlon ls an lmporLanL de-
LermlnanL of SL parLlclpaLlon only for non-
adopLees - whlch of course has meanlng for
Lhe lnLerpreLaLlon of Lhe Lwo klnds of lnflu-
ences. 1hls flndlng ls echoed by lug and
vl[verberg (2003) who model parenLal ablllLy
as havlng boLh a dlrecL effecL on Lhelr chll-
dren's educaLlonal aLLalnmenL as well as an
lndlrecL effecL Lhrough lncome. 1hey argue
LhaL hlgher ablllLy parenLs earn hlgher ln-
comes and Lhese lncomes are also lmporLanL
lnpuLs lnLo a chlld's educaLlon. 1hey flnd LhaL
Lhe effecLs of parenLal ablllLy are reduced
from abouL 70-73 per cenL of Lhe LoLal ablllLy
Lransfer Lo some 33-60 per cenL when Lhe
lndlrecL effecLs of ablllLy on lncome are ln-
cluded ln Lhe model. 8esLuccla and urruLla
(2004) show LhaL mosL of Lhe lnLergenera-
Llonal perslsLence ln educaLlonal aLLalnmenL
can be accounLed for by lnnaLe ablllLy, buL
LhaL Lhe quallLy of early educaLlon can en-
large Lhese exogenous dlfferences ln ablllLy
over Llme.
AJJloq nlqb 5cbool CtoJes to tbe MoJel
nexL we Lurn Lo analyslng Lhe lnfluence
on access Lo SL of Lhe grades LhaL sLudenLs
recelved ln hlgh school aL Lhe Llme of Lhe flrsL
wave of Lhe survey ln 2000 when Lhese
young people were 13-years old. 1he resulLs
of Lhese esLlmaLlons are presenLed ln a con-
densed form ln 1ables 2a and 2b (males and
females, respecLlvely). 1he basellne resulLs
from Lhe full model esLlmaLed ln 1able 1 are
also lncluded Lo faclllLaLe Lhe relevanL com-
parlsons.
1he overall hlgh school average grade
(numerlcal score) has a negaLlve effecL on
college parLlclpaLlon and a poslLlve effecL on
unlverslLy parLlclpaLlon. 1he esLlmaLes sug-
gesL LhaL a Len-percenLage polnL hlgher grade
average wlll resulL ln abouL a Lhree-
percenLage polnL decrease ln college aLLen-
dance, buL abouL a 21 percenLage polnL ln-
crease ln unlverslLy parLlclpaLlon. 1hus,
hlgher grades shlfL Lhe dlsLrlbuLlon of SL
parLlclpaLlon from college Lo unlverslLy, wlLh
a sLrong neL lncrease ln SL parLlclpaLlon.
9 LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and 1lme
Lo uegree CompleLlon
1hese resulLs are almosL ldenLlcal for boLh
females and males.
S

Plgh school grades ln each of maLh, lan-
guage, and sclence (also numerlcal scores),
also yleld coefflclenLs of Lhe same slgn, al-
Lhough of smaller magnlLudes, suggesLlng
LhaL lL ls overall grades, raLher Lhan any lndl-
vldual grade, LhaL ls lmporLanL ln deLermlnlng
parLlclpaLlon ln unlverslLy or college. lndeed,
when Lhe model ls esLlmaLed wlLh all grades
lncluded (column 3), lL ls Lhe overall grade
LhaL ls of paramounL lmporLance ln deLermln-
lng unlverslLy aLLendance.
An lnLeresLlng parLlcular resulL ls LhaL Lhe
maLh grade ls numerlcally Lhe leasL lmporLanL
ln deLermlnlng unlverslLy aLLendance,
wheLher Lhe varlable enLers Lhe model lndl-
vldually or [olnLly wlLh Lhe oLher grade varl-
ables. 1he lack of numeracy skllls (aL leasL as
reflecLed by Lhe maLh grade) does noL appear
Lo be as lmporLanL as Lhe language grade ln
any of Lhe speclflcaLlons for elLher sex. Cf
course Lhls resulL could change lf we were Lo
look aL speclflc area of sLudy, wlLh maLh
grades llkely Lo be relaLed Lo enLry lnLo Lhe
sclences, englneerlng, and oLher dlsclpllnes
whlch depend more on Lhe relaLed sklll seLs.
AnoLher lnLeresLlng resulL ls LhaL Lhe ef-
fecLs of parenLal lncome remaln relaLlvely
unchanged when hlgh school grades are ln-
cluded, whereas Lhe effecL of parenLal educa-
Llon ls aLLenuaLed, ln some cases slgnlflcanLly
so, ln boLh Lhe cases of males and females.
lor example, Lhe effecL of havlng parenLs
who are a unlverslLy graduaLe on males' unl-
verslLy aLLendance decllnes from 43.1 per-
cenLage polnLs ln Lhe model wlLh no grade
varlables (Lhe 8asellne" resulLs), Lo 24.2
percenL when Lhey are all lncluded (column
3). lor women, Lhe change ls from 37.2 per-
cenL Lo 19.9 percenL. 1hese resulLs suggesL
LhaL Lhe lnfluence of parenLal educaLlon
works, aL leasL ln parL, Lhrough hlgh school
grades, whereas lncome ls capLurlng some-
Lhlng oLher Lhan Lhls - lndeed, perhaps a
pure" lncome effecL (an lnference whlch ls
sLrengLhened by Lhe resulLs whlch follow).
1hls flndlng ls lnLeresLlng Lo compare Lo
whaL ls reporLed ln Carnelro and Peckman,
2002 (and ln oLher work by Peckman and
varlous co-auLhors). 1hey flnd LhaL Lhe slg-
nlflcanL effecLs of famlly lncome on u.S. col-
lege aLLendance (unlverslLy" for us) are
largely ellmlnaLed once an lC LesL score ob-
Lalned when Lhe person was ln hls or her mld
Leens ls lncluded ln Lhe model. 1helr lnLer-
preLaLlon - ln a conLexL where Lhey do noL
lnclude parenLal educaLlon ln Lhelr models -
ls LhaL famlly lncome ls, ln Lhe absence of Lhe
LesL score, proxylng a famlly's lnpuLs Lo Lhe
chlld's schoollng and oLher such lnfluences.
Cur flndlng of a reduced parenLal educaLlon
effecL as hlgh school grades are added pre-
sumably sLems from a slmllar seL of relaLlon-
shlps, excepL LhaL lL ls Lhe effecLs of parenLal
educaLlon on a chlld's hlgh school ouLcomes
whlch ls belng capLured when grades are
omlLLed from Lhe model, raLher Lhan Lhe ef-
fecLs of famlly lncome on Lhese. Meanwhlle,
our flndlng of an endurlng lncome effecL sug-
MLSA - Measurlng Lhe LffecLlveness of SLudenL Ald 10
S CaLegorlcal grades were also Lrled ln place of Lhe conLlnuous numerlcal grades ln Lhese Lwo Lables (e.g., 30-60, 60-70, eLc.). 1he re-
sulLs were essenLlal Lhe same as Lhose presenLed here, Lhose wlLh hlgher grades were less llkely Lo parLlclpaLe ln college buL more llkely Lo
parLlclpaLe ln unlverslLy. 1he numerlcal grades used here were seL aL Lhe means of Lhese caLegorlcal varlables (e.g., 60-70 per cenL equals
63 per cenL, eLc.). Appendlx 1able A-2 conLalns deLalls of Lhe grade dlsLrlbuLlons. Crades were also enLered as a quadraLlc ln Lhe model.
Agaln, Lhere were no lmporLanL dlfferences wlLh Lhe resulLs presenLed here.
gesLs money mlghL be maLLerlng aL Lhe polnL
of enLry lnLo SL ln a way Peckman and Car-
nelro dld noL flnd.
AJJloq tbe 5cole votlobles
1he ?l1S also lncludes a seL of scale"
varlables, whlch are lndlces derlved from baL-
Lerles of quesLlons deslgned Lo measure varl-
ous aspecLs of a person's engagemenL ln hlgh
school, self-esLeem, parenLal behavlours, eLc.
A full descrlpLlon of Lhese varlables, whlch
are based on daLa collecLed ln Lhe flrsL survey
when Lhe lndlvlduals ln Lhe sample were age-
13, can be found ln Appendlx 1able A-3. Ac-
companylng summary sLaLlsLlcs are conLalned
ln 1able A-4.
1hese varlables, wlLh Lhe excepLlon of
readlng ablllLy (see below), are normallzed aL
mean zero and a sLandard devlaLlon of one.
1hls ls lmporLanL Lo keep ln mlnd when lnLer-
preLlng Lhe resulLs below.
6
1he lnfluence of
Lhese scale varlables on college and unlver-
slLy aLLendance ls reporLed ln 1ables 3a and
3b for males and females. 1hese are agaln
enLered lnLo Lhe baslc model lndlvldually and
Lhen [olnLly ln Lhe flnal column of Lhese
Lables.
7

Lach of Lhe Lhree hlgh school engage-
menL varlables (as well as Lhe Lwo subcaLego-
rles LhaL comprlse Lhe academlc engagemenL
varlable - academlc ldenLlflcaLlon and aca-
demlc parLlclpaLlon) ls poslLlvely relaLed Lo
unlverslLy aLLendance, buL has llLLle lnfluence
on college aLLendance.
1he mosL lmporLanL of Lhese ls academlc
parLlclpaLlon (a measure of aLLendlng school,
dolng homework, eLc.) whlch ls lmporLanL on
lLs own and also as comprlslng one-half of
Lhe academlc engagemenL varlable, whlch ls
a slmple average of academlc parLlclpaLlon
and academlc ldenLlflcaLlon, Lhe laLLer belng
baslcally a measure of valulng and belonglng
aL school. lor females, belng one sLandard
devlaLlon above Lhe mean on Lhls academlc
parLlclpaLlon scale lncreases unlverslLy aLLen-
dance by abouL 13.9 percenLage polnLs, and
for males Lhe flgure ls abouL 10.6 percenLage
polnLs. 1hus, sLudenLs who aLLend class regu-
larly, compleLe asslgnmenLs on Llme, and
spend more Llme sLudylng, are much more
llkely Lo aLLend unlverslLy Lhan Lhose who do
noL have Lhese hablLs. 1he academlc ldenLlfl-
caLlon effecLs are slgnlflcanL, buL a llLLle
weaker, ln Lhe 6-7 percenL effecL range.
ln addlLlon, Lhe lnfluence of parenLal
educaLlon decllnes ln lmporLance once aca-
demlc parLlclpaLlon ls lncluded, suggesLlng
LhaL Lhe Lwo varlables are poslLlvely corre-
laLed, presumably because parenLs wlLh
more educaLlon Lend Lo Leach Lhelr chlldren
Lo have good work hablLs and so on - or be-
cause Lhey oLherwlse purchase Lhe lnpuLs
(exLracurrlcular acLlvlLles, clubs, eLc.) LhaL
bulld Lhese behavlours, or because Lhey oLh-
erwlse Lransfer or generaLe Lhe aLLrlbuLes
LhaL bulld academlc parLlclpaLlon.
Soclal engagemenL reflecLs Lhe degree Lo
whlch an lndlvldual feels accepLed, respecLed
11 LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and
1lme Lo uegree CompleLlon
6 1he varlables also Lend Lo have subsLanLlal dlsLrlbuLlons (Appendlx 1able A-4), lndlcaLlng LhaL Lhey are lndeed capLurlng someLhlng,
whlch ls of course reflecLed ln Lhe slgnlflcanL effecLs Lhey seem Lo carry as seen ln Lhe resulLs whlch follow.
7 1o check for non-llnearlLles, separaLe esLlmaLes were conducLed uslng boLh a quadraLlc speclflcaLlon of Lhe scale varlables as well as
caLegorlcal dummy varlables. As wlLh Lhe case of hlgh school grades, Lhere are no lmporLanL dlfferences beLween Lhese resulLs and Lhe
resulLs reporLed here.
and lncluded ln Lhe hlgh school envlronmenL,
and alLhough lL ls poslLlvely relaLed Lo unlver-
slLy parLlclpaLlon for boLh males and females,
Lhe effecLs are relaLlvely weak. 1hose who
are one sLandard devlaLlon above Lhe mean
on Lhls measure, for example, are abouL Lwo
percenLage polnLs more llkely Lo aLLend unl-
verslLy. 1he measure has no slgnlflcanL effecL
on college aLLendance.
Self-percepLlon would also seem Lo be an
lmporLanL deLermlnanL of SL parLlclpaLlon.
Self-percepLlon ls capLured by Lhree separaLe
varlables: self-esLeem ls a measure of self-
worLh and self-accepLance, self-efflcacy ls Lhe
sLudenL's own percepLlon of hls or her com-
peLence and confldence ln performlng class
work, and self-masLery ls a measure of belng
ln conLrol of one's own desLlny. SLudenLs who
scored hlgh on any or all of Lhese measures
mlghL be more prepared Lo enLer SL, and
Lhe resulLs ln facL show LhaL all Lhree are
poslLlvely and slgnlflcanLly correlaLed wlLh
unlverslLy - buL noL college - aLLendance, aL
leasL when enLered lndlvldually. Self-efflcacy,
however, has Lhe largesL effecL for boLh
males and females, abouL Lwlce Lhe lmpacL of
Lhe oLher Lwo measures.
Soclal supporL ls sLaLlsLlcally lmporLanL for
males, buL Lhe coefflclenL esLlmaLes are rela-
Llvely small, and Lhe varlable ls noL slgnlflcanL
for females. 1hese resulLs Lhus suggesL LhaL
Lhose who look Lo lmprovlng Lhese elemenLs
of a young person's slLuaLlon as a means of
causlng more of Lhem Lo enrol ln SL would,
Lherefore, appear Lo have Lhe odds sLacked
agalnsL Lhem ln Lerms of geLLlng slgnlflcanL
resulLs wlLh any such sLraLegles.
arenLal behavlour ls dlvlded lnLo Lhree
subcaLegorles: monlLorlng behavlour" ad-
dresses how well parenLs feel lnformed abouL
Lhe acLlvlLles of Lhelr chlldren, whlle nurLur-
ance behavlour" and lnconslsLenL dlsclpllne"
are boLh self-explanaLory. Cf Lhese, monlLor-
lng behavlour ls poslLlvely relaLed Lo unlver-
slLy aLLendance for boLh males and females,
whlle nurLurlng ls lmporLanL only for males.
nelLher has any lnfluence on college parLlcl-
paLlon. llnally, lnconslsLenL dlsclpllne ls nega-
Llvely relaLed Lo parLlclpaLlon ln unlverslLy for
boLh sexes buL agaln ls noL relaLed Lo college
parLlclpaLlon. 1he magnlLudes of Lhese lnflu-
ences lay ln Lhe 2-3 percenLage range (unl-
verslLy aLLendance) where slgnlflcanL, less
where noL (noL surprlslngly).
8eadlng ablllLy ls an exLraordlnarlly lm-
porLanL correlaLe of SL parLlclpaLlon. 1hls
varlable has a mean of 300 and a sLandard
devlaLlon of 100 and was creaLed from Lhe
cross-naLlonal rogramme for lnLernaLlonal
SLudenL AssessmenL (lSA) readlng LesL re-
sulLs. 1he polnL esLlmaLes show LhaL females
who were one-sLandard devlaLlon above Lhe
mean were, on average, 23-percenLage
polnLs more llkely of aLLendlng unlverslLy
(and [usL sllghLly less llkely Lo aLLend college).
lor males, Lhe correspondlng flgure ls 18-
percenLage polnLs.
Cf noLe here ls Lhe facL LhaL Lhe effecLs of
parenLal educaLlon do noL change very much
when Lhe scale varlables are added, lmplylng
LhaL Lhese varlables are noL hlghly relaLed ln
Lhelr effecLs. 1he excepLlons Lo Lhls are aca-
demlc parLlclpaLlon, self-efflcacy, and readlng
ablllLy ln Lhe case of unlverslLy parLlclpaLlon.
ln Lhe flrsL Lwo lnsLances Lhe effecLs of paren-
Lal educaLlon decllnes Lo a moderaLe degree,
MLSA - Measurlng Lhe LffecLlveness of SLudenL Ald 12
whlle Lhe lncluslon of Lhe readlng score de-
creases Lhe value of parenLal educaLlon aL
Lhe bachelor's and graduaLe levels (for exam-
ple) by approxlmaLely one-Lhlrd for boLh
males and females (compare columns 1 and
13 ln 1ables 3a and 3b). 1he lSA readlng
score ls a sLandardlzed LesL and Lhese resulLs
are agaln comparable Lo Lhose obLalned by
Carnelro and Peckman (2002) who also use
sLandardlzed LesL scores Lo explaln dlffer-
ences ln college aLLendance ln Lhe unlLed
SLaLes.
lL ls lmporLanL Lo noLe LhaL Lhe lSA
scores are based on acLual LesLs, whereas Lhe
hlgh school grades (such as Lhose shown ln
1ables 2a and 2b) are self-reporLed. llnnle
and Meng (2003) have shown LhaL Lhese
Lypes of LesL score measures of sklll perform
beLLer Lhan self-assessmenLs of sklll. ln par-
Llcular, Lhey use llLeracy as an example uslng
boLh Lypes of measures (l.e., LesL scores and
self-reporLed ablllLy), and flnd LhaL Lhe self-
assessed measure Lends Lo lead Lo a slgnlfl-
canL underesLlmaLlon of Lhe effecL of llLeracy
on employmenL compared Lo Lhe LesL meas-
ure. ln Lhe presenL work, Lhls blas may also
be presenL, as lndlcaLed by Lhe greaLer esLl-
maLed effecL and assoclaLed decrease ln Lhe
magnlLude of Lhe parenLal educaLlon varlable
(malnly aL Lhe 8A and graduaLe school levels)
when Lhe sLandardlzed LesL score ls lncluded
(1ables 3) versus when Lhe self-assessed
measure of ablllLy are lncluded ln Lhe models
(1ables 2).
8
When all Lhese scale varlables are con-
sldered LogeLher (column 13 ln Lhe Lables),
Lhe resulLs conLlnue Lo suggesL relaLlvely llLLle
lnfluence on college parLlclpaLlon - hardly
surprlslng glven Lhe general lack of lmpacL
when enLered lndlvldually.
lor unlverslLy aLLendance, however, and
desplLe Lhe large number of regressors ln-
cluded, some of whlch would be expecLed Lo
be correlaLed and have somewhaL slmllar ef-
fecLs, cerLaln sLrong lnfluences conLlnue Lo be
seen. ln parLlcular, academlc parLlclpaLlon,
self-efflcacy and readlng ablllLy all remaln
lmporLanL, alLhough Lhe relaLlve magnlLudes
of Lhese varlables conLlnue Lo dlffer. A one-
sLandard devlaLlon lncrease from Lhe mean
of academlc parLlclpaLlon ralses female unl-
verslLy parLlclpaLlon raLes by abouL 8.8 per-
cenLage polnLs on average. lor males, Lhe
correspondlng flgure ls 6.8 percenLage
polnLs. Cne sLandard devlaLlon above Lhe
mean on Lhe self-efflcacy scale means almosL
a flve-percenLage polnL lncrease ln unlverslLy
aLLendance for boLh females and males.
8eadlng ablllLy clearly has Lhe largesL ef-
fecL. lor females, a score one-sLandard devla-
Llon above Lhe mean ls assoclaLed wlLh Lhe
woman belng 19 percenLage polnLs more
llkely Lo aLLend unlverslLy, and for males Lhe
correspondlng flgure ls 13 percenLage polnLs.
WlLh Lhe mean unlverslLy parLlclpaLlon raLes
for males and females ln our sample belng
30.9 and 44.7 per cenL, respecLlvely, readlng
ablllLy ls clearly a ma[or deLermlnanL of who
goes and who does noL. Slnce readlng ablllLy
13 LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and
1lme Lo uegree CompleLlon
8 Cne only needs Lo compare Lhe decllne ln Lhe 8A and graduaLe school coefflclenL values ln 1ables 2 and 1ables 3 when measures of
readlng ablllLy are lncluded. ln Lhe former case (column 3 of 1ables 2a and 2b), lL ls Lhe maln language grade of Lhe lasL year ln hlgh school
(self-reporLed) whlch ls lncluded whereas ln Lhe laLLer case (column 13 of 1ables 3a and 3b) lL ls Lhe admlnlsLraLlve lSA readlng LesL score
whlch ls used. lncluslon of Lhe self-reporLed measure resulLs ln a modesL decllne ln Lhe lnfluences of parenLal educaLlon, whereas ln Lhe
laLLer case Lhese drops are qulLe dramaLlc.
ls a sklll derlved over a perlod of Llme, Lhls
resulL ls also conslsLenL wlLh Lhe earller work
on Lhls sub[ecL by Cameron and Peckman
(1998, 2001) and Carnelro and Peckman
(2002), all of whlch polnL Lo Lhe lmporLance
of long-Lerm famlly facLors ln deLermlnlng
success ln SL Lhrough Lhe lnvesLmenLs faml-
lles make ln Lhelr chlldren's early cognlLlve
and behavloural developmenL.
CLher varlables LhaL conLlnue Lo show
some slgnlflcance, alLhough ln a more mlxed
fashlon (men versus women) and lndlcaLlng
generally smaller effecLs, lnclude soclal sup-
porL, whlch conLlnues Lo have Lhe curlous
effecL of sllghLly decreaslng unlverslLy par-
LlclpaLlon (men and women) and lncreaslng
college parLlclpaLlon ln Lhe case of women,
parenLal monlLorlng, whlch has a small posl-
Llve effecL on unlverslLy aLLendance for fe-
males, and lnconslsLenL parenLal dlsclpllne,
whlch has a small negaLlve effecL on unlver-
slLy parLlclpaLlon for males and females allke.
1he general resulLs presenLed above are
also conslsLenL wlLh Lhe sparse Canadlan llL-
eraLure whlch aLLempLs Lo lnclude addlLlonal
background facLors ln models of access Lo
SL. ln parLlcular, llnnle, Lascelles and
SweeLman (2003) also flnd LhaL lncludlng a
varleLy of hlgh school grades and oLher back-
ground varlables reduces Lhe lnfluence of pa-
renLal educaLlon on access (Lhey do noL have
a famlly lncome varlable). 1hey also flnd LhaL
parenLal educaLlon ls sLlll an lmporLanL lnflu-
ence afLer Lhese grades and background varl-
ables are Laken lnLo conslderaLlon, especlally
aL Lhe unlverslLy level. Cur resulLs are slmllar.
Ovetvlew of tbe lloJloqs
WhaL can we make of Lhese resulLs Laken
LogeLher? 1he facL LhaL academlc parLlclpa-
Llon aL Lhe hlgh school level and hlgher lSA
readlng ablllLy resulLs are lmporLanL deLerml-
nanLs of unlverslLy educaLlon seems obvlous.
1he lesson here ls LhaL LalenL, worklng hard
and belng responslble abouL one's sLudles ls
lmporLanL. 1hese effecLs are also clearly re-
laLed Lo parenLal educaLlon (especlally par-
enLs wlLh a 8A or above) as Lhe coefflclenLs
on Lhese varlables lndlcaLe LhaL Lhey are less
lmporLanL (dlrecL) deLermlnanLs of unlverslLy
parLlclpaLlon when Lhese oLher background
varlables are added Lo Lhe model.
1he Lransmlsslon mechanlsm from parenL
Lo chlld, however, cannoL be ascerLalned
from our esLlmaLes. uo hlghly educaLed par-
enLs push Lhelr chlldren harder, provlde Lhem
wlLh more and/or beLLer developmenLal ln-
puLs, or are readlng ablllLy and academlc
work hablLs LransmlLLed by some oLher
mechanlsm? lndeed, ls Lhere a pure (geneLlc)
herlLablllLy componenL Lo Lhese relaLlon-
shlps? undersLandlng Lhese relaLlonshlps and
mechanlsms would, of course, be a useful
llne for furLher research, uslng Lhe ?l1S daLa
or oLher klnds of analysls.
1he generally small and ofLen lnslgnlfl-
canL coefflclenLs on Lhe parenLal behavlour
varlables seem Lo provlde some supporLlng
evldence for Lhe klnds of explanaLlon LhaL lay
beyond parenLs slmply pushlng Lhelr chlldren
Lo do beLLer or provldlng Lhem wlLh rlcher
developmenLal experlences. ln any case, our
resulLs do underllne Lhe lmporLance of famlly
background and naLural sklll endowmenLs as
emphaslzed by Cameron and Peckman
(2001), keane and Wolpln (2001), and Car-
MLSA - Measurlng Lhe LffecLlveness of SLudenL Ald 14
nelro and Peckman (2002), Lo name buL
Lhree of Lhe relevanL recenL sLudles.
llnally, Lhe models all polnL Lo a smaller,
buL robusL, dlrecL/lndependenL effecL of fam-
lly lncome on access Lo SL, regardless of
whaL oLher explanaLory varlables are added
Lo Lhe model - afLer Lhe esLlmaLed lnfluence
decllne sharply wlLh Lhe addlLlon of Lhe pa-
renLal educaLlon varlables. lamlly lncome ls
noL, lL seems, [usL capLurlng lnpuLs Lo a
chlld's developmenL or Lhe oLher klnds of ln-
fluences whlch our fuller models are able Lo
conLrol for. Agaln, Lhe preclse manner ln
whlch famlly lncome dlrecLly affecLs access Lo
SL remalns beyond Lhe scope of Lhls paper.
Conc|us|ons and o||cy Imp||cat|ons
1hls research has addressed how varlous
background facLors are relaLed Lo college and
unlverslLy ln Canada, and a number of lnLer-
esLlng resulLs have been found. MosL lmpor-
LanL, probably, ls LhaL LhaL our flndlngs agree
wlLh Lhe lncreaslngly common resulL found ln
Lhe llLeraLure LhaL Lhe lmpacL of parenLal ln-
come ls greaLly reduced once a broader seL
of explanaLory varlables are added Lo Lhe
model. ln our case, Lhe greaLesL parL of Lhe
decllne occurs when parenLal educaLlon ls
lncluded. ln oLher words, parenLal educaLlon
- and Lhe varlous lnfluences Lo whlch paren-
Lal educaLlon appears Lo be relaLed (such as a
chlld's readlng ablllLy ln hlgh school) - and
noL parenLal lncome ls whaL largely drlves
young people Lo aLLend SL ln general, and
unlverslLy ln parLlcular. 1haL belng sald, a
smalllsh pure" (dlrecL) lncome effecL re-
malns. Cur oLher ma[or flndlngs are as fol-
lows.
llrsL, a ma[or deLermlnanL of unlverslLy
parLlclpaLlon ls Lhe lndlvldual's score on Lhe
readlng porLlon of Lhe cross-counLry lSA LesL
- probably our besL measure of overall/
general ablllLy," whlle also represenLlng a
parLlcular sklll seL (l.e., llLeracy).
Second, overall hlgh school grades, as
well as Lhe Lhree sub[ecL grades under con-
slderaLlon, also Lend Lo be poslLlvely corre-
laLed wlLh unlverslLy aLLendance, and more
weakly, negaLlvely correlaLed wlLh college
aLLendance (as sLudenLs wlLh hlgher grades
evldenLly shlfL Lhelr SL parLlclpaLlon from
college Lo unlverslLy). lurLhermore, lL ls Lhe
overall hlgh school grade, raLher Lhan any
lndlvldual sub[ecL grade, whlch has Lhe larg-
esL lnfluence. 1hls ls an lnLeresLlng resulL
slnce lL ls ofLen assumed LhaL language arLs
and maLhemaLlcs grades are Lhe mosL lmpor-
LanL deLermlnanLs of academlc success. SLlll,
our resulL makes sense ln llghL of Lhe facL
LhaL mosL sLudenLs Lake a general sLudles
program upon enLerlng unlverslLy, so lL ls noL
surprlslng LhaL a comparably general creden-
Llal ls Lhe mosL lmporLanL deLermlnanL of SL
parLlclpaLlon aL LhaL level. 1hese resulLs also
make sense glven Lhe sLrucLure of Lhe unlver-
slLy and college sysLems ln Canada, Lhe for-
mer ls more excluslve, wlLh admlsslon belng
based on hlgh school grades, whlle Lhe laLLer
ls generally characLerlsed by more open ad-
mlsslon pollcles.
1hlrd, engagemenL aL hlgh school, espe-
clally academlc parLlclpaLlon, whlch essen-
Llally relaLes Lo an lndlvldual's work hablLs, ls
also a slgnlflcanL deLermlnanL of unlverslLy
(buL noL college) parLlclpaLlon. ln facL, lL ls
Lhe mosL lmporLanL of all Lhe engagemenL
varlables, alLhough self-efflcacy (or a feellng
13 LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and
1lme Lo uegree CompleLlon
of compeLence and confldence aL school) ls
also slgnlflcanL.
lourLh, when hlgh school grades, aca-
demlc parLlclpaLlon, or Lhe readlng score on
Lhe lSA are added Lo Lhe baslc model, Lhe
dlrecL effecL of parenLal educaLlon ls dlmln-
lshed, buL ls far from ellmlnaLed, and remalns
an lmporLanL (lndependenL) deLermlnanL of
access Lo SL, especlally unlverslLy. 1he larg-
esL drop ln Lhe parenLal educaLlon effecL
comes from Lhe lncluslon of Lhe lSA readlng
ablllLy score, probably Lhe mosL rellable lndl-
caLor of ablllLy among Lhe background varl-
ables lncluded ln Lhe model (many of whlch
are self-reporLed).
ln oLher words, parenLal educaLlon ap-
pears Lo work Lhrough Lhese oLher seLs of
varlables (readlng ablllLy, course grades, sLu-
denL behavlours, eLc.) Lo lnfluence access Lo
SL. We cannoL, however, ldenLlfy from Lhese
esLlmaLes Lhe preclse paLh LhaL Lhls lnfluence
Lakes. lL could be Lhe resulL of hlghly edu-
caLed parenLs expecLlng more of Lhelr chll-
dren, Leachlng Lhelr chlldren beLLer work
hablLs, provldlng Lhem wlLh more and beLLer
developmenLal lnpuLs, shaplng Lhelr prefer-
ences for SL and Lhe sorLs of careers and
llfesLyles lL enLalls, or lL could be due Lo some
oLher lnpuLs or oLher characLerlsLlc passed on
from parenL Lo chlld whlch are correlaLed
wlLh parenLal educaLlon, buL noL observed
nor conLrolled for ln Lhe esLlmaLes. lurLher
dlsenLangllng Lhese relaLlonshlps would
clearly an lmporLanL avenue for fuLure re-
search.
1he pollcy lmpllcaLlons of Lhls research
are noL sLralghLforward. arenLal educaLlon ls
correlaLed wlLh oLher background varlables
LhaL are Lhemselves lmporLanL deLermlnanLs
of SL parLlclpaLlon. lL ls unreallsLlc Lo expecL
pollcy Lo change Lhe exlgenL level of parenLal
educaLlon, alLhough pollcy could lnfluence lLs
level for Loday's young people for when Lhey
Lhemselves are parenLs. 1herefore, Lhe shorL-
Lerm pollcy focus musL be on Lhe facLors Lo
whlch parenLal educaLlon ls relaLed, as well
as Lhe oLher dlrecL lnfluences on access Lo
SL ldenLlfled ln our models, such as readlng
ablllLy, academlc parLlclpaLlon, and hlgh
school grades.
SLlll, Lhls may noL be as slmple as lL
seems. CurrenL research ln Lhls area (Cunha,
eL al., 2006, Cunha and Peckman, 2007) re-
[ecLs Lhe ldea LhaL economlc ouLcomes such
as access Lo hlgher educaLlon can be ade-
quaLely explalned uslng an addlLlve naLure/
nurLure dlchoLomy. 8aLher Lhe acqulslLlon of
skllls ls complex and lnvolves Lhe lnLeracLlon
beLween cognlLlve, non-cognlLlve and envl-
ronmenLal lnfluences. lurLhermore, lnvesL-
menL ln sklll formaLlon feeds on pasL lnvesL-
menLs ln Lhe same. ?eL alLhough lL ls lnvesL-
menLs early ln llfe whlch have Lhe hlghesL
raLe of reLurn, Lhe lack of Lhese lnvesLmenLs
can (aL leasL ln parL) be overcome by lnvesL-
menLs laLer ln llfe.
9

Accordlng Lo Peckman (2000) cognlLlve
ablllLles (as measured by lC) may peak early
ln llfe and remaln sLable afLer abouL age 10,
buL oLher ablllLles such as moLlvaLlon, self-
MLSA - Measurlng Lhe LffecLlveness of SLudenL Ald 16
9 Peckman and MasLerov (2007: 6) noLe:" Caps ln college aLLendance across socloeconomlc groups are largely shaped by ablllLles formed
ln Lhe early years. Caps ln chlld ablllLy across famllles of dlfferenL lncome levels are assoclaLed wlLh parenLal envlronmenLs and parenLlng
pracLlces. Larly lnLervenLlons can parLlally remedlaLe Lhese deflclLs. LaLer lnLervenLlons are much less effecLlve." Peckman (2007) also
exLenslvely dlscusses Lhls facLors assoclaLed wlLh Lhls parLlclpaLlon gap beLween socloeconomlc groups.
dlsclpllne, and soclal skllls can be enhanced
aL laLer ages. 1hese are facLors whlch more or
less correspond Lo Lhe varlables lncluded ln
Lhe presenL research whlch are poslLlvely re-
laLed Lo unlverslLy access. 1hls lmplles LhaL
even Lhose lndlvlduals wlLh a low probablllLy
of SL parLlclpaLlon based on famlly back-
ground may be able Lo overcome Lhls obsLa-
cle wlLh Lhe developmenL of complemenLary
skllls as Lhey move Lhrough chlldhood and
lnLo adolescence. lL ls Lhe processes by whlch
Lhese complemenLary skllls are acqulred afLer
Len years of age, buL before Lhe SL declslon
ls made, LhaL would provlde a mosL useful
avenue of research ln Lhe medlum-Lerm, as
pollcy makers ponder Lhe quesLlon of brlng-
lng Loday's posL-prlmary school sLudenLs lnLo
SL.
And lL ls perhaps here where new re-
sources need Lo be prlnclpally LargeLed lf
access paLLerns are Lo be slgnlflcanLly
changed, raLher Lhan - for example - con-
Llnulng Lo focus on sLudenL flnanclal ald.
AdequaLe sLudenL ald ls cerLalnly a crlLlcal
elemenL of any access pollcy, buL lL ls only
one of many, and sLudenLs need Lo l) be pre-
pared for SL, and ll) wanL Lo go Lo SL be-
fore any ald pollcy can faclllLaLe Lhe resulLlng
cholces for enLerlng Lhe sysLem. And lL seems
llkely from Lhls research and whaL oLhers
have been reporLlng of laLe LhaL any galns ln
Lhls respecL wlll have Lo focus on more fun-
damenLal lssues, and developmenLs earller ln
a person's llfe, raLher Lhan slmply provlde
enough flnanclal supporL aL Lhe polnL a per-
son may choose Lo enLer Lhe sysLem.
17 LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and
1lme Lo uegree CompleLlon
keferences
8arr, nlcholas. 1993. AlLernaLlve lundlng 8esources for Plgher LducaLlon," 1993. 1he Lconomlc !ournal
103(418), 718-728.
Cameron, SLephen v., and !ames !. Peckman. 1998. Llfe Cycle Schoollng and uynamlc SelecLlon 8las:
Models and Lvldence for llve CohorLs of Amerlcan Males," !ournal of ollLlcal Lconomy 106(2),
262-333.
Cameron, SLephen v., and !ames !. Peckman. 2001. 1he uynamlcs of LducaLlonal ALLalnmenL for 8lack,
Plspanlc, and WhlLe Males, !ournal of ollLlcal Lconomy 109(3), 433-99.
Carnelro, ablo, and !ames !. Peckman. 2002. 1he Lvldence on CredlL ConsLralnLs ln osL-Secondary
Schoollng," 1he Lconomlc !ournal 112(482), 703-34.
Chapman, 8ruce. 1997. ConcepLual lssues and Lhe AusLrallan Lxperlence wlLh lncome ConLlngenL
Charges for Plgher LducaLlon," 1he Lconomlc !ournal 107(442), 738-31.
ChrlsLophldes, Louls n., !lm Clrello, and Mlchael Poy. 2001. lamlly lncome and osL-Secondary Lduca-
Llon ln Canada," Canadlan !ournal of Plgher LducaLlon, 31(1), 177-208.
Coelll, Mlchael. 2003. 1ulLlon, 8aLlonlng and lnequallLy ln osL-Secondary LducaLlon ALLendance," unl-
verslLy of 8rlLlsh Columbla Worklng aper.
Corak, Mlles, CarLh Llpps, and !ohn Zhao. 2003. lamlly lncome and arLlclpaLlon ln osL-Secondary
LducaLlon," SLaLlsLlcs Canada, AnalyLlcal SLudles, 8esearch aper no. 210.
Cunha, llavlo, !ames !. Peckman, Lance Lochner, and ulmlLrly v. MasLerov. 2006. lnLerpreLlng Lvldence
of Llfe Cycle Sklll lormaLlon," ln Pandbook of Lhe Lconomlcs of LducaLlon, vol. 1. Ld. Lrlc A. Pa-
nushek and llnls Welch. new ?ork: Llsevler, 697-812.
Cunha, llavlo, and !ames !. Peckman. 2007. 1he 1echnology of Sklll lormaLlon," Amerlcan Lconomlc
8evlew, apers and roceedlngs 97(2), 31-47.
uroleL, Marle. 2003. arLlclpaLlon ln osL-secondary LducaLlon ln Canada: Pas Lhe 8ole of arenLal ln-
come and LducaLlon Changed over Lhe 1990s?" SLaLlsLlcs Canada, AnalyLlcal SLudles 8ranch 8e-
search aper Serles no. 243.
Lrmlsch, !ohn, and Marco lrancesconl. 2001. lamlly MaLLers: lmpacLs of lamlly 8ackground on Lduca-
Llonal ALLalnmenLs," Lconomlca 68(270), 137-36.
llnnle, 8oss, Lrlc Lascelles, and ArLhur SweeLman. 2003. Who Coes? 1he ulrecL and lndlrecL LffecLs of
lamlly 8ackground on Access Lo osLsecondary LducaLlon," ln Plgher LducaLlon ln Canada. Ld.
8each, Charles M., 8obln W. 8oadway, and 8. Marvln Mclnnls. MonLreal and klngsLon: McClll-
Cueen's unlverslLy ress, 293-338.
llnnle, 8oss, and 8onald Meng. 2003. LlLeracy and Labour MarkeL CuLcomes: Self-assessmenL versus
1esL Score Measures," Applled Lconomlcs 37(17), 1933-31.
lreneLLe, Marc. 2003. 1he lmpacL of 1ulLlon lees on unlverslLy Access: Lvldence from a Large-scale
rlce ueregulaLlon ln rofesslonal rograms," SLaLlsLlcs Canada, AnalyLlcal SLudles 8esearch aper
Serles no. 263.
MLSA - Measurlng Lhe LffecLlveness of SLudenL Ald 18
lreneLLe, Marc. 2007. Why are ?ouLh from Lower-lncome lamllles Less Llkely Lo ALLend unlverslLy?
Lvldence from Academlc AblllLles, arenLal lnfluences, and llnanclal ConsLralnLs" SLaLlsLlcs Can-
ada, AnalyLlcal SLudles 8esearch aper Serles no. 293.
Peckman, !ames !. 2000. ollcles Lo losLer Puman CaplLal," 8esearch ln Lconomlcs 34(1), 3-36.
Peckman, !ames !. 2007. 1he Lconomlcs, 1echnology and neurosclence of Puman CapablllLy lorma-
Llon," n8L8 Worklng aper no. 13193.
Peckman, !ames !., and ulmlLry v. MasLerov. 2007. 1he roducLlvlLy ArgumenL for lnvesLlng ln ?oung
Chlldren," n8L8 Worklng aper no. 13016.
!unor, Sean, and Alex usher. 2004. 1he rlce of knowledge 2004: Access and SLudenL llnance ln Can-
ada. MonLreal: 1he Canadlan Mlllennlum Scholarshlp loundaLlon.
keane, Mlchael . 2002. llnanclal Ald, 8orrowlng ConsLralnLs, and College ALLendance: Lvldence from
SLrucLural LsLlmaLes," Amerlcan Lconomlc 8evlew, apers and roceedlngs 92(2), 293-7.
keane, Mlchael ., and kenneLh l. Wolpln. 2001. 1he LffecL of arenLal 1ransfers and 8orrowlng Con-
sLralnLs on LducaLlon ALLalnmenL," lnLernaLlonal Lconomlc 8evlew 42(4), 1031-1103.
Murray, Charles. 2002. lC and lncome lnequallLy ln a Sample of Slbllng alrs from AdvanLaged lamlly
8ackgrounds," Amerlcan Lconomlc 8evlew, apers and roceedlngs 92(2), 339-43.
lug, Lrlk, and Wlm vl[verberg. 2003. Schoollng, lamlly 8ackground, and AdopLlon: ls lL naLure or ls lL
nurLure?" !ournal of ollLlcal Lconomy 111(3), 611-41.
8esLuccla, ulego, and Carlos urruLla. 2004. lnLergeneraLlonal erslsLence ln Larnlngs: 1he 8ole of Larly
and College LducaLlon," Amerlcan Lconomlc 8evlew 94(3), 1334-78.
SacerdoLe, 8ruce. 2002. 1he naLure and nurLure of Lconomlc CuLcomes," Amerlcan Lconomlc 8evlew,
apers and roceedlngs 92(2), 344-48.
19 LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and
1lme Lo uegree CompleLlon
1ab|es and I|gures
1ab|e 1. MulLlnomlal LoglL LsLlmaLes of Access Lo College and unlverslLy
Ma|es Iema|es
1 2 1 2
Co||ege Un|vers|ty Co||ege Un|vers|ty Co||ege Un|vers|ty Co||ege Un|vers|ty
nS |ocat|on - Urban (8ural) -0.0486 0.102 -0.0443 0.0733 -0.0944 0.0686 -0.0823 0.0331*
[0.016] [0.019] [0.016] [0.018] [0.016] [0.019] [0.016] [0.018]
nS rov|nce (Cn)
newfoundland and Labrador -0.111 0.127 -0.112 0.109 -0.130 0.142 -0.124 0.137
[0.023] [0.028] [0.023] [0.026] [0.022] [0.026] [0.022] [0.023]
rlnce Ldward lsland -0.133 0.238 -0.130 0.200 -0.173 0.216 -0.163 0.188
[0.021] [0.029] [0.022] [0.027] [0.019] [0.023] [0.019] [0.024]
nova ScoLla -0.143 0.242 -0.137 0.193 -0.169 0.219 -0.138 0.193
[0.020] [0.026] [0.021] [0.023] [0.018] [0.023] [0.019] [0.023]
new 8runswlck -0.163 0.172 -0.163 0.137 -0.163 0.178 -0.137 0.160
[0.020] [0.027] [0.019] [0.023] [0.019] [0.023] [0.019] [0.023]
ManlLoba -0.213 0.0898 -0.211 0.0768 -0.148 0.0784 -0.147 0.0780
[0.016] [0.029] [0.016] [0.027] [0.020] [0.026] [0.020] [0.023]
SaskaLchewan -0.170 0.0823 -0.174 0.0673 -0.139 0.0437* -0.140 0.0378
[0.018] [0.023] [0.018] [0.023] [0.019] [0.023] [0.019] [0.023]
AlberLa -0.141 -0.0244 -0.143 -0.0269 -0.0909 -0.0904 -0.0937 -0.0820
[0.018] [0.021] [0.018] [0.020] [0.020] [0.023] [0.019] [0.021]
8rlLlsh Columbla -0.0868 0.00949 -0.0882 -0.00638 -0.0392 -0.0369! -0.0616 -0.0396
[0.020] [0.022] [0.020] [0.020] [0.020] [0.022] [0.020] [0.021]
Irench m|nor|ty outs|de C (All
CLhers)
0.0374* -0.0133 0.0363* -0.0118 0.0464 0.0101 0.0320* 0.0123
[0.031] [0.036] [0.030] [0.032] [0.029] [0.032] [0.029] [0.031]
Iam||y 1ype (1wo arenLs)
MoLher only -0.0308 0.0107 -0.0289 -0.00389 0.0133 0.0216 0.0183 0.0124
[0.024] [0.033] [0.023] [0.030] [0.024] [0.029] [0.024] [0.028]
laLher only 0.00461 -0.0346 0.0219 -0.0719 -0.0116 -0.046 -0.0121 -0.0412
[0.033] [0.061] [0.036] [0.033] [0.033] [0.066] [0.032] [0.060]
CLher -0.0394 -0.0843 -0.0299 -0.0363 -0.0303 -0.186 -0.0368 -0.161!
[0.038] [0.084] [0.064] [0.10] [0.033] [0.068] [0.033] [0.063]
V|s|b|e m|nor|ty (All oLhers) -0.0103 0.187 -0.00329 0.174 -0.00087 0.142 0.0111 0.114
[0.027] [0.032] [0.027] [0.030] [0.027] [0.031] [0.027] [0.031]
Canad|an by |mm|grat|on (by
blrLh)
-0.0239 0.130 -0.0026 0.0316 -0.037 0.182 -0.0297 0.111
[0.038] [0.046] [0.040] [0.044] [0.033] [0.041] [0.037] [0.041]
V|s|b|e M|nor|ty & Canad|an by
|mm|grat|on (oLhers)
-0.0122 -0.112! -0.0188 -0.0682 0.000963 -0.0898 -0.00631 -0.0468
[0.033] [0.046] [0.034] [0.033] [0.063] [0.062] [0.062] [0.062]
arenta|]guard|an's Lducat|on
(PS compleLed)
Less Lhan PS -0.0691! -0.0676! -0.0333 -0.118
[0.029] [0.027] [0.029] [0.032]
Some SL 0.0119 0.0383* 0.0409 0.0434
[0.031] [0.033] [0.033] [0.033]
1rade/College 0.0212 0.0761 -0.0103 0.0618!
[0.021] [0.022] [0.020] [0.024]
unlverslLy-below 8A degree
-0.00123 0.163 -0.0811! 0.302
[0.039] [0.046] [0.034] [0.040]
unlverslLy-8A -0.0488! 0.312 -0.0807 0.299
[0.022] [0.030] [0.022] [0.027]
unlverslLy-Crad -0.0922 0.431 -0.143 0.372
[0.026] [0.033] [0.023] [0.032]
CLher/unknown - - - -
- - - -
ConLlnued
MLSA - Measurlng Lhe LffecLlveness of SLudenL Ald 20
1able 1 conLlnued
Ma|es Iema|es
1 2 1 2
College
unlverslLy
College unlverslLy College unlverslLy College
unlverslLy
arenta| Income Leve|
($30000 Lo $73000)
LxLremely low ($0-$3000) 0.013 -0.0209 0.0139 0.00268 -0.101! -0.0773 -0.0931! -0.087
[0.071] [0.080] [0.071] [0.069] [0.046] [0.097] [0.047] [0.093]
$3000 Lo $23000 0.016 -0.108 0.03 -0.0481 -0.0226 -0.191 -0.0288 -0.102
[0.033] [0.034] [0.033] [0.037] [0.029] [0.036] [0.029] [0.038]
$23000 Lo $30000 0.0198 -0.0367 0.0228 -0.0134 -0.00074 -0.124 -0.00816 -0.0813
[0.020] [0.022] [0.020] [0.021] [0.019] [0.023] [0.020] [0.023]
$73000 Lo $100000 0.00143 0.110 0.0167 0.0401* -0.0121 0.0639 0.00336 0.0203
[0.019] [0.023] [0.020] [0.023] [0.020] [0.024] [0.020] [0.023]
$100000 and up -0.0103 0.191 0.0191 0.0601! -0.0326! 0.198 -0.0143 0.103
[0.022] [0.028] [0.023] [0.026] [0.022] [0.023] [0.023] [0.026]
Cbservat|ons 7832 7832 8311 8311
Notes. Avetoqe motqlool effects ote sbowo. OmltteJ coteqotles ote lo poteotbesls. 5tooJotJ ettots ote lo btockets. p<0.01, ! p<0.05, *
p<0.1.
21 LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and
1lme Lo uegree CompleLlon
1ab|e 2a. LffecLs of Plgh School Crades on Access Lo College and unlverslLy, Males
8ase||ne 1 2 3 4 S
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
arenta|]guard|an's
Lducat|on (PS com-
pleLed)
Less Lhan PS -0.0691! -0.0676! -0.0624! -0.0473* -0.0721! -0.0671! -0.0717! -0.0391! -0.0394* -0.0327* -0.0331
*
-0.0473*
[0.029] [0.027] [0.0309] [0.0248] [0.0297] [0.0273] [0.0299] [0.0276] [0.0311] [0.0269] [0.0333][0.0239]
Some SL 0.012 0.0383* 0.019 0.013 0.014 0.044 0.016 0.043 0.024 0.030 0.028 0.008
[0.031] [0.033] [0.0317] [0.0249] [0.0313] [0.0296] [0.0313] [0.0297] [0.0326] [0.0312] [0.0332][0.0237]
1rade/College 0.021 0.0761 0.023 0.0317 0.018 0.0709 0.018 0.0641 0.018 0.0736 0.021 0.0483!
[0.021] [0.022] [0.0213] [0.0194] [0.0212] [0.0212] [0.0212] [0.0209] [0.0214] [0.0209] [0.0220][0.0196]
unlverslLy-below
8A degree
-0.001 0.163 0.004 0.1146 -0.004 0.1318 0.004 0.1366 -0.007 0.1601 -0.002 0.1161
[0.039] [0.046] [0.0396] [0.0397] [0.0396] [0.0441] [0.0407] [0.0423] [0.0400] [0.0423] [0.0410][0.0417]
unlverslLy-8A -0.0488! 0.312 -0.031 0.1981 -0.0428* 0.2707 -0.0438! 0.2319 -0.0466! 0.2369 -0.031 0.1692
[0.022] [0.030] [0.0224] [0.0232] [0.0223] [0.0287] [0.0220] [0.0278] [0.0223] [0.0263] [0.0230][0.0242]
unlverslLy-Crad -0.09220.431 -0.0373! 0.2869 -0.09000.3909 -0.0723! 0.3433 -0.0704! 0.3293 -0.0310
*
0.2416
[0.026] [0.033] [0.0286] [0.0306] [0.0268] [0.0333] [0.0282] [0.0349] [0.0284] [0.0329] [0.0298][0.0304]
CLher/unknown - - - - - - - - - - - -
- - - - - - - - - - - -
arenta| Income Leve|
($30000 Lo $73000)
LxLremely low ($0-
$3000)
0.016 0.003 0.023 -0.026 0.014 -0.004 0.017 -0.010 -0.039 -0.004 -0.033 -0.032
[0.071] [0.069] [0.0703] [0.0496] [0.0706] [0.0629] [0.0716] [0.0614] [0.0624] [0.0708] [0.0613][0.0349]
$3000 Lo $23000 0.030 -0.048 0.043 -0.047 0.036 -0.034 0.034 -0.032 0.026 0.008 0.047 -0.013
[0.033] [0.037] [0.0366] [0.0341] [0.0336] [0.0363] [0.0361] [0.0380] [0.0336] [0.0393] [0.0386][0.0371]
$23000 Lo $30000 0.023 -0.013 0.026 0.000 0.023 0.000 0.018 -0.002 0.022 -0.003 0.027 0.010
[0.020] [0.021] [0.0201] [0.0183] [0.0201] [0.0206] [0.0199] [0.0201] [0.0203] [0.0200] [0.0209][0.0183]
$73000 Lo $100000 0.017 0.0401* 0.013 0.0410! 0.013 0.0423* 0.013 0.0339* 0.002 0.0308! 0.001 0.0438!
[0.020] [0.023] [0.0194] [0.0190] [0.0197] [0.0221] [0.0193] [0.0210] [0.0193] [0.0211] [0.0194][0.0188]
$100000 and up 0.019 0.0601! 0.020 0.0629 0.022 0.0669 0.019 0.0370! 0.016 0.0319! 0.018 0.0608
[0.023] [0.026] [0.0228] [0.0223] [0.0234] [0.0248] [0.0231] [0.0243] [0.0233] [0.0239] [0.0233][0.0221]
Cvera|| grade of |ast
year nS (numerlcal)
-0.00330.0211 -0.0022
*
0.0134
[0.0008] [0.0006] [0.0012][0.0011]
Math grade of |ast
year nS (numerlcal)
-0.0012! 0.0101 0.000 0.001
[0.0006] [0.0006] [0.0007][0.0007]
Ma|n |anguage grade
of |ast year nS (nu-
mer|ca|)
-0.00240.0137 -0.0019
!
0.0029
[0.0007] [0.0006] [0.0008][0.0008]
Sc|ence grade of nS
(numerlcal)
-0.00120.0123 0.000 0.0033
[0.0004] [0.0003] [0.0003][0.0006]
Cbservat|ons 7832 7677 7626 7643 7223 6924
Notes. Avetoqe motqlool effects ote sbowo. OmltteJ coteqotles ote lo poteotbesls. coottols locloJe oll tbose lo 1oble 1. 5tooJotJ ettots
ote lo btockets. p<0.01, ! p<0.05, * p<0.1. loll tesolts ote ovolloble opoo tepoest.
MLSA - Measurlng Lhe LffecLlveness of SLudenL Ald 22
1ab|e 2b. LffecLs of Plgh School Crades on Access Lo College and unlverslLy, lemales
8ase||ne 1 2 3 4 S
Co||ege
Un|ver-
s|ty
Co||ege
Un|ver-
s|ty
Co||ege
Un|ver-
s|ty
Co||ege
Un|ver-
s|ty
Co||ege
Un|ver-
s|ty
Co||ege
Un|ver-
s|ty
arenta|]guard|an's
Lducat|on
(PS compleLed)
Less than nS
-0.033 -0.118 -0.037 -0.0673! -0.039 -0.1092

-0.033 -0.0972

-0.043 -0.0714! -0.043 -0.030


[0.029] [0.032] [0.0293] [0.0319] [0.0292] [0.0324] [0.0298] [0.0329] [0.0284] [0.0342] [0.0294] [0.0334]
Some SL
0.041 0.043 0.038 0.047 0.033 0.048 0.033 0.027 0.043 0.042 0.031 0.049
[0.033] [0.033] [0.0326] [0.0332] [0.0328] [0.0333] [0.0327] [0.0336] [0.0327] [0.0332] [0.0328] [0.0330]
1rade]Co||ege
-0.010 0.0618! -0.010 0.0436! -0.012 0.0623 -0.012 0.0434* -0.006 0.0393* -0.011 0.0363*
[0.020] [0.024] [0.0201] [0.0217] [0.0203] [0.0238] [0.0202] [0.0223] [0.0201] [0.0223] [0.0203] [0.0216]
Un|vers|ty-be|ow
8A degree
-0.0811! 0.302 -0.044 0.1868 -0.0772! 0.2821 -0.0708! 0.2478 -0.038 0.2037 -0.034 0.1623
[0.034] [0.040] [0.0369] [0.0332] [0.0344] [0.0390] [0.0330] [0.0380] [0.0362] [0.0363] [0.0368] [0.0342]
Un|vers|ty-8A
-0.0807

0.299 -0.0322! 0.1933 -0.0791

0.2802 -0.0644

0.2238 -0.0434! 0.2117 -0.0399* 0.1690


[0.022] [0.027] [0.0224] [0.0249] [0.0223] [0.0263] [0.0228] [0.0262] [0.0223] [0.0236] [0.0227] [0.0230]
Un|vers|ty-Grad
-0.143 0.372 -0.0993

0.2424 -0.1329

0.3407 -0.1216

0.2792 -0.0933

0.2318 -0.0822

0.1991
[0.023] [0.032] [0.0284] [0.0301] [0.0238] [0.0320] [0.0272] [0.0323] [0.0283] [0.0314] [0.0302] [0.0293]
Cther]unknown
- - - - - - - - - - - -
- - - - - - - - - - - -
arenta| Income
Leve| ($30000 Lo
$73000)
Lxtreme|y |ow (50-
5S000)
-0.0931! -0.087 -0.1049! -0.064 -0.1099! -0.106 -0.1091! -0.034 -0.1042! -0.033 -0.1237

-0.041
[0.047] [0.093] [0.0461] [0.0933] [0.0446] [0.0932] [0.0443] [0.0897] [0.0438] [0.0913] [0.0407] [0.1000]
5S000 to 52S000
-0.029 -0.102 -0.040 -0.0728! -0.041 -0.0892! -0.040 -0.0670* -0.036 -0.049 -0.042 -0.041
[0.029] [0.038] [0.0293] [0.0363] [0.0290] [0.0391] [0.0286] [0.0332] [0.0290] [0.0361] [0.0297] [0.0360]
52S000 to 5S0000
-0.008 -0.0813

-0.013 -0.0613

-0.012 -0.0801

-0.013 -0.0601

-0.013 -0.0649

-0.021 -0.0344

[0.020] [0.023] [0.0189] [0.0203] [0.0193] [0.0224] [0.0192] [0.0218] [0.0192] [0.0214] [0.0192] [0.0203]
57S000 to
5100000
0.003 0.021 -0.001 0.030 -0.001 0.012 0.002 0.024 -0.006 0.030 -0.009 0.032
[0.020] [0.023] [0.0191] [0.0199] [0.0201] [0.0221] [0.0200] [0.0218] [0.0193] [0.0208] [0.0191] [0.0196]
5100000 and up
-0.014 0.103 -0.006 0.0893 -0.014 0.0974 -0.014 0.1041 -0.012 0.0900 -0.010 0.0842
[0.023] [0.026] [0.0226] [0.0236] [0.0232] [0.0260] [0.0231] [0.0234] [0.0223] [0.0238] [0.0223] [0.0233]
Cvera|| grade of |ast
year nS (numerlcal)
-0.0034

0.0219 -0.0019* 0.0134


[0.0008] [0.0007] [0.0011] [0.0012]
Math grade of |ast
year nS (numerlcal)
-0.0010* 0.0080 0.000 0.000
[0.0003] [0.0007] [0.0006] [0.0007]
Ma|n |anguage grade
of |ast year nS
(numerlcal)
-0.0013* 0.0149 0.000 0.0038
[0.0007] [0.0007] [0.0008] [0.0009]
Sc|ence grade of nS
(numerlcal)
-0.0030

0.0133 -0.0024

0.0073
[0.0004] [0.0006] [0.0003] [0.0007]
Cbservat|ons 8311 8204 8136 8160 7869 7642
Notes. Avetoqe motqlool effects ote sbowo. OmltteJ coteqotles ote lo poteotbesls. coottols locloJe oll tbose lo 1oble 1. 5tooJotJ ettots
ote lo btockets. p<0.01, ! p<0.05, * p<0.1. loll tesolts ote ovolloble opoo tepoest.
23 LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and
1lme Lo uegree CompleLlon
1ab|e 3a. LffecLs of SLudenL 8ackground on Access Lo College and unlverslLy, Males
8ase||ne 1 2 3 4 S 6 7
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
arenta|]guard|an's Lducat|on (PS com-
pleLed)
Less than nS -0.0691
!
-0.0676
!
-0.0683
!
-0.0399
!
-0.0662
!
-0.0333*-0.0668
!
-0.0323*-0.0688
!
-0.0637
!
-0.0673
!
-0.0391
!
-0.0338
*
-0.0396
!
-0.0643
!
-0.0333
!
[0.029] [0.027] [0.0292
]
[0.0273] [0.0293
]
[0.0278] [0.0292
]
[0.0274] [0.0292
]
[0.0276] [0.0292
]
[0.0279] [0.0306
]
[0.0292] [0.0298
]
[0.0274]
Some SL 0.012 0.0383* 0.0123 0.0331* 0.0098 0.0324* 0.0111 0.0319* 0.0116 0.0398* 0.0133 0.0378* 0.016 0.0344 0.0113 0.0431
[0.031] [0.033] [0.0313
]
[0.0321] [0.0311
]
[0.0308] [0.0312
]
[0.0308] [0.0313
]
[0.0334] [0.0313
]
[0.0326] [0.0317
]
[0.0333] [0.0313
]
[0.0314]
1rade]
Co||ege
0.021 0.07610.0212 0.07670.0188 0.06770.0196 0.07130.0206 0.07770.0218 0.07690.0248 0.07310.0196 0.0774
[0.021] [0.022] [0.0209
]
[0.0214] [0.0209
]
[0.0203] [0.0209
]
[0.0207] [0.0209
]
[0.0217] [0.0209
]
[0.0212] [0.0211
]
[0.0217] [0.0210
]
[0.0213]
Un|vers|ty-
be|ow 8A
degree
-0.001 0.163 -0.0019 0.1637-0.0011 0.1417-0.0007 0.1493-0.0001 0.13990.0003 0.13130.0024 0.14990.003 0.1438
[0.039] [0.046] [0.0393
]
[0.0433] [0.0393
]
[0.0447] [0.0393
]
[0.0441] [0.0393
]
[0.0461] [0.0396
]
[0.0447] [0.0400
]
[0.0449] [0.0396
]
[0.0433]
Un|vers|ty-
8A
-0.0488
!
0.312 -0.0487
!
0.3063-0.0484
!
0.2643-0.0483
!
0.2779-0.0306
!
0.3133-0.0491
!
0.3001-0.0466
!
0.3008-0.0460
!
0.2794
[0.022] [0.030] [0.0217
]
[0.0293] [0.0219
]
[0.0282] [0.0218
]
[0.0283] [0.0213
]
[0.0297] [0.0216
]
[0.0289] [0.0220
]
[0.0292] [0.0220
]
[0.0286]
Un|vers|ty-
Grad
-0.0922

0.431 -0.0921

0.4430-0.0907

0.4196-0.0911

0.4249-0.0922

0.4309-0.0911

0.4403-0.0893

0.4308-0.0804

0.3960
[0.026] [0.033] [0.0264
]
[0.0336] [0.0268
]
[0.0342] [0.0266
]
[0.0348] [0.0264
]
[0.0334] [0.0264
]
[0.0332] [0.0266
]
[0.0336] [0.0273
]
[0.0332]
Cther]
unknown
- - -0.2232

-0.034 -0.2244

-0.0463 -0.2236

-0.0316 -0.2234

-0.0413 -0.2223

-0.013 -0.2189

-0.0231 -0.2227

-0.0247
- - [0.0133
]
[0.1313] [0.0133
]
[0.1336] [0.0138
]
[0.1279] [0.0130
]
[0.1483] [0.0138
]
[0.1637] [0.0130
]
[0.1623] [0.0136
]
[0.1470]
arenta| Income Leve| ($30000 Lo
$73000)
Lxtreme|y
|ow (50-
5S000)
0.016 0.003 0.0134 -0.0001 0.0222 -0.002 0.0191 -0.0049 0.0173 -0.0001 0.0206 -0.0037 -0.0284 0.0131 0.019 0.0091
[0.071] [0.069] [0.0703
]
[0.0633] [0.0716
]
[0.0388] [0.0703
]
[0.0388] [0.0714
]
[0.0680] [0.0722
]
[0.0634] [0.0620
]
[0.0746] [0.0720
]
[0.0634]
5S000 to
52S000
0.030 -0.048 0.0311 -0.0311 0.0314 -0.0431 0.0323 -0.0489 0.0296 -0.043 0.0308 -0.0418 0.0313 -0.0393 0.0333 -0.041
[0.033] [0.037] [0.0333
]
[0.0373] [0.0331
]
[0.0339] [0.0336
]
[0.0363] [0.0348
]
[0.0377] [0.0333
]
[0.0373] [0.0336
]
[0.0382] [0.0333
]
[0.0379]
52S000 to
5S0000
0.023 -0.013 0.022 -0.0093 0.0213 -0.0044 0.0209 -0.0044 0.022 -0.0119 0.0204 -0.0061 0.0186 -0.0069 0.0213 -0.0021
[0.020] [0.021] [0.0199
]
[0.0208] [0.0198
]
[0.0204] [0.0198
]
[0.0202] [0.0199
]
[0.0213] [0.0198
]
[0.0207] [0.0201
]
[0.0212] [0.0198
]
[0.0202]
57S000 to
5100000
0.017 0.0401* 0.0134 0.0436! 0.0166 0.0273 0.0133 0.0377* 0.0131 0.0416* 0.0142 0.0422* 0.0093 0.0483! 0.0139 0.0382*
[0.020] [0.023] [0.0197
]
[0.0223] [0.0197
]
[0.0216] [0.0196
]
[0.0214] [0.0197
]
[0.0230] [0.0197
]
[0.0223] [0.0199
]
[0.0227] [0.0197
]
[0.0219]
5100000 and
up
0.019 0.0601! 0.0198 0.0383! 0.0187 0.0439* 0.0197 0.0308! 0.02 0.0370! 0.0191 0.0307! 0.0169 0.0327! 0.0233 0.0431*
[0.023] [0.026] [0.0233
]
[0.0236] [0.0233
]
[0.0230] [0.0233
]
[0.0247] [0.0233
]
[0.0263] [0.0234
]
[0.0237] [0.0237
]
[0.0239] [0.0236
]
[0.0232]
(cootlooeJ)
MLSA - Measurlng Lhe LffecLlveness of SLudenL Ald 24
1able 3a conLlnued
8ase||ne 1 2 3 4 S 6 7
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
nS Lngagements
Academ|c
|dent|f|cat|on
0.0018 0.0670
[0.007
2]
[0.0073]
"#$%&'(#
)$*+(#()$+(,-
0.0074 0.1033
[0.007
3]
[0.0076]
Academ|c
engagement
0.0034 0.1048
[0.007
2]
[0.0073]
Soc|a| en-
gagement
-0.006 0.0249
[0.006
7]
[0.0077]
Cvera|| en-
gagement
-9L-04 0.0729
[0.007
0]
[0.0073]
Se|f-percept|on:
Se|f-esteem 0.0071 0.0647
[0.007
0]
[0.0077]
Se|f-eff|cacy -0.007 0.1081
[0.006
8]
[0.0069]
Se|f-mastery
Soc|a| Support:
arents' 8ehav-
|ours:
Mon|tor|ng
behav|our
Nurturance
behav|our
Incons|stent
d|sc|p||ne
(8e[ecLlon-
orlenLed
behavlour)
kead|ng Ab|||ty:
Cbservat|ons 78S2 78S0 78S0 78S0 7846 7846 7S97 7749
23 LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and
1lme Lo uegree CompleLlon
1able 3a conLlnued
8 9 10 11 12 13 14 1S

Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
arenta|]guard|an's Lducat|on
(PS compleLed)
Less Lhan PS
-0.0333* -0.0614! -0.0703! -0.0631! -0.0639! -0.0643! -0.0686! -0.0632! -0.0677! -0.0634! -0.0620! -0.0339 -0.0397! -0.0319 -0.043 -0.0317
[0.0310] [0.0293] [0.0293] [0.0278] [0.0293] [0.0278] [0.0292] [0.0278] [0.0292] [0.0276] [0.0296] [0.0282] [0.0297] [0.0273] [0.0312] [0.0281]
Some SL
0.0137 0.0323 0.014 0.0378* 0.0119 0.0393* 0.0099 0.0398* 0.0104 0.0383* 0.0136 0.0237 0.0123 0.0211 0.0121 0.0209
[0.0317] [0.0337] [0.0317] [0.0333] [0.0313] [0.0334] [0.0312] [0.0333] [0.0313] [0.0329] [0.0311] [0.0287] [0.0310] [0.0272] [0.0313] [0.0286]
1rade/
College 0.0238 0.0739 0.0181 0.0776 0.0213 0.0733 0.0214 0.0761 0.0213 0.0790 0.0213 0.0430! 0.0201 0.0419! 0.021 0.0463!
[0.0212] [0.0220] [0.0209] [0.0217] [0.0209] [0.0213] [0.0209] [0.0216] [0.0209] [0.0216] [0.0206] [0.0192] [0.0203] [0.0184] [0.0209] [0.0191]
unlverslLy-
below 8A
degree 0.0004 0.1320 -0.0044 0.1370 -0.0014 0.1348 -0.0021 0.1662 -0.0022 0.1613 0.0034 0.0916! 0.0049 0.0893! 0.003 0.0914!
[0.0399] [0.0447] [0.0393] [0.0433] [0.0397] [0.0438] [0.0394] [0.0463] [0.0393] [0.0439] [0.0398] [0.0386] [0.0393] [0.0372] [0.0398] [0.0377]
unlverslLy-
8A -0.0494! 0.3062 -0.0310! 0.3089 -0.0491! 0.3091 -0.0490! 0.3141 -0.0492! 0.3083 -0.0373* 0.2214 -0.0403* 0.2008 -0.0377* 0.1941
[0.0220] [0.0297] [0.0217] [0.0296] [0.0217] [0.0296] [0.0216] [0.0297] [0.0217] [0.0293] [0.0217] [0.0238] [0.0214] [0.0246] [0.0220] [0.0231]
unlverslLy-
Crad
-0.0910
0.4343
-0.0930
0.4302
-0.0923
0.4309
-0.0931
0.4334
-0.0918
0.4422 -0.0393! 0.2899 -0.0363* 0.2799 -0.0346* 0.2700
[0.0266] [0.0336] [0.0264] [0.0333] [0.0262] [0.0349] [0.0262] [0.0348] [0.0266] [0.0349] [0.0287] [0.0321] [0.0290] [0.0314] [0.0293] [0.0324]
CLher/
unknown
-0.2196
-0.0241
-0.2231
-0.0744
-0.2238
-0.0632
-0.2233
-0.0331
-0.2223
-0.0283
-0.2227
0.0272
-0.2238
0.029
-0.2199
0.0473
[0.0131] [0.1629] [0.0129] [0.1201] [0.0129] [0.1297] [0.0132] [0.1368] [0.0141] [0.1373] [0.0140] [0.1790] [0.0133] [0.1636] [0.0147] [0.1687]
arenta| Income Leve|
($30000 Lo $73000)
LxLremely
low ($0-
$3000)
-0.0264 -0.0044 0.0139 -0.0013 0.0213 0.0124 0.0139 0.0014 0.019 -0.0048 0.0093 0.0117 0.0166 0.0003 -0.0173 0.0182
[0.0618] [0.0728] [0.0721] [0.0680] [0.0703] [0.0698] [0.0710] [0.0693] [0.0708] [0.0663] [0.0687] [0.0624] [0.0686] [0.0339] [0.0619] [0.0372]
$3000 Lo
$23000
0.0313 -0.0403 0.0303 -0.049 0.0333 -0.0324 0.0289 -0.046 0.0282 -0.0483 0.0234 -0.0036 0.0246 -0.0093 0.0313 -0.0116
[0.0333] [0.0384] [0.0349] [0.0374] [0.0347] [0.0368] [0.0344] [0.0372] [0.0344] [0.0370] [0.0333] [0.0331] [0.0343] [0.0337] [0.0333] [0.0346]
$23000 Lo
$30000
0.0207 -0.0048 0.0226 -0.0133 0.0224 -0.0133 0.0231 -0.0136 0.0226 -0.0133 0.0196 0.0066 0.0173 0.0126 0.0178 0.0197
[0.0203] [0.0216] [0.0200] [0.0210] [0.0198] [0.0213] [0.0199] [0.0214] [0.0199] [0.0212] [0.0196] [0.0193] [0.0193] [0.0186] [0.0199] [0.0192]
$73000 Lo
$100000
0.0092 0.0313! 0.0143 0.0439* 0.0137 0.0393* 0.0163 0.0387* 0.0161 0.0438* 0.0173 0.0336* 0.013 0.0337* 0.0079 0.0369*
[0.0199] [0.0232] [0.0197] [0.0229] [0.0197] [0.0229] [0.0198] [0.0230] [0.0197] [0.0228] [0.0194] [0.0203] [0.0192] [0.0193] [0.0193] [0.0197]
$100000 and
up
0.0172 0.0347! 0.0216 0.0360! 0.019 0.0601! 0.0193 0.0381! 0.0182 0.0623! 0.0214 0.0446* 0.0208 0.0408* 0.0243 0.0403*
[0.0237] [0.0262] [0.0236] [0.0261] [0.0234] [0.0261] [0.0233] [0.0264] [0.0234] [0.0261] [0.0229] [0.0233] [0.0228] [0.0223] [0.0234] [0.0231]
MLSA - Measurlng Lhe LffecLlveness of SLudenL Ald 26
1able 3a conLlnued
8 9 10 11 12 13 14 1S
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
nS Lngagements
Academ|c
|dent|f|ca-
t|on
-0.0006 0.0232 -0.0014 0.0113
[0.0081] [0.0074] [0.0087] [0.0081]
"#$%&'(#
)$*+(#()$+(,-
0.0064 0.0681 0.0071 0.0677
[0.0079] [0.0078] [0.0082] [0.0082]
Academ|c
engagement
Soc|a| en-
gagement
-0.0098 0.006 -0.0160! 0.0034
[0.0070] [0.0070] [0.0082] [0.0082]
Cvera|| en-
gagement
Se|f-percept|on:
Se|f-esteem
0.0099 0.0081
[0.0107] [0.0103]
Se|f-eff|cacy
-0.0127 0.0486
[0.0081] [0.0077]
Se|f-mastery
0.0083
0.0313
0.011 0.001
[0.0071]
[0.0074]
[0.0097] [0.0092]
Soc|a| Support: -0.0029 0.0439 -0.0036 -0.0131*
[0.0071] [0.0078] [0.0090] [0.0088]
arents'
8ehav|ours:
MonlLorlng
behavlour
0.0149* 0.0302 0.0130* 0.0091
[0.0077] [0.0084] [0.0079] [0.0081]
nurLurance
behavlour
-0.001 0.0223 -0.0074 0.0013
[0.0071] [0.0079] [0.0076] [0.0076]
lnconslsLenL
dlsclpllne
(8e[ecLlon-
orlenLed
behavlour)
-0.0067 -0.0337 -0.0073 -0.0128*
[0.0071] [0.0073] [0.0073] [0.0074]
kead|ng Ab|||ty: -0.0002! 0.0018 -0.0002 0.0017 -0.0002 0.0013
[0.0001] [0.0001] [0.0001] [0.0001] [0.0001] [0.0001]
Cbservat|ons 7369 7774 7846 7830 7844 7838 7838 7426
Notes. Avetoqe motqlool effects ote sbowo. OmltteJ coteqotles ote lo poteotbesls. coottols locloJe oll tbose lo 1oble 1. 5tooJotJ ettots
ote lo btockets. p<0.01, ! p<0.05, * p<0.1. loll tesolts ote ovolloble opoo tepoest.
27 LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and
1lme Lo uegree CompleLlon
1ab|e 3b: LffecLs of SLudenL 8ackground on Access Lo College and unlverslLy, lemales
8ase||ne 1 2 3 4 S 6 7
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
arenta|]guard|an's Lducat|on (PS com-
pleLed)
Less than nS -0.0333 -0.118 -0.0346 -0.1142 -0.0333 -0.1091 -0.0334 -0.1090 -0.0333 -0.1139 -0.0334 -0.1090 -0.036 -0.1036 -0.0368 -0.1028
[0.029] [0.032]
[0.0290] [0.0321] [0.0288] [0.0323] [0.0291] [0.0326] [0.0289] [0.0318] [0.0289] [0.0323] [0.0282] [0.0323] [0.0294] [0.0337]
Some SL
0.0409 0.0434
0.0398 0.0444 0.0439 0.0328 0.0406 0.0384 0.041 0.0439 0.0398 0.044 0.0402 0.0446 0.0381 0.0374
[0.033] [0.033]
[0.0330] [0.0347] [0.0323] [0.0328] [0.0327] [0.0332] [0.0330] [0.0331] [0.0328] [0.0340] [0.0328] [0.0330] [0.0326] [0.0340]
1rade]
Co||ege
-0.0103 0.0618!
-0.0109 0.0630 -0.0047 0.0448* -0.0084 0.0343! -0.0104 0.0626 -0.0106 0.0610 -0.009 0.0360! -0.0094 0.0494!
[0.020] [0.024]
[0.0203] [0.0237] [0.0200] [0.0231] [0.0201] [0.0232] [0.0203] [0.0239] [0.0202] [0.0233] [0.0204] [0.0241] [0.0202] [0.0233]
Un|vers|ty-
be|ow 8A
degree
-0.0811! 0.302 -0.0773! 0.2923 -0.0632* 0.2342 -0.0679* 0.2663 -0.0813! 0.3032 -0.0790! 0.2929 -0.0839! 0.2903 -0.0714! 0.2728
[0.034] [0.040]
[0.0346] [0.0391] [0.0331] [0.0404] [0.0330] [0.0391] [0.0340] [0.0394] [0.0341] [0.0389] [0.0341] [0.0407] [0.0349] [0.0392]
Un|vers|ty-
8A
-0.0807

0.299
-0.0780
0.2884
-0.0632
0.2496
-0.0687
0.2621
-0.0811
0.2968
-0.0772
0.2819
-0.0729
0.2806
-0.0730
0.2673
[0.022] [0.027]
[0.0223] [0.0268] [0.0223] [0.0264] [0.0226] [0.0266] [0.0224] [0.0267] [0.0223] [0.0263] [0.0229] [0.0271] [0.0222] [0.0266]
Un|vers|ty-
Grad
-0.143 0.372
-0.1390
0.3368
-0.1226
0.3107
-0.1279
0.3226
-0.1428
0.3708
-0.1380
0.3311
-0.1424
0.3333
-0.1334
0.3208
[0.023] [0.032]
[0.0230] [0.0324] [0.0263] [0.0323] [0.0262] [0.0328] [0.0243] [0.0318] [0.0231] [0.0322] [0.0240] [0.0321] [0.0234] [0.0321]
Cther]
unknown
- -
-0.1217 -0.2236! -0.1133 -0.1434 -0.116 -0.1964* -0.126 -0.2228! -0.1233 -0.2313! -0.1023 -0.2421 -0.1193 -0.1361
- -
[0.1290] [0.0963] [0.1273] [0.1346] [0.1289] [0.1111] [0.1273] [0.1000] [0.1274] [0.0927] [0.1330] [0.0913] [0.1283] [0.1336]
arenta| Income Leve| ($30000 Lo
$73000)
Lxtreme|y
|ow (50-
5S000)
-0.0931! -0.087
-0.0969! -0.0746 -0.1028! -0.0303 -0.1002! -0.0363 -0.0923* -0.0741 -0.0973! -0.0313 -0.0928* -0.0817 -0.1003! -0.0694
[0.047] [0.093]
[0.0471] [0.0931] [0.0431] [0.0938] [0.0460] [0.0930] [0.0476] [0.0933] [0.0467] [0.0931] [0.0476] [0.0989] [0.0473] [0.0933]
5S000 to
52S000
-0.0288 -0.102
-0.0268 -0.1130 -0.0277 -0.1043 -0.0249 -0.1173 -0.0283 -0.0986 -0.0279 -0.1031 -0.0219 -0.0962! -0.0318 -0.0946!
[0.029] [0.038]
[0.0293] [0.0368] [0.0293] [0.0367] [0.0293] [0.0360] [0.0294] [0.0376] [0.0294] [0.0371] [0.0301] [0.0374] [0.0294] [0.0376]
52S000 to
5S0000
-0.00816 -0.0813
-0.0068 -0.0891 -0.0093 -0.0788 -0.0064 -0.0883 -0.0077 -0.0787 -0.0087 -0.0784 -0.0126 -0.0678 -0.0137 -0.0719
[0.020] [0.023]
[0.0193] [0.0223] [0.0193] [0.0223] [0.0194] [0.0222] [0.0193] [0.0228] [0.0193] [0.0224] [0.0194] [0.0228] [0.0193] [0.0222]
57S000 to
5100000
0.00336 0.0203
0.004 0.0193 0.0031 0.0183 0.0037 0.0183 0.0037 0.0204 0.0036 0.0193 -0.0007 0.0262 0.0008 0.0274
[0.020] [0.023]
[0.0202] [0.0222] [0.0200] [0.0218] [0.0200] [0.0218] [0.0202] [0.0226] [0.0202] [0.0223] [0.0202] [0.0223] [0.0200] [0.0217]
5100000 and
up
-0.0143 0.103
-0.0134 0.1038 -0.0117 0.0936 -0.0147 0.0989 -0.0142 0.1037 -0.0166 0.1070 -0.0177 0.1094 -0.0178 0.1083
[0.023] [0.026]
[0.0231] [0.0233] [0.0230] [0.0249] [0.0229] [0.0230] [0.0233] [0.0237] [0.0230] [0.0233] [0.0234] [0.0238] [0.0230] [0.0233]
(cootlooeJ)
MLSA - Measurlng Lhe LffecLlveness of SLudenL Ald 28
1able 3b conLlnued
8ase||ne 1 2 3 4 S 6 7
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
nS Lngagements
Academ|c
|dent|f|cat|on
0.004 0.0643
[0.0076] [0.0082]
"#$%&'(#
)$*+(#()$+(,-
-0.0163
! 0.1389
[0.0082] [0.0089]
Academ|c
engagement
-0.007 0.1183
[0.0079] [0.0084]
Soc|a| en-
gagement
0.0114* 0.0229
[0.0067] [0.0079]
Cvera|| en-
gagement
0.0063 0.0736
[0.0071] [0.0080]
Se|f-percept|on:
Se|f-esteem
0.0007 0.0639
[0.0070] [0.0081]
Se|f-eff|cacy
-0.0093 0.1134
[0.0072] [0.0076]
Se|f-mastery
Soc|a| Support:
arents' 8ehav-
|ours:
Mon|tor|ng
behav|our
Nurturance
behav|our
Incons|stent
d|sc|p||ne
(8e[ecLlon-
orlenLed
behavlour)
kead|ng Ab|||ty:
Cbservat|ons 8311 8311 8309 8309 8311 8309 8171 82S2
29 LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and
1lme Lo uegree CompleLlon
1able 3b conLlnued
8ase||ne 8 9 10 11 12 13 14
Co||ege
Un|ver-
s|ty
Co||ege
Un|ver-
s|ty
Co||ege
Un|ver-
s|ty
Co||ege
Un|ver-
s|ty
Co||ege
Un|ver-
s|ty
Co||ege
Un|ver-
s|ty
Co||ege
Un|ver-
s|ty
Co||ege
Un|ver-
s|ty
arenta|]guard|an's Lducat|on
(PS compleLed)
Less than nS
-0.0346 -0.1087 -0.0383 -0.1187 -0.0333 -0.1161 -0.0339 -0.1174 -0.0344 -0.1194 -0.0334 -0.0732! -0.0323 -0.0667! -0.0363 -0.0602*
[0.0284] [0.0321] [0.0289] [0.0321] [0.0289] [0.0318] [0.0289] [0.0318] [0.0288] [0.0319] [0.0292] [0.0324] [0.0287] [0.0322] [0.0283] [0.0331]
Some SL
0.0391 0.0483 0.0387 0.0468 0.0412 0.0448 0.0406 0.0463 0.0404 0.0427 0.0498 0.0123 0.0318 0.0066 0.0472 0.0137
[0.0329] [0.0333] [0.0331] [0.0337] [0.0332] [0.0333] [0.0330] [0.0334] [0.0330] [0.0333] [0.0319] [0.0304] [0.0316] [0.0291] [0.0318] [0.0297]
1rade]
Co||ege -0.0087 0.0332! -0.0131 0.0613! -0.0117 0.0624 -0.0103 0.0626 -0.0103 0.0399! 0.0016 0.0226 0.0041 0.0166 0.0017 0.0143
[0.0204] [0.0241] [0.0203] [0.0241] [0.0203] [0.0240] [0.0203] [0.0240] [0.0203] [0.0240] [0.0198] [0.0212] [0.0193] [0.0207] [0.0197] [0.0210]
Un|vers|ty-
be|ow 8A
degree -0.0831! 0.2911
-0.0874
0.3028 -0.0808! 0.3012 -0.0820! 0.3029 -0.0801! 0.2930 -0.0276 0.1633 -0.0136 0.1402 -0.0239 0.1420
[0.0342] [0.0402] [0.0339] [0.0397] [0.0342] [0.0396] [0.0340] [0.0398] [0.0342] [0.0396] [0.0370] [0.0371] [0.0377] [0.0369] [0.0381] [0.0367]
Un|vers|ty-
8A
-0.0736
0.2836
-0.0849
0.2967
-0.0793
0.2931
-0.0808
0.2990
-0.0790
0.2897 -0.0380* 0.1703 -0.0287 0.1438 -0.0291 0.1422
[0.0229] [0.0271] [0.0224] [0.0270] [0.0223] [0.0268] [0.0224] [0.0268] [0.0226] [0.0269] [0.0226] [0.0232] [0.0223] [0.0246] [0.0229] [0.0231]
Un|vers|ty-
Grad
-0.1437
0.3624
-0.1324
0.3738
-0.1443
0.3736
-0.1432
0.3731
-0.1410
0.3617
-0.0942
0.2166
-0.0820
0.1788
-0.0934
0.1793
[0.0238] [0.0318] [0.0229] [0.0313] [0.0244] [0.0318] [0.0243] [0.0319] [0.0248] [0.0322] [0.0280] [0.0304] [0.0287] [0.0300] [0.0278] [0.0302]
Cther]
unknown -0.0998 -0.2443 0.1242 0.032 -0.1169 -0.2089* -0.1168 -0.2030* -0.1277 -0.2467 -0.1241 -0.087 -0.1349 -0.0837 0.1213 0.0134
[0.1341] [0.0863] [0.1802] [0.2022] [0.1316] [0.1083] [0.1313] [0.1106] [0.1309] [0.0874] [0.1202] [0.1304] [0.1131] [0.1294] [0.1633] [0.1348]
arenta| Income Leve|
($30000 Lo $73000)
Lxtreme|y
|ow (50-
5S000)
-0.0881* -0.0697 -0.0924* -0.0861 -0.0931! -0.0808 -0.0937! -0.0867 -0.0961! -0.0673
-0.1149
0.0082
-0.1137
0.03 -0.1106! 0.037
[0.0493] [0.0989] [0.0482] [0.0933] [0.0473] [0.0990] [0.0467] [0.0933] [0.0464] [0.0936] [0.0410] [0.0923] [0.0403] [0.0988] [0.0437] [0.1009]
5S000 to
52S000
-0.0213 -0.0963! -0.0236 -0.0980 -0.0288 -0.1021 -0.0283 -0.1022 -0.0284 -0.1042 -0.0424 -0.0333 -0.0404 -0.0434 -0.0291 -0.046
[0.0301] [0.0378] [0.0299] [0.0376] [0.0294] [0.0374] [0.0294] [0.0376] [0.0294] [0.0372] [0.0281] [0.0344] [0.0279] [0.0332] [0.0293] [0.0333]
52S000 to
5S0000
-0.0098 -0.0720 -0.0137 -0.0779 -0.0086 -0.0799 -0.0078 -0.0820 -0.0078 -0.0806 -0.0138 -0.0463! -0.0143 -0.0303 -0.02 -0.0431!
[0.0193] [0.0229] [0.0194] [0.0230] [0.0193] [0.0230] [0.0193] [0.0230] [0.0193] [0.0229] [0.0189] [0.0197] [0.0186] [0.0194] [0.0188] [0.0198]
57S000 to
5100000
0.0007 0.0233 0.0031 0.0232 0.0014 0.0227 0.0032 0.0207 0.0023 0.0222 -0.0017 0.0331* -0.0007 0.0323 -0.0061 0.0386*
[0.0202] [0.0223] [0.0202] [0.0227] [0.0202] [0.0223] [0.0202] [0.0226] [0.0202] [0.0223] [0.0191] [0.0202] [0.0189] [0.0198] [0.0189] [0.0198]
5100000 and
up
-0.0166 0.1023 -0.0137 0.1009 -0.0133 0.1034 -0.0139 0.1027 -0.0149 0.1043 -0.0096 0.0924 -0.0074 0.0880 -0.0129 0.0933
[0.0233] [0.0238] [0.0233] [0.0239] [0.0233] [0.0237] [0.0233] [0.0238] [0.0233] [0.0237] [0.0224] [0.0237] [0.0221] [0.0232] [0.0223] [0.0233]
MLSA - Measurlng Lhe LffecLlveness of SLudenL Ald 30
1able 3b conLlnued
8ase||ne 8 9 10 11 12 13 14
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
Co||ege Un|ver-
s|ty
nS Lngagements
Academ|c
|dent|f|ca-
t|on
"#$%&'(#
)$*+(#()$+(,-
Academ|c
engagement
Soc|a| en-
gagement
Cvera|| en-
gagement
Se|f-percept|on:
Se|f-esteem
-0.0014 0.0137
[0.0098] [0.0104]
Se|f-eff|cacy
-0.0101 0.0466
[0.0084] [0.0087]
Se|f-mastery -0.0038 0.0348
-0.0102 0.0023
[0.0073] [0.0083]
[0.0093] [0.0102]
Soc|a| Support: 0.0170! 0.0126
0.0189! -0.0296
[0.0070] [0.0083]
[0.0082] [0.0091]
arents' 8ehav-
|ours:
Mon|tor|ng
behav|our
-0.0032 0.0383
-0.0038 0.0233
[0.0080] [0.0096]
[0.0079] [0.0088]
Nurturance
behav|our
-0.0043 0.0047
-0.0034 -0.0096
[0.0073] [0.0083]
[0.0076] [0.0082]
Incons|stent
d|sc|p||ne
(8e[ecLlon-
orlenLed
behavlour)
-0.0044 -0.0333
-0.0047 -0.0161!
[0.0068] [0.0076]
[0.0069] [0.0072]
kead|ng Ab|||ty:
-0.0003 0.0023 -0.0003 0.0021 -0.0003 0.0019
[0.0001] [0.0001] [0.0001] [0.0001] [0.0001] [0.0001]
Cbservat|ons 8131 8234 8306 8306 8306 8304 8302 8049
Notes. Avetoqe motqlool effects ote sbowo. OmltteJ coteqotles ote lo poteotbesls. coottols locloJe oll tbose lo 1oble 1. 5tooJotJ ettots
ote lo btockets. p<0.01, ! p<0.05, * p<0.1. loll tesolts ote ovolloble opoo tepoest.
31 LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and
1lme Lo uegree CompleLlon
1ab|e A1. Sample SelecLlon
Male Female

% of obs. in
the starting
sample
% of lost in
each stage
of exclusion
# of obs.
left
% of obs. in
the starting
sample
% of lost in
each stage
of exclusion
# of obs.
left
Starting sample (YITS-A participants over all 3 cycles) 10226 10521
QC, Territories or Outside Canada
PS institution in QC 13.79 17.19
Last year of high school in QC 23.36 22.53
HS attended at cycle1 is in QC 23.56 22.70
Ever resident in QC 23.74 22.99
Ever resident, taken HS or PSE in QC 24.13 23.72
PS institution in Territories or outside Canada 0.66 0.89
Last year of high school in Territories or outside Canada 0.39 0.57
Ever resident in Territories or outside Canada 0.24 0.52
Any of the above 25.00 25.00 8415 25.03 25.03 8713
HS continuer or status unknown 6.54 6.05 7961 4.01 3.39 8458
Non-Canadian citizen / immigrant status unknown 0.75 0.70 7931 0.99 0.93 8422
Missing values
Unknown visible minority status 0.34 0.32 7909 0.40 0.51 8395
PSE
Unknown level of PSE program 0.49 0.93
Unknown type of PSE institution 0.19 0.48
Unknown PSE 0.64 0.69 7852 1.15 1.03 8311
Missing values in scale variables
HS Engagement
Academic identification -
Academic participation - -
Academic engagement - -
Social engagement -
Overall engagement - -
Self-perception
Self-esteem 3.47 2.20
Self-efficacy 1.32 0.77
Self-mastery 3.96 2.40
Social Support 1.14 0.89
Parents' Behaviour
Monitoring behaviour - -
Nurturance behaviour - -
Inconsistent discipline (Rejection-oriented behaviour) - -
Reading Ability 0.14 0.19
Missing value of any scale variable 5.64 4.91 7426 3.76 3.70 8049
Missing values in last year HS grades
Overall grade in last year HS 2.77 1.97
Math grade in last year HS 2.87 2.42
Main language grade in last year HS 2.80 2.10
Science grade in HS at cycle 1 8.79 6.39
Missing value of any HS grade 12.82 9.42 6685 9.46 6.62 7486
MLSA - Measurlng Lhe LffecLlveness of SLudenL Ald 32
1ab|e A2. ulsLrlbuLlon of varlous Plgh School Crades, Males and lemales
Categorical Grades (%)
Mean of Nu-
merical Grades
90% to
100%
80% to
89%
70% to
79%
60% to
69%
55% to
59%
50% to
54%
Less
than
50%
Total
Males
HS Overall Grade 5.9 26.2 41.8 21.4 3.3 0.9 0.6 100.0 75.58
HS Math Grade 9.6 22.2 30.3 25.9 7.3 4.1 0.6 100.0 73.99
HS Main Language Grade 5.8 27.7 36.2 23.1 4.4 2.5 0.2 100.0 75.14
HS Science Grade 11.4 25.7 26.9 19.2 8.0 5.2 3.5 100.0 73.54
Females
HS Overall Grade 9.0 38.6 37.9 12.1 1.7 0.5 0.3 100.0 78.90
HS Math Grade 10.3 23.7 31.0 22.7 6.0 5.8 0.6 100.0 74.46
HS Main Language Grade 13.3 39.4 31.4 13.2 1.6 1.1 0.1 100.0 79.71
HS Science Grade 13.3 30.5 26.8 16.3 6.2 4.2 2.6 100.0 75.67
Note: 1o calculaLe means, caLergorlcal grades are converLed Lo numerlcal grades as follows: 90 Lo 100 = 93, 80 Lo 89 = 83, 70 Lo
79 = 73, 60 Lo 69 = 63, 33 Lo 39 = 37, 30 Lo 34 = 32, and less Lhan 30 = 23.
33 LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and
1lme Lo uegree CompleLlon
1ab|e A3. LxplanaLlon of Scale varlables
All of Lhe varlous scales used ln Lhe ?l1S 13-year-old cohorL, and ln Lhe ?l1S 18-20 year-old cohorL are modeled afLer Lhe
LlkerL Scale (LlkerL, 1932). Scores released for ?l1S scales were based on an lLem response Lheory (l81) approach. 1he l81
scores and Lhelr respecLlve sLandard errors were esLlmaLed uslng welghLed maxlmum llkellhood (see Warm, 1989) and ap-
plylng a generallzed parLlal credlL model. 1he generallzed parLlal credlL model ls an exLenslon of Lhe Lwo parameLer loglsLlc
dlsLrlbuLlon Lo polyLomous (caLegorlcal) daLa (Murakl, 1997). lor esLlmaLlng l81 scores, Lhe populaLlon dlsLrlbuLlon of Lhe
scores was speclfled Lo have a mean of zero and a sLandard devlaLlon of one. Cnce sLandardlzed, Lhe respondenL's esLl-
maLed score, ln Lhls case, can be lnLerpreLed as Lhe number of sLandard devlaLlons of Lhe populaLlon of lnLeresL above (lf
poslLlve) or below (lf negaLlve) Lhe mean.
n|gh Schoo| Lngagement Sca|e
Ovetoll scbool eoqoqemeot
Measures a respondenL's overall engagemenL for Lhe sLudenL's presenL school year, focuslng upon examlnlng behavloural
facLors.
8elaLed CuesLlons: uerlved by a slmple average of Lhe varlables academlc engagemenL" and soclal engagemenL."
5oclol eoqoqemeot
ueflned as Lhe ldenLlflcaLlon wlLh and behavloural lnvolvemenL ln Lhe soclal aspecLs of school (Lhe school soclal llfe). lL ln-
volves boLh a feellng of belonglng Lo Lhe school's soclal envlronmenL and a sense of flL beLween Lhe lndlvldual and Lhe
school. 1hls connecLlon reflecLs Lhe exLenL Lo whlch sLudenLs feel personally accepLed, respecLed, lncluded and supporLed
by oLhers ln Lhe school's soclal envlronmenL.
ke|ated uest|ons
?SA9k eople aL school are lnLeresLed ln whaL l have Lo say,
?SA9C l have frlends aL school whom l can Lalk Lo abouL personal Lhlngs,
?SA9 l have frlends aL school who can help me wlLh school work, lf needed,
S131C01 My school ls a place where l feel llke an ouLslder,
S131C02 My school ls a place where l make frlends easlly,
S131C03 My school ls a place where l feel llke l belong,
S131C04 My school ls a place where l feel awkward and ouL of place,
S131C03 My school ls a place where oLher sLudenLs seem Lo llke me,
S131C06 My school ls a place where l feel lonely.
AcoJemlc eoqoqemeot
ueflned as Lhe ldenLlflcaLlon wlLh and behavloural lnvolvemenL (parLlclpaLlon) ln Lhe academlc aspecLs of school. Academlc
aspecLs of school lnclude Lhe sLudenLs' deallngs wlLh Leachers, currlcula, and Lhe school governance.
8elaLed CuesLlons: uerlved by a slmple average of Lhe varlables academlc parLlclpaLlon" and academlc ldenLlflcaLlon".
AcoJemlc pottlclpotloo
locuslng on Lhe flrsL Lhree levels of Laxonomy Lo academlc parLlclpaLlon: Lhe acqulescence Lo Lhe need Lo aLLend school,
Lo be prepared and Lo respond Lo dlrecLlons and quesLlons, sLudenLs demonsLraLlng lnlLlaLlve-Laklng behavlours, and
parLlclpaLlon ln Lhe soclal, exLracurrlcular, and aLhleLlc aspecLs of school llfe ln addlLlon Lo or as a subsLlLuLe for exLenslve
parLlclpaLlon ln academlc work.
8elaLed CuesLlons:
?SA6 hours on homework ouLslde of class durlng free perlods and aL home,
?SA7 number of Llme l cuL or sklpped a class wlLhouL permlsslon,
?SA88 l compleLed my asslgnmenLs,
S132C01 l compleLed homework on Llme,
S133C01 S133C02
S133C03
Cn average, Llme spenL each week on homework and sLudy ln Lhese sub[ecL areas: LesL language,
maLhemaLlcs and sclence, respecLlvely.
MLSA - Measurlng Lhe LffecLlveness of SLudenL Ald 34
n|gh Schoo| Lngagement Sca|e: AcoJemlc oqoqemeot (conLlnued)
AcoJemlc pottlclpotloo
Measures a respondenL's academlc ldenLlflcaLlon wlLh hlgh school, Lhe focus of aLLenLlon ls on Lwo componenLs of lden-
LlflcaLlon, valulng and belonglng. A sLudenL who falls Lo ldenLlfy wlLh school ls expecLed Lo have a lack of valulng for Lhe
school and a lack of feellngs of belonglng Lo Lhe school.
8elaLed CuesLlons:
?SA8l l geL along well wlLh Leachers,
?SA8! l am lnLeresLed ln whaL l am learnlng ln class,
?SA9L School ls one of Lhe mosL lmporLanL Lhlngs ln my llfe,
?SA9l Many of Lhe Lhlngs we learn ln class are useless,
?SA9C MosL of my Leachers don'L really care abouL me
?SA9P MosL of Lhe Llme, l would llke Lo be any place oLher Lhan ln school,
?SA9! MosL of whaL l learn ln school wlll be useful when l geL a [ob,
?SA9L School ls ofLen a wasLe of Llme,
?SA9M School ls more lmporLanL Lhan mosL people Lhlnk,
?SA9n MosL of my Leachers do a good [ob of Leachlng,
S130C03 MosL of my Leachers really llsLen Lo whaL l have Lo say,
S130C04 lf l need exLra help, l wlll recelve lL from my Leachers,
S130C03 MosL of my Leachers LreaL me falrly,
S131C07 My school ls a place where l do noL wanL Lo go,
S132C06 l am glvlng lnLeresLlng homework.
Se|f-percept|on
5elf-esteem
1he self-esLeem scale LhaL was chosen for ?l1S ls Morrls 8osenberg's22 self-esLeem scale (8SL) (8osenberg, 1963, p.17). 8o-
senberg deflnes selfesLeem as favourable or unfavourable aLLlLudes Lowards self and proposes a serles of Len quesLlons Lo
measure lL.WlLhln Lhe conLexL of ?l1S, 8SL aLLempLs Lo measure adolescenLs' global feellngs of self-worLh or self-accepLance.
ke|ated uest|ons
?Sl1A l feel l am a person of worLh, aL leasL on an equal basls wlLh oLhers,
?Sl18 l feel LhaL l have a number of good quallLles,
?Sl1C All ln all, l Lend Lo feel LhaL l am a fallure,
?Sl1u l am able Lo do Lhlngs as well as mosL oLher people,
?Sl1L l feel l do noL have much Lo be proud of,
?Sl1l l have a poslLlve aLLlLude Loward myself,
?Sl1C Cn Lhe whole, l am saLlsfled wlLh myself,
?Sl1P l wlsh l could llke myself more,
?Sl1l l cerLalnly feel useless aL Llmes,
?Sl1! AL Llmes l Lhlnk l am no good aL all.
5elf-efflcocy
ueflnes academlc self-efflcacy as Lhe sLudenL's compeLence and confldence ln performance of class work as percelved by
Lhe sLudenL. 1hls concepL should be dlsLlngulshed from global self-efflcacy or masLery whlch ls Lhe bellef LhaL one has con-
Lrol over one's own desLlny.
ke|ated uest|ons
?SA8k l am cerLaln l can undersLand Lhe mosL dlfflculL maLerlal presenLed ln LexLs,
?SA8L l am confldenL l can undersLand Lhe mosL complex maLerlal presenLed by Leacher,
?SA8M l am confldenL l can do an excellenL [ob on asslgnmenLs and LesLs,
?SA8n l am cerLaln l can masLer Lhe skllls belng LaughL
33 LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and
1lme Lo uegree CompleLlon
Se|f-percept|on (conLlnued)
5elf-mostety
1he powerlessness scale chosen by ?l1S ls based upon Lhe work of earlln and Schooler (1978). 1hls scale, referred Lo as Lhe
MasLery scale23, assesses a feellng of powerlessness wlLhouL reference Lo concreLe llfe slLuaLlons. MasLery can be deflned
as a measure LhaL assesses Lhe exLenL Lo whlch one regards one's llfechances as belng under one's own conLrol ln conLrasL
Lo belng faLallsLlcally ruled" (earlln and Schooler, 1978). Pence, lf one scores hlgh on Lhe masLery scale, one does noL feel
powerless.
ke|ated uest|ons
?Sl2A SomeLlmes l feel l'm belng pushed around ln llfe,
?Sl28 WhaL happens Lo me ln Lhe fuLure mosLly depends on me,
?Sl2C 1here ls really no way l can solve some of Lhe problems l have,
?Sl2u 1here ls llLLle l can do Lo change many of Lhe lmporLanL Lhlngs ln my llfe,
?Sl2L l ofLen feel helpless ln deallng wlLh Lhe problems of llfe,
?Sl2l l have llLLle conLrol over Lhe Lhlngs happen Lo me,
?Sl2C l can do [usL abouL anyLhlng l really seL my mlnd Lo.
Soc|a| support
Measures Lhe avallablllLy of soclal supporLs, vla frlends, famlly and oLher sources for Lhe youLh. 1hree aspecLs are lncluded:
rellable alllance (Lhe assurance LhaL oLhers can be counLed upon for pracLlcal help), aLLachmenL (emoLlonal closeness) and
guldance (advlce or lnformaLlon). 1hese aspecLs are mosL dlrecLly relaLed Lo problem-solvlng wlLhln Lhe conLexL of sLress.
1wo lLems were proposed Lo measure each of Lhese aspecLs for a LoLal of slx lLems.
ke|ated uest|ons
?Su1A lf someLhlng wenL wrong, no one would help me,
?Su18 l have famlly and frlends who help me feel safe, secure and happy,
?Su1C 1here ls someone l LrusL whom l would Lurn Lo for advlce lf l were havlng problems,
?Su1u 1here ls no one l feel comforLable Lalklng abouL problems wlLh,
?Su1L 1here ls no one l feel close Lo,
?Su1l 1here are people l can counL on ln Llmes of Lrouble
arents' behav|ours
arenLs who are supporLlve of Lhelr youLh's educaLlon, who are lnvolved ln Lhelr youLh's school and who have a flrm buL
responslve parenLlng sLyle have a poslLlve lnfluence on Lhelr youLh's achlevemenL and educaLlonal aLLalnmenL. 1he parenL-
lng pracLlces scales are deslgned Lo measure Lhree faceLs of parenLlng: nurLurance, lnconslsLenL re[ecLlon-orlenLed dlsclpllne
(re[ecLlon) and monlLorlng. An overall parenLlng scale was noL formed from Lhe Lhree subscales.
Mooltotloq bebovloot
Measures parenLs' monlLorlng behavlour. A monlLorlng parenL ls deflned as one who belleves LhaL he or she ls knowledge-
able abouL hls or her chlld's acLlvlLles, whereabouLs and frlends.
ke|ated uest|ons
817A know where chlld goes aL nlghL,
817u know whaL chlld ls dolng when he/she goes ouL,
817C know who chlld spends Llme wlLh when he/she goes ouL.
Nottotooce bebovloot
Measures parenLs' nurLurlng behavlours. nurLurance represenLs chlld-cenLered effecLlve parenLlng pracLlces such as nurLur-
ance, lnvolvemenL, and poslLlve relnforcemenL.
ke|ated uest|ons
817C ralse chlld,
817l LlsLen Lo chlld's ldeas and opLlons,
817! Make sure chlld knows LhaL Lhey are appreclaLed,
817M Speak of good Lhlngs Lhose chlldren does,
817C Seem proud of Lhe Lhlngs chlld does.
MLSA - Measurlng Lhe LffecLlveness of SLudenL Ald 36
arents' behav|ours (conLlnued)
locooslsteot Jlsclplloe (8e[ecLlon-orlenLed behavlour)
ke|ated uest|ons
8178 Soon forgeL a rule LhaL Lhey have made,
817L nag chlld abouL llLLle Lhlngs,
817P keep rules only when lL sulLs Lhemselves,
817l CeL angry and yell aL chlld,
817L 1hreaLen punlshmenL more ofLen Lhan uslng lL,
817n Lnforce or do noL enforce rules dependlng on Lhelr mood
Student's performance score |n read|ng
WelghLed llkellhood esLlmaLe ln readlng ablllLy, whlch ls provlded for all sLudenLs who answered aL leasL one readlng ques-
Llon. lL was Lransformed Lo a scale wlLh a mean of 300 and a sLandard devlaLlon of 100 by uslng Lhe daLa for Lhe parLlclpaL-
lng CLCu counLrles only (excepL Lhe neLherlands).
37 LffecLs of unlverslLy CharacLerlsLlcs and Academlc 8egulaLlons on SLudenL erslsLence, uegree CompleLlon, and
1lme Lo uegree CompleLlon
1ab|e A4: Summary SLaLlsLlcs, Scale varlables, Males and lemales
Ma|es Iema|es
Min. Max Mean S.D. Min. Max Mean S.D.
HS Engagement
Academic identification -4.57 4.35 -0.08 0.98 -5.04 3.97 0.12 0.96
Academic participation -4.84 2.83 -0.09 1.00 -4.84 2.85 0.28 0.90
Academic engagement -5.37 3.66 -0.10 0.99 -4.74 3.62 0.24 0.93
Social engagement -3.84 2.74 -0.13 1.03 -3.84 2.74 0.06 1.01
Overall engagement -5.55 3.22 -0.14 0.99 -4.14 3.60 0.17 0.96
Self-perception
Self-esteem -3.52 2.86 0.03 1.00 -3.67 2.04 -0.07 0.96
Self-efficacy -2.55 2.21 0.09 1.03 -2.55 2.31 -0.14 0.98
Self-mastery -3.71 2.92 0.00 0.99 -3.70 2.89 -0.07 0.95
Social Support -3.11 1.77 -0.26 0.99 -2.85 1.77 0.20 0.95
Parents' Behaviour
Monitoring behaviour -7.96 1.26 0.03 0.94 -7.96 1.26 0.11 0.84
Nurturance behaviour -5.08 1.78 -0.04 1.01 -5.47 1.78 0.09 0.98
Inconsistent discipline -3.93 3.54 0.11 0.96 -3.93 4.56 0.00 0.98
Reading Ability 166.01 887.31 519.29 97.04 120.56 909.52 550.01 88.54
MLSA - Measurlng Lhe LffecLlveness of SLudenL Ald 38

You might also like