This question is both about statistical methods and the spss interface (also posted under SPSS for Beginners). I have an incomplete block design with two independent variables:
1) Payment source (3 levels: Company A vs. B vs. C)
2) Cost size (2 levels: High vs. Low cost)
The dependent measure is customer satisfaction with a certain product.
I'd like to test the interaction, specifically the hypothesis that customers are more satisfied by Company A than the other companies, but perhaps only when the cost of the product is Low. The problem is that I don't have cost size data for Company C. So, what is the most appropriate test for comparing the available conditions? Should I still conduct a standard univariate anova, and simply ignore the fact that one of the comparisons is missing? Must I exclude C from the analysis? Should I randomly assign company C's unspecified cost scores to two dummy cost levels? Or should I specify a custom model within the univariate anova window? If so, I'm unclear how the drop-down options map on to the structure of my hypothesis, and would really appreciate step-by-step instructions.
Thank you for helping a beginner!