Academic journal article The Journal of Research in Business Education

Analysis of Nonresponse Bias in Research for Business Education

Academic journal article The Journal of Research in Business Education

Analysis of Nonresponse Bias in Research for Business Education

Article excerpt


This research examined the issue of nonresponse bias and how it was reported in nonexperimental quantitative research published in the Delta Pi Epsilon Journal between 1995 and 2004. Through content analysis, 85 articles consisting of 91 separate samples were examined. In 72.5% of the cases, possible nonresponse bias was not examined in the respective study. Of those who reported addressing the issue, comparing respondents to nonrespondents was the technique most frequently employed to understand possible systematic differences between the groups. Chi-square analyses did not reveal statistically significant differences between sampling technique (probability and nonprobability), response rate, and nonresponse bias checking. Response rates for the studies ranged from 12%-100%, with a mean of 62.2%. The implication for the external validity of business education research is discussed and future research is recommended.


WHEN A RESEARCHER CONDUCTS EMPIRICAL RESEARCH, the implementation and thorough reporting of scientific methods distinguishes between mediocre and superior research. Typically, when reporting research methods, researchers include sections on participants (population/ sample), instrumentation, data collection (including followup procedures), and data analysis. When conducting quantitative research, the process should include developing research questions, selecting participants, identifying specific methods to answer the questions, selecting analysis tools, analyzing the data, and subsequently interpreting the results (Holton & Burnett, 2005). Each step in this process is vital to creating reliable and valid research results.

Figure 1 presents the conceptual framework to guide the collection of reliable and valid data in terms of defining the population, sampling, collecting data, following-up in the data collection phase, and analyzing for nonresponse bias, often an issue in survey research (Rogelberg & Luong, 1998).

Because one of the strengths of quantitative research is the ability to make statements about larger groups from smaller samples, it is crucial that the collected data are representative of the larger group. While the particular research methods are crucial to producing findings that are reliable and valid, selecting participants and collecting data are key components of getting findings that are generalizable, i. e. having external validity. Without external validity, researchers cannot appropriately make generalizations about a smaller sample to predict some attribute of the larger target population.

One of the major issues associated with sample data representativeness is possible systematic bias introduced through low response rates to surveys (Rogelberg & Luong, 1998). Not only do low response rates mean smaller sample sizes, thereby negatively influencing statistical power and increasing the size of confidence intervals around sample statistics, low response rates can also produce biased samples of respondents and reduce the perceived integrity of the survey results (Rogelberg, Conway, Sederburg, Spitzmuller, Aziz, & Knight, 2003). Thus, one goal of researchers should be to optimize survey response rates. Although there are other possible sources of systematic bias in social science research, for example, purposive sampling (Passmore & Baker, 2005), this study focuses on survey nonresponse bias. Survey nonresponse bias refers to the possible bias introduced into a study when nonrespondents differ systematically from the respondents in one or more ways. For example, when studying employee satisfaction, dissatisfied employees may be more likely to participate, thereby distorting the true level of employee satisfaction. Thus, for generalization purposes, it is important to have an understanding of how respondents and nonrespondents compare on the study's respective research variables.

Purpose and Significance of the Study

Much of the progress of the behavioral sciences has been, in part, due to the efforts of researchers to produce reliable and valid techniques to measure social variables (Linder, Murphy, & Briers, 2001). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.