The dichotomous nature of null hypothesis significance testing (NHST) is just another way for psychology to gain much needed credibility by trying to be more scientific in its methods. The popularity of NHST over the last 50 years, and the reluctance of researchers to accept and use confidence intervals to report their research (despite the fact they provide more detail than p-values) are all attempts at trying to have psychology be taken seriously as a science, by taking behaviour (too often a single facet of it) and quantifying it. The over-reliance on reporting p-values only serves to support this accusation. Analysing data and producing p-values to report results is a good attempt at being scientific on psychology’s part, but it is a misguided one. In trying to be more like a hard science, psychology has gone too far, so far that using p-values is more of a detriment to research, and the advancement of the discipline. Confidence intervals provide information on the size of an effect, but because this gives the appearance of making the data seem less precise (and therefore less credible), they are not often used in research.
Another problem with psychology is that the link between it and biology is played up in order to gain much-needed credibility (McPherson, 1992). Psychology in the past has often tried to ‘borrow’ outdated models from other scientific disciplines (Valentine, 1991), and is sometimes accused of being less concerned with behaviour, than taking a small component of it and studying it at the expense of complexities of the mind and behaviour (McPherson, 1992). Empirical research often lives up to this explanation. Contemporary studies in psychology are quite adept at taking a grey area (in this case human behaviour) and breaking it down to convert it into something black and white (such as a single hypothesis which is either accepted or rejected on the basis of the data collected). When a hypothesis is rejected, the alternative hypothesis is accepted, usually without there being any evidence to support such a conclusion. This means those who reach conclusions that their data does not support abuse the process of understanding the events or phenomena being studied, which is another violation of the aims of the science.
This in turn presents problems with the generalisability of research findings in psychology. As well as the methodological problems I mentioned which hamper the generalization of findings, psychologists too often conduct research with small samples, or samples which are not representative of the population (Wallace & Sechrest, 1973). Rats and first-year university students are heavily relied on as participants in psychological research, which ultimately means that generalizations can only be made to other rats and first year university students, which is rarely the case. Many researchers think it appropriate to generalize to the entire population, something that needs to be rectified by using more people from different age groups, as well as different minority groups.
The nature of the phenomena being studied in psychological research makes research difficult due to the fact that the mental world cannot be directly observed (Valentine, 1991). Psychology can not measure cognitive or affective aspects of behaviour, such as anger and guilt, using direct observation techniques. It can only measure the overt behaviour itself. The best that can be done is guesswork, based on observable behaviour which results, or self-report, which is subjective and therefore unreliable. Mental constructs can only be tied, through inference, to observable behaviour (an outburst caused by anger for example), which is another example of science neglecting proper scientific methods in favour of convenience.
Making inferences about behaviour is not the only problem. Before any study is conducted in psychology, researchers must apply for approval from the relevant ethics committee or committees. Only when ethics approval is granted can a study be commenced. Ethics committees make it increasingly difficult to obtain approval for research, and are so protective of the interests of potential participants in psychological research that many studies of human behaviour can no longer be done. This is in contrast to psychological research in the past, where controversial and damaging experiments were often done, without much regard for the welfare of the participants (take for example Watson’s little Albert study).
These days, careful consideration has to be given to the impact a study will have on it’s participants, and must take into account risk of physical, mental, or social harm, which includes the administration of pain, treatments or therapeutic techniques (including the withholding of treatment from a control group for example), and emotional distress. Informed consent, deception, and issues to do with privacy, anonymity, and confidentiality for participants in research are all things that need to be taken into account before a study goes ahead. Gone are the days of leading someone to believe they are administering electric shocks (which are gradually increasing in intensity) to someone in another room who is screaming in pain (albeit acting the part) in order to study how far someone will go when taking orders.
Another factor which can affect the validity of psychological research is the presence of variables, which can be hard to control for. Laboratory settings are used in order to control for extraneous variables, but it is well known that observation affects that which is being observed – the experiment is a social situation in itself (Valentine, 1991). It follows that results gained through observation in un-natural (a laboratory) setting cannot be relied upon. Naturalistic observations where the observer remains hidden at all times, so as to minimize or eradicate the effect on the participants, are the only ones which can be relied upon. But naturalistic observations are rarely used, because of the fact that these variables cannot be controlled for. This makes investigation of the relevant variables and uncovering their exact influence impossible.
So just how far has psychology taken us? Even with all the studies conducted to investigate the mind and human behaviour since psychology first became a discipline in 1879, there is still so much that psychologists do not know. They cannot predict or control behaviour, as their observations and understanding of their subject material is either severely limited, or flawed, or both.
So this begs the question, should psychology be a science? I think psychology is too concerned with trying to be a science, which is a big part of its problem. ‘Science’ is just a label that shouldn’t matter as much as the state that psychology as a discipline is in. A desperate scramble for that label by any means necessary is not going to bring any kind of credibility to psychology, nor should it. What is more important, and what psychologists should focus on, is improving the way studies are conducted, how research is reported, and what implications this has for the wider population. This is much more likely to bring credibility to the discipline, something which it seems to be sorely lacking in its current state. I do not think that different areas of psychology need different answers to the questions addressed in this essay. Anything that operates under the umbrella of psychology should be open to the same sorts of criticisms as psychology in general. I do not think the way things are done, with regard to methodology, etc., differs greatly, if at all, within different areas of psychology. Labeling some areas of psychology science is not at all justified at this stage.
So in conclusion, psychology is not a science for several reasons, the main reason being that psychological research is conducted in a way that makes understanding, prediction and control close to impossible. There are so many problems that still need to be overcome by those working in the profession before psychology comes close to resembling a science. Methodological issues such as sample sizes and characteristics need to be addressed, as do methods used to analyze and report research findings. Serious consideration should be given to including confidence intervals, and not just p-values in research. Psychologists should stick to material which they can investigate in a scientific manner if they wish to call themselves scientists. Investigating the subject matter in a scientific way will bring credibility to the discipline, and improve its standing. This should be the main aim for psychologists, not scrambling to make their methods appear scientific in order to have psychology be considered a science.
References
Malim, T., Birch, A., & Wadeley, A. (1992). Perspectives in Psychology. Hong Kong: MacMillan.
McPherson, M. W. (1992). Is Psychology the Science of Behaviour? American Psychologist, 47,329-335.
Plucker, J. (2003). The Cyril Burt Affair. Retrieved May 12, 2004 from http://www.indiana.edu/~intell/burtaffair.shtml.
Valentine, E. R. (1991). Conceptual issues in psychology. (2nd ed.).
London: Routledge
Wallace, J., & Sechrest, L. (1973). The Nature and Study of Psychology. Itasca: F. E. Peacock.
Webster. (1992). Webster Comprehensive Dictionary Encyclopedic Edition (Volume 2). Chicago: J. G. Ferguson