No announcement yet.
  • Filter
  • Time
  • Show
Clear All
new posts

  • Survey data analysis

    I am new to STATA (15.1) and hope someone can help me with how to "Declare my survey design for my data set". I am using the results of very old surveys (tuberculin skin test surveys) done between 1988 and 2000 in a certain country. The data was collected from in totaal 6 of the 60 provinces in the country, whereby 2 provinces from each region (North, central and south) were (randomly?) chosen. In each province, schools were randomly selected from ALL districts within the chosen provinces such that the number of children eventually included from each district was proportional to the district population (at the time). My data has district names and school names but I have no idea how many schools there were in each district at the time, nor the population numbers per district. When coding "svyset" I have no idea what to fill in for fpc or sampling weight. Do I have to fill them in? Can I just fill in the primary sampling units: region, province, district, school? Or is that totally incorrect and will the clustering at school level not be taken into account anymore? I hope someone can help: I have spent hours reading and trying to understand manuals, videos etc on this topic but I am unsure about what I am doing and what I might be doing wrong.