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.
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!
Thank you so much on behalf of the first year of public administration at the Erasmus university. But for which kind off regressions can you use this. I've noticed that when you have to create only one other 'dummy variable' it gives the wrong output.
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.
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.
Finally!!!!! 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!
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 view.....
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 😊
Thank you for creating a great video! I have run into an issue when trying to replicate it. I create my variable, but it does not recognize the data. For example, I created a variable "Asian" and set up values of 1 for "Asian" and 0 for "others", but my data under the new variable heading "Asian" are zeroes when I should have two 1s. Any thoughts?
This is a great video. Please help, do I need to create dummy variables for all 16 items in my questionnaire where participants either chose a narcissistic response or non narcissistic response?
+ken hz Not sure if you still need advice, but it depends on how you asked the question. If the question is "yes/no" or something like that - basically, if there are only two possible answers per question - then they are already dichotomous and you are good to go.
But what if some of the responses for the categorical variables are "don't know" or "NA" or "IAP", do we still count those as responses and recode them with the k-1 rule? or is there a way to eliminate those responses completely for the multiple regression if I'm only interested in looking at those who responded with the necessary categories?
Thank you for this great video:) I am wondering if it should be 4 dummy variables not 5(since there were 5 choices in the categorical variable) , does it mean that we don't need to make "others" as the standard value number 1?
Is there a way to take care of missing/refuse responses during this process? If not, does it matter if I take care of it before or after? I am transitioning from SAS and having some difficulty.
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.
I can't believe I spent 1 hour reading how to do this and I learned it in 3 min here..
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!
I love you good sir. Following this: Analyse -> Regression -> Linear (put all dummies in block 1) - SPSS automatically uses 1 dummy as reference.
Thank you SO much!!! Your comment has saved me so much time and hassle!
Hats off to you sir... I dreaded thinking I had to do this 46 times for my 46 dummies. This helped a lot. Thank you!
Thank you so much on behalf of the first year of public administration at the Erasmus university. But for which kind off regressions can you use this. I've noticed that when you have to create only one other 'dummy variable' it gives the wrong output.
OH MY GOD THANK YOU
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.
Project due tomorrow. This has completely saved me. God bless you.
This is life saving. Thanks so much. Using this for logistic regression for my MD thesis
Thank you so much for this. I spent probably a good five hours trying to find information and your video was the clearest
Thank you, BrunelASK. Do you also have a tutorial for conducting the multiple regression with this dummy variables?
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.
Are you saving my life? You are saving my life. THANK YOU
Thanks for saving our university semester !!!!!
This was quick and simple to understand. Thank you very much!!!
What if my categorical variable also has its own "other" value label. How do I dummy code it?
Thank you so much! Informative and helpful, made my day!
Thank you for your clear explanation!!
Thank you! So simple, and effective explanation 👌
Finally!!!!! 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!
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 view.....
Awesome! Thank you for your help. :)
Great tutorial! Very helpful!
Super helpful, thanks a lot!
Thank you BrunelASK - how would you do this with gender i.e. male, female and prefer not to say?
This was great. Thank you!! Thank you!! Thank you!!
Thanks a lot for the help!
do we do the same way for ordinal variables?
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 😊
THANK YOU SO MUCH for this video =)
Thank you! Very helpful!
Thank you for creating a great video! I have run into an issue when trying to replicate it. I create my variable, but it does not recognize the data. For example, I created a variable "Asian" and set up values of 1 for "Asian" and 0 for "others", but my data under the new variable heading "Asian" are zeroes when I should have two 1s. Any thoughts?
I transformed my variables from string to numeric type variables using autorecode, and this resolved the problem ;).
Omg thank you 💜💜💜💜
Thank you 😊
After creating a dummy variables then how we can correlation dependent variable and dummy variables ?
This is a great video. Please help, do I need to create dummy variables for all 16 items in my questionnaire where participants either chose a narcissistic response or non narcissistic response?
+ken hz Not sure if you still need advice, but it depends on how you asked the question. If the question is "yes/no" or something like that - basically, if there are only two possible answers per question - then they are already dichotomous and you are good to go.
But what if some of the responses for the categorical variables are "don't know" or "NA" or "IAP", do we still count those as responses and recode them with the k-1 rule? or is there a way to eliminate those responses completely for the multiple regression if I'm only interested in looking at those who responded with the necessary categories?
Thank you for this great video:) I am wondering if it should be 4 dummy variables not 5(since there were 5 choices in the categorical variable) , does it mean that we don't need to make "others" as the standard value number 1?
Yes, four dummy variables. Google dummy variable trap in regression models. One categorical variable should be dropped.
This helped out a lot.
Thank you so much.
i am not getting results from this recoding it is not working.
Is there a way to take care of missing/refuse responses during this process? If not, does it matter if I take care of it before or after? I am transitioning from SAS and having some difficulty.
What if your participant is both WhiteEuropean and Asian (or other biracial)? Can you enter 1 and 1?
Thank you!
really... thanks!
TYYYYYY
like ur voice
I hate this program so goddamn much