Many social influence experiments have been criticised for not having experimental validity. This means that it is possible that the participants were not convinced by the procedures, that they thought they were actually, for example, not giving electric shocks in the Milgram study or obeying an order given by someone who was a fake doctor in the Hofling study. High experimental validity would occur if the Ps actually believed that the procedures were real.
Ecological validity is rarely found in laboratory experiments as the term means that the situation could easily be found in the real world. The procedures in the Milgram study, giving electric shocks to complete strangers would b unlikely to be an everyday situation. However, it can be argued that nurses, as in the Hofling study, may come across a situation where they are expected to follow orders given by a doctor in a hospital, during the working day and this would demonstrate ecological validity.
In the Milgram study involving the teacher giving electric shocks to the learner when he gave a wrong answer to a question, it has been suggested that there was high experimental validity. This has been suggested because of the high incidence of obedience found during the experiment as well as the responses from participants when asked if they were convinced that they were giving electric shocks to the learners. It was found that a high percentage of Ps were indeed convinced that they were administering the shocks. However this has been questioned by research carried out by Orne and Holland (1968) who argued that the experiments lacked experimental validity and that the Ps were in fact not deceived at all, that they were just going along with the act.
The Milgram study, however, did lack ecological validity as the experiments took place in a laboratory and the Ps were in no doubt that they were taking part in an experiment although they were not told the true nature of the experiment. However, the procedures have been replicated in other cultures and similar results have been found, but again, Orne and Holland disagree with this saying that the lab surroundings are too ‘set up’ to be realistic and that people would not obey to this extent in real life situations.
In contrast to the Milgram study on obedience, Hofling (196?) carried out an obedience study in a hospital. He telephoned nurses, on the ward, posing as a doctor who the nurses did not know, and gave instructions over the phone to administer 20mg of Astroten when the label on the bottle clearly said that 10mg was the maximum dose. Despite this being against the rules, 21 out of the 22 nurses obeyed the order. The study had high ecological validity as it took place in a hospital during a normal shift and the procedures were not unusual. Nurses take orders from doctors all the time.
The study also had high experimental validity as none of the nurses guessed that they were being tested for obedience to authority and all said the they were often expected to take orders over the phone. However, Rank and Jacobson argued that the nurses may not have been so obedient if they had been asked to administer an overdose of a drug they were familiar with, such as Valium. This may have led them to question the order as they would be familiar with the outcome of an overdose of Valium.
When looking at these two studies it would seem that the Hofling results support the results found by Milgram in that people do obey authority figures without question, whether in the real world or in the laboratory and that experimental validity is the most important issue in experiment.