I run a regression where I created a set of dummies to represent different intervals. I had set of 4 dummies, only included 3 in the regression so the one with the highest interval (experience) acts as a benchmark. How do I interpret those if my dependent variable is in ln form? I have no clue how to interpret this also in relation to the benchmark variable when I have multiple sets of dummy variables.
For example, if the benchmark is 'bigger AS size' and the AS 5000m is negative -,764 does that mean that having only 'AS 5000m' decreases the output but by how much? 76,4% ? or only 0,764? I am not sure how to interpret the constant since it is in log and it also represents the benchmark for other dummies in the regression...
I have shortened the model, but there are in essence two 'sets' of dummies (locations and the size intervals) and two normal dummies.
variable B coeff. (Constant) 6,281 Car ,389 Inclination (dummy) -,125 Fallow Land (dummy) -,072 ln Input (USD/ha) ,180 No_Agroforestry -,210 **AS_5000m2 -,764** ln Land (ha) -,337 P2: Sacaba -,042 P3:Tiquipaya -,130 P7:Tapacari -,846