Social desirability bias of questionnaire survey

Social desirability bias (SDB) arises because respondents like to appear to be other than they are. This can occur consciously, because respondents want to manage the impression that they are giving of themselves in terms of social responsibility, or subconsciously, because they believe themselves to be other than they are, possibly a form of denial. Thus SDB can manifest itself both in stated behaviour, with, say, an over-claiming of environmen­tally friendly behaviour, or in the attitudes that someone expresses.

Sudman and Bradburn (1982: 32-33) identified the following topics as being desirable and therefore areas in which behaviour is likely to be over-reported:

• Being a good citizen:
  • registering to vote and voting;
  • interacting with government officials;
  • taking a role in community activities;
  • knowing the issues.
• Being a well-informed and cultured person:
  • reading newspapers, magazines and books, and using libraries;
  • going to cultural events such as concerts, plays and exhibitions;
  • participating in educational activities.
• Fulfilling moral and social responsibilities:
  • giving to charity and helping friends in need;
  • actively participating in family affairs and child rearing;
  • being employed.

They also quote examples of conditions or behaviour that may be under­reported in an interview:

• Illness and disabilities:
  • cancer;
  • venereal diseases;
  • mental illness.
• Illegal or contranormative behaviour:
  • committing a crime, including traffic violations;
  • tax evasion;
  • drug use;
  • consumption of alcoholic products;
  • sexual practices.
• Financial status:
  • income;
  • savings and other assets.

Until relatively recently, SDB was seen as an issue mainly affecting social research, as the above list suggests. Thus, it has been a problem in health care, where people might claim to lead a healthier lifestyle than is the case. It has been an issue for social researchers in a range of issues such as immi­gration, attitudes to minority groups, housing, public transport and the environment. If it has affected market researchers, it has been an issue mainly for a small number of specific categories in which there is a perceived element of social responsibility, or perceived social irresponsibil­ity. In certain markets, such as tobacco, alcohol and gambling, both atti­tudes and behaviour are likely to be misrepresented. Many respondents will deliberately under-report their consumption in these markets in order to appear socially responsible, while others may over-report their consumption, particularly in the alcohol market, in order to appear more ‘macho’ to the interviewer. Researchers working in these fields have learnt that they cannot ignore SDB as an influence on the data that they collect.

More recently, though, the rise in the association between many types of businesses and the impact that they have on both the physical and social environments has meant that this has become an issue for researchers working in many more fields:

• For consumer goods companies and retailers it can arise with consumer concerns about the impact on the environment of excessive or inappropriate packaging.
• The social responsibility of food and confectionery manufacturers to their customers and suppliers has become a global issue.
• For manufacturers of consumer durables the impact of the disposal of their products can be a social concern.
• Cause-related marketing has been adopted by many organizations in recent years, in which the brand is linked to a good cause, such as supporting schools.
• Issues such as ‘fair trade’ products arise in individual markets.

It can no longer be assumed that SDB is an issue only for social researchers. Researchers in commercial markets now have to be equally aware of it.

In many areas of commercial market research, if the questionnaire writer and researcher fail to recognize that SDB may be influencing responses, then they may come to false conclusions from the research data.

Types of SDB

1. Impression management

Possibly the most common form of SDB is the need for approval, known as ‘impression management’. This is partly a function of the individual and partly a function of the question, and its occurrence varies depending on a combination of the two. Some people will answer honestly certain questions but will not do so other questions where they feel the need for approval. The questions or topics on which people feel the need for approval may vary between respondents. However, within any one study it is most likely that if impression management occurs, it will do so on a small and consistent set of questions.

2. Ego defence and self-deception

Maintaining one’s own esteem is a further cause of bias. Here respon­dents’ intentions are not to manage the impression that they give to someone else, such as the interviewer or the researcher, but to convince themselves that they think and behave in socially responsible ways. This is less likely to be a conscious activity than is the need for approval, but can result in the same exaggeration of claimed socially responsible behav­iour and attitudes. This type of behaviour may particularly affect future projections of likely behaviour, where the respondents convince them­selves that they will behave in a responsible fashion in the future even if they do not do so currently. When this is carried out consciously it is known as ‘ego defence’; when it is carried out subconsciously it is known as ‘self-deception’.

3. Instrumentation

A further type of bias, and one that is totally conscious, is instrumentation (Nancarrow, Brace and Wright, 2000). This means that respondents give answers designed, in their own view, to bring about a socially desirable outcome. Respondents may say that they will participate in a scheme or purchase a product, for example, although they know that it is unlikely that they will. They do so because they believe the introduction of that scheme or product is desirable. A survey of attitudes to how lottery money should be divided between good causes and lottery administra­tors may suffer from this effect, for example. Respondents may deliber­ately give low estimates of the proportion that should be allocated for administration because they believe that if it is seen that the public wants a higher proportion to go to charities this could have an impact on the decisions of the regulatory body. This may be in addition to or in place of impression management, in which the respondent wishes to be seen by the interviewer to be generous to charities. Many respondents are rela­tively sophisticated with regard to marketing and to market research, and know that they have an opportunity to influence decision making through their responses to the survey.

Source: Brace Ian (2018), Questionnaire Design: How to Plan, Structure and Write Survey Material for Effective Market Research, Kogan Page; 4th edition.