Sampling procedure
Samples were designed to represent E-firms and non-E-firms. E-firms were defined as businesses that used the internet during the last month before the interview date to conduct transactions, such as buying and/or selling of goods or services. Samples were designed to complete 250 interviews with E-firms and 500 interviews with non-E-firms.
Since no data were available about the proportion of E-firms in any of the study countries, a two-phase sampling design were used. Data from the first phase was used to guide the sampling procedures in the second phase to achieve the target sample size, especially for E-firms.
For Egypt: a systematic sample of 9,087 businesses were selected in Phase 1. To align the sample distribution across business sectors with the distribution in the frame distribution, the selection was done after sorting the frame according to the business sectors. The sample of Phase 1 composed of two samples: main sample of 5,066 businesses and a supplement sample of 4,021 businesses. Businesses with website available on Yellow Pages (YP) were oversampled to increase the chance of finding E-firms. This has been accounted for during the weight calculation to retrieve the actual distribution in YP frame. In Phase 2, a systematic sample of 18,901 businesses were selected from businesses with website available on Yellow Pages.
For Jordan: a stratified sample of 3,134 businesses were selected in Phase 1. The sample was stratified according to 17 business sectors. The sample was selected in two stages; in the first stage, 316 pages were selected from the Kinz website, where the selected pages were proportionally distributed to the distribution of all pages by business sector. In the second stage, all businesses in selected pages were contacted. In average, there are 10 businesses per page. In Phase 2, a stratified sample of 3,160 businesses were selected using the same approach used in Phase 1.
For Morocco: a stratified sample of 9,099 businesses were selected in Phase 1. The sample was stratified according to three main geographic regions of Morocco, Casa-Rabat, North, and South. The sample of Phase 1 composed of two samples: main sample of 4,388 businesses and a supplement sample of 4,711 businesses. In the two samples, 158 pages were selected for each sample, and 28-30 businesses were contacted from each page. In Phase 2, a stratified sample of 25,120 businesses were selected in two stages, where 314 pages were selected in the first stage, and about 80 businesses were contacted per each selected page.
Weighting
Inverse probability weighting was undertaken to account for the sampling strategy and non-response. Since non-E-firms were not eligible for interviews in Phase 2, the weight was post-stratified to retrieve the actual percentage distribution of E-firms and non-E-firms in the population. The weighted distribution from Phase 1 was used for post-stratification.
Note: there are more details on the weights and sampling in the “Digitalization in MENA _Sampling design and weighting” document in the documentation tab.