One of the critical tasks we have as empirical political scientists is to assess the validity of the comments people around us make about the socio-political world. For this assignment, you are asked to assess whether there is truth to an empirical claim by analyzing data from the Canadian Election Studies of 2011.

You can choose to write your paper on one of the following five empirical claims:

1. Most Ontarians would not want to see big changes in welfare spending.

2. The average immigrant to Canada arrived before 1970.

3. Canadians considered health care by far the most important issue in the 2011 Canadian elections.

4. The average Canadian man does not like feminists. 

5. More than a third of all Canadians over 50 would like to see the gun registry scrapped entirely.

Guidelines for the data analysis  

In order to conduct a systematic investigation of the empirical claim you are analyzing, make sure to take the following steps:  

- Find appropriate variables

(Which of the questions in the Canadian Election Studies will be of most use to you? What kind of variables are they? Are there other questions that might be of use?)

- Groom your data

(Is it necessary to recode your variables? Are you sure missing data are coded as missing?)

- Select the right cases

(Is it necessary to employ a filter? Are you sure you are including all the cases you are interested in, and nothing more?)

- Visualize the data

(What does a first view at the data tell you? Is it necessary to combine some answer categories to get a better overview? Can you already make an estimated guess of whether there is support for the empirical claim?)

- Calculate the appropriate measures of central tendency and dispersion

(Taking into account the level of measurement, what are the most appropriate measures of central tendency and dispersion?)

- Calculate a confidence interval for a 95 percent confidence level (Are you calculating a confidence interval for a proportion or a mean?  Knowing that a confidence interval with a confidence level of 95 percent ranges from your point estimate plus and minus 1.96 standard errors, what is your confidence interval?)