I'm hoping someone will be able to help me with this.
I'm trying to replicate some analysis that was done in a published paper to see if I get the same results with my own data.

In both sets of data we have 1 categorical variable that relates to confidence (no confidence, low confidnece, quite high confidence, very high confidence) and a dependent measure which is a continuous variable. One would expect a linear increase in the dependent measure as a function of an increase in confidence so we are trying to find out if this is the case. The authors in the paper I'm reading describe their analysis as follows:

"For each participant we calculated the regression coefficient bewtween our dependent variable and a dummy vector representing the 3 confidence levels. The null hypothesis states there will be no difference between confidence and the dependent variable and as such, predicts that the across subject mean of the regression coefficient is zero. Contrary to this, we found the derived coefficient was significantly different from zero."

It is important that I replicate what these authors have done but I'm not used to doing regression analysis. I'm falling at the first hurdle because the categorical variable is within subjects so I am NOT comparing differnt groups. I don't understand how you can create dummy variables based on a categorical variable that is within subjects. Did these authors just 'pretend' that each confidnece level represented a differnt group of participants to do this analysis? I'm also not sure why they have done this instead of a repeated measures ANOVA with follow up T tests - unless the AVOVA / t test method has less power and would therfore be less likley to produce significant results??

Any help would be much appreciated.