Applying survey weights using survey package in R
12:50 28 Jul 2023

I need some guidance in using survey weights in RStudio using the survey package.

table_1015 consists of survey data from the years 2010 to 2015 where each row is one response from a participant including information such as age, race, sex, and education level/grade (all are categorical variables). It also includes responses to questions in the survey labeled as q1, q2, q3, etc. where 1 = Yes and 0 = No and 'Missing' = NA/No Response.

I used the below code to calculate the unweighted frequencies and proportion by year and age.

age_1015 <- table_1015 %>%
  group_by(year, age) %>%
  summarise(count = n()) %>%
  mutate(proportion = count / sum(count))

Example of part of the output for age_1015:

year age count proportion
2010 12yo 5 0.185
2010 13yo 8 0.296
2010 14yo 14 0.519
2011 12yo 3 0.115
2011 13yo 7 0.269
2011 14yo 16 0.615
2012 12yo 2 0.118
2012 13yo 4 0.235
2012 14yo 11 0.647

How would I go about applying the following survey design object to the above? I want to calculate the weighted frequency (weighted_count) and the weighted proportion (weighted_prop).

srvy_des_1015 <- <- svydesign(ids = ~PSU, data = table_1015, weights = ~weight, strata = ~stratum, nest = TRUE)
r frequency survey weighted