Critical Analysis of Sampling Designs in Human Geography Research by Focusing on Selection Bias in Probability Samples

• Main Authors: Masoud Safaei Pour, Alireza Rahimi Chamkhani
• Format: Article
• Language: fas
• Published: Afarand Scholarly Publishing Institute 2018-03-01
• Series: تحقيقات جغرافيايی
• Subjects: probability sampling; sampling design; selection bias; reliability and generalizability; human geography research
• Online Access: http://georesearch.ir/article-1-365-en.html

Using of words " probability and random" has such a special meaning and in research sampling method from the point of view of research method principles, it means respecting the principles and rules of sampling theory and lack of selection bias in sampling design that guaranties reliability and generalizability of estimates. By assuming the existence of selection bias sample in probability sampling designs in human geography research, this research evaluates this assumption by using of a critical approach and qualitative method. In  this respect, to explore and understand the issue in detail, 25 samples of articles  with probability or random sampling method were selected by using of targeted and criterion-oriented sampling method through articles published in scientific geographical research journal of country in the last five years, but there is evidence and indications that oppose the above claim, and indicate not respecting the principles and rules of probability  sampling (the possibility  of existence of selection bias) in these samplings. Then aspects of the problem were studied by content analysis and citation analysis in each sampling separately. The results showed that in all studied samples, the special scientific claim of the samples being "probable or random" were not well founded, and applied sampling method was practically not based on probability in most of samplings, maybe because of lack of adequate knowledge of principles and rules of sampling theory. Some of them had considered the idiomatic and popular meaning of words " probability and random or accidental", and some despite being aware of scientific importance and meaning of these concepts, had not used them properly and had mentioned using of these method very briefly, and did not present a clear image of their mechanism, while there are in some cases the evidences indicating the reversal of above claim in their content. Finally, there is possibility of existence of selection bias in all of them and their estimates did not have reliability and generalizability. Due to extension of this issue in journals, research and different academic institutions of production science all around of country, serious consideration of this issue seems necessary in order to defend the prestige and the status of geography science as programming and development science.