I am currently analyzing the results of my first ACBC study. I am especially interested in how certain personality characteristics (e.g. level of
extraversion) influence the preferences for my attributes. Therefore, I am including theses personality characteristics as covariates in my HB model. The article by Orme & Howell (2009) was very helpful for setting up the model. In order to determine which covariates have a significant impact on the utilities, I am checking the alpha file based on the rule by Orme & Howell: “When the percent of draws that are positive is >95% or <5%, it suggests that these covariate weights are significantly different from zero, at or better than the 90% confidence level (two-sided test).”
However, I do not understand how the authors come to that conclusion. What kind of test was done, and how? Are there also benchmarks for stricter confidence levels (e.g. 95 %)?
Thanks and best,