I did a between subject design..

Im testing whether adding a label to product packaging of food products, that is displaying the total amount of protein within the product affects the respondents attitude, intention to buy and health perception of the product.

I created a radomization variable, that divided the respondents Into two groups. Each group was presented with 8 pages in a survey, and on each page there was a product and the following 3 questions (measured on a 7 point bipolar scale ranking from -3 to 3):

1) Attitude (how much they liked the product)

2) Intention to buy (to which degree they would buy the product)

3) health perception (how healthy they found the product to be).

Now both groups each saw 8 pages with a picture of a product and the 3 questions on each page. The groups saw the same products and answered the same questions , but one group were exposed to the pictures where a label displaying the protein content was photoshopped onto the packaging.

I wish to test whether the group who saw and answered the questions about products with and with protein label on answered differently regarding their attitude, intention to buy and health perception for each individuel product as well as overall (differences between their total attitude, intention to buy and health perception scores) for alle 8 products combined.

Would this best solved the best with a 2 way between subject anova, or do Better options exist?

An example would be, that i want to compare differences like this:

Grouping variable: seen with label or without label

DV: attitude regarding chicken with protein label.

Dv: attitude towards chicken without protein label.

Kind regards