Veljavnost osebnega in spletnega anketiranja v geografskem raziskovanju // Validity of face-to-face and web surveys in geographical research

摘要

Osebno anketiranje, ki predstavlja tradicionalno metodo zbiranja podatkov, je v zadnjem desetletju za?elo zamenjevati spletno anketiranje. Do danes se je njegova uporaba mo?no raz?irila in je ena izmed vodilnih metod zbiranja podatkov nove generacije. ?eprav je njegova uporaba vse pogostej?a, pa se poraja dvom o njegovi veljavnosti in s tem uporabnosti, saj so v spletne ankete v nasprotju od tradicionalnih oblik anketiranja najpogosteje vklju?eni neverjetnostni vzorci, pri katerih je vklju?itev v tovrstne raziskave odvisna od udele?encev samih. Preverili smo veljavnost osebnega in spletnega anketiranja ter ugotavljali, ali se s spletnim anketiranjem v obliki, kot se v dana?njih raziskavah vse bolj pojavlja, lahko pribli?amo reprezentativnosti vzorca v enaki meri kot pri tradicionalni metodi osebnega anketiranja, izdelani na podlagi verjetnostnih vzorcev. Rezultati so pokazali, da med izbranima metodama obstajajo zna?ilne razlike v reprezentativnosti vzorca, pri ?emer gre v najve?ji meri za razlike v demografski in izobrazbeni strukturi vzorcev.Face-to-face surveys, which represent a traditional method of data gathering, are gradually being replaced by web surveys during the last decade. Their use has widely spread and today it represents one of predominant data gathering methods of the new generation. Although their use is becoming increasingly common, doubts have been arising about their usefulness. Namely, as opposed to traditional forms of surveying, web surveys include non-probability samples in which inclusion in surveys depends on participants themselves. We checked their validity and tried to establish whether sample representativeness in the types of web surveys that are nowadays increasingly used in research can be approached to such an extent as in the traditional method of face-to-face surveys based on probability samples. The results show that significant differences in sample representativeness exist between the selected methods. They are largely a consequence of demographic structure of samples.