Uganda MIS 2018-19 urban/rural sample size discrepancies [message #24809] |
Thu, 14 July 2022 06:25 ![Go to next message Go to previous message](/theme/default/images/down.png) |
andrew.M4M
Messages: 5 Registered: June 2022
|
Member |
|
|
Hello. There appear to be discrepancies in the urban/rural sample sizes presented in the Uganda MIS 2018-19 household recode and its Excel PCA results files. The discrepancies seem to stem from two separate issues - (1) an unexplained reclassification of some rural dwellers as urban and (2) the addition of refugee households as their own distinct residence classification. Within the household dataset you can see that there are three residence categories (hv025): urban, rural, and refugee.
From household recode dataset hv025
Urban - 2,149
Rural - 6,202
Refugee - 606
Total - 8,957
Confusingly, the overall sample size on the common tab of the PCA Excel document does correspond to the total dataset sample size (n=8957), but its urban (n=2390) and rural (n=5961) samples are different from those described in the data and also do not equal the total sample. In fact, the difference between the total sample and the combined urban and rural samples is 606, the refugee residence category. Of course, these discrepancies mean that the descriptive statistics provided in the urban and rural tabs of the Excel document do not match those found in the data.
I attempted to review the SPSS syntax normally posted with the Excel file, but it won't load. I also looked back at the 2016 Uganda DHS to see how refugee populations may have been previously handled, but neither the dataset nor the report for that survey refer to refugees. If refugee populations were included in that survey they weren't differentiated from the overall Ugandan population. It is only in this MIS that they do so. Can you provide guidance on how to handle these discrepancies? Thank you.
|
|
|