replacing variable value via a loop conditioning on the contents of other variables

[Daniel Sutcliffe]

Hello,

I have dataset which has 14 variables relating ethnic group: eth1 eth2 eth3...eth14.

The values are probabilities between 0-1 and an individual observation in the dataset can have a value in more than one ethnicity , for example, 0.5 for eth1, 0.25 for eth10 and 0.25 for eth14, and all the other ethnic groups would be 0. This is an oversimplication - but gives an idea of how the data is arranged.,

Now for some observations, around 60%, I know the actual ethnicity, not the modelled probability, and I want to overwrite the model probabilities with a 1 where I know the ethnicity, and then set all the other ethnicities to 0. So in the above example, eth1 would be 1 and eth10 and eth14 and all the other ethnic groups would be 0. The individual, actual ethnicities are contained in a variable called Ethnic_Group.

Having created a copy of the 14 eth variables I have overwritten them as follows:

replace eth1= 1 if Ethnic_Group=="A"
replace eth1 = 1 if Ethnic_Group=="a"

having done this for each eth, from eth1 to eth14, I now want to write some code so as to set all the other eth variables to 0 if one of them is 1:

for each var of varlist eth* {

replace `var' = 0 if `var' <1

}

BUT, only doing the above, if one of the variables is indeed 1, if none of the eth variables are 1, then make no changes to them.

I am not sure how to amend my code to take into account, that I don't want to replace the values with 0, if none of the eth variables are 1.

[William Lisowski]

Very often a simple description really isn't clear without more detail, or at a minimum it is too difficult to guess at a good answer from what has been shared. When that happens other members will decide not to answer the question, or ask for an improved presentation that could have been provided to begin with.

Please help us help you. Show example data, at least one with mixed ethnicity and one with actual ethnicity. Explain what results you want for each of the observations you show.The Statalist FAQ provides advice on effectively posing your questions, posting data, and sharing Stata output.

[Tom Scott]

egen certaintyind = anymatch(eth*), v(1)

would equal 1 if any eth* variable had a value of 1 and 0 otherwise. I would think you could modify your code as

replace `var' = 0 if `var' <1 & certaintyind==1

[Daniel Sutcliffe]

egen certaintyind = anymatch(eth*), v(1)

would equal 1 if any eth* variable had a value of 1 and 0 otherwise. I would think you could modify your code as

replace `var' = 0 if `var' <1 & certaintyind==1

Thank you that works perfectly