Abstract:

The paper describes briefly how respondent different scale usage may affect the analysis of results derived from discrete scales such as those commonly used in marketing research. Those effects are essentially due to yea / nay / middle-of-the road saying - use of upper / lower ends or middle-points of the scale. It is demonstrated how the use of the Maximum Difference Scaling (MaxDiffs) method overcomes this difficulty, showing as an example a customer-satisfaction like study evaluating the importance of non-financial attributes in the analysis of investment in the capital market.