Wyniki wyszukiwania - BazTech

Ograniczanie wyników

Znaleziono wyników: 1

Liczba wyników na stronie

Wyniki wyszukiwania

Sortuj według:

Ogranicz wyniki do:

Statystyka a prawda

Czerwiński Z.

Badania Operacyjne i Decyzje

2001

nr 1

41-50

Omówiono przydatność statystyki do opisu i badania złożonych zjawisk masowych. Podkreślano rolę konwencji zarówno w opisie, jak i wnioskowaniu statystycznym. Pytanie, czy przyjmowane w badaniach statystycznych konwencje sprzyjają odkrywaniu prawdy jest trudnym do rozstrzygnięcia problemem.

The language of statistics is hardly understable to the man in the street because statistical statements describe mass phenomena and not individual facts observed by the man, often seemingly inconsistent with statistics. Statistical description is the less accessible to the public, the more complex (multidimensional) phenomena are subject to the statistical investigation. Statistical indicators used to measure the intensity of complex phenomena are based on conventions. Different conventions result in different conclusions. It is hardly possible to agree upon one ("proper") indicator to measure the intensity of a given phenomenon (welfare, level of development, etc.). Statistical interference (from random sample to population) is fallible. It may be to some extent justified by referring to probability theory. Quality of statistical inference is judged by means of some conventional characteristics, e.g., the probability of the 1-st and 2-nd type error (which is meaningful only is sampling is repeatable). The latter, however, can be estimated only in exceptional cases. Statistics (both description of populations and rules of statistical inference) is a set of conventions. To what extent the conventions contribute to discovering the truth - this is an open question.