Showing papers on "False positive paradox published in 1992"
TL;DR: This paper briefly reviews the past promise, current results, and problems that have resulted from application of the risk group approach to understanding the etiology of the major psychoses.
Abstract: (1992). Multiple-Risk Cohorts and Segmenting Risk as Solutions to the Problem of False Positives in Risk for the Major Psychoses. Psychiatry: Vol. 55, No. 4, pp. 370-381.
112 citations
TL;DR: This work constructed a perceptron type neural network for E. coli promoter prediction and improved its ability to generalize with a new technique for selecting the sequence features shown during training and showed the correlation between the prediction rates of neural networks achieved by previous researchers and the information content of their data sets.
Abstract: We have constructed a perceptron type neural network for E. coli promoter prediction and improved its ability to generalize with a new technique for selecting the sequence features shown during training. We have also reconstructed five previous prediction methods and compared the effectiveness of those methods and our neural network. Surprisingly, the simple statistical method of Mulligan et al. performed the best amongst the previous methods. Our neural network was comparable to Mulligan's method when false positives were kept low and better than Mulligan's method when false negatives were kept low. We also showed the correlation between the prediction rates of neural networks achieved by previous researchers and the information content of their data sets.
82 citations
TL;DR: In this article, a two-stage procedure for estimating item bias was examined with six indexes of item bias and with the Mantel-Haenszel (MH) statistic; the sample size, the number of biased items, and the magnitude of the bias were varied.
Abstract: A two-stage procedure for estimating item bias was examined with six indexes of item bias and with the Mantel-Haenszel (MH) statistic; the sample size, the number of biased items, and the magnitude of the bias were varied. The second stage of the procedure did not identify substantial numbers of false positives (unbiased items identified as biased). However, the identification of true positives in the second stage was useful only when the magnitude of the bias was not small and the number of biased items was large (20% or 40% of the test). The weighted indexes tended to identify more true and false positives than their unweighted item response theory counterparts. Finally, the MH statistic identified fewer false positives, but did not identify small bias as well as the item response theory indexes
73 citations
TL;DR: The results suggest that while the frequency of false negatives may be larger than thefrequency of false positives, the latter error may nevertheless produce large errors in prevalence estimates, particularly if the prevalence rate is small.
Abstract: Response errors in self-reported drug questionnaires are evaluated, with particular reference to incidence and prevalence estimates. A simple model of response errors is defined, and some analytical results are obtained from this model. By assuming independence of response errors at different points in time, estimates for response errors are obtained from panel data. The results suggest that while the frequency of false negatives may be larger than the frequency of false positives, the latter error may nevertheless produce large errors in prevalence estimates, particularly if the prevalence rate is small.
64 citations
Journal Article•
TL;DR: A model of a system subject to catastrophic failure deteriorates according to a delayed Markov process and is subjected to a series of binary tests that may yield false negative and false positive outcomes is developed.
Abstract: A system subject to catastrophic failure deteriorates according to a delayed Markov process and is subjected to a series of binary tests that may yield false negative and false positive outcomes. A corrective action is carried out when a true positive is observed, thereby reducing the chance of system failure. Costs of inspections, false positives, the corrective action, and failure are incurred, and dynamic programming is used to compute the optimal inspection schedule. Two tractable computational methods are developed. The model, which is suited for medical screening, is applied to the problems of post-operative periumbilical pruritis and breast cancer.
45 citations
TL;DR: The Captia Toxo IgG test gave the fewest discordant results on initial screening, and is a simple and reliable method for screening provided that at least two dilutions are used.
Abstract: AIMS: To evaluate the suitability of five commercial assays (Toxoreagent, DA, Captia Toxo IgG, Toxenz-G, Toxonostika-G) for screening large numbers of sera for antibodies to Toxoplasma gondii. METHODS: Sera from 1000 pregnant women booking for antenatal care at a London hospital were screened in parallel by each test. Sera giving discordant results were retested. RESULTS: The Captia Toxo IgG enzyme immune assay gave the best specificity on initial screening, with 0/773 false positives and only 2/218 false negatives. The Toxoreagent latex agglutination test performed well provided sera were tested at several dilutions to prevent prozone effects; 0/218 false negatives (greater than 12 IU/ml). Only one evidently false positive result was seen in the 1000 samples tested. The DA test gave no false negative results but produced 23/773 false positives. After repeat testing there were 9/1000 sera which gave equivocal results which were negative by the Captia Toxo IgG test (less than 12 IU/ml) but with low titres of 16 in the Toxoreagent test or 4 IU/ml in the DA test. In this situation women would have been asked for a follow up sample for repeat testing. Only 300 sera were tested by Toxenz-G; initial screening produced 4/58 false negative results and 4/242 false positives. CONCLUSIONS: The Captia Toxo IgG test gave the fewest discordant results on initial screening. Results could be readily expressed in international units using a programmable plate reader, and this may be useful for epidemiological studies. The Toxoreagent test is considerably cheaper, and is a simple and reliable method for screening provided that at least two dilutions are used.
17 citations
TL;DR: Evaluating the efficacy of four state nursing home preadmission screening instruments highlights differences in eligibility restrictiveness across states as well as possible tradeoffs of cost containment and access in designing a preadmissions screen.
Abstract: This study evaluates the efficacy of four state nursing home preadmission screening instruments. Using data from three community-based services demonstration experiments, the screens were compared on measures of sensitivity, specificity, and proportions of false negatives and false positives. Results indicate that the more restrictive screens tended to have lower sensitivity but higher specificity, and to generate a higher proportion of false negative decisions than the more liberal screens. Overall rates of correct prediction were higher for the more restrictive screens. Results highlight differences in eligibility restrictiveness across states as well as possible tradeoffs of cost containment and access in designing a preadmission screen.
16 citations
TL;DR: The potentially high incidence of false positive tests with some criteria has important implications for the use of signal-averaged electrocardiography as a screening test in patient populations with various arrhythmia risks.
13 citations
01 Jan 1992
TL;DR: The purpose of antifungal screening is not to define the therapeutic properties of an individual compound, but rather to indicate which of a large number of compounds or samples are worthy of further study.
Abstract: The purpose of antifungal screening is not to define the therapeutic properties of an individual compound, but rather to indicate which of a large number of compounds or samples are worthy of further study. The purpose is simply to provide a YES/NO answer to the question, “Is there sufficient interaction between this sample and fungi to warrant further investigation?” The attempt is to eliminate those samples which are of no definite interest, rather than to be dogmatic about samples which may be of interest. One wishes to reduce the size of a batch of samples considerably, so that those which remain can be subject to more careful study. Screening is not testing a handful of compounds for activity; it is essentially playing the numbers game. If compounds identified as interesting by the screen subsequently prove to be of no value (false positives), that does not matter; it is important that the screen does not reject compounds which really are of interest (false negatives). Once the preliminary process of screening is complete, the few samples which have been identified as possibly of interest can be evaluated in more comprehensive tests to determine their real potential and limitations. It is important to recognize the distinction between screening and evaluation. Screening is the detection of potentially interesting activity; evaluation is the quantitation and characterization of that activity.
11 citations
TL;DR: A clinical criterion based on the ART levels at adjacent frequencies is proposed, which further reduces the false positive rate and could also prove effective in the differential diagnosis of cochlear from retro-cochlear lesions.
Abstract: In this study several methods of defining the relationship between the magnitude of cochlear hearing loss and the acoustic reflex threshold (ART) levels were examined, with a view to defining the appropriate upper limits of ART that would ease clinical applicability and reduce the false positive rate. The 90th percentile, two standard deviations (SD) from the mean, the regression method and an empirical method based on the scatter plots of the ART at each activator frequency were all applied to the results of 99 patients with cochlear and/or peripheral vestibular pathology. The upper limits of ART defined on the basis of the scatter plots provided the most appropriate upper limits for different ranges of hearing loss, with relatively few false positives. A clinical criterion based on the ART levels at adjacent frequencies is proposed, which further reduces the false positive rate and could also prove effective in the differential diagnosis of cochlear from retro-cochlear lesions.
9 citations
TL;DR: The Bayesian classifier produced about half as many false positive errors as the box classifier, with the final false positive rate being in the region of one candidate dicentric chromosome in every four cells.
Abstract: Classification of dicentric chromosomes in a practical automatic screening system comprises three stages. The first generates plausible centromere candidates from each chromosome in an automatically segmented metaphase, and uses contextual knowledge to generate distributions of “probably true” and “probably false” centromeres, thus adapting to the conditions within a particular metaphase. The second stage classifier uses these distributions to re-classify the candidates as centromeres or non-centromeres. From this classification, likely dicentrics are found by counting centromeres; a third classifier attempts to reject false positives among the likely dicentric chromosomes, by comparing the feature values of the proposed centromeres of a chromosome and rejecting chromosomes for which these values do not satisfy certain similarity criteria. The second stage classifier may be a simple box classifier, or may use a variety of parametric Bayesian methods. The performance of these alternatives has been tested both on reference data sets comprising about 600 metaphases, and on larger data sets when embedded in a practical fully automatic dicentric pre-screening system. When operating parameters were such that a similar number of true positives were found by both classifiers, the Bayesian classfier produced about half as many false positive errors as the box classifier, with the final false positive rate being in the region of one candidate dicentric chromosome in every four cells.
11 Mar 1992
TL;DR: The algorithm described demonstrates the feasibility of automated detection of microcalcification clusters using edge detection, graph searching, linear discriminant analysis and cluster detection techniques with a significant improvement over other methods with 100% true-positives, 0% false positives using the re-substitution method.
Abstract: The algorithm described demonstrates the feasibility of automated detection of microcalcification clusters using edge detection, graph searching, linear discriminant analysis and cluster detection techniques. It represents a significant improvement over other methods with 100% true-positives, 0% false positives using the re-substitution method and 98% true-positives, 0% false positives using the leave-one-out method. The leave-one-out method resulted in one abnormal region being misclassified. The error occurred not in the detection of potential microcalcifications, but in their classification prior to cluster detection. A number of possible reasons for this misclassification are given.
TL;DR: This paper examined false positive rates obtained for two normal adult age groups, using three different methods for interpreting performances on the Luria-Nebraska Neuropsychological Battery (LNNB).
Abstract: This study examined false positive rates obtained for two normal adult age groups, using three different methods for interpreting performances on the Luria-Nebraska Neuropsychological Battery (LNNB). Among normal subjects ages 18 to 30 and ages 65 to 85, false positive rates were found to range from 3.9% to 7.7% for younger subjects and 13.5% to 32.7% for older subjects. Significant differences were found between the proportion of false positives in each age group on each method of interpretation. A lack of agreement was found between the three different methods of interpretation concerningwhich subjects should be classified as impaired. Further analysis was performed by dividing older subjects into two groups, ages 65 to 70 and over 70. False positive rates for subjects over age 70 were found to range from 20.7% to 41.4%. Significant differences were also found between the proportion of false positives in each of these two groups.
01 Aug 1992
TL;DR: In this article, a statistical image technique was used for semiconductor in-process wafer inspection to minimize the frequency of false positives, which is particularly important for the later stages of device manufacture where the thin films contain random texture which results in the potential for many false positives.
Abstract: The detection of defects in sub-micron semiconductor devices has reached new limits in recent years. Detection sensitivity limits are now at 0.25 microns and are being driven to 0.1 microns for 256M DRAM production. At these levels of defect detection sensitivity the need for discrimination between true defects and false defects becomes extremely important. Users of such systems cannot afford to manually sort through the large numbers of total reported defects, true and false, and thereby isolate which defects will cause yield loss. This paper discusses a system that has been developed for semiconductor in-process wafer inspection which incorporates a proprietary `statistical image'' technique and other user adjustable system parameters to minimize the frequency of false positives. The results of this study indicate that this is particularly important for the later stages of device manufacture where the thin films contain random texture which results in the potential for many false positives.
01 Oct 1992
TL;DR: In countries such as Guatemala with low prevalence of HIV, it is necessary to confirm positive ELISA results in order to eliminate false positives and positive screening results should not be communicated until confirmation by Western blot or other test.
Abstract: Laboratory tests are essential for diagnosis and management of patients infected with HIV for screening of blood donors and for epidemiologic surveillance of HIV infection and AIDS. Diagnosis of HIV infection is based on demonstration of antibodies against the virus detection of viral antigens isolation of the virus or detection of genetic material in infected cells. Tests that detect antibodies are the most frequently used because of their ease of use and relatively low cost. The ELISA test is the most widely used for screening although a variety of other tests are available. In countries such as Guatemala with low prevalence of HIV it is necessary to confirm positive ELISA results in order to eliminate false positives. Factors to consider in selection of screening tests include availability of skilled personnel to interpret the results the average life and storage conditions of the tests the desired sensitivity and specificity of the test cost and possible existence of HIV-2 in the region. Confirmatory tests have a high degree of specificity. The Western blot is the best known although indirect immunofluorescence and radioimmunoprecipitation tests are also acceptable. There is agreement on what constitutes a negative Western blot but the US Centers for Disease Control Food and Drug Administration and Red Cross have different criteria for positive and indeterminate results. 128 of the 183 Western blots at Guatemalas Center for Orientation Diagnosis and Treatment of Sexually Transmitted Diseases between August 1991 and June 1992 were positive 31 indeterminate and 24 negative. Protocols for serodiagnosis should include pre- and post-test counseling and should assure confidentiality. Positive screening results should not be communicated until confirmation by Western blot or other test.
TL;DR: This article showed that goal-setting alone alone will not improve performance and that reinforcement with goal-setters alone will neither enhance performance nor significantly improve performance in proofreading task, concluding that goals must be combined with reinforcement to increase performance.
Abstract: Locke, Shaw, Saari, and Latharn (2) concluded that goal setting has a positive effect on performance regardless of the presence of reinforcing stimuli Others have concluded that goal setting alone will nor enhance performance (1) The current research assessed whether goal setting separately or with reinforcement were treatments that would significantly improve performance Proofreading accuracy was measured by the number of true positives (identifying true rnistakes) and false positives (identifying false mistakes) in a 2 x 2 factorial design The material read for the task were two coUege level manuscripts on goal setting, the first being 15 typed pages and the second being 30 typed pages Eighty-seven undergraduate students were randomly assigned to one of four groups: reinforcement (extra class points), goal setting (goal = 100% accuracy), goal setting and reinforcement, and control (only received feedback on location of mistakes) The main effect of reinforcement was not significant for either dependent variable The main effect for the goal-setting condition was not significant for true positives but was for false positives (F,,,, = 458, p<05) A Newman-Keuls comparison test showed that only goal setting combined with reinforcernenr had significanrly fewer false positives than the control group (Ms = 9 vs 17, p< 01) The current study provides additional evidence for the position that goals must be combined with reinforcement to increase performance The lack of significant findings for true positives may be related to a ceiling effect inherent in this proofreading task Given the time limit for the proofreading cask, there may have been a practical Limit to the number of actual errors that could be detected Neverrheless, accuracy was best when reinforcement and goal setting were combined