2024-03-29T12:46:28Z
https://nagoya.repo.nii.ac.jp/oai
oai:nagoya.repo.nii.ac.jp:02002386
2023-01-16T04:57:38Z
323:350:957:1649916810434
Unobserved Errors in Web Surveys : Focusing on the Social Stratification of Registered Monitors
Web調査における非観測誤差 : 登録モニターの社会階層に着目して
澤田, 涼
SAWADA, Ryo
古殿, 真大
FURUDONO, Shinta
藤川, 寛之
FUJIKAWA, Kanno
島袋, 海理
SHIMABUKURO, Kairi
内田, 良
UCHIDA, Ryo
Web surveys are an effective way to approach educational issues. However, there is not much published about errors in the surveys. In this article, we focus on the unobserved error, which is one of the most frequently discussed error in a web survey. We then examine the biases existent in the attributes of the monitors used by marketing research companies that obtain basic information for estimating the unobserved error. Previous studies on this subject have compared marketing research companies and summarized the characteristics of respondents to surveys, putting an emphasis on the disclosure of monitor information relegated to future tasks. Therefore, examining the characteristics of the monitors owned by marketing research companies can be considered an urgent issue in the promotion of reliable web surveys. For example, an examination of Macromill Inc., one of the leading marketing research companies in Japan, finds the number of monitors has more than doubled in 10 years. Therefore, updating the information regarding the monitors should be addressed. In addition, information regarding the monitors of partner companies that have added data to Macromill monitors should also be clearly stated. As a specific task, for variables related to social stratification (occupation and personal income) and variables representative of respondents’ attributes (age and gender), we compared the monitors of Macromill (“Macromill alone”) and the monitors when affiliated companies are included with those of Macromill (“Macromill, including affiliated companies”) with the national survey conducted by the government (“National Census.”) Our results found, in terms of age, the proportion of those surveyed “under 12” and “60 and over” of “Macromill alone” and “Macromill, including affiliated companies” was lower; also, “20-59 years old” was higher than that of “The National Census.” In terms of gender, the percentage of “female” respondents for “The National Census” was higher than that of “Macromill alone” and for “Macromill, including affiliated companies.” On the other hand, there was no significant difference among the three groups in terms of occupation and high-income bracket. Thus, this research discovered gaps in these research surveys that have existed for more than a decade. Results showed that a wide range of people, some consisting of those who participated in a web survey for purposes other than “rewards,” are registered as monitors. The characteristics of Web survey participants are not only the attributes revealed in this article, but also suggest differences relating to the purpose of participating in a survey as a monitor.
departmental bulletin paper
名古屋大学大学院教育発達科学研究科
The Graduate School of Education and Human Development, Nagoya University
2022-03-31
application/pdf
名古屋大学大学院教育発達科学研究科紀要. 教育科学
2
68
167
177
Bulletin of the Graduate School of Education and Human Development. Educational Sciences
1346-0307
https://nagoya.repo.nii.ac.jp/record/2002386/files/nueduca6802_12.pdf
jpn