Psychometric results

This communication displays three distinct parts. The first part presents a brief history of the psychometric paradigm and its value for the risk manager. The second part exposes the general methodology of psychometric studies. The last part outlines presuppositions and criticisms the risk manager should be aware of when using psychometric results. 

Before using psychometric results, it is necessary to check if the samples are representative enough. In this respect, it has to be noted that, in the pioneer psychometric works, the samples were small and not randomly selected. However, the main trends of these first studies have been confirmed in later works, based on more representative samples. 

Sexton et al. reported favourable psychometric results for the SAQ, with strong evidence for reliability and validity. Subsequent studies (Etchegaray and Thomas 2012 Profit et al. 2012 Schwendimann et al. 2012 Weng et al. 2012) have yielded similar results at the individual and unit level. In addition to psychometric findings, much research exists about the SAQ as a measure of intervention success. Integrated within this area of research is examination of the SAQ as a predictor of outcomes. We review both of these areas of research. 

Because the principal focus of pharmacoeconomics is to incorporate outcomes in the demonstration of the value of a therapy, the quality of life that results from a therapy has received a great deal of attention in recent years. Most frequently, quality of life is measured using standardized instruments to collect patient-reported quality of life. These instruments are traditionally developed using patient or expert interviews to establish the areas of inquiry or domains of the instrument The use of these instruments in pharmacoeconomics is heavily dependent on the demonstrated psychometric properties of the instrument, such as its reliability, validity and responsiveness, or on sensitivity to the changes in the disease to be studied. 

Each subject wrote a subjective account of his experience within a week after the experimental session. Approximately six weeks after the session, subjects were administered questionnaires that related to (1) the effects of the session on postsession creative ability and (2) the validity and acceptance of solutions conceived during the session. These data were in addition to the psychometric data comparing results of the two testing periods. 

The aim of applying mathematical and statistical methods is first of all to detect and secondly to describe the interrelations between influencing forces on the one hand, and the interrelations between these influences and the resulting effects on the other hand. Subsequently, new scientific subdisciplines such as biometrics, psychometrics, medical statistics, econometrics, etc. have been developed. These developments have been reflected by the publication of corresponding journals, for example Biometrika (1901), Psycho-metrika (1936), Technometrics (1959) [GELADI, 1995]. Chemometrics was, therefore, not the first discipline of its kind. 

Whereas there has been a huge interest in multimode calibration in the theoretical chemometrics literature, there are important limitations to the applicability of such techniques. Good, very high order, data are rare in analytical chemistry. Even three-way calibration, such as in DAD-HPLC, has to be used cautiously as there are frequent experimental difficulties with exact alignments of chromatograms in addition to interpretation of the numerical results. However, there have been some significant successes in areas such as sensory research and psychometrics. 

Tata et al. (1994) used psychometric tests to follow up 21 patients 6 months after abstinence from long-term therapeutic doses of prescribed BZs. They also examined 21 normal matched controls. Pre- and postwithdrawal and 6 months afterward, the results demonstrated significant impairment in patients in verbal learning and memory, psychomotor, visuo-motor and visuo-conceptual abilities, compared with controls. Lagnaoui et al. (2002) found increased dementia in elderly patients treated with BZs in a community setting. 

These simple psychometric screening tests do, however, have certain disadvantages, which have to be considered when results are being assessed, (s. tab. 10.5)... 

Based on the results in the literature (which also correspond to our own experience), psychometric test programmes have been developed for use in the hospital and the doctor s surgery. 

The transition from the latent stage (SHE) to stage I of clinically manifest HE can be gradual, as demonstrated by the results of the psychometric tests. In this transitional zone (SHEI), there may occasionally be minor neurological findings, such as hyperreflexia, fine tremor or asterixis. (s. p. 272) (s. tab. 15.5)... 

L-dopa, a precursor of the neurotransmitters norepinephrine and dopamine, was introduced into HE therapy by J.D. Parkes et al. in 1970. The results were good. As yet, there is still no accepted opinion on the use of this substance. Piracetam, as a nootropic substance, led to a clear improvement in typical electrical brain activities in animals displaying hepatic damage and symptoms of encephalopathy. Similarly, a double-blind randomized cross-over study with the nootropic agent centropheno-xine partly showed positive effects in psychometric testing. Bromocriptine, an agonist of the dopamine receptor, was also used in 1980 for chronic hepatic encephalopathy. (146,163) Application of L-carnitine (6 g/day orally, divided into two doses, for 4 weeks) leads to a marked reduction of hyperammonaemia and a clear improvement in clinical symptoms of HE in cirrhotic patients. (119) (s. p.49)... 

If indices of brain function are associated with prevailing levels of psychometric intelligence one might expect them to become impaired when mental test scores are affected by prevailing blood glucose levels. In the case of reaction time (decision time and movement time) and inspection time indices this occurs, but not for nerve conduction velocity. The Hick reaction time results are now considered in more detail. Forty trials were completed at each of four levels of stimulus uncertainty (1,... 

All of the tasks are similar in format and use the same stimuli. Tasks are presented by computer and responses are made on a touch screen. The tasks were thoroughly piloted before being administered as a battery. The average reliability of the measures obtained from the tasks is around 0.80. This battery, in my opinion, is both psychometrically and cognitively fairly crude. I believe it could be substantially improved upon and so the results I present here are conservative. 

Datafrom rats show that systems that contribute togare independent. The findings from human data are supported by studies with rats.Thompson et al (1990) lesioned rats in various brain areas and then tested them on a number of tasks. When they factor analysed the resulting data, they found brain areas that they thought corresponded to psychometric. While some of these areas were in the cortex, others were in lower centres of the brain. Most interestingly, though, each portion of the brain implicated in psychometric involved a different brain system. 

BPS examinations are administered with the assistance of an educational testing firm, resulting in a process that is psychometrically sound and legally defensible. Each of the five specialties has its own eligibility criteria, examination specifications, and recertification processes. All... 

---