frq() problem with NA value and weight with percent calculation

[caayala]

This issue could be a regression from #77.

df <- data.frame(val = c(1, 2, NA),
                 wgt = c(1, 1, 1))

sjmisc::frq(df, val)
#> 
#> val <numeric>
#> # total N=3  valid N=2  mean=1.50  sd=0.71
#> 
#>  val frq raw.prc valid.prc cum.prc
#>    1   1   33.33        50      50
#>    2   1   33.33        50     100
#>   NA   1   33.33        NA      NA
sjmisc::frq(df, val, weights = wgt)
#> 
#> val <numeric>
#> # total N=2  valid N=2  mean=1.50  sd=0.71
#> 
#>  val frq raw.prc valid.prc cum.prc
#>    1   1      50        50      50
#>    2   1      50        50     100
#>   NA   0       0        NA      NA

^{Created on 2020-01-07 by the reprex package (v0.3.0)}

I expect that both frequencies tables are equal.

• sjmisc [2.8.2]

[frankcsliu]

I have the same question: I was expecting that the number of NA in the output of using weights should be remained consistent with that before using weights. However, the manual help page says that "the weight will be applied to weight all observations," which implies those labelled with NA maybe moved to other categories?

My replicable dataset is here: https://github.com/frankcsliu/R4surveyresearch/blob/master/tscs2013.rda
I was comparing these two results and felt confuesed.
frq(tscs2013$v65r)
frq(tscs2013$v65r, weights = tscs2013$wr)

Thank you.

[strengejacke]

I'm not sure how to consistently do this. When there's just one NA, the weighting would work. But multiple NA values, possibly with missing weights as well, no longer works. How to best treat those cases?

df <- data.frame(val = c(1, 2, NA, 5),
                 wgt = c(.7, 1.2, 1.1, NA))

df
#>   val wgt
#> 1   1 0.7
#> 2   2 1.2
#> 3  NA 1.1
#> 4   5  NA
xtabs(wgt~val, data = df, addNA = TRUE)
#> val
#>    1    2    5 <NA> 
#>  0.7  1.2       1.1


df <- data.frame(val = c(1, 2, NA, 5, NA),
                 wgt = c(.7, 1.2, 1.1, NA, NA))

df
#>   val wgt
#> 1   1 0.7
#> 2   2 1.2
#> 3  NA 1.1
#> 4   5  NA
#> 5  NA  NA
xtabs(wgt~val, data = df, addNA = TRUE)
#> val
#>    1    2    5 <NA> 
#>  0.7  1.2

^{Created on 2021-11-27 by the reprex package (v2.0.1)}

[strengejacke]

Maybe only consider complete cases.