Can the expss package process multiple answer questions with weighted data?

Question

Can the expss package solve the question asked in this link?, it's about multiple response question with a weighting variable in the dataset

How to use the R survey package to analyze multiple response questions in a weighted sample?

Let's suppose we have this dataset:

demo <- tribble(
~dummy1, ~dummy2, ~dummy3, ~survey_weight,
      1,       0,       0,          1.5,
      1,       1,       0,          1.5,
      1,       1,       1,           .5,
      0,       1,       1,          1.5,
      1,       1,       1,           .5,
      0,       0,       1,           .5,
)

I need to calculate the percentage based on total respondents who answered the question, and not on total responses

score 3 · Accepted Answer · answered Dec 14 '19 at 10:50

Yes, it is quite easy:

library(expss)
demo = text_to_columns("
    dummy1 dummy2 dummy3 survey_weight
    1        0        0           1.5
    1        1        0           1.5
    1        1        1            .5
    0        1        1           1.5
    1        1        1            .5
    0        0        1            .5
")


demo %>% 
    tab_cells(mdset(dummy1 %to% dummy3)) %>%  # 'mdset' designate that with have multiple dichotomy set
    tab_weight(survey_weight) %>% # weight
    tab_stat_cpct() %>% # statistic
    tab_pivot() 

# |              | #Total |
# | ------------ | ------ |
# |       dummy1 |   66.7 |
# |       dummy2 |   66.7 |
# |       dummy3 |   50.0 |
# | #Total cases |    6.0 |

# shorter notation with the same result
calc_cro_cpct(demo, mdset(dummy1 %to% dummy3), weight = survey_weight)

But note that expss use simple frequency weights in the SPSS style while 'survey' package can provide more accurate weighting schemes.

Can the expss package process multiple answer questions with weighted data?

1 Answers1