0

Kind people,

I want to measure if there is any difference in the use of newsframes in 198 news articles between the 9 countries who own nucleair weapons. I have created these frames with a factor analysis.

5 Dependent variables: score of each of the five frames (scale level) (responsibility, human interest, conflict, morality, economic)

9 Indepenent variables: the nine countries (us, russia, great britain, france, china, north korea, india, pakistan, israel)

I have measured the presence of each of the countries with dichotomous(binary) data (0=country not mentioned in story, 1=country mentioned in story), with a multiple response question. Now I have 9 variables (one for each country), which for every article notes if the country is mentioned or not. This way in one article more then one country might be mentioned.

The problem: Now, how can I measure (in SPSS) if there are any significant differences in the use of frames between the countries? I.e. is the conflict frame being used significantly more often when we talk about North-Korea then with any other country?

I can't seem to find a conclusive answer online (while I have looked for one extensively), so my hope goes out to you.

Stan
  • 1
  • 1
  • When the groups partly overlap by case instances the differences still can be tested for significance. SPSS Custom Tables procedure can do it for you automatically: it accepts multiple response sets as grouping variables (as table columns) at doing a comparison. – ttnphns May 07 '18 at 10:41
  • This, however, won't be MANOVA (multivariate analysis) - Custom Tables will compare one dependent variable at a time. – ttnphns May 07 '18 at 10:43
  • Thank you very much, this has done the trick. Great help! – Stan May 07 '18 at 11:46
  • This seems to be an interesting problem amenable to analysis using Multiple Correspondence Analysis. Primarily descriptive rather than inferential. – David Marso May 08 '18 at 16:19

0 Answers0