The hypothesis being tested seems to have been that Americans would not be willing to inflict pain on one another simply because they were told to do so.
The dependent variable in this experiment was the voltage of shock that the "teachers" in the experiment were willing to administer to the "learners."
In the original experiment, there really was not an independent variable. The researchers were simply trying to find out how far Americans would go in obeying authority.
You could argue that nationality was the independent variable. However, I do not think this is correct because the original experiment was not comparing results of various nationalities.
As a Social psychologist, Stanley Milgram was particularly interested in the affects of interaction on behavior. Interestingly, this experiment was born to test the limits of people's compliance and obedience under conditions of potentially extreme cognitive dissonance, and authority. Essentially, Milgram wanted to understand, how far would people obey orders given from an authority figure, under circumstances that contradicted their beliefs. This was his hypothesis.
It is important to note that this experiment was conducted with four variations. Overall, each variation involved the same basic test. Respondents, or "Teachers", were told that the effect of punishments on learning ability were being testing. "Learners", or test subjects, were to be shocked when they have an incorrect answer to a question. "Teachers" were also told to treat silence as an incorrect response. With each incorrect response, the degree of shock administered was to increase based on a set of levers ranging from 75 volts, or minimal shock, to 450 volts or, XXX.
With each variation of the experiment, the independent and dependent variables changed slightly, as the experiment varied. Please see a brief description of the independent variables through each variation of the experiment, and how they were tested.
Experiment Variation I:
Independent Variable: Immediate proximity to authority.
Test: Respondents or "Teachers" were in a room with the authority figure, they could see the test subject, and were nudged along with statements like,
The experiment requires you to continue.
Experiment Variation II:
Independent Varaible: Rebellious interveners (This variation tested rather introducing people who questioned authority, would change the obedience of others.)
Test: Respondents were kept together in a room, and some individuals who fit the category of rebellious or questions authority, were kept amongst a group of average responders. The test was to see how many people would follow the rebellion and not comply with the experiment.
Experiment Variation III:
Independent Variable: Proximity to Learner (This variation tested rather being in the room with the test subject would change the obedience level.)
Test: Respondents were in the same room with the learner, they could not see the authority, and were forced to place the learner's hand on the shock plate, and administer the shock.
Experiment Variation IV:
Independent Varaibale: Perception of authority (This variation tested how each responder's idea of the authority figure, changed the level of obedience.)
Test: Responder's were paired with lesser authorities and told to administer shocks to the learner. This could have been two responder pairings, or a lab tech and a responder.
The hypothesis that was tested in Stanley Milgram's famous "Milgram Experiment" was "the degree of pain an individual is willing to inflict upon another individual just because he was ordered by an authority figure."
In the experiment, the "teacher" was made to ensure the learning of word pairs by a "learner" by use of progressively increasing electric shocks. The maximum level of shock a teacher was willing to administer to the learner was used to test this hypothesis.
The dependent variable of the experiment was the amount or level of electric shock that was delivered to the learner to ensure compliance.
Milgram performed a series of these experiments with some variations and based on these series, the independent variables tested included proximity of administrator and teacher, perceived proximity of teacher and learner, number of teachers in the same room (0-2), among others.