CODEWITHSUNDEEP
rquantile
user24610
user24610

Reputation:

Using deciles to define strata

I have a survey data set and want to use the deciles to define strata, U1 = (all units in data set is between D0=min and D1), U2=(between D1 and D2).....U10=(between D9 to D10=max).

How do I use the deciles to define strata?

Upvotes: 1

Views: 397

Answers (1)

Glen_b
Glen_b

Reputation: 8252

An example of using the quantile function to compute deciles of one variable, and then the cut function to compute a factor based on those deciles, and then using that factor in other calculations, via tapply:

# Let's set up some data:

 y <- rnorm(30, 100, 20)
 x <- rpois(30, 25-y/20)  # make x depend on y a little
 surveyres <- data.frame(y=y,x=x)

# set up the deciles of one variable

 yd <- cut(y, breaks=c(-Inf,quantile(y,seq(0.1,0.9,by=0.1)),Inf) )

# compute means of another variable over deciles of the first:

 tapply(surveyres$x, yd, mean)      
(-Inf,84.2] (84.2,88.8] (88.8,93.8] (93.8,97.5]  (97.5,100]   (100,104] 
   23.66667    28.00000    22.33333    20.00000    20.33333    17.33333 
  (104,110]   (110,114]   (114,123]  (123, Inf] 
   20.66667    19.33333    21.00000    20.33333

See also the by function which should work with a variable like yd.

Upvotes: 4

Related Questions