Postby **curren3** » Tue Jan 15, 2013 5:24 pm

Sorry about that : )

I tried the search function but was having a hard time coming up with terms to describe the problem- but if there is something out there I missed, a link would be great.

I am using SPSS 20 on a windows 7 PC.

The experiment involves subjects performing multiplication or division math problems that are either symbolic (with numbers) or nonsymbolic (with dots) where they choose from 5 possible answers. Each math problem has 7 possible answers (ex: 2 x 3 = 3 4 5 6 7 8 9) (with the 4th one, in the case 6, being the correct) but only 5 are presented in each trial. Low means the first 5 choices are presented (6 x 3 = 3 4 5 6 7) and therefore the 4th one, 6, is correct. High means the last 5 choices were presented ( 6 x 3 = 5 6 7 8 9) and therefore the 2nd one, 6, is correct.

Types of variables:

var1 'Notation'- nominal, values are 1 'Symbolic' or 2 'Nonsymbolic'

var2 'Operation'- nominal, values are 1 'Multiplication' or 2 'Division'

condition- nominal, 1 'High' or 2 'Low' (range of answer choices presented)

Response- ordinal, 1,2,3,4 or 5

So the data is organized something like this:

subject problem var1 var2 condition response

1 6x3 1 1 1 4

1 5x10 2 1 2 1

1 10/2 1 2 1 3

.

.

.

For nonsymbolic (var1=2), the graphs look fine because there are a range of responses (ie subjects get some wrong), but with the symbolic (var1=1), in some cases they are all correct so there might only be 1 point on the graph...I would like to have points at the rest of the responses, but just show 0%.

I used the legacy dialog, but have also tried the chart builder and had similar issues.

If I didn't use the split file, is there a way to represent these 4 groups without creating a new variable (although I could do that fairly easily)?

I am not so familiar with using SPSS for graphs- I usually just use excel, but would like to move away from that and do everything in SPSS, so perhaps I am not using the correct procedure.