There is a multiway crosstable saved as a frequency table like this:
df = data.frame(
Correct = c('yes','no','yes','no','yes','no','yes','no',
'yes','no','yes','no','yes','no','yes','no'),
Type = c('t1','t1','t2','t2','t1','t1','t2','t2',
't1','t1','t2','t2','t1','t1','t2','t2'),
Subtype = c('st1','st1','st1','st1','st2','st2','st2','st2',
'st1','st1','st1','st1','st2','st2','st2','st2'),
Level = c('a','a','a','a','a','a','a','a',
'b','b','b','b','b','b','b','b'),
Freq = c(115,99,140,81,104,100,156,52,61,160,59,164,41,160,48,159)
)
How can one recalculate Freq into proportions conditioned on all variables but `Correct', resulting in this new variable:
df$Prop = c(0.54,0.46,0.63,0.37,0.51,0.49,0.75,0.25,
0.28,0.72,0.26,0.74,0.20,0.80,0.23,0.77)
df
# Correct Type Subtype Level Freq Prop
# 1 yes t1 st1 a 115 0.54
# 2 no t1 st1 a 99 0.46
# 3 yes t2 st1 a 140 0.63
# 4 no t2 st1 a 81 0.37
# 5 yes t1 st2 a 104 0.51
# 6 no t1 st2 a 100 0.49
# 7 yes t2 st2 a 156 0.75
# 8 no t2 st2 a 52 0.25
# 9 yes t1 st1 b 61 0.28
# 10 no t1 st1 b 160 0.72
# 11 yes t2 st1 b 59 0.26
# 12 no t2 st1 b 164 0.74
# 13 yes t1 st2 b 41 0.20
# 14 no t1 st2 b 160 0.80
# 15 yes t2 st2 b 48 0.23
# 16 no t2 st2 b 159 0.77
In base R:
df$Prop <- with(df, ave(x = Freq, Type, Level, Subtype, FUN = \(x) x / sum(x)))
# or
df$Prop <- with(df, Freq / ave(x = Freq, Type, Level, Subtype, FUN = sum))
# Correct Type Subtype Level Freq Prop
# 1 yes t1 st1 a 115 0.5373832
# 2 no t1 st1 a 99 0.4626168
# 3 yes t2 st1 a 140 0.6334842
# 4 no t2 st1 a 81 0.3665158
# 5 yes t1 st2 a 104 0.5098039
# 6 no t1 st2 a 100 0.4901961
# 7 yes t2 st2 a 156 0.7500000
# 8 no t2 st2 a 52 0.2500000
# 9 yes t1 st1 b 61 0.2760181
# 10 no t1 st1 b 160 0.7239819
# 11 yes t2 st1 b 59 0.2645740
# 12 no t2 st1 b 164 0.7354260
# 13 yes t1 st2 b 41 0.2039801
# 14 no t1 st2 b 160 0.7960199
# 15 yes t2 st2 b 48 0.2318841
# 16 no t2 st2 b 159 0.7681159