The oroclinal tests uses bootstrapped (Gaussian or Standard) linear regressive techniques to determine the relationship between bi-dimensional data (e.g. strike and declination).
Data file must contain tab delimited Declination, Declination Uncertainty, Strike, Strike Uncertainty, (code). An example input file is here.
Data are bootstrapped and randomly sampled from a confidence interval (95%) and mean.
Figure 1 - Oroclinal test showing the sampled declinations and strikes (blue dots) with measurement uncertainties (black bars). The red line shows the total least squares regression for the data. The surrounding shaded red area illustrates the confidence interval for 1000 bootstrapped regressions. For comparison the average bootstrap shown in green.
Figure 2a, b - Showing the slope (left) and intercept (right) of the total least squares regression for the data (blue vertical) with shaded blue bootstrapped confidence interval. The green vertical indicates the average bootstrapped value. The black vertical indicates the result from a weighed regression (see text).
Residuals to Regression
Figure 3 - Illustrating the data residuals in blue dots including lines indicating 1σ, 2σ, and 3σ. The bars are histograms for the data and bootstraps in blue and orange respectively. These should approximate a normal distributed. The length of each bin is represented by a sample (scaled by -1000 for the bootstraps).
Figure 4 - Quantile - Quantile plot for the selected residuals. A normal distribution of residuals would approximate a straight line.
Figure 5 - Oroclinal foldtest showing the circular variance as a function of unfolding. A smaller circular variance indicates a tighter clustering of the data. The bolded red line represents the actual data and is surrounded by the first 25 bootstraps shaded in blue.
Awesome Tip: Click the thick red line to view the declination distribution for that percentage of unfolding.