Creating an index using survey data

[Eddie1]

Hi,

I am new to R and want to create an index based on answers to 5 survey questions.

I want to then use the index as the dependent variable in some regressions later on.

Are these functions compatible in r or should i try something else?

Thanks

[andresrcs]

You can do all those things in R (and much more) for sure, but if you need more specific help, please provide a proper REPRoducible EXample (reprex) illustrating your issue.

[Eddie1]

Thanks andresrcs.

I am trying to recode answers to the survey for the index and I keep getting:

Error: Can't convert < character> to < double >.
Run rlang::last_error() to see where the error occurred.

The data i am trying to recode is

Labels:
value label
1 Very good
2 Good
3 So so (not good nor bad)
4 Bad
5 Very bad
7 Do not understand the question
8 Can't choose
9 Decline to answer

and the code i have tried to use is:
w3q1.1 <- recode(ABS3.1$q1, "1='5. Very good'; 2='4. Good'; 3='3. So so (not good or bad)'; 4='2. Bad'; 5='1. Very bad'; 7='7. Do not understand the question'; 8='8.Can't choose'; 9='9. Decline to answer'")

When i run:

rlang::last_error()

The console says this:
'Backtrace:

1. car::recode(ABS3.1$q1, "1='5. Very good'; 2='4. Good'; 3='3. So so (not good or bad)'; 4='2. Bad'; 5='1. Very bad'; 7='7. Do not understand the question'; 8='8.Can't choose'; 9='9. Decline to answer'")
2. vctrs:::[<-.vctrs_vctr(*tmp*, var == val, value = "9. Decline to answer")
3. vctrs::vec_cast(value, x)
4. haven:::vec_cast.haven_labelled.character(...)
5. haven:::vec_cast.haven_labelled.haven_labelled(x, to, ...)
6. vctrs::vec_cast(...)
7. vctrs::vec_default_cast(...)
8. vctrs::stop_incompatible_cast(...)
9. vctrs::stop_incompatible_type(...)
10. vctrs:::stop_incompatible(...)
11. vctrs:::stop_vctrs(...)'

Any advice?

[startz]

Do you have semicolons where there should be commas? Also, I think quotes are in the wrong place.

[Eddie1]

Removed the semicolons and it says:
Error in recode(ABS3.1$q1, "1='5. Very good', 2='4. Good', 3='3. So so (not good or bad)', 4='2. Bad', 5='1. Very bad'") :

in recode term: 1='5. Very good', 2='4. Good', 3='3. So so (not good or bad)', 4='2. Bad', 5='1. Very bad'
message: Error in parse(text = strsplit(term, "=")[[1]][2]) :
:1:15: unexpected ','
1: '5. Very good',
^

where should the quotation marks go?

[startz]

recode() is kind of confusing. It will probably be replaced one day. In the meantime, you don't want the double quotes. And the syntax is old value = new value, unlike most other functions.

You may want to look at the help for recode()

?recode

[Eddie1]

Thanks for the tip.

I think i solved the issue.

Had to change from character vector to a numeric.

Used this video if anyone else wants to know:

Data Type Conversion in R (2 Examples) | Convert Character & Factor to Numeric Class