Ghana - National Industrial Census 2003

A one-stage sample design was used for the survey with the primary sampling units (PSU) being the individual establishment. The sampling frame of establishments had two levels of stratification in addition to implicit stratification from ordering the establishments within each stratum. The first level of stratification was by the 4-digit ISIC activities, which represented the smallest domains of analysis. Within each of these individual activities, the establishments were further stratified by the number of persons engaged, which is also correlated with industrial production, revenue, expenditures and other aggregates to be measured in the survey.


All of the establishments in the Construction, Electricity and Water, Mining and Quarrying sub-sectors were included in the survey regardless of size. Given that reliable estimates are required for each 4-digit ISIC group, activities with few establishments were also identified as certainty strata.


Establishments are classified into mining and quarrying, manufacturing, electricity and production and construction sectors. Under manufacturing we considered establishments which engaged less than 10 persons and that which engaged 10 or more persons.

The frame for establishments engaging 1 - 9 persons is 22,404
The frame for establishments engaging 10+ persons is 4,230

The number of establishments sampled:
1 - 9 =1,263
10+ =4,233


Number of establishment completed:
1 - 9 = 895
10+ = 2,918

Number of establishment out of scope + Association
1 - 9 = 53
10+ = 765

Number of establishment not completed
1 - 9 = 304
10+ = 484

Number of establishment closed down
1 - 9 = 43
10+ = 81

Response rate for :
1 - 9 = 0.77
10+ = 0.86

Response rate is equal to the ratio of number of completed to the adjusted number of not completed establishment

Adjusted number of establishments excludes out of scopes and association, groups and cooperative, closed down and not located establishments

* the difference in the sample and the frame for 10+ persons engaged establishment was due to the migration of 1-9 persons engaged establishment into the 10+ category


Sample weight were calculated for the manufacturing and mining data files. The sample weight variable is called weights. The sample weight was computed as the inverse of the probability of selection. The sample weight were adjusted to take into account the non-interview (non response) within each stratum including the certainty strata