Using Categorical (Nominal) independent variables in Linear Regression?

[convivial1]

Hi,

can I use categorical (nominal) variables in regression as independent variables?

I have read that upto two levels of categorical variables (e.g. yes or no) we can use but in case of more than two levels, one has to do some changes. I am not sure if SPSS does those changes automatically?

[Link]

Usually, when there are more than two levels, dummy indicator variables are created and added to the model as separate parameters.

[convivial1]

yes that's what I have read but I am not sure that even in case of SPSS I do need to do this? I tried without doing anything and still SPSS gave me the results. Now, I am not sure if SPSS did wrong or not.

[fed1]

I was working with someone recently who said they had to do the dummy coding by hand in SPSS and was very concerned with doing it by hand in SAS.

Is it true that SPSS makes you code your own design matrix. I havent used it in a while, but it seems to me this is probably not the case. Why would someone make stats software where you had to "do it yourself"?

My bet is that SPSS takes carry of all coding issues internally, and you should let it do so.

[convivial1]

I was thinkign the same that SPSS does it automatically. But I am still not sure. Now how come will I come to know whether SPSS does it automatically or not?