Ghana - Public Expenditure Tracking Surveys 2007

The survey covered all education and health facilities; exit polling of patients attending health facilities as well as pupils/students attending basic school (primary and junior high school); regional offices of the education and health sectors; metropolitan, municipal and distict assemblies; central/national and regional nodes in the budgetary flows and release mechanisms for the selected resources (funds and materials) being tracked.

The PETS was based on a two-stage stratified nationally representative sample design. At the first stage of sampling, 48 districts (PSUs) were selected using systematic random sampling with probability proportional to size (PPS) method. Determination of the sampling rates by size strata used proportional allocation based on each region's share of the 2006/2007 number of public primary schools in the country.

At the second stage, for education facilities: 15 primary schools, 10 Junior High schools and one Senior High school were selected randomly for each district that had been selected earlier to produce a total of 720 primary schools, 480 JHS and 48 SHS. Additionally, were randomly selected for the entire country. It is worth noting that there are only 38 Teacher Training Colleges and 24 technical /vocational schools in Ghana. For health facilities: seven health facilities (government and mission - hospitals, clinics, CHPS etc), including the district hospital were randomly selected in each of the 48 selected districts. Also, all teaching hospitals and the 10 regional hospitals were included with certainty in the survey.


- Sample size: Primary sampling units (PSUs) were 48 districts.
Education Sector: secondary sampling units (SSUs) were, primary schools 720, Junior High Schools 480, Senior High Schools 48, Teacher training colleges 10, and Technical/vocational schools 7.
In aducation, Regional Education offices were 48, Distrct Education offices 48 and District assemblies 48.
Health Sector: SSUs were, CHIPS/Health centres & clinics/Mission hospital 336, District hospitals 48. Regional hospitals 10, Regional medical stores 10, Regional health offices 10, teaching hospitals 3.

- Selection process was probability proportional to size method.
- Stratification into homogeneous groups was explicit and carried out on the basis of adminstrative region and by type of facility.
- Stages of sample selection was based on a two-stage stratified nationally representative design.
- Level of representation: the survey was representative at the national and regional levels.
- A supplementary list was provide to cater for absent respondents or those not found.
- Sampling frame used: the list of schools from the 2006/07 EMIS data of the Ministry of Education was used as the sampling frame for the Education Sector, while the list of the 2006 health facilities of the Ministry of Health constituted the Health sector sampling frame. All the education (schools) and health facilities had unique identification codes, with district and region information. The primary sampling units (PSUs) were the administrative districts (metropolitan, municipal and district assemblies - MMDAs), which constitute well-defined basic administrative and geographical units for which education and health service facilities data are readily available. Schools and health facilities within the districts constituted the secondary sampling units (SSUs).


Response rates forthe PETS 2007 are: Regional Education Offices (90%); District Education Offices (100%); District Assemblies (100%); Primary schools (96.4%); JHS (100%); SHS (91.7%); Teacher training colleges (90%); Technical/vocational schools (85.7%);CHPS/Health centres & clinics/Mission hospitals (94.3%); Regional Health Offices (90%); Regional Medical Stores (100%); District Hospitals (75%); District Health Offices (100%); Regional Hospitals (90%); Teaching Hospitals (100%)