The simple ransoms technique is a probability technique.so you simply randomy select any say 500 samples from your population.You could also go for systematic probability technique if you have a list of consumers you are surveying.Here we could select every second or third sample so that the bias in Selection is removed

First, you need to avoid bias, such as causing certain landowners to respond (or not) simply due to the nature of the questions.

Then you determine sample size using the 95% confidence level Z score (1.96), estimated standard deviation (usually 0.5) and margin of error (usually 5%), which corresponds to 385 respondents using the formula:

Necessary Sample Size = (Z-score)^2 * StdDev*(1-StdDev) / (margin of error)^2

If you want greater power, then you can increase the sample size.

It's very true what McKenzie says.

I used tidyverse package in R to sample required number and let R do random sampling instead of worrying how I should do. This gives one subset out of several subsets using combination. R Code provided below writes resulting data into .csv file.:

write.csv(

# # sample_n (

# # tbl = select(

# # read_csv(

# # "/file.csv"

# # ),

# # 'Var 1',

# # 'Var 2',

# # 'Var 3',

# # 'Var 4',

# # ),

# # size = ,

# # replace = FALSE,

# # weight = NULL,

# # .env = NULL,

# # ... = NULL

# # ),

# # file = "output.csv"

# # )

Please remove # in each line before using it. Thank you all for your input!!! They are helpful to understand theoretical aspects.