# chi square for trend - SPSS

I have followed fourth-year university students for ten years, recording their attitudes to a particular lecture. The outcome is binary (good-bad). I have 10 rows and 2 columns, meaning year and attitude, respectively. The number of responses vary from 60 to 90 depending on the year. How should I perform an analysis to look for a significant change in their attitudes? Can I use chi-square for trend using Crosstabs with each year in the rows and good-bad in the columns, looking at linear-by-linear association for a possible significant change in their attitudes?

I understand this is perhaps a simple question, but being a novice I would very much appreciate your explanations nonetheless!

• Could you elaborate on what you mean by "chi-square for trend" and "linear-by-linear association"? What exactly are these procedures?
– whuber
Commented May 16, 2015 at 19:51
• I have 10 rows and 2 columns, meaning year and attitude, respectively. How should I perform an analysis to look for a significant change in their attitudes? Commented May 16, 2015 at 19:59
• That's a good way to ask your question (+1). I have taken the liberty of copying your comment into the post, because not everybody will read the comments.
– whuber
Commented May 16, 2015 at 20:19

## 1 Answer

In SPSS I tried to run a chi-square trend analysis, but this linear-by-linear association only outputs the p value. Odds ratios for each level do not appear. I recommend doing it on Epi-Info - CDC, as it has chi- square for trend there.