Create dummy variables in SPSS

Create dummy variables from one categorical variable in SPSS. This technique is used in preparation for multiple linear regression when you have a categorical variable with more than two groups which you want to use as an independent variable.

For example, say I wanted to use a participants Ethnicity and there were 5 different ethnic groups. This variable is nominal and has 5 groups so I cannot use it as it is for linear regression. I would need to create 4 dummy variables.

I do not explain the theory of creating dummy variables in this video. For that, please see a good stats textbook.

ASK SPSS Tutorial Series

[khepron]

It is so refreshing when an explanation is concise. All I wanted to know was "How to do it." The example was exactly what I wanted to do. Thank you. This was perfect.

[WhenMoonWhenLambo]

I can't believe I spent 1 hour reading how to do this and I learned it in 3 min here..

[dylanlaugharne]

Project due tomorrow. This has completely saved me. God bless you.

[SinaiShur]

Clear presentation, thanks. Reminds me that the rule is k-1; so we only need 4 dummy variables for those 5 response items. I heard one presentation noted that we can use all k-dummies but not use the intercept of the model.

[KongDenisMD]

This is life saving. Thanks so much. Using this for logistic regression for my MD thesis

[nazzy_knows]

Thank you so much for this. I spent probably a good five hours trying to find information and your video was the clearest

[pablolundie]

Thank you for this video, it was easy to understand and useful for my class. I have been having difficulties understanding the teaching method in class and I cannot rely on that. I am lucky to found your video.

[SquareWarren]

That's a good way, especially if you need to collapse categories, but I'm about to save you hours of work. At least for SPSS version 20 and above, in the top menu...
(1) Click "Transform > Create Dummy Variables" (underneath "Automatic Recode", also handy).
(2) Select the variable(s) and move them/select them into the left top box.
(3) Then add in a name you want as the variable prefix (e.g., "Race_ethnicity") in the box titled "Root Names (One Per Selected Variable):".
(4) Click ok.
As long as you leave "Dummy Variable Labels" as "Use value labels", SPSS will automatically assign the each new variable the value name from your previous labels/coding... for example, if you use that procedure on the example above, you'll have a new variable "Race_ethnicity_Asian", a new variable "Race_ethnicity_White", etc... whatever your original value labels were. Welcome to making categorical data into dichotomous variables in 6 seconds.
Hope this helps!

[romyt]

Are you saving my life? You are saving my life. THANK YOU

[augustinahickinson]

This was quick and simple to understand. Thank you very much!!!

[marieleveille]

Thanks for saving our university semester

[roxannemax]

Thank you! So simple, and effective explanation 👌

[Zasi]

Thank you so much! Informative and helpful, made my day!

[rohansully]

an explanation that clearly shows how to have more than one recoded value!
SPSS has a terrible non intuitive demand that one must hit 'reset' before adding another recoded variable. terrible logic!

[pepersajou]

Thank you, BrunelASK. Do you also have a tutorial for conducting the multiple regression with this dummy variables?

[joannemamo]

Thank you for your clear explanation!!

[DrKMSmithPhD]

This was great. Thank you!! Thank you!! Thank you!!

[tubabuday]

What if my categorical variable also has its own "other" value label. How do I dummy code it?

[huigan]

Hi, thank you for your sharing on the issue of dummy variables, it was very fantastic! However, I have a doubt where some sources indicated that dummy variables should be created based on the formula c-1. This means that 4 dummy variables are created if there are 5 categories. Can I have your advice on this issue? Thank you in advance 😊

[amyrobinson]

What about doing this process the other way? I have separate already and want to combine into two larger variables rather than separating out into smaller ones. Have been instructed to create dummy categorical variable but when I do this, there is no data in the data