Discuss the issues of Reliability and Validity in the Diagnosis of Mental Abnormality Diagnosis is the process of identifying a disease, and allocating it to a category on the basis of symptoms and signs. A classification system such as the DSM-IV version can be used to diagnose mental abnormality. Categorisation therefore is very important, as diagnosis using the classification system will bring about certain treatments, and if the diagnosis is wrong, the person may be receiving inappropriate treatment, which may do more harm than good. Clearly any classification system will be of little value unless psychologists can agree with one another when trying to reach a diagnosis. However if for example a person is diagnosed with a mental illness such as manic depression, the clinical psychologist will know what symptoms to expect and possible ways of treating the illness. By clinical psychologists all using the same classification system, diagnosis should be less bias, because they use objective behaviours and sets of symptoms.

However a clinical psychologist may interpret a person's behaviour in a different way to another clinical psychologist, so diagnosis can be seen as quite sceptical. A diagnosis is considered reliable if more than one psychologist would give the same diagnosis to the same individual. Early studies consistently showed poor diagnostic reliability. Cooper et al (1972) conducted the US-UK diagnostic project, where American and British psychiatrists watched clinical interviews on video, and were asked to make a diagnosis. The results showed that American psychiatrists diagnosed schizophrenia twice as often, while British psychiatrists diagnosed mania and depression twice as often. This shows how diagnosis can be very unreliable, and how cultural differences can affect reliability.

Other studies such as that of Beck (1961) showed that diagnosis can be unreliable, after his experiment where two psychiatrists were given the same 153 patients to diagnose, but the two only agreed on their diagnosis 54% of the time. In contrast Mehl (1977) said that with enough information and by sticking to thorough descriptions and major categories, diagnosis could be completely reliable. It is possible that patients are affected by demand characteristics, and give different psychiatrists different information, which makes diagnosis unreliable. An institution only admits patients with certain diagnoses, and the mental health professional wants them all to be admitted, which again reduces the reliability of diagnosis. Unstructured interviews are considered less reliable than structured interviews in diagnosis of mental abnormality, but structured interviews are not used in situations where they would be useful. There are some mental illnesses that are very hard to diagnose, and some psychiatrists do not have the time to gather all the information needed to make an accurate, worthwhile diagnosis.

However the reliability of diagnosis is ever improving, as agreement between psychiatrists can be improved if they are trained to use standardised interview schedules, such as the present state examination (PSE; Wing et al, 1974) and Endicott's and Spitzer's (1978) schedule of affective disorders and schizophrenia (SAYS). These are intended to specify sets of symptoms which must be enquired about, and define the symptoms precisely, giving instructions on rating their severity. There have undoubtedly been improvements in the reliability of diagnosis since the publication of DSM (1980), aided by the use of 'decision trees' and computer programs that lead psychiatrists through the 'tree'. Sartorius et al (1993) concluded that ICD-10's clinical guidelines were suitable for widespread international use, and showed good reliability. The primary purpose of making a diagnosis is to enable a suitable programme of treatment to be chosen.

Identifying a path for a mental illness and what is likely to happen is needed to make a diagnosis valid. Treatment must work to some extent for the diagnosis to be called valid. According to Heather (1976) there is only a 50% chance of predicting correctly what treatment a patient will receive on the basis of diagnosis. Robins and Gaze (1970) gave five criteria for deciding whether a diagnosis is valid: . There is a clinical description that goes beyond the actual symptoms, for example if people of the patient's age suffer from the disease...

Laboratory studies support the conclusions, so that particular group of symptoms usually go together... This particular set of problems is different from other disorders, and the differences are noticeable... Follow-up studies have indicated reliability, for example by using a test-retest method... Family studies suggest that a particular disorder is hereditary.

As long as one or more of these five criteria can be adopted, there should be some validity in the diagnosis. Rosen ham (1973) conducted an experiment to illustrate the poor validity of the diagnostic classification system for mental disorders. He used 8 sane people, who said they could hear voices (pseudo-patients), to try and get admitted to a mental institution, and when in the hospital they should stop simulating their symptoms and respond normally to instructions. Their task was to seek release by convincing the staff they were perfectly sane.

The results showed that the patients sanity was never detected by the staff, only by other patients, and they were kept in the hospital for an average of 19 days. All except 1 participant was diagnosed as 'schizophrenic in remission', supporting they were never thought of as sane by the staff. Participants noted that even normal behaviour become distorted as 'schizophrenic', as waiting outside the cafeteria was seen by a member of staff as 'oral acquisitive nature of the syndrome'. This illustrates how the validity of the classification system used for diagnosis is extremely low, although only the experimental condition was conducted and the experiment was based more on qualitative data, rather than quantitative data. It is possible that there are other factors other than diagnosis that may be equally important in deciding on a particular treatment such as socio-economic factors. Some psychologists believe it is the psychiatrists who are the ones at fault and not the method of diagnosis.

According to Clare (1980), the criticisms of psychiatric diagnosis should be aimed at the psychiatrists and not the process of diagnosis in general. Validity is more difficult to assess than reliability of diagnosis, because for most disorders there is no absolute standard against which diagnosis can be compared. However much we improve reliability, there is still no guarantee that the patient has received the correct diagnosis, and if diagnosed falsely can have serious implications, by harming the patient even more physically and mentally. According to Winter (2000), 'diagnostic systems are only aids to understanding, not necessarily descriptions of real disease entities'.