Ghana - Transport Indicator Database Survey 2012, Second Round

The sample was representative of all households in Ghana. To achieve the study objectives, the sample size chosen was based on the type of variables under consideration, the required precision of the survey estimates and available resources.
Taking all of these into consideration, a sample size of 6,000 households was deemed sufficient to achieve the survey objectives. This was enough to yield reliable estimates of all the important survey variables as well as being manageable to control and minimize non-sampling errors.

Stratification and Sample Selection Procedures
The total list of the Enumeration Areas (EAs) from the demarcation for the 2010 Population and Housing Census formed the sampling frame for the Phase II of the Transport Indicators Survey.
The sampling frame was stratified into urban/rural residence and the 10 administrative regions of the country for the selection of the sample. The sample was selected in two stages.

The first stage selection involved the systematic selection of 400 EAs with probability proportional to size, the measure of size being the number of households in each EA.
The second stage selection involved the systematic selection of 15 households from each EA. See Appendix A for more details on the sample design.


No deviations


National:
(5996/6000)*100=99.93%

By Regions:
Western=99.8%
Central= 100.0%
Greater Accra= 100.0%
Volta = 99.5%
Eastern=100.0%
Ashanti = 100.0%
Brong Ahafo = 100.0%
Northern = 100.0%
Upper East = 100.0%
Upper West = 100.0%

Region Hhs completed Hhs Expected Response rate
Western 569 570 99.8
Central 510 510 100.0
Greater Accra 855 855 100.0
Volta 567 570 99.5
Eastern 705 705 100.0
Ashanti 1,125 1,125 100.0
Brong Ahafo 585 585 100.0
Northern 615 615 100.0
Upper East 285 285 100.0
Upper West 180 180 100.0
Total 5,996 6,000 99.9



Causes of non response
Region
Result of Interview Western Volta Total
Refused 1 0 1
No HHold Member at Home 0 2 2
Other 0 1 1
Total 1 3 4


The weight variable used in the data sewt was weight