Good Excellent question!
Depends on how you define it.
What we are really interested in here isn't the absolute number of minutes each candidate gets in debates. That would get very complicated very quickly given differing over all lengths of the debates and times spent asking questions and covering procedural matters.
Easiest to compare and determine biases by tracking percentages of time democrat vs republican.
Set the scale at democrat time-republican time. With a perfectly balanced debate time, the times would be equal: d-r=0.
Democrat bias>0, Republican bias<0.
Now your chart can be symmetric.
We both agree it isn't, we're quibbling about the degree and objective verifiability of the bias.
One can always hope someone, somewhere will think "Hey this guy knows control charts, maybe we can hire him..."
But since we know the rats got more time all we need to do is map those on an xy and there is your trend.