Figures & data
Fig. 1 The upper graph presents the model at some typical θ value and the observed data value ; the lower graph records in addition the p-value and the likelihood for that θ value, and also in dots the density using the maximum likelihood
value for the parameter.
![Fig. 1 The upper graph presents the model at some typical θ value and the observed data value yobs; the lower graph records in addition the p-value and the likelihood for that θ value, and also in dots the density using the maximum likelihood θ=θ̂obs value for the parameter.](/cms/asset/be58ead4-7503-4b1d-9490-3581e556e498/utas_a_1556735_f0001_c.jpg)
Fig. 2 The upper graph presents the p-value function and the lower graph the likelihood function, for the simple Normal example; the median estimate of θ is which here is the maximum likelihood value
.
![Fig. 2 The upper graph presents the p-value function and the lower graph the likelihood function, for the simple Normal example; the median estimate of θ is θ̂0.50 which here is the maximum likelihood value θ̂obs=10.](/cms/asset/d9995bbf-c52a-4c87-a34b-166cd27cf1e0/utas_a_1556735_f0002_c.jpg)
Fig. 3 The upper graph (a) indicates the median estimate and the one-sided 0.975 and 0.025 confidence bounds; the lower graph (b) records the observed likelihood function.
![Fig. 3 The upper graph (a) indicates the median estimate and the one-sided 0.975 and 0.025 confidence bounds; the lower graph (b) records the observed likelihood function.](/cms/asset/3e83963f-b786-47e5-a794-c4b5088a91b0/utas_a_1556735_f0003_c.jpg)
Fig. 5 The sample space for a sample of 2 from the uniform distribution on . The observed data point is
; the square is the sample space for
. Only θ-values in the range 8.9–9.1 can put positive density at the observed data point. The short line through this point illustrates the small data range consistent with the observed data.
![Fig. 5 The sample space for a sample of 2 from the uniform distribution on (θ−1/2,θ+1/2). The observed data point is (9.4,8.6); the square is the sample space for θ=θ̂obs=y¯obs=9. Only θ-values in the range 8.9–9.1 can put positive density at the observed data point. The short line through this point illustrates the small data range consistent with the observed data.](/cms/asset/50923a79-8b60-4b7c-b235-6b5187e2f184/utas_a_1556735_f0005_b.jpg)
Fig. 9 For the simple exponential model, the p-value function is plotted using the Normal approximation for r (dotted), using the third-order approximation (dashed), and compared to the exact p-value function (solid).
![Fig. 9 For the simple exponential model, the p-value function is plotted using the Normal approximation for r (dotted), using the third-order approximation (dashed), and compared to the exact p-value function (solid).](/cms/asset/c79363ea-a3ae-4c5f-8e03-67ed459c3500/utas_a_1556735_f0009_c.jpg)
Fig. 10 A log-likelihood function with the log-likelihood ratio and the standardized maximum likelihood departure
identified.
![Fig. 10 A log-likelihood function with the log-likelihood ratio r2/2 and the standardized maximum likelihood departure q=(φ−φ^)Jφφ1/2 identified.](/cms/asset/65eab86a-915d-42d5-9f3f-58cc0f2f6fca/utas_a_1556735_f0010_b.jpg)
Fig. 11 The p-value function and log-likelihood function for the mean of the Gamma model data from Gross and Clark (Citation1975).
![Fig. 11 The p-value function and log-likelihood function for the mean of the Gamma model data from Gross and Clark (Citation1975).](/cms/asset/1c56b094-8a68-41c2-ae19-356267e9491a/utas_a_1556735_f0011_c.jpg)