Profiles, Judges performance

Graphs of the individual scores

The software proposes a set of 2 or 3 dimensional representations allowing the comparison of individual taster scores to the statistics obtained from the whole panel. The statistics to be displayed can be chosen.

Example of a histogram of individual scores

Panel and judge performance

Judge performance graphs and tables allow to analyze repeatability and discrimination of the judges and the panel from profile data.

Example of a judge performance graph

Descriptive statistics

Means, ranges, variation coefficients, standard deviations, confidence intervals. The software allows customizing the editions (data to compute, presentation by product or descriptor, summary tables...).

Example of descriptive statistics results

Profile graphs

The software proposes several types of representations: profile or spider plot, 3 dimensional bar chart, 2 dimensional bar chart by product or descriptor. The statistics to be displayed can be chosen.

Example of a spider plot

Distribution

Edition and graphic representation of the distributions, with or without Gaussian distribution, normality tests (Chi-Square, Kolmogorov-Smirnov), medians, quartiles.

Example of a distribution graph

Box-whisker plots

Graphic representation of box-whisker plots (median, quartile, outliers...)

Example of a box-whisker plot

Analysis of Variance

1 to 3 factors with interactions. Fixed, random or mixed model. Computation of missing data. Choice of the interactions to integrate in the error term. Interaction details. Multiple mean comparison tests: Duncan, Newman-Keuls, L.S.D., Bonferroni, Scheffé, Tukey, Dunnett (comparison of the means to a control). These tests may be done at 1, 5 and 10% risk levels. Their triggering threshold is customizable. The software also includes the computation of the ANOVA with Balanced Incomplete Blocks.

Analysis of Variance

Student T test

The software computes the Student T test on paired or independent samples. It computes the normality and the homogeneity of the variances of the compared populations.

Non parametric comparison tests

The software computes the Wilcoxon (comparison of two paired samples), Mann-Whitney (comparison of two independent samples), Kruskal-Wallis (multiple comparison of independent samples), and Friedman tests. McNemar and Cochran tests allow to analyze related series with two categories.