论文信息 - Parametric and Nonparametric Signal Detection Measures: Pros and Cons

Parametric and Nonparametric Signal Detection Measures: Pros and Cons

Parametric measures of signal detection sensitivity (d') and response criterion (β) require somewhat restrictive distributional assumptions. Nonparametric analogs are available with no distributional assumptions. This paper describes several signal detection measures, discussing the advantages and limitations of each. A concurrent analysis of parametric and nonparametric estimates in a signal detection experiment is presented to demonstrate relationships and intercorrelations among the measures.

C. Marlin Brown

[1] D. M. Green,et al. Signal detection theory and psychophysics , 1966 .

[2] Larry Hochhaus,et al. A table for the calculation of d' and BETA. , 1972 .

[3] J. Swets,et al. A decision-making theory of visual detection. , 1954, Psychological review.

[4] H. J. Jerison. Signal Detection Theory in the Analysis of Human Vigilance1 , 1967, Human factors.

[5] J. Grier,et al. Nonparametric indexes for sensitivity and bias: computing formulas. , 1971, Psychological bulletin.

[6] A Craig,et al. Nonparametric Measures of Sensory Efficiency for Sustained Monitoring Tasks , 1979, Human factors.

[7] William Hodos,et al. Nonparametric index of response bias for use in detection and recognition experiments. , 1970 .

[8] Gerald M. Long,et al. Limitations on the Practical Applicability of d′ and β Measures , 1981 .

[9] James P. Egan,et al. Signal detection theory and ROC analysis , 1975 .

[10] J. B. Sidowski,et al. Experimental methods and instrumentation in psychology , 1966 .

[11] Donald A. Norman,et al. A non-parametric analysis of recognition experiments , 1964 .

[12] R. E. Pastore,et al. SIGNAL DETECTION THEORY: CONSIDERATIONS FOR GENERAL APPLICATION , 1974 .

[13] Allen Parducci,et al. Limits on the applicability of signal detection theories , 1970 .