20.10.2014 Views

The Role and Impact of Public-Private Partnerships in Education

The Role and Impact of Public-Private Partnerships in Education

The Role and Impact of Public-Private Partnerships in Education

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

What Do We Know About <strong>Public</strong>-<strong>Private</strong> <strong>Partnerships</strong> <strong>in</strong> <strong>Education</strong>? 39<br />

Table 3.4<br />

Studies <strong>of</strong> private management<br />

Empirical<br />

strategy Country <strong>and</strong> study Data: type <strong>and</strong> year Outcome variables Results<br />

R<strong>and</strong>omization<br />

Difference<br />

<strong>in</strong> difference<br />

Propensity<br />

<strong>and</strong> match<strong>in</strong>g<br />

United States,<br />

Charter schools,<br />

Chicago (Hoxby<br />

<strong>and</strong> Rock<strong>of</strong>f 2004)<br />

United States,<br />

Charter schools,<br />

New York (Hoxby<br />

<strong>and</strong> Murarka 2007)<br />

United States,<br />

Charter schools,<br />

Texas (Booker et al.<br />

2008)<br />

United States,<br />

Charter schools,<br />

Michigan (Bett<strong>in</strong>ger<br />

2005)<br />

United States,<br />

Charter schools,<br />

North Carol<strong>in</strong>a<br />

(Bifulco <strong>and</strong> Ladd<br />

2006)<br />

United States,<br />

Charter schools,<br />

Texas (Hanushek<br />

et al. 2007)<br />

Colombia,<br />

Concession schools<br />

(Barrera-Osorio<br />

2007)<br />

Venezuela, Fe y<br />

Alegría (Allcott <strong>and</strong><br />

Ortega 2007)<br />

Sources: Authors’ compilations; Gill et al. 2007.<br />

Adm<strong>in</strong>istrative data,<br />

cross-section, 2000–<br />

2002, student level<br />

Adm<strong>in</strong>istrative data,<br />

cross-section, 2000–<br />

2005, student level<br />

Adm<strong>in</strong>istrative data,<br />

longitud<strong>in</strong>al 1995–<br />

2002, student level<br />

Adm<strong>in</strong>istrative data,<br />

panel, school level<br />

Longitud<strong>in</strong>al data,<br />

1996-2002<br />

Adm<strong>in</strong>istrative data,<br />

longitud<strong>in</strong>al panel,<br />

1996–2002, student<br />

level<br />

Panel at school level,<br />

two years, 1999–2003,<br />

student level<br />

Cross-section, 2003,<br />

student <strong>and</strong> school<br />

level<br />

St<strong>and</strong>ardized test<br />

(math, read<strong>in</strong>g),<br />

1st–8th grades<br />

St<strong>and</strong>ardized test<br />

(math, read<strong>in</strong>g),<br />

1st–12th grades<br />

St<strong>and</strong>ardized test<br />

(math, read<strong>in</strong>g),<br />

3rd–8th <strong>and</strong> 10th<br />

grades<br />

Test scores<br />

St<strong>and</strong>ardized test<br />

scores (math,<br />

read<strong>in</strong>g), 4th–8th<br />

grades<br />

St<strong>and</strong>ardized test<br />

scores (math,<br />

read<strong>in</strong>g), 4th–7th<br />

grades<br />

Dropout rates,<br />

st<strong>and</strong>ardized test<br />

scores (math,<br />

read<strong>in</strong>g), 11th grade<br />

St<strong>and</strong>ardized test<br />

scores (math,<br />

read<strong>in</strong>g), 11th grade<br />

Positive effects for lower grades: read<strong>in</strong>g, 11 percentile po<strong>in</strong>ts<br />

<strong>and</strong> math, 10 percentile po<strong>in</strong>ts. None or a negative impact (–4<br />

percentile po<strong>in</strong>ts <strong>in</strong> math) for higher grades (6–8).<br />

Positive effects on math (0.09 st<strong>and</strong>ard deviation), read<strong>in</strong>g<br />

(0.04 st<strong>and</strong>ard deviation), 3rd–8th grades, achievement<br />

positively correlated with the length <strong>of</strong> academic year.<br />

Initially (one year), students perform worse <strong>in</strong> charter schools<br />

than <strong>in</strong> public ones; after three years, students have similar<br />

scores to those <strong>in</strong> public schools.<br />

No statistical differences between (nearby) public schools<br />

<strong>and</strong> charters.<br />

Charter students score 0.1 (read<strong>in</strong>g) <strong>and</strong> 0.16 (math) st<strong>and</strong>ard<br />

deviation lower than public students.<br />

Initially, charter students score lower than regular public<br />

students (0.17 st<strong>and</strong>ard deviation), but this depends on how<br />

long the school has been operat<strong>in</strong>g; after three years <strong>of</strong><br />

operation there are no differences between them.<br />

Positive effects on math (0.19 st<strong>and</strong>ard deviation), read<strong>in</strong>g<br />

(0.27 st<strong>and</strong>ard deviation).<br />

Positive effects on dropout rates (1.7 percentage po<strong>in</strong>ts).<br />

Some evidence <strong>of</strong> completion effects on nearby public<br />

schools.<br />

Positive effect on math scores (0.08 st<strong>and</strong>ard deviation),<br />

verbal (0.1 st<strong>and</strong>ard deviation).<br />

for Colombia <strong>and</strong> Venezuela. <strong>The</strong>y found<br />

that the private management <strong>of</strong> schools<br />

yielded higher test scores for students at<br />

the end <strong>of</strong> their basic education (grade 11)<br />

than public <strong>in</strong>stitutions did. <strong>The</strong>se two<br />

studies used propensity score match<strong>in</strong>g<br />

estimators with limited data <strong>and</strong>, therefore,<br />

their results should be viewed with<br />

caution.<br />

Subsidies<br />

<strong>The</strong>re is only limited robust empirical<br />

evidence on subsidies (table 3.5). Kim,<br />

Alderman, <strong>and</strong> Orazem (1999) evaluated a<br />

subsidy program <strong>in</strong> Pakistan’s Balochistan<br />

prov<strong>in</strong>ce. <strong>The</strong> budget allocation for the<br />

program was smaller than the resources<br />

needed to cover all the target population.<br />

For this reason, the authorities decide to<br />

run a lottery to decide who should benefit.<br />

Kim, Alderman, <strong>and</strong> Orazem (1999) found<br />

that the program had had a positive impact<br />

on girls’ enrollment rates. However, because<br />

the basel<strong>in</strong>e treatment <strong>and</strong> control groups<br />

had important differences, it is unclear<br />

whether the differences <strong>in</strong> enrollment<br />

between beneficiaries <strong>and</strong> nonbeneficiaries<br />

can be attributed to the program or to other<br />

differences <strong>in</strong> their characteristics.<br />

Uribe et al. (2006) <strong>in</strong>vestigated the differences<br />

between public <strong>and</strong> private schools<br />

Delivered that by receive <strong>The</strong> World subsidies, Bank e-library especially to: those<br />

unknown<br />

concerned IP : 192.86.100.35 with the use <strong>of</strong> school <strong>in</strong>puts.<br />

<strong>The</strong>y Mon, reached 30 Mar 2009 several 12:16:23 conclusions. First,<br />

after controll<strong>in</strong>g for such factors as family<br />

background <strong>and</strong> teacher characteristics,<br />

(c) <strong>The</strong> International Bank for Reconstruction <strong>and</strong> Development / <strong>The</strong> World Bank

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!