Here are some notes for the categorical how tos. There will be two data sets. One where Gaskell and colleagues (1993) asks people what comes to mind when science is mentioned. The second was an experimental Carlucci and colleagues (2011) performed on South Beach.

The Gaskell data are in two files: mind.sav and mindexpand.sav. The first has them in a table like format. It is the frequencies, in the last column, that you will actually graph for a bar chart. The second has a row for each person. These are data for what comes to mind when science is mentioned.

For SPSS, if you use the table format you have to tell SPSS to weight by frequency (go to DATA, WEIGHT CASES, and then put freq in the box). SPSS will now treat these like one row per person. If you were having to type all the data in, this would be a lot quicker. The SPSS bar chart can be either simple, for one variable, or clustered, for multiple variables. Play around with these.

For R, have a look at this code. This loads the file, and gives some output.

For the 2x2, we will use data from Carlucci et al. (2011). Using the Crosstab procedure in SPSS you get:

The Crosstabs procedure also allows various statistics to be calculated. Here a confederate approached a group of people, and asked when time it was. He went away, and an RA asked either the person he talked with, or the other person, to identify the person from a 6 person lineup. Then the RA asked the other person. Carlucci was interested in whether the second person conformed. If they were just guessing, what percentage of time would they have said the same thing? Carlucci was specifically interested in whether the person conformed more if they were the SAME person who interacted with the confederate, or was the OTHER person.

For SPSS, put these data in and use the weight function. The odds ratio is in RISK in crosstabs, but show how to calculate it by hand.

For R,

conform <- rep(c(0,1,0,1),c(70,23,46,37))

approach <- rep(c(0,1),c(93,83))

table(approach,conform)

The main thing is to calculate the odds ratio. Do this by hand. You may also write a function to do this. Try writing one where you just put the four numbers in and it returns the odds ratio.