Good day everyone
May I ask a question about my research?
I've been thinking of this for quite a while but I am a little confused on what method I should use. I am posting this thread to verify if what I am doing is correct (I believe it is correct, but I would like to hear your thoughts on it)
My study looks into how electronic voting has affected people's satisfaction and trust in the election results. I will be making use of these http://www.sws.org.ph/pr20100728.htm surveys to compare the 2010 (which was the automated elections) and the 2004 (non-automated)
My Hypothesis is that the 2010 automated electin results generated a higher score (basically positive feedback) than the 2004 elections. (So automation is good!) And I'm planning to test this
I'm planning on using
a) (One-tailed?) z-test
Basically hypothesis testing. I want to see if the 2010 results (automated) is -significantly higher- than that of the 2004 (non-automated)
Ho: mu = (the 2004 results)
Ha: mu > (the 2004 results)
I'm also having a hard time looking for studies that made use of the z-test to compare election results between two national elections. Most of the studies I have found were related to medicine and etc. Such was the study which included a standard set of questions asked on a) people who have encountered cancer b) people who did not encounter cancer
So I thought the logic would be the same wherein; the "set of standard questions" will be the criteria in the link (ex. Satisfaction with the general conduct, etc) and the two sets would be the two years (2010 and 2004)
I'm a little hesitant on how I will go about with the given charts/results though.
Another method I was planning to do was
1) Look for a survey asking if people were for or against the automation of elections
2) Use the chi-square to look into the correlation between the automation and the survey results
However I have come to my senses that this will not make sense as the samples were different.
I am also looking into the Wilcoxon test, but I am a little unfamiliar with it.
Any suggestion will be appreciated
Advertise on Talk Stats