Strictly speaking, bandgap corresponds to transition between valence band and conduction band. In this sense, doping ion at rather low content will not form the conduction band and its contribution to optical absorption cannot be considered as modification of the bandgap. If you work with thin samples, you will always see difference between fundamental absorption and doping ion one. However, if you have thick sample containing impurity with allowed (e.g. s-p) transition at energy below bandgap, then, with the account for dynamic range of your spectrometer, you may see something that may force you to think of 'narrowing of the bandgap'. E.g., 10% of trivalent bismuth can make an impression of bangap shift to 300 or 350 nm at 1 mm thick sample and OD limit of 4.5. But it is not true bandgap, to my opinion.
Strictly speaking, bandgap corresponds to transition between valence band and conduction band. In this sense, doping ion at rather low content will not form the conduction band and its contribution to optical absorption cannot be considered as modification of the bandgap. If you work with thin samples, you will always see difference between fundamental absorption and doping ion one. However, if you have thick sample containing impurity with allowed (e.g. s-p) transition at energy below bandgap, then, with the account for dynamic range of your spectrometer, you may see something that may force you to think of 'narrowing of the bandgap'. E.g., 10% of trivalent bismuth can make an impression of bangap shift to 300 or 350 nm at 1 mm thick sample and OD limit of 4.5. But it is not true bandgap, to my opinion.