Original articleComparability of epidemiological information between self- and interviewer-administered questionnaires
Introduction
Self-administered questionnaires and interviewer-administered questionnaires have been widely used to assess health status and develop policies for disease prevention and health promotion in epidemiological studies. Although a number of studies have reported good agreement between data obtained by questionnaire and by personal interview 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, it has been recently recognized that respondents tend toward more socially desirable responses (i.e., the tendency to respond in such way as to avoid criticism) in interviews than with self-administered questionnaires 3, 5, 9, 13, 14, 15, 17, 19, 21, 22, 23, 26, 27. It has also been reported that there is a lower chance of biased responses, a higher complete assurance of confidentiality, and greater validity and accuracy with self-administered questionnaires [19]. These findings indicate that the extent of the effect to which social desirability of responses is a biasing factor [21] in an interview needs to be understood and minimized. To date, however, most epidemiological studies have evaluated interview data without consideration of the effect of social desirability. Few studies have systematically and quantitatively examined the extent of the effect of social desirability on responses concerning lifestyle factors in the community [10], although this would be valuable information when interpreting interview data.
Moreover, although many studies have examined the factors that influence socially desirable responses, most have focused on respondent-related factors such as gender 9, 13, 16, 21, 25, 28 and age 14, 28, 29, 30, 31. The style, training, age, and gender of the interviewer may also be factors; interviewers trained to have detached and dispassionate attitudes deter respondents from answering sensitive items [10]. Choi et al. reported that the performance and characteristics of interviewers affect the responses when evaluating subjects' mood [32]. Siemiatycki indicated that interviewers should strive to make a relaxed and less stressful atmosphere [2].
It is also indicated that selection of interviewers with similar backgrounds and characteristics helps to minimize the interviewer effect [32]. The effect of interviews, especially the relationship between interviewer and respondent with socially desirable responses, must inevitably be considered. However, few studies have quantitatively assessed the strength of the effect of the relationship between interviewer and respondent with regard to social desirability.
Therefore, the purposes of this study were (1) to show quantitatively the extent of social desirability on self-reported lifestyle factors and (2) to clarify the factors related to socially desirable responses, using univariate and multivariate analyses adjusted for the interaction of effects among all variables added in the model.
Section snippets
Methods
The subjects were 234 (169 male and 65 female) workers at a steel company near Nagoya City, aged 19–67 years (mean age, 43.8 years; standard deviation [SD], 12.4 years). All subjects were randomly assigned to receive either the self-administered questionnaire followed approximately 2 weeks later by the interview or the reverse order. For the self-administered questionnaire, subjects were asked to complete the questionnaires and return them within 2 weeks. When a questionnaire was not returned
Results
The response rate was 100% with both methods. Table 2 shows the distribution of the characteristics of the subjects according to the data collection methods. There was no significant difference in demographic characteristics between the two method groups.
Table 3 shows the agreement for each lifestyle factor between the self-administered and interview questionnaires. Percentage agreement exceeded 75% for all lifestyle factors. Kappa values ranged from a minimum of 0.43 (eating breakfast) to a
Discussion
The purposes of this study were (1) to examine agreements and discrepancies between responses about lifestyle factors from the self-report and interview data, (2) to examine gender and age differences in the discrepancies between the two data collection methods, and (3) to examine the effect of the respondent/interviewer relationship (i.e., age differences between respondents and interviewers). Overall, the agreements of responses obtained with the two data collection methods were good or
References (48)
- et al.
The assessment of life eventsself-administered questionnaires versus interviews
J Affect Disord
(1986) - et al.
Interviewer versus self-administered questionnaires in developing a disease specific, health-related quality of life instrument for asthma
J Clin Epidemiol
(1993) - et al.
Reports of smoking in a national surveydata from screening and detailed interviews, and from self-and interviewer-administered questions
Ann Epidemiol
(1998) - et al.
Reports of smoking in a national survey:data from screening and detailed interviews, and from self- and interviewer-administered questions
Ann Epidemiol
(1998) - et al.
A comparison of alcohol, tobacco and drug use as determined from household and school surveys
Drug Alcohol Depend
(1985) - et al.
A comparison of observed and reported restraint use by children and adults
Accid Anal Prev
(1985) - et al.
Relationship of health status and health practices
Prev Med
(1972) - et al.
Response error current smoking frequency by black and white established smokers
Ann Epidemiol
(1996) - et al.
Validity of residents' self-reported cardiovascular disease prevention activitiesthe Preventive Medicine Attitudes and Activities Questionnaire
Prev Med
(2000) - et al.
Response effect surveya review and synthesis
(1974)
A comparison of mail, telephone, and home interviews strategies for household health surveys
Am J Public Health
Apparent validity of alternative survey methods
J Community Health
Obtaining information on patient satisfaction with hospital caremail versus telephone
Health Serv Res
Comparison of mail questionnaires and telephone interviews as data gathering strategies in a survey of attitudes toward restriction on cigarette smoking
Can J Public Health
Comparison between data obtained through questionnaires and interviewslife-style habits of pregnant women
Scand J Soc Med
Assessing alcohol's consumption in general practice patentsa comparison between questionnaires and interviews
Alcohol Alcohol
Sex differences, social desirability, masculinity, and the tendency to worry
J Genet Psychol
Modes effect in survey researcha comparison of mail, telephone, and face to face survey
Bull Method Sociol
A comparison of mail and telephone interviews strategies for mental health surveys
Can J Psychiatry
Gender-controlled measures of socially desirable responses
J Clin Psychol
Relationship of self-concept and social desirability tendency of Hong Kong Chinese adult with physical disabilities
Percept Mot Skills
Influence of social desirability response set on self-report for assessing the outcome of treated alcoholics
Alcohol Clin Exp Res
Social desirability bias in dietary self- report may compromise the validity of dietary intake measures
Int J Epidemiol
Are women heavier drinkers than we thought they were
J Stud Alcohol
Cited by (72)
Public perceptions, knowledge, responsibilities, and behavior intentions on marine litter: Identifying profiles of small oceanic islands inhabitants
2023, Ocean and Coastal ManagementCitation Excerpt :This means that participants can answer in a way they consider socially accepted by others to gain their approval (King and Bruner, 2000). To avoid this effect, the questionnaires were completed without the presence of the researchers, thus circumscribing the putative inflated answers (Okamoto et al., 2002). A random sampling of individuals was undertaken during this study.
Childhood maltreatment class and sexually violent behavior among university men in Vietnam
2022, SSM - Population HealthCitation Excerpt :Both the childhood maltreatment measure and Sexual Experiences Survey use multiple, behaviorally-specific items to capture experiences, which improves content validity (Diamantopoulos, Sarstedt, Fuchs, Wilczynski, & Kaiser, 2012), and tends to stimulate recall and reduce under-reporting (Khare & Vedel, 2019). Moreover, self-administration of the questionnaire allowed for greater privacy, which may have enhanced disclosure of potentially sensitive or shameful experiences (Okamoto et al., 2002). Despite these efforts, the potential for biased reporting may not be fully mitigated; however, under such circumstances, our findings may provide a lower bound for the effects of childhood maltreatment class on sexually violent behavior.
Effect of mode of administration on Edinburgh Postnatal Depression Scale in the South Indian population: A comparative study on self-administered and interviewer-administered scores
2021, Asian Journal of PsychiatryCitation Excerpt :Though EPDS was originally developed as a self-administered tool, it has been validated for an interviewer/expert administration (Upadhyay et al., 2017; de Figueiredo et al., 2015; January and Chimbari, 2018). Earlier research studies on socially and self-stigmatized conditions such as depression suggest differences in how respondents answer sensitive questions dependent on the mode of assessment, mainly self-administered mode in comparison with an interviewer-administered mode (Krumpal, 2013; Moum, 1998; Okamoto et al., 2002). Differences in the mode of administration are highly relevant in a clinical assessment, especially in linguistically and culturally diverse countries such as India where research studies in perinatal mental health have been very limited.
Marine litter: A review of educative interventions
2021, Marine Pollution BulletinA minimally processed dietary pattern is associated with lower odds of metabolic syndrome among Lebanese adults
2018, Public Health Nutrition