A new method is proposed that sharply increases the amount of statistical information extracted from experimental data. The new statistical technology is particularly suited for scarce sample populations, when it increases with one order of magnitude the extracted information. Hidden features would remain inaccessible otherwise. The method could prove useful also for signal processing, when a high level of accuracy in details rendition is required. The efficiency of the method is demonstrated by the experimental measurement of the combustion heat of rocket propellants. Populations as small as of six elements are thus successfully processed and statistically characterised, while the regular statistics could only be applied to much larger populations.