Taking part demonstration
Research question: Does having paid work have an effect on taking part in a lawful public demonstration within the last 12 months?
weight by pspwght.
fre pbldmn.
RECODE pbldmn (1=1)(2=0) into pbldmn_2cat.
VARIABLE LABELS pbldmn_2cat ‘ taking part or not in lawful public demonstration last 12 months?’.
VALUE LABELS pbldmn_2cat 1’yes’ 0’no’.
fre pbldmn pbldmn_2cat.
fre domicil.
RECODE domicil (1 2=1)(3 thru 5=0) into domicil_2cat.
VARIABLE LABELS domicil_2cat ‘domicil=big city’.
VALUE LABELS domicil_2cat 1’big city or outskirts’ 0’not big city’.
fre domicil domicil_2cat.
fre pdwrk.
fre eduyrs.
LOGISTIC REGRESSION pbldmn_2cat WITH pdwrk domicil_2cat eduyrs.
Step 1
The following table shows the effect of having a paid work on taking part in a demonstration.
b1: The log odds of taking part in a lawful demonstration among those who had a paid work in the last 7 days is higher by 1.060 than among those who did not have a paid work in the last 7 days.
b0: The log odds of taking part in a lawful demonstration among those who did not have paid work in the last 7 days is -4.038.
Exp(b1): The odds of taking part in a lawful demonstration among those who had a paid work in the last 7 days are 2.886 times as high as among those who did not have a paid work in the last 7 days.
Exp(b0): The odds of taking part in a lawful demonstration among those who did not have paid work in the last 7 days are 0.018.
Conclusion: Because b1 is a positive number and Exp(b1) is larger than 1 and the p-value is lower than 0,05, we conclude that having a paid work increases the likelihood of taking part in a lawful demonstration. So, this is our original relationship, when we have only one independent variable. But, we are curious whether this relationship still stands when we introduce a new variable in our model, so we will check this in the next step.
Step 2
So, here we introduce a new independent variable: domicile. If we divide the people by domicil, so if we see this effect among big-city residents and among small city residents will this effect still be there?
Note: the interpretation is different when you have multiple independent variables and not just one independent variable (as in the previous example). In this case, you have to add that “everything else held constant”. So, you interpret b1 in the case when all the other independent variables are 0.
everything else held constant = all the other variables take up the value of 0.
b0: The log odds of taking part in a lawful demonstration among those who did not have paid work in the last 7 days and live in a small city is -4.521. (Note: this is the case when all the independent variables are 0.)
Exp(b0): The odds of taking part in a lawful demonstration among those who did not have paid work in the last 7 days and live in a small city are 0.018. (Note: this is the case when all the independent variables are 0.)
b1: The log odds of taking part in a lawful demonstration among those who had a paid work in the last 7 days everything else held constant are higher by 1.064 than among those who did not have a paid work in the last 7 days.
Exp(b1): The odds of taking part in a lawful demonstration among those who had a paid work in the last 7 days everything else held constant are 2.898 times as high as among those who did not have a paid work in the last 7 days. (Everything else held constant means this is true for those who are not living in a big city, domicile=0.)
In other words: The odds of taking part in a lawful demonstration are 2,898 times as high among those who had a paid job in the last 7 days as among those who did not have a paid job, but only among small city residents. (So, this is only true for those who live in a small city, because small city refers to the 0 category of domicil.)
b2: The log odds of taking part in a lawful demonstration among those who live in a big city everything else held constant is higher by 1,208 than among those who live in a small city.
Exp(b2): The odds of taking part in a lawful demonstration among those who live in a big city everything else held constant are 3,347 times as high as among those who live in a small city.
In other words: The odds of taking part in a lawful demonstration among those who live in a big city are 3,347 times as high as among those who live in a small city, but this difference refers to only those who do not have a paid work (because you have to regard the 0 category of the other independent variable).
Conclusion: Here you have to focus on the significance level of the original variable. Here we see that the p-value is still significant (p=0,001) and the Exp(b1) did not change too much, it is almost the same. This means that involving domicile does not change the effect of doing a paid work on taking part in lawful demonstration. This is good news for us since we know that doing a paid job does have an effect on taking part in a lawful demonstration.
Step 3
In this step we add one more independent variable: years of completed full-time education.
b0: The log odds of taking part in a lawful demonstration among those who did not have paid work in the last 7 days and live in a small city and completed 0 years of full-time education is -5,429. (This is a hypothetical situation because we do not have in our sample people who have completed 0 years of education.)
Exp(b0): The odds of taking part in a lawful demonstration among those who did not have paid work in the last 7 days and live in a small city and completed 0 years of full-time education are 0.004. (Note: this is the case when all the independent variables are 0.)
b1: The log odds of taking part in a lawful demonstration among those who had a paid work in the last 7 days everything else held constant are higher by 1,021 than among those who did not have a paid work in the last 7 days.
b2: The log odds of taking part in a lawful demonstration among those who live in a big city everything else held constant is higher by 1,194 than among those who live in a small city.
Exp(b1): The odds of taking part in a lawful demonstration among those who had a paid work in the last 7 days everything else held constant are 2.776 times as high as among those who did not have a paid work in the last 7 days.
Exp(b2): The odds of taking part in a lawful demonstration among those who live in a big city everything else held constant is 3,301 times as high as among those who live in a small city.
b3: The log odds of taking part in a lawful demonstration is higher by 0,070 for each year of completed full-time education, everything else held constant. However, the effect is not statistically significant.
Exp(b3): The odds of taking part in a lawful demonstration is multiplicated by 0,004 for each year of completed full-time education, everything else held constant. However, the effect is not statistically significant.
Conclusion: It seems that after controlling for domicile and completed years of education the effect of having a paid job on taking part in a demonstration remains significant. So, it seems like having a paid job influences taking part in a lawful demonstration.