# Survey Results & Regression Analysis..Dummy Variable?

#### StatsGuy112233

##### New Member
Hello All.

I have just received results back from a survey that I launched and I need help on how to analyze the results. There may be some need for dummy variables and if so, I am not sure if it is necessary.

The dependent variable in the survey was:
How much time do you spend on youtube on an average visit:
A= 0-15 minutes
B= 16-30 minutes
C= 31-45 minutes...etc

Then, the independent variables were:
How often do you look at the following videos on an average visit?
There were five values to choose from on a 1-5 scale. 1= Never 5=Always

Looking at funny videos:
Looking at scary videos:
Looking at videos friends sent to you:
Looking at at videos with babies:..etc.

We have received the results and are now wondering how properly interpret the results. We believe there are two potential answers, not sure which one to choose.

We ran a stepwise regression with all the variables and got back the following:

I know how to analyze this, reading the output is not the issue. The issue is...

We are not sure if we should use this model or, we need to create DUMMY VARIABLES for each of the variables. Since we have a total of 10 variables with 5 choices (4 dummy variables) for each, we have a total of 40 variables. The output is the following:

We are wondering what the right thing is to do. Should we continue with the analysis with a regular stepwise regression or do we need to create dummy variables for each variable and then run the stepwise regression?

Thank you for any help!

#### krytellan

##### New Member
I'm not sure I'm going to have the answer you want to hear. Looking at what you have, I am not sure I see a regression and if there is one, cannot tell what you are trying to predict.

Are you trying to say that what kind of videos people watch predict how long they are on youtube? If not, what is your research question.

Multiple regression is really not designed to be used with a categorical IVs or DVs. Categorical DVs cannot be used without transformations which I am not qualified to talk about (maybe someone else can). You can dummy code categorical IVs, though again, it is not the intention of MR.

Can I ask why you chose MR as the analysis for this data?