Dummy Linear Regression
A dummy variable is a variable that can take two values, 1 and 0. Usually 1 represents the presence of an attribute and 0 the absence of it.
Our variables do not need to be recoded if they are already coded as 0 and 1. If this would not be the case then it is recommended to recode them using Transform > Recode into different variables, select the variable, name the new variable and recode the old values with the new values (0 and 1).
We want to examine the effect of importance of work on the power over life. Out hypothesis is that the level of importance that people give to work does not have an effect on how much power do they feel over their life. We check our data in the Variable View window and we see that this is an ordinal variable. We want to create a dummy linear regression, so we have to recode the variable into a dummy. We want to examine only two categories, so we merge some of the values of the original variable creating a dummy. You should always check how many cases each category have and preferably merge the less frequent categories.
First we check the Frequency table.
Here we see that most of our respondents marked that work is very important in their life. So, we create the dummy variable in such a way, that 1-very important, 0 – not very important
Old values: 1- very important; New value: 1 – very important
Old values: 2-4: not very important; New value: 0-not very important
We will name our new variable as v1_rec_imp_work and we will label it in the Variable View in the Values column.
After every recoding it is very important that you run a frequency and you should double check if the recoding was correct.
In the frequency table you can see that the same numbers appear in both the table. You compare the categories and the Totals. They are all the same, so the recoding was correct.
Now we can run the linear regression.
Our model is not statistically significant (p=0,061). Thus, we don’t know weather the importance of work has an effect or not on how much power people feel they have over their life.
The result supports our hypothesis.