Nepal - Survey of Small Manufacturing Establishments 2008-2009

The sampling design of SSME 2008-2009 comprised of complete as well as sample enumeration. Actually, there were three modes of selecting samples and the probability of selection of a sample is unequal. The sampling frame for the design was obtained by conducting one year listing operation throughout the country. In that operation, there were 32326 small industries found active. Out of them, 3737 had been selected for the survey.

At first, the sampling frame was sorted and split by NSIC and development regions. Then, a cut-off point (<5 establishments) was determined and in turn, all establishments under a NSIC with frequency less than 5 in a development region are selected. The total number of such establishments was 152. In fact, they were chosen for complete enumeration. This was done to achieve more reliable and better representative sampling distribution. The probability of selection of these establishments was considered one.

Secondly, the remaining establishments of the sampling frame had been further divided into two groups. The first group contained all establishments within NSIC 1531, 1810, 2811, 3610 and 3691; and the next group belonged to remaining NSICs. The first group again classified into 3 strata by number of persons engaged. There were altogether 24837 establishments within this group. Out of them, 2485 simple random samples were selected with probability 0.10 proportional among the strata.

Thirdly, the next group had 7337 establishments. It was also divided into 3 strata by number of persons engaged. Among them, 1100 simple random samples had been selected with probability 0.15 proportionally.

Finally, about 12 percent of total industries were selected. Out of total about 45 percent belong to Central Development region. Eastern and Western Development Region contains nearly 27 percent and 16 percent samples respectively.


The sample design of the SSME 2008-2009 mentioned a sample of 3737 establishments and the estimated data of the survey was also based on all of these 3737 industries. Hence the response rate was 100 percent. It means, the survey questionnaire was administered to each and every sample successfully.


There were unequal probabilities of selection of samples in the sample design of SSME 2008-2009. In fact, there were three modes of selection of samples. In terms of sampling units, they were units selected under the cut-off point (less than 5 establishments within a development region), Sampling Units with Selection Probability 0.10, and Sampling Units with Selection Probability 0.15.

All establishments selected below the cut-off point were given sampling weight of 1, and those selected by simple random sampling were given weights equal to the inverse of their probability of selection. Hence, the weight of a sample establishment having selection probability 0.10 had become either 10 or 11 according to the size corresponding strata. Likewise, the weight of the sampled establishment having selection probability 0.15 varied from 4 to 9. In this way, by using SPSS Software, the weight of each sample establishment had been obtained at the time of sample selection; and aforementioned three types of weights were merged into a consolidated sample weight file and sorted them by "Form No"; the unique identification number given for each establishment. Thus, to estimate any variable under study, to the corresponding weight by "Form No" should be multiplied with it.