Lack Internal Validity
Psychological research is fraught with problems such as investigator effects and demand characteristics, which compromise the internal validity of the research. This is an disadvantage because the observed effects may be due to variables other than the research. For example, a questionnaire may involve leading questions, which means that the findings are not valid.
Lack External Validity
Findings from psychology experiments are not always supported by real-life everyday observations. This suggests that the findings of psychological research cannot be generalised beyond the particular settings in which they were conducted. For example, Mandle said that Milgram's findings bore very little relationship to the behaviour of the real world, because its not everyday that obedience is tested through the harming of others.
In order to conduct research, behaviour must be reduced to a set of individual operationalised variables. This is true in experiments and also in observational studies where behavioural categories are operationalised. The result of this reductionism is that we may oversimplify something that cannot be simplified and in doing so, are no longer studying what we meant to study. An example of this is R.D Laing who, when discussing the causes of schizophrenia, said its inappropriate to view a person suffering distress as a physical-chemical system gone wrong. He claimed that treatment can only succeed if each patient is treated as an individual.
Ignores Individual Differences
Science takes the nomothetic approach, looking to make generalisations about people and find similarities. This is a disadvantage because the result is that gender, culture, age and other individual differences are overlooked. For example, much research into psychology has involved American participants with the results being generalised to whole populations. However, this group of people may have unique characteristics.
This is a disadvantage because the issue is to consider whether the benefits of research outweigh the ethical costs. Even if the ethical costs are 'excusable' the end result is that particpants may have been harmed in some way. For example, in Milgram's study individual participants may have felt that their rights have been infringed despite the fact that some knowledge about obedience was gained.v