Tuesday, December 3, 2019

Response set and bias in surveys

Response set is a tendency to respond similarly to all or many questions such as frequently chosing "somewhat agree" on scale options ranging from "strongly agree" to "strongly disagree."

Response bias occurs when respondents deliberately give false responses.

There are many forms of response bias.

Acquiescence bias occurs when respondents select only positive answers. This is also called "yea-saying."

Demand characteristics influence answers to survey items when the respondent attempts to provide answers according to the way they think an ideal participant should respond.

Extreme bias occurs when respondents frequently choose the extreme options on survey items such as the "Strongly Agree" and "Strongly Disagree" options.

Hostility bias occurs when respondents feel provoked by items in the survey. Researchers must take care in wording items that may be sensitive. Explanations and instructions might help.

Nay-saying is the opposite of Acquiescence bias. Respondents select only, or mostly, negative responses.

Nonresponse bias refers to suspected differences between the people who respond to take a survey and those who do not.


Ad. Learn more about creating surveys available on AMAZON.















Prestige bias is like social desirability bias focused on a specific aspect of a tendency to want to appear as having a higher social status in terms of a culture's values such as education, wealth, or social power.

Primacy bias or primacy effects occur when respondents choose the first available response to each item.

Recall bias occurs whenever respondents rely on their memory to respond to survey items. Human memory is not infallible and subject to many biases.

Recency bias, or recency effect, is the tendency of respondents to choose a response based on their previous--most recent--response. Participants may disengage after a long survey.

Response order bias occurs when respondents do not carefully weigh all the options but chose one that comes easily to mind. Context can make a difference. Contrast effects can be seen when the order of questions reveals large differences in the responses obtained. Assimilation effects occur when the order of survey items leads to more similar responses.



Self-selection bias occurs when people participate who were not chosen to be a part of the sample.

Social response bias, also called social desirability bias, refers to a tendency of respondents to over-report socially desirable or "good" responses.

Sponsorship bias occurs when respondents are aware of who is sponsoring the survey and their perception of that organization influences their responses.

Stereotype bias occurs when items evoke a personal response, which activates a respondent's stereotypes. Stereotypes are widely held  simplistic and relatively fixed beliefs about groups of people such as "all men or all women" and "all Blacks or all Whites." Stereotypes can also exist about companies, things, and ideas.

Straight lining occurs when respondents choose the same answer. Sometimes this can be avoided by using reverse scaling of items or identified by including items that people would rarely endorse.

Related Issues

Satisficing is a term referring to the degree to which a respondent processes the survey item. Some may quickly respond and others may think carefully about the item.

Commonly misreported topics in surveys include abilities and skills, personality characteristics, sexual behavior, religion and spirituality, income, and unlawful behavior.

Response bias is difficult to eliminate.

Connections

   My Page    www.suttong.com
   My Books   AMAZON     GOOGLE PLAY STORE
   FACEBOOK   Geoff W. Sutton
   TWITTER  @Geoff.W.Sutton

Publications (many free downloads)
  Academia   Geoff W Sutton   (PhD)     

  ResearchGate   Geoffrey W Sutton   (PhD)

If you are a counselor, you may find this book helpful. It is also available on AMAZON.




No comments:

Post a Comment

Death and Treatment Need Statistics 2020 Pandemic

Statistical models are needed to guide government leaders and health service providers concerned with maximizing the number of infec...