Dear colleagues,
I am running analysis and regressions on a dataset from a laboratory experiment. In my dataset, each row represents one choice made by one subject on one round. Each subject has therefore 10 rows as the game runs for 10 rounds. The subjects, on each round of the game, see two wage offers but are automatically matched to one of the two. Sometimes these wage offers are the same, but sometimes they differ and it can happen that a subject is matched to the lower wage offer. After being matched, subjects choose a level of work effort to provide (as a reciprocal behaviour), which can be from 0 to 1.
The hypothesis I want to test is whether being matched on one round to the lower wage offer between the available two leads the subject to provide lower level of effort (perhaps because of disappointment for not being matched to the higher offer, hence reciprocate less). However, this might strongly dependent on the preferences of the subject. I want to test empirically if this is the case. Here's the model I am running:
First I tell STATA that my dataset is panel and the effects to be fixed are at the subject level:
xtest subjectID
Then I create a dummy variable for whether the subject on a specific round is matched to the lower offer: matched_lower_dummy
Then I run the following model: xtreg Effort matched_lower_dummy, fe
The coefficient of the dummy is negative (as expected) and significant. But I want to know if what I am doing is right and makes sense, or if I missing something here. Also, how do I interpret the coefficient of this model? If the value is - 0.178, how do I explain what this means in plain English?
Lastly, if I then were to interact the dummy variable with other continuous variables, is this a problem for FE?
I know some of these questions may sound basic, so apologies if that is the case, but I just want to make sure I am doing the right thing here...
Many thanks,
Bill
I am running analysis and regressions on a dataset from a laboratory experiment. In my dataset, each row represents one choice made by one subject on one round. Each subject has therefore 10 rows as the game runs for 10 rounds. The subjects, on each round of the game, see two wage offers but are automatically matched to one of the two. Sometimes these wage offers are the same, but sometimes they differ and it can happen that a subject is matched to the lower wage offer. After being matched, subjects choose a level of work effort to provide (as a reciprocal behaviour), which can be from 0 to 1.
The hypothesis I want to test is whether being matched on one round to the lower wage offer between the available two leads the subject to provide lower level of effort (perhaps because of disappointment for not being matched to the higher offer, hence reciprocate less). However, this might strongly dependent on the preferences of the subject. I want to test empirically if this is the case. Here's the model I am running:
First I tell STATA that my dataset is panel and the effects to be fixed are at the subject level:
xtest subjectID
Then I create a dummy variable for whether the subject on a specific round is matched to the lower offer: matched_lower_dummy
Then I run the following model: xtreg Effort matched_lower_dummy, fe
The coefficient of the dummy is negative (as expected) and significant. But I want to know if what I am doing is right and makes sense, or if I missing something here. Also, how do I interpret the coefficient of this model? If the value is - 0.178, how do I explain what this means in plain English?
Lastly, if I then were to interact the dummy variable with other continuous variables, is this a problem for FE?
I know some of these questions may sound basic, so apologies if that is the case, but I just want to make sure I am doing the right thing here...
Many thanks,
Bill