Cultural and Interviewer Effects on Interview Privacy: Individualism and National Wealth

Abstract

Privacy (or the lack of it) is an important feature of the interview. Researchers rely on interviewers to ensure a private setting. In reality, interviewers are guests in respondents’ homes and might find it difficult to achieve privacy. Thus, a substantial proportion of interviews are conducted in the presence of a third party. We investigate whether there are cultural and interviewer variations in interview privacy and whether respondent characteristics associated with interview privacy vary by culture. We analyzed data from 14 countries. The results highlight the direct and indirect role the country-level characteristics (level of individualism and wealth) play in explaining cultural differences in rates of interview privacy and the significant between-interviewer variation in interview privacy.

Keywords

interviewer variance cultural variation national wealth individualism third-person presence privacy cross-cultural

Introduction

Survey interviewing is a dynamic process that is affected by the setting of the interview. In face-to-face surveys, the setting encompasses physical features as well as events and interactions during the interview, and is usually jointly determined by the interviewer, the respondent, other household members, and the culture they live in. Surveys designed to collect sensitive information usually call for a private interview setting. Although interview privacy is a broad concept and connotes different meanings, in this article, we define interview privacy as the absence of a third person during the whole interview.

Privacy is an important feature of the interview due to its possible effect on reporting, especially the reporting of sensitive information. The literature on the effects of third-party presence on survey reporting is mixed (see Tourangeau & Yan, 2007, for a meta-analysis). Although some studies found that third-party presence is associated with reduced reporting of undesirable outcomes (W. A. Aquilino, 1993; W. S. Aquilino, Wright, & Supple, 2000; Moskowitz, 2004; Taietz, 1962), others observed increased reporting of such outcomes (Bulck, 1999; Edwards, Slattery, & Ma, 1998; Hoyt & Chaloupka, 1994). Such mixed findings could be attributed to several factors, including differential effects by respondent characteristics and cultural variations. A recent publication (Mneimneh, Tourangeau, Pennell, Heeringa, & Elliott, 2015) showed that the effect of third-party presence varied by respondent’s need for social approval and the income level of respondent’s country of residence. Other factors that could potentially explain the variation in the effects of third-party presence on reporting sensitive information are the type of question, study design features (e.g., the mode of data collection), as well as the relationship of the third party to the respondent. Regardless of the direction of the effect on reporting sensitive information, variation in interview privacy within a sample and across cross-cultural samples changes the interview experience across respondents, leading to measurement differences in the information collected and compromising comparability in multicultural surveys.

But, how common are nonprivate interviews?

Table 1 summarizes published rates of nonprivate interviews across different countries. Most studies conducted in the United States reported that a third party was present in more than 40% of the interviews with rates ranging from 21% to 59% (Anderson & Silver, 1987; Moskowitz, 2004; Pollner & Adams, 1994, 1997; Silver, Abramson, & Anderson, 1986; T. W. Smith, 1997). Most often, the third person present is a spouse or other partner (Hartmann, 1995; Pollner & Adams, 1997; Reuband, 1992; T. W. Smith, 1997). Similar ranges of rates have also been reported in several countries in Europe, including Belgium (where 48%-52% had a third party present; Bulck, 1999; Welkenhuysen-Gybels & Billiet, 2001), France (38%-53%; Festy, Gaymu, & Thévenin, 2014), Germany (24%-48%; Chadi, 2013; Hartmann, 1995; Reuband, 1992), and the United Kingdom (40%-50%; Zipp & Toth, 2002). In developing countries, rates as high as 82% (Nepal) and 67% (Nigeria) have been reported. Partners were also the most common type of third party to be present during the interview in these countries.

Table 1.

Percentage of Interviews Conducted in the Presence of a Third Party During the Interview.

Study	Country	Sample type	Presence of third party during interview
The United States
Silver, Abramson, and Anderson (1986)	The United States	Adult population in 11 National Election Surveys	Range between 41% to 57% (any third party)
Anderson and Silver (1987)	The United States	Couple immigrants from Soviet Union to the United States	52% (any third party including spouse and children)
W. A. Aquilino (1993)	The United States	Married couples from the 1987-1988 National Survey of Families and Households	36% (spouse)
Hoyt and Chaloupka (1994)	The United States	Youths (14-21) from National Longitudinal Survey of Youth in 1984 and 1988	1984: 5% (parents); 2% (friends); 9% (other household members); 3% (others) 1988: 3% (parents); 2% (friends); 12% (other household members); 3% (others)
Pollner and Adams (1994)	The United States	Adult population in Los Angeles site of the Epidemiological Catchment Area Program	48% (any third party)
W. A. Aquilino (1997)	The United States	Couples 18-45 years old	25% (spouse)
Pollner and Adams (1997)	The United States	Couples living together in Los Angeles site of the Epidemiological Catchment Area Program	48% (any third party)
T. W. Smith (1997)	The United States	Leisure Study, 1975 National Health and Social Life Survey,1992 General Social Survey, 1994	26% (any third party) 21% (any third party) 37% (any third party)
W. S. Aquilino, Wright, and Supple (2000)	The United States	Adolescents and Adults (12-34) living with parents	Parent presence varied by Interview Section from 6% to 10% Sibling presence varied by Interview section from 2% to 3%
Moskowitz (2004)	The United States	Adolescents sample ages 12-17 years old	CATI mode: 59% (any third party) T-ACASI: 42% (any third party)^a
Brenner (2012)	The United States	Adult population in 5 National Election Surveys	40% (any third party excluding children younger than 6 years old)
Europe
Taietz (1962)	The Netherlands	Elderly in two municipalities	78% (child and/or partner)
Reuband (1992)	West Germany	Adult population	28% (any third person), 17% (spouse), 7% (children), 4% (other family members), 2% (others)
Hartmann (1995)	West Germany	Adult population of German General Social Survey from 1980 to 1990	Range between 24% to 35% (any third party)
Chadi (2013)	Germany	German Socio-Economic Panel (SOEP), 1985-1996	48% (any third person)
Bulck (1999)	Belgium	Adult population in Flanders	52% (any third party)
Welkenhuysen-Gybels and Billiet (2001)	Belgium	Flemish adult population ages 18-65 years old	48% (any third party)
Zipp and Toth (2002)	United Kingdom	Couples in the British Household Panel Study	40% among Females; 50% among Males; (spouse)
Festy, Gaymu, and Thévenin (2014)	France	Couples 50-79 years old in the first wave of the Generation and Gender Survey 2005	38% among females; 53% among Males; (partner)
Developing countries
Podmore, Chaney, and Golder (1975)	Hong Kong	Household sample of young population ages 15-29 years old	51% had one or both parents present
Casterline and Chidambaram (1984)	Countries in Asia, Latin America, the Caribbean, and Africa	Females less than 45 years old	Range between 17% (Paraguay) to 82% (Nepal); Majority of countries (17 out of 24) > 40%
Diop, Le, and Traugott (2015)	Morocco	Females age 15-49 years old in Demographic and Health Survey, 2003	24% (any third person)
Herrera, Benjet, Méndez, Casanova, and Medina-Mora (2017)	Mexico	Youth ages 12-17 years old residing in Mexico City Metropolitan Area	49% (at least one adult), 4.8% (both children and adult), 3.4% (children only)
Lau et al. (2017)	Ghana, Nigeria, Uganda, Brazil, and Guatemala	Adults age 15-64 in each of the countries	Range between 36% (Brazil) to 67% (Nigeria); any third person

Note. All the studies summarized above are quantitative in nature and the majority explicitly mention that the design required interviewers to conduct their interviews in private. For the rest, it was not explicitly mentioned in the article, but because they deal with sensitive topics (such as drug use), include a more private mode, or aim to measure the respondent’s knowledge, it was assumed that interview privacy was required in their design. Moreover, none of the studies experimentally manipulated third-party presence. All studies summarized above, except two (Anderson & Silver, 1987; Welkenhuysen-Gybels & Billiet, 2001), included sensitive questions on a variety of desirable and undesirable outcomes such as reading books, voting behavior, attitudes toward marriage, attitudes toward extended family, TV viewing, mental health, smoking behavior, alcohol and drug use, and intercourse frequency. As the biggest majority included such sensitive questions and as the range of topics greatly varied within a study, any conclusion about patterns of nonprivate interviews across studies was avoided.

Interview privacy measures were collected by asking the respondent the following question over the phone: “For our final question, we would like to know if a parent could hear you being interviewed at any time?”

Thus, even when the study design calls for privacy, a significant percent of interviews is still conducted in the presence of a third person. Researchers and practitioners who wish to reduce such rates or adjust for them need to understand when such nonprivate interviews occur. We suspect that interviewers, respondents, and the culture they come from play an important role in determining the privacy of an interview.

Interviewer’s Role

During training, interviewers are generally “instructed” that interviews need to be conducted in a private setting (if the study protocol calls for it). Unfortunately, the need for interview privacy does not seem to get much attention in interviewer trainings (W. A. Aquilino, 1993; T. W. Smith, 1997; Taietz, 1962). T. W. Smith (1997) pointed out that though privacy and confidentiality are emphasized in training materials, interviewing manuals do not explicitly mention avoiding the presence of a third party. It seems likely that interviewers are mainly left to their own judgment and skills in enforcing privacy. In view of the limited attention given to interview privacy in interviewer training, one would expect interviewers to vary in their attitudes toward the importance of privacy and their ability to achieve it. Our search of the literature found one study that investigated the relationship between interviewer characteristics (such as age, gender, marital status, and years of experience) and the presence of others during the interview (Lau et al., 2017). However, that study did not estimate interviewer variation in interview privacy. If interviewers vary significantly in their rates of interview privacy, after controlling for respondent variations, this will have important practical implications related to training interviewers on requesting, negotiating, achieving, and maintaining interview privacy.

Respondent’s Role

The presence of a third party during all or part of the survey interview is not a random occurrence. Respondent characteristics and motivations and household characteristics may all play a role in the presence of others during the interview. Third parties could include the respondent’s spouse or partner, parents, children, siblings, and others. Most of the literature on predictors of third-party presence focuses on partner presence and any third-party presence.

Hartmann (1995) described a number of conditions that need to be met for a spouse to be present during the interview. Given that the respondent is married, a number of motivational, demographic, and socioeconomic factors could contribute to his or her presence.

Older respondents are more likely to have their spouse present during the interview (W. A. Aquilino, 1993; Casterline & Chidambaram, 1984; Welkenhuysen-Gybels & Billiet, 2001). This may reflect the tendency for older couples to spend more time at home or the perceived need for assistance in the interview process. In Germany and the United States, women were more often present when their husband was being interviewed than vice versa (W. A. Aquilino, 1993, 1997; W. S. Aquilino et al., 2000; Hartmann, 1995). This may be because women are more likely to be at home when their husband is being interviewed, and they also may be more interested in the interview process than men are. Whether husbands are present when their wives are interviewed may be a function of the gender of the interviewer. Hartmann (1995) found that, though women were less likely to have their spouse present than men were, when women were interviewed by males, their husbands were more likely to be present.

Respondents who are not working are usually available during the day when their spouses are likely to be at work and thus are more likely to be interviewed in private (W. A. Aquilino, 1993; Hartmann, 1995). However, spouses who are not working may spend more time at home and are more likely to be around when the respondent is being interviewed (W. A. Aquilino, 1993; Hartmann, 1995). Lower education level and lower income both increase the likelihood of having a spouse present during the interview (W. A. Aquilino, 1993, 1997). These socioeconomic differences also apply to the presence of any third party and not just spouse presence (Casterline & Chidambaram, 1984; Diop, Le, & Traugott, 2015; Festy et al., 2014). Yet none of the published work investigated whether such socioeconomic effects differ across a wide range of cultures. Understanding such differences is essential for cross-cultural comparisons.

Culture’s Role

Anthropologists have pointed out that there are considerable differences in how interview privacy is viewed across cultures (Streib, 1973). Such differences may reflect cultural dimensions that shape people’s social motivations. One key dimension is the culture’s level of collectivism (or individualism).

In collectivist cultures, the self is defined in terms of relationships with others. To maintain harmony and interdependence in such cultures, close attention is given to others in the social context, especially if they belong to the in-group circle—the individual’s family and friends and others who are concerned with his or her welfare (P. K. Smith, Bond, & Kağıtçıbaşı, 2006; Triandis, 1995). Family is usually the most important in-group and individuals are expected to obey in-group authorities (Triandis, 1989). Hofstede, Hofstede, and Minkov (2010) discussed how it is considered normal in a collectivist culture for a member of one’s in-group to invade one’s privacy. Moreover, members of the in-group distrust members of the out-group and are usually unwilling to cooperate with them (Triandis, 1972). Thus, although one is normally prosocial toward in-group members, one can be quite rude to out-group members (Triandis, 1989). Gabrenya and Barba (1987) reported that collectivists are not as effective in meeting strangers as are individualists, who meet people easily and cooperate with them even when they do not know them well.

As a result of these differences, in collectivist cultures, interviewers might find it difficult to request privacy if the respondent insists that a family member be present or if family members insist on staying during the interview. This is in contrast to individualist societies where the right to privacy is a central theme. In their article, Lau et al. (2017) briefly discussed how the collectivist orientation in developing countries might affect the type of third-person presence. However, their study focuses on five countries (Ghana, Nigeria, Uganda, Brazil, and Guatemala) and does not include a mix of collectivist and individualist cultures. This did not allow the authors to estimate differences in interview privacy due to variations in the country’s collectivist orientation.

One of the major antecedents to individualism is affluence (Triandis, 1989). As societies become more affluent, individuals become less dependent on their in-group and become more interested in self-expression. Hofstede et al. (2010) reported that Gross National Income (GNI) per capita was found to explain 71% of the nations’ individualism scores. Thus, an increase in the wealth of the country could affect interview privacy not only through the possible increase in physical opportunities available for individuals (such as more living space) but also through the shift in their values toward more individualistic views that emphasize the protection of privacy.

Although there are other cultural dimensions that could be also investigated including uncertainty avoidance and indulgence (Hofstede et al., 2010), we mainly explore collectivism because the core of its definition revolves around giving attention to in-group members and specifically family members, who are the ones most likely to be present during the interview.

Thus, cultural and sociodemographic characteristics might make it easier or harder for the interviewer to negotiate the privacy of the interview setting. Using data from a cross-national survey, the World Mental Health (WMH) Initiative, we investigate between-country and between-interviewer variations in the presence of third party during the interview. We also test the association of third-party presence and respondent sociodemographics, country’s level of individualism, and income. Our analyses examine the following six hypotheses:

Hypothesis 1: There is significant variation across countries in the presence of any third party during the interview, controlling for respondent-level characteristics.

Hypothesis 2: The presence of any third party during the interview is inversely associated with the respondent’s country’s level of individualism, controlling for respondent-level characteristics.

Hypothesis 3: The relationship between respondent’s sociodemographics and third-party presence is moderated by respondent’s country’s level of individualism.

Hypothesis 4: Controlling for the country’s level of GNI per capita reduces the effect of the country’s individualism on third-party presence.

Hypothesis 5: Respondents interviewed in low- and middle-income countries have increased odds of being interviewed in the presence of a third party compared with respondents interviewed in high-income countries, controlling for respondent-level characteristics and country’s level of individualism.

Hypothesis 6: There is significant variation across interviewers within a country in the presence of any third party during the interview, controlling for respondent-level characteristics, the country’s level of individualism, and the country’s income level.

From a theoretical point of view, this investigation will help survey researchers conducting cross-cultural research to better understand the dynamics of the interview and the possible effects of interview privacy (or lack of it) on reporting sensitive information. From a practical point of view, it will help better design surveys and inform interviewer training practices, certification, and materials to deal with and minimize third-person presence.

Method

WMH Surveys

The WMH initiative is a large-scale cross-national survey initiative. The survey design and implementation across participating countries are coordinated by two central organizations. All participating countries are required to follow a standard survey protocol that includes a probability sample design, a shared core instrument, a specific translation protocol, a set of quality control measures, and specific interviewer training protocols. However, countries were allowed to adapt certain features of the design and survey implementation including computerization of the interview, respondent incentives, and field team structure. This mix of centralization and local control allowed for establishing a well-adapted survey protocol while maintaining comparability of the data collected (Pennell, Harkness, Levenstein, & Quaglia, 2010).

Sample Design for WMH Surveys

The WMH surveys were administered in 24 countries. Of these countries, 14 have scores on the nation’s level of individualism published by Hofstede et al. (2010). These 14 countries have also collected observations on interview privacy. The countries include Belgium, Brazil, Bulgaria, Germany, India, Italy, Japan, Lebanon, Mexico, Nigeria, China, Romania, Spain, and the United States. All of the surveys targeted the adult population and most of them featured nationally representative household samples (Heeringa et al., 2008). See Pennell et al. (2008) for detailed information on the survey design and methods.

To reduce interview length, the WMH interviews were designed to be administered in two parts: Part 1 included core questionnaire sections and Part 2 included noncore sections. All respondents completed Part 1; Part 2 was administered to a subsample of Part 1 respondents. The current article focuses on Part 2 respondents because a number of the independent variables were collected only in Part 2. Country-specific Part 2 completed interviews range from a low of 1,031 (Lebanon) to a high of 5,692 (the United States). Table 2 shows the number of completed interviews in each country and the year(s) during which the data were collected.

Table 2.

WMH Survey Interview Mode, Supervisor-to-Interviewer Ratio, Completed Interviews, and Response Rates.

Country	Data collection year	Mode	Supervisor-to-interviewer ratio	Completed Part 2 interviews	Response rate
Belgium	2001-2002	CAPI	1:7	1,015	51
Brazil	2005-2007	PAPI	—	2,942	81
Bulgaria	2003-2007	PAPI	—	2,232	72
China	2001-2002	PAPI	1:5	1,628	75
Germany	2002-2003	CAPI	—	1,315	58
India	2003-2005	PAPI	—	1,368	99
Italy	2000-2003	CAPI	1:20	1,768	71
Japan	2002-2006	CAPI	1:20	1,374	56
Lebanon	2002-2003	PAPI	1:5	1,031	70
Mexico	2001-2002	CAPI	1:6	2,350	77
Nigeria	2002-2003	PAPI	1:4	2,141	79
Romania	2005-2006	CAPI	—	2,356	71
Spain	2001-2002	CAPI	1:20	2,121	79
The United States	2001-2003	CAPI	1:10	5,327	71

Note. WMH = World Mental Health; CAPI = Computer Assisted Personal Interview; PAPI = Paper-And-Pencil Interview; — = information not available.

Questionnaire

The WMH surveys use the World Health Organization (WHO) Composite International Diagnostic Interview (CIDI) Version 3.0. The CIDI 3.0 is a structured interview that generates diagnoses for a wide range of mental health disorders. Detailed questions on social and family life, employment history, finances, and childhood experiences are also included in the WMH. The questionnaire was translated into each country’s local language using the WHO translation guidelines (Harkness, Pennell, Villar, Gebler, & Aguilar-Gaxiola, 2008; Kessler & Üstün, 2004).

Questionnaire Administration

All interviews in the 14 WMH countries considered here were conducted face-to-face by trained interviewers using either paper-and-pencil interviewing (PAPI) or computer assisted personal interviewing (CAPI). Interviewer training in each country was modeled after a centralized training session attended by project staff from each country. Table 2 shows the mode of the interview, supervisor-to-interviewer ratio, and response rates in each of the countries.

Measures

Interview privacy

Information on interview privacy was collected in an interviewer observation section at the end of the questionnaire. Interviewers made the following observations: (a) whether a third party was present during any part of the interview and (b) the relationship of the third party to the respondent (parent, partner, child, youth, or “other adults”).

Respondent-level measures

Respondent-level measures included gender, age (18-34, 35-49, 50-64, older than 64 years), marital status (never married, currently married or cohabiting, previously married), relative educational level (high, middle, low, very low), current employment status (employed, studying, taking care of home, other), relative household income level (high, middle, low, very low), household size (one, two, three, more than three), and having any 12-month mental health disorder. Respondents who reported symptoms related to any mental health disorders (including mood, anxiety, impulse control, alcohol, or substance use disorder) in the last 12 months before the interview were coded as having a disorder. Those who did not report such symptoms were coded as not having a disorder.

Interviewer-level measures

Interviewer identification numbers were available for all countries except India. This enabled the modeling of random effects (intercepts) for the individual interviewers in each country. However, no information was available on the characteristics of the interviewers across countries.

Country-level measures

Dimension scores for individualism for each of the 14 countries are available from the book Culture and Organizations: Software of the Mind (Hofstede et al., 2010). For each country, a standardized individualism score was calculated based on the average score and the standard deviation across the 14 countries. Higher scores indicated higher levels of individualism. The scores were based on survey data from Geert Hofstede’s studies in the 1970s and their replication by Hoppe (1998), Shane (1995), Merritt (2000), de Mooij and Hofstede (2001), Mouritzen and Savara (2002), and Van Nimwegen (2002). The position of nations, relative to one another, with respect to cultural dimensions, including individualism, has been shown to be stable over time (Hofstede et al., 2010; Ingelhart, 1997; Schwartz, Bardi, & Bianchi, 2000).

Countries’ economic strength and standard of living was measured by its GNI per capita. GNI measures in nominal dollar values that were calculated according to the Atlas Method for the year the data were collected are used in this article. Because of the wide variation in GNI across the 14 countries, and the big gaps in the measure, the countries were divided into three categories: high-GNI per capita, middle-GNI per capita, and low-GNI per capita. This classification matches the World Bank classification of countries by income.

Analysis

Multilevel logistic regression was used to estimate between-country and between-interviewer variances in the probability that any third party would be present during the interview and to investigate the predictors of third-party presence. Although information on the relationship between the third person and the respondent are available, we focus on any party presence to understand the contextual effects of the overall lack of privacy during the interview and help guide the future interviewer training protocols.

To predict the presence of any third party, two models were used. The first (Model 1) is a three-level random intercept logistic regression with respondents at Level 1, interviewers at Level 2, and countries at Level 3. Predictors include respondent-level characteristics, country’s individualism score, and significant interactions between the country’s individualism score and respondent characteristics.

The second model (Model 2) includes the same predictors as Model 1 but adds the country’s level of GNI per capita. The purpose of this addition was to investigate changes in the effect of the country’s individualism score after GNI was controlled for.

All multilevel models presented here are unweighted. To explore the effect of weights on the findings, the analyses were replicated using weighted and unweighted single-level logistic regression models with dummy variables representing countries. The weights accounted for within-country differential probability of selection, poststratification adjustments to the country’s sociodemographic distributions, and subsampling of respondents to specific questionnaire sections. Adjusted and unadjusted point estimates were consistent suggesting that the use of weights to adjust for the sample design was not necessary for the relationships investigated. All analyses were conducted using PROC GLIMMIX SAS version 9.2.

Results

Presence of Third Party in WMH Interviews

Third-party presence during the WMH survey interviews ranged from 12.6% in Japan to 70.0% in India. Partners were the most common third parties present in the majority of the countries, followed by “other adults.” The least common type of third-party presence was a parent (see Table 3). Controlling for respondent-level characteristics, as hypothesized (Hypothesis 1), we find significant variation between countries in third-party presence during interviews; the country variance estimate is 0.41 (the standard error is 0.17).

Table 3.

Percent (SE) of World Mental Health Interviews Conducted in the Presence of a Third Party.

Country	n	Any third person	Parent	Partner	Youth	Adult
Belgium	1,015	25.4 (1.4)	2.1 (0.5)	18.7 (1.2)	2.7 (0.5)	3.2 (0.6)
Brazil	2,942	59.0 (0.9)	8.2 (0.5)	26.7 (0.8)	18.8 (0.7)	23.9 (0.8)
Bulgaria	2,232	39.3 (1.0)	3.9 (0.4)	26.1 (0.9)	4.4 (0.4)	9.0 (0.6)
China	1,628	36.1 (1.2)	8.7 (0.7)	19.9 (1.0)	4.1 (0.5)	8.8 (0.7)
Germany	1,315	25.9 (1.2)	2.4 (0.4)	16.4 (1.0)	2.7 (0.5)	6.0 (0.7)
India	1,368	70.0 (1.2)	12.4 (0.9)	31.0 (1.3)	7.3 (0.7)	31.4 (1.3)
Italy	1,768	36.0 (1.1)	6.4 (0.6)	21.2 (1.0)	2.1 (0.4)	9.4 (0.7)
Japan	1,374	12.6 (0.9)	1.4 (0.3)	6.2 (0.7)	1.0 (0.3)	4.5 (0.6)
Lebanon	1,031	66.4 (1.5)	11.8 (1.0)	31.7 (1.5)	15.8 (1.1)	32.1 (1.5)
Mexico	2,350	34.6 (1.0)	7.0 (0.5)	11.4 (0.7)	10.7 (0.6)	11.1 (0.7)
Nigeria	2,141	32.0 (1.0)	3.1 (0.4)	8.4 (0.6)	9.1 (0.6)	19.4 (0.9)
Romania	2,356	35.4 (1,0)	3.7 (0.4)	20.7 (0.8)	5.0 (0.5)	10.6 (0.6)
Spain	2,121	41.2 (1.1)	6.7 (0.5)	23.4 (0.9)	4.6 (0.5)	10.9 (0.7)
The United States	5,327	30.2 (0.6)	2.8 (0.2)	17.3 (0.5)	7.6 (0.4)	9.1 (0.4)

Note. Results based on Part 2 sample with valid observations on interview privacy. SE = standard error.

Predictors of Third-Person Presence

Cultural variations

Before we controlled for the country’s level of GNI per capita (Model 1, Table 4), the country’s level of individualism was found to be inversely associated with third-party presence as predicted by Hypothesis 2. There were also significant interactions between the country’s level of individualism and several respondent-level characteristics in predicting third-party presence providing support for Hypothesis 3. There were main effects for being a female, below 64 years (as compared with 65 or older), or a homemaker (as compared with employed) that reduced the odds of having a third person present during the interview (odds ratios [ORs] = 0.87, 0.70, and 0.89, respectively, at the mean level of individualism across all countries). The significant interactions showed that these differences by gender, age, and employment status mattered more as the country’s level of individualism increased. For example, in a country with a level of individualism one standard deviation above the mean, females had 0.83 times the odds to have any other person present during an interview compared with males, but in a country with a mean level of individualism females had 0.87 times the odds of males (Table 4, Model 1).

Table 4.

OR (95% Confidence Interval) from Multilevel Logistic Model Predicting Presence of Any Third Party During the Interview (Total N = 27,573).

	Model 1	Model 2
Sex
Females	0.87 (0.82-0.92)**	0.87 (0.82-0.92)**
Males	1.00	1.00
Age in years
18-34	0.70 (0.62-0.79)**	0.70 (0.62-0.79)**
35-49	0.70 (0.62-0.78)**	0.69 (0.62-0.78)**
50-64	0.69 (0.63-0.77)**	0.69 (0.63-0.77)**
65 and older	1.00	1.00
Marital status
Married/cohabiting	1.17 (1.06-1.28)**	1.17 (1.06-1.28)**
Divorced/widowed/separated	0.73 (0.64-0.82)**	0.73 (0.64-0.82)**
Single	1.00	1.00
Education level
Very low	2.27 (2.05-2.52)**	2.27 (2.05-2.52)**
Low	1.78 (1.62-1.94)**	1.78 (1.63-1.94)**
Mid	1.39 (1.28-1.51)**	1.39 (1.28-1.51)**
High	1.00	1.00
Employment status
Homemaker	0.89 (0.80-0.98)*	0.89 (0.80-0.98)*
Student	1.05 (0.88-1.25)	1.05 (0.88-1.25)
Unemployed/retired/other	1.11 (1.03-1.21)**	1.11 (1.03-1.20)**
Employed	1.00	1.00
Income level
Very low	1.14 (1.03-1.26)*	1.14 (1.03-1.26)*
Low	1.07 (0.98-1.17)	1.07 (0.98-1.17)
Mid	1.01 (0.92-1.10)	1.01 (0.93-1.10)
High	1.00	1.00
Household size
HH size two	2.80 (2.49-3.14)**	2.80 (2.49-3.15)**
HH size three	3.10 (2.74-3.50)**	3.10 (2.74-3.50)**
HH size more than three	3.27 (2.89-3.71)**	3.27 (2.89-3.71)**
HH size one	1.00	1.00
Any 12-month disorder
Have a disorder vs. no disorder	1.18 (1.10-1.26)**	1.18 (1.10-1.26)**
Standardized individualism score (IND)	0.63 (0.41-0.96)*	1.20 (0.61-2.35)
Level of GNI per capita
Low	—	4.13 (0.76-22.49)
Middle	—	4.96 (1.44-17.06)*
High	—	1.00
Female × IND	0.95 (0.90-1.00)*	0.95 (0.90-1.00)*
Age 18-34 × IND	0.95 (0.86-1.04)	0.95 (0.86-1.04)
Age 35-49 × IND	0.94 (0.85-1.02)	0.93 (0.85-1.02)
Age 50-64 × IND	0.91 (0.84-1.00)*	0.91 (0.84-1.00)*
Married × IND	1.09 (1.01-1.17)*	1.09 (1.01-1.17)*
Divorced/Widowed × IND	1.12 (1.01-1.24)*	1.12 (1.01-1.24)*
Very Low Education × IND	1.10 (1.01-1.20)*	1.10 (1.01-1.20)*
Low Education × IND	1.06 (0.98-1.14)	1.06 (0.98-1.14)
Mid Education × IND	1.05 (0.98-1.13)	1.05 (0.98-1.13)
HH Size Two × IND	1.33 (1.20-1.47)**	1.33 (1.20-1.47)**
HH Size Three × IND	1.34 (1.20-1.48)**	1.34 (1.20-1.48)**
HH Size More Than Three × IND	1.35 (1.22-1.50)**	1.35 (1.21-1.50)**
Between-country variance (SE)	0.41 (0.18)	0.25 (0.12)
Between-interviewer variance (SE)	0.59 (0.04)	0.59 (0.04)

Note. Dashes “—” indicate variables not included in the model. Models exclude India because no interviewer identification was available. Interpretation of OR for variables with interactions: The odds ratio of having a third person present among females (compared with males) in a country with an average level of individualism is 0.87 (OR for female dummy variable in Table 4). The odds ratio for females in a country with a level of individualism that is 1 SD larger than the mean is 0.83 (0.87 × 0.95, where 0.95 is the OR for the interaction term between females and IND). OR = odds ratio; IND = individualism.

p value < .05 and **p value < .01.

Interactions were also found between the country’s level of individualism and the respondent’s marital status, educational level, and household size. Being married, unemployed, having very low education, very low income, and living in larger families increased the odds of having a third person present during the interview (ORs = 1.17, 1.11, 2.27, 1.14, and 3.27, respectively, at the mean level of individualism across all countries), but these marital, educational, and family size differences in third-party presence during the interview were more pronounced among respondents living in more individualized countries (Table 4, Model 1).

GNI per capita effect

The main effect of the country’s level of individualism was no longer significant when the country’s GNI per capita was taken into account (partially supporting Hypothesis 4), though all interactions between respondent-level characteristics and the country’s level of individualism remained significant (Table 4, Model 2). As hypothesized in Hypothesis 5, the country’s level of GNI per capita significantly predicted higher rates of third-party presence during the interview, but only among middle-income countries. Respondents interviewed in a country with middle-level GNI per capita had 5 times the odds, OR = 4.96, of having a third person present during the interview than respondents interviewed in high-income countries. Third-party presence was also more common in countries with a low-level of GNI per capita than in high-income countries, but the difference was not significant. The country’s GNI per capita explained a big proportion of the between-country variation in third-party presence. The between-country variance dropped from 0.41 to 0.25 after the country’s GNI per capita was added to the model.

Interviewer variations

In addition to between-country variation in third-party presence, significant between-interviewer variation was found, as we hypothesized (Hypothesis 6). As the dependent variable was binary, it was not possible to directly examine the total variance decomposition. However, comparing the magnitude of between-country variance to the between-interviewer variance, before controlling for the level of individualism and country’s GNI, we found that the latter is larger than the former (0.41 vs. 0.59). The average between-interviewer variance across all countries was 0.59. The lowest country-specific between-interviewer variance was 0.22. The interviewer-level rate of third-party presence in this latter country ranged from 11% to 64%. The country with the highest between-interviewer variance (1.22) had an interviewer-level third-party presence rate that ranged from 0% to 69%.

Discussion

Cultural Variation Associated With Interview Privacy

As in past studies, we find in every country that a substantial percentage of interviews was conducted in the presence of a third party. The lowest rate was in Japan, where one in every seven interviews had a third-party present. This rate increased to five in seven interviews in India. As hypothesized (Hypothesis 1), the variation in nonprivate interviews between countries was significant, even after we controlled for respondent-level characteristics. As we expected, both the country’s level of individualism and wealth, as measured by GNI per capita, were significantly associated with the presence of a third party during the interview; third-party presence was less likely as the country’s level of individualism and wealth increased, specifically for middle-income compared with high-income countries (Hypotheses 2 and 5). However, once we took the country’s GNI per capita into account, the main effect of individualism was no longer significant, suggesting that wealth drives the relationship between individualism and interview privacy. This was partially in line with our fourth hypothesis which predicted a reduction not a disappearance in the effect of the country’s level of individualism on third party. This is also consistent with the literature showing that the country’s level of wealth leads to individualism and not vice versa (Hofstede et al., 2010). When a country’s wealth increases, its citizens might give more attention to self-expression and personal choice. Moreover, as wealth increases, resources become more available, allowing citizens to become more independent rather than interdependent. Other than its high correlation with the level of individualism, the country’s level of wealth as measured by GNI might also encompass other country-level variations such as educational differences. Thus, the effect of the country’s GNI level on privacy not only reflects differences due to level of individualism but also other differences, including access to larger dwellings that could affect privacy.

In addition to its main effect on interview privacy, country-level individualism also moderated the effect of respondent-level characteristics giving support for Hypothesis 3. For example, the sharper difference in interview privacy between female and male respondents in less individualistic countries could reflect the general lack of privacy given to females by their family members, including their parents or in-laws in collectivist countries (Casterline & Chidambaram, 1984). Individualism also moderated the effect of the respondent’s educational level on third-person presence. Again, the effect of the respondent’s educational level on interview privacy gets larger as the country’s level of individualism increases. Low education respondents in a country with China’s level of individualism had twice the odds of being interviewed with a third party present than high education respondents. The OR increased to 2.8 in countries with an individualism level equal to that of the United States. Thus, it seems that the behavior of respondents and other household members and the social interaction between them is greatly influenced by their shared cultural values. To understand when third parties may be present during an interview, researchers should pay closer attention to the dynamics of the relationship between the respondent and other household members and how it is shaped by prevailing cultural values. This is important information for cross-cultural researchers who investigate methodological and substantive cross-cultural differences and who want to control for potential contextual effects that could affect comparability.

Although there was a large reduction in the between-country variation in interview privacy after controlling for the country’s level of wealth and individualism (the between-country level variation was reduced by 39%), there was still significant residual variation across countries that could be explained by other cultural factors or by potential variations in the survey implementation across the different survey firms contracted by the different countries.

Interviewer Variation Associated With Interview Privacy

Researchers rely on interviewers to ensure privacy. Achieving and maintaining privacy usually requires some negotiation between the interviewer and household members. Interviewers are essentially guests, and household members do not normally expect them to dictate the interview setting. Interviewers who have relevant social skills (e.g., who are friendly negotiators and inspire trust) might be better at achieving privacy. Interviewers’ attitudes, traits, and skills, however, could vary, and indeed the estimated between-interviewer variance in the level of interview privacy was significant as hypothesized (Hypothesis 6) and is even larger than the estimated between-country variation (0.59 compared with 0.41, before we control for any country-level characteristics).

Moreover, the magnitude of the country-specific between-interviewer variance did not greatly differ from country to country. This is not surprising given that the standard training protocols that were administered in all WMH surveys did not put a great emphasis on interview privacy, like most interviewer trainings (T. W. Smith, 1997). Still, the range of interviewer-privacy rates varied across countries, ranging from 0% to 69% in the country with the highest between-interviewer variance and from 11% to 64% in the country with the lowest variance.

Variations in interview privacy—whether brought about by interviewers or by respondent characteristics and the culture they live in—could affect reporting. In fact, the effect of interview privacy on reporting sensitive information has been mixed. Although some studies found that third-party presence is associated with reduced reporting of undesirable outcomes (W. A. Aquilino, 1993, 1997; W. S. Aquilino et al., 2000; Moskowitz, 2004; Taietz, 1962), others observed increased reporting of such outcomes (Bulck, 1999; Edwards et al., 1998; Hoyt & Chaloupka, 1994). This inconsistency could be partially attributed to respondent and cultural differences and their interaction in the effect of third-party presence on reporting sensitive information. For example, Mneimneh et al. (2015) had shown that the effect of third-party presence varied by respondent’s need for social approval and the income level of respondent’s country of residence. At any rate, to reduce such effects, researchers sometimes rely on more private interview modes, such as some form of self-administration questionnaires. Although such modes have been effective in reducing interviewer effects on reporting sensitive information (Tourangeau & Smith, 1996; Tourangeau & Yan, 2007), their success in mitigating the effect of third-party presence on reporting has not been established. In fact, there are only five published studies that looked at the benefit of using such modes to counter the effect of third-party presence on reporting sensitive information (W. A. Aquilino, 1997; W. S. Aquilino et al., 2000; Chauchard, 2013; Couper, Singer, & Tourangeau, 2003; Moskowitz, 2004). These studies found mixed results. Aquilino attributes this to the possible psychological presence created by the physical presence of the third person even when a more private mode is being used. The interviewing experience still seems to vary among respondents who completed their interviews in private and those who did not, even when the questions are self-administered.

Another approach to minimize variations in interview privacy within a sample is to reduce any presence of third parties during the interview. Interviewer trainings need to strongly underscore the importance of interview privacy, an issue that has been raised by a number of researchers (Lau et al., 2017; T. W. Smith, 1997). Trainings ought to also teach interviewers how to request, negotiate, achieve, and maintain privacy. Practice on possible techniques tailored to the cultural norms need to be incorporated in interviewer training. For example, having interviewers remind the third person about study confidentiality requirements or having them induce the respondent to request privacy might be appropriate in certain cultures. In other cultures, however, interviewers might have to repeatedly satisfy a third person’s curiosity by answering their questions; or the study design may have to incorporate features (such as gender matching of interviewers to respondents or asking the third party to complete a self-administered questionnaire designed to collect additional information) that reduce the likelihood that a third party will be present during the interview.

Limitations

Although the number of countries included in the analysis varied on the level of income, only three countries fell in the low-income category, compared with five countries in the middle-income group and six in the high-income group. The relatively low number of countries in the low-income group reduced our power to detect a significant difference in the rate of interview privacy between this income group and the high-income group. This is reflected in the large confidence interval for the low-income country group. As for the other country-level measure, the level of individualism, this came from data collected several years ago. One concern is the applicability of those indices to data collected between 2000 and 2004. Still, a number of researchers have shown that while the values of many nations have been changing, the relative positioning of nations has been maintained (Hofstede et al., 2010; Ingelhart & Baker, 2000; Schwartz et al., 2000). Another important consideration is whether the definition and the translation of the interview privacy questions vary by country. Given our simple definition of privacy, whether or not a third person was present during the interview, we don’t expect much cultural variation in how this variable was operationalized. The information on third-party presence came from an observation section filled by the interviewer. All of the observation questions were discussed during train-the-trainer sessions among research staff from each of the countries and trainers of the English source language of the instrument. No concerns were raised about the phrasing of these questions. In addition, all participating countries had to follow a highly specific translation protocol set by the central coordinating institutions.

With respect to between-interviewer variance, two important limitations are worth noting. First, as an interpenetrated design was not implemented and as the information on the assignment of interviewers to geographically clustered areas or segments was not available, isolating the possible geographic clustering component from the estimated interviewer variance was not possible in our study. It is conceivable that households clustered in certain geographical areas accept and respect interview privacy more than other households clustered in areas, contributing to the large between-interviewer variance. Our models, however, attempt to control for some household characteristics, such as income and household size, to account for potential variation in household characteristics across interviewers. Second, interview privacy measures were based on interviewer observations. Interviewers are regularly monitored and evaluated in the field. If the study protocol explicitly calls for interview privacy, it is possible that some interviewers might underreport the presence of a third person to avoid a negative evaluation. Such systematic misreporting in addition to the possibility of unintentional measurement error in recording the type of third person present might contribute to some of the observed variation.

Conclusion

This study highlights the role of country’s individualism and wealth in explaining interview privacy variations across a wide range of countries. Such country-level characteristics play a direct role in explaining cultural differences in rates of interview privacy and also shape the effect of the respondent’s characteristics on privacy. Our results also shed light on the interviewers’ role. The findings call for practical measures related to training interviewers on requesting, negotiating, achieving, and maintaining interview privacy. However, before such measures are implemented, more quantitative and qualitative data about the interviewer’s behavior and attitude toward privacy; the dynamics of the interaction between the respondent, other household members, and the interviewer; and other characteristics of the interview setting should be collected and studied. Such data could be directly collected by the interviewer or, ideally, through recording the interaction. Investigating those factors will help researchers better understand interviewer variations and cultural influences and help them to develop tailored methods to shape interviewers’ attitudes and behaviors toward achieving interview privacy. Most important, collecting more detailed information on the interview setting and the process through which the interpersonal context of the interview is assembled will allow survey researchers to better understand the effect of third-party presence on the interviewer’s behavior and on the respondent’s cognitive processes when answering survey questions.

Footnotes

Acknowledgements

We thank the staff of the WMH Data Collection and Data Analysis Coordination Centers for assistance with instrumentation, fieldwork, and consultation on data analysis.

Authors’ Note

Zeina N. Mneimneh is an assistant research scientist at the Michigan Program in Survey Methodology, Survey Research Center, Institute for Social Research, University of Michigan. None of the funders had any role in the design, analysis, interpretation of results, or preparation of this article. A complete list of all within-country and cross-national WMH publications can be found at

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The collection of some of the data used in the manuscript and for some of the research conducted from the World Mental Health Initiative Data: The World Health Organization World Mental Health (WMH) Survey Initiative is supported by the National Institute of Mental Health (NIMH; R01 MH070884), the John D. and Catherine T. MacArthur Foundation, the Pfizer Foundation, the U.S. Public Health Service (R13-MH066849, R01-MH069864, and R01 DA016558), the Fogarty International Center (FIRCA R03-TW006481), the Pan American Health Organization, Eli Lilly and Company, Ortho-McNeil Pharmaceutical, GlaxoSmithKline, and Bristol-Myers Squibb. The São Paulo Megacity Mental Health Survey is supported by the State of São Paulo Research Foundation (FAPESP) Thematic Project Grant 03/00204-3. The Bulgarian Epidemiological Study of common mental disorders EPIBUL is supported by the Ministry of Health and the National Center for Public Health Protection. The ESEMeD project is funded by the European Commission (Contracts QLG5-1999-01042, SANCO 2004123, and EAHC 20081308), the Piedmont Region (Italy), Fondo de Investigación Sanitaria, Instituto de Salud Carlos III, Spain (FIS 00/0028), Ministerio de Ciencia y Tecnología, Spain (SAF 2000-158-CE), Departament de Salut, Generalitat de Catalunya, Spain, Instituto de Salud Carlos III (CIBER CB06/02/0046,RETICS RD06/0011 REM-TAP), and other local agencies and by an unrestricted educational grant from GlaxoSmithKline. The World Mental Health Japan (WMHJ) Survey is supported by the Grant for Research on Psychiatric and Neurological Diseases and Mental Health (H13-SHOGAI-023, H14-TOKUBETSU-026, H16-KOKORO-013) from the Japan Ministry of Health, Labor and Welfare. The Lebanese National Mental Health Survey (L.E.B.A.N.O.N.) is supported by the Lebanese Ministry of Public Health, the WHO (Lebanon), National Institute of Health/Fogarty International Center (R03 TW006481-01), Sheikh Hamdan Bin Rashid Al Maktoum Award for Medical Sciences, anonymous private donations to IDRAAC, Lebanon, and unrestricted grants from AstraZeneca, Eli Lilly, GlaxoSmithKline, Hikma Pharm, Pfizer, Roche, Sanofi-Aventis, Servier, and Novartis. The Mexican National Comorbidity Survey (MNCS) is supported by The National Institute of Psychiatry Ramon de la Fuente (INPRFMDIES 4280) and by the National Council on Science and Technology (CONACyT-G30544- H), with supplemental support from the Pan American Health Organization (PAHO). The Nigerian Survey of Mental Health and Wellbeing (NSMHW) is supported by the WHO (Geneva), the WHO (Nigeria), and the Federal Ministry of Health, Abuja, Nigeria. The Chinese World Mental Health Survey Initiative is supported by the Pfizer Foundation. The Shenzhen Mental Health Survey is supported by the Shenzhen Bureau of Health and the Shenzhen Bureau of Science, Technology, and Information. The Romania WMH study projects “Policies in Mental Health Area” and “National Study regarding Mental Health and Services Use” were carried out by National School of Public Health and Health Services Management (former National Institute for Research & Development in Health, present National School of Public Health Management & Professional Development, Bucharest), with technical support of Metro Media Transilvania, the National Institute of Statistics—National Center for Training in Statistics, SC. Cheyenne Services SRL, Statistics Netherlands and were funded by Ministry of Public Health (former Ministry of Health) with supplemental support of Eli Lilly Romania SRL. The U.S. National Comorbidity Survey Replication (NCS-R) is supported by the National Institute of Mental Health (NIMH; U01-MH60220) with supplemental support from the National Institute of Drug Abuse (NIDA), the Substance Abuse and Mental Health Services Administration (SAMHSA), the Robert Wood Johnson Foundation (RWJF; Grant 044708), and the John W. Alden Trust.

Author Biographies

Zeina N. Mneimneh is an assistant research scientist in the Survey Methodology Program within the Survey Research Center, University of Michigan. She is also an affiliated assistant research professor in the School of Public Policy at Georgetown University. She is the director of the World Mental Health Data Collection Coordinating Center that supports the design and implementation of national mental health surveys in more than 30 countries. She is also the chair of the executive committee for the International Comparative Survey Design Initiative, an annual workshop that provides a forum for international researchers involved in research relevant to comparative survey methods. She has published more than 35 peer-reviewed publications and book chapters. Her research investigates factors affecting the reporting of sensitive information including interviewer, respondent, and question characteristics, and contextual factors related to the interview setting. Her recent work examines the use of pardata to monitor interviewer behavior and the use of social media data for social science research.

Michael R. Elliott is a professor of biostatistics at the University of Michigan School of Public Health and Research Scientist at the Institute for Social Research. He received his PhD in biostatistics in 1999 from the University of Michigan. Prior to joining the University of Michigan in 2005, he held an appointment as an assistant professor at the Department of Biostatistics and Epidemiology at the University of Pennsylvania School of Medicine, and prior to that as a visiting professor of biostatistics at the University of Michigan School of Public Health and as a visiting research scientist at the University of Michigan Transportation Research Institute. His statistical research interests focus around the broad topic of “missing data,” including the design and analysis of sample surveys, casual and counterfactual inference, and latent variable models. He has worked closely with collaborators in injury research, pediatrics, women’s health, and the social determinants of physical and mental health. He serves as an associate editor for the Journal of the American Statistical Association and the Journal of Survey Statistics and Methodology.

Roger Tourangeau is a vice president and associate director at Westat and a research professor emeritus at the University of Michigan. He has a PhD in psychology from Yale University. He is the lead author of two books—The Psychology of Survey Response (with Lance Rips and Kenneth Rasinski) and The Science of Web Surveys (with Frederick Conrad and Mick Couper)—and has authored or co-authored more than 80 research articles. He is the past president of the American Association for Public Opinion Research.

Steven G. Heeringa is a senior research scientist at the University of Michigan Institute for Social Research (ISR). He is a member of the Faculty of the University of Michigan Program in Survey Methods and the Joint Program in Survey Methodology. He is a fellow of the American Statistical Association and elected member of the International Statistical Institute. He is the author of many publications on statistical design and sampling methods for research in the fields of public health and the social sciences. He is the lead author of Applied Survey Data Analysis, Second Edition (Chapman & Hall, 2017), a comprehensive new text on methods for the statistical analysis of complex sample survey data. He has more than 40 years of statistical sampling experience in the development of the SRC National Sample design, as well as research designs for ISR’s major longitudinal and cross-sectional survey programs. Since 1985, he has collaborated extensively with scientific colleagues in the design and conduct of major studies in aging, psychiatric epidemiology, and physical and mental health. He has been a teacher of survey sampling and statistical methods to U.S. and international students and has served as a sample design consultant to a wide variety of international research programs based in countries such as Russia, the Ukraine, Uzbekistan, Kazakhstan, India, Nepal, China, Egypt, Iran, the United Arab Emirates, Qatar, South Africa, and Chile.

References

Anderson

B. A.

Silver

B. D.

(1987). The validity of survey response: Insights from interviews of married couples in a survey of Soviet emigrants. Social Forces, 66, 537-554.

Aquilino

W. A.

(1993). Effects of spouse presence during the interview on survey responses concerning marriage. Public Opinion Quarterly, 57, 358-376.

Aquilino

W. A.

(1997). Privacy effects on self-reported drug use: Interactions with survey mode and respondent characteristics. In Harrison

Hughes

(Eds.), The validity of self-reported drug use: Improving the accuracy of survey estimates (Monograph 167, pp. 383-415). Rockville, MD: National Institute on Drug Abuse.

Aquilino

W. S.

Wright

D. L.

Supple

A. J.

(2000). Response effects due to bystander presence in CASI and paper-and-pencil surveys of drug use and alcohol use. Substance Use & Misuse, 35, 845-867.

Brenner

P. S.

(2012). Overreporting of voting participation as a function of identity. The Social Science Journal, 49, 421-429.

Bulck

(1999). Does the presence of a third person affect estimates of TV viewing and other media use? Communications, 24, 105-115.

Casterline

Chidambaram

V. C.

(1984). The presence of others during the interview and the reporting of contraceptive knowledge and use. In Ross

J. A.

McNamara

(Eds.), Survey analysis for the guidance of family planning programs (pp. 267-298). Liege, Belgium: Ordina Editions.

Chadi

(2013). Third person effects in interview responses on life satisfaction. Schmollers Jahrbuch, 133, 323-333.

Chauchard

(2013). Using MP3 Players in Surveys: The Impact of a Low-Tech Self-Administration Mode on Reporting of Sensitive Attitudes, Public Opinion Quarterly, 77(S1), 220-231.

10.

Couper

M. P.

Singer

Tourangeau

(2003). Understanding the effects of audio-CASI on self-reports of sensitive behavior. Public Opinion Quarterly, 67, 385-395.

11.

de Mooij

Hofstede

(2001). Convergence and divergence of customer behavior: Implications for international retailing. Journal of Retailing, 78, 61-69.

12.

Diop

K. T.

Traugott

(2015). Third-party presence effect with propensity score matching. Journal of Survey Statistics and Methodology, 3, 193-215.

13.

Edwards

S. L.

Slattery

M. L.

K. N.

(1998). Measurement errors stemming from nonrespondents present at in-person interviews. Annals of Epidemiology, 8, 272-277.

14.

Gabrenya

W. K.

Jr. Barba

(1987). Cultural differences in social interaction during group problem solving. Paper presented at the meetings of the Southeastern Psychological Association, Atlanta.

15.

Festy

Gaymu

Thévenin

(2014). Assessing the household’s financial situation, alone with the interviewer or in the partner’s presence. Population, 69, 81-102.

16.

Harkness

Pennell

B. E.

Villar

Gebler

Aguilar-Gaxiola

(2008). Translation procedures and translation assessment in the World Mental Health Survey Initiative. In Kessler

R. C.

Üstün

T. B.

(Eds.), The WHO world mental health surveys: Global perspectives on the epidemiology of mental disorders (pp. 91-113). New York, NY: Cambridge University Press.

17.

Hartmann

(1995). Response behavior in interview settings of limited privacy. International Journal of Public Opinion Research, 7, 383-390.

18.

Heeringa

S. G.

Wells

E. J.

Hubbard

Mneimneh

Z. N.

Chiu

W. T.

Sampson

N. A.

Berglund

P. A.

(2008). Sample designs and sampling procedures. In Kessler

R. C.

Üstün

T. B.

(Eds.), The WHO world mental health surveys: Global perspectives on the epidemiology of mental disorders (pp. 14-32). New York, NY: Cambridge University Press.

19.

Herrera

A. V.

Benjet

Méndez

Casanova

Medina-Mora

M. E.

(2017). How mental health interviews conducted alone, in the presence of an adult, a child or both affects adolescents’ reporting of psychological symptoms and risky behaviors. Journal of Youth and Adolescence, 46, 417-428.

20.

Hofstede

G. J.

Minkov

(2010). Culture and organizations: Software of the mind (3rd ed.). New York, NY: McGraw-Hill.

21.

Hoppe

M. H.

(1998). Validating the masculinity/femininity dimensions on elites from nineteen countries. In Hofstede

and Associates (Eds.), Masculinity and femininity: The taboo dimension of national cultures (pp. 29-43). Thousand Oaks, CA: Sage.

22.

Hoyt

G. M.

Chaloupka

F. J.

(1994). Effect of survey conditions on self-reported substance use. Contemporary Economic Policy, 7, 109-121.

23.

Ingelhart

(1997). Modernization and postmodernization: Cultural, economic, and political change in 43 societies. Princeton, NJ: Princeton University Press.

24.

Ingelhart

Baker

W. E.

(2000). Modernization, cultural change, and the persistence of traditional values. American Sociological Review, 65, 19-51.

25.

Kessler

R. C.

Üstün

T. B.

(2004). The World Mental Health (WMH) Survey Initiative version of the World Health Organization (WHO) Composite International Diagnostic Interview (CIDI). International Journal of Methods in Psychiatric Research, 13, 93-121.

26.

Lau

C. Q.

Baker

Fiore

Greene

Lieskovsky

Matu

Peytcheva

(2017). Bystanders, noise, and distractions in face-to-face surveys in Africa and Latin America. International Journal of Social Research Methodology, 20, 469-483.

27.

Merritt

(2000). Culture in the cockpit: Do Hofstede’s dimensions replicate? Journal of Cross-Cultural Psychology, 31, 283-301.

28.

Mneimneh

Z. M.

Tourangeau

Pennell

B. E.

Heeringa

S. G.

Elliott

M. R.

(2015). Cultural variations in the effect of interview privacy and the need for social conformity on reporting sensitive information. Journal of Official Statistics, 31, 673-697.

29.

Moskowitz

J. M.

(2004). Assessment of cigarette smoking and smoking susceptibility among youth: Telephone computer-assisted self-interviews versus computer-assisted telephone interviews. Public Opinion Quarterly, 68, 565-587.

30.

Mouritzen

P. E.

Savara

J. H.

(2002). Leadership at the apex: Politicians and administrators in western local governments. Pittsburgh, PA: University of Pittsburgh Press.

31.

Pennell

B. E.

Harkness

J. A.

Levenstein

Quaglia

(2010). Challenges in cross-national data collection. In Harkness

J. A.

et al et al. (Eds.), Survey methods in multinational, multiregional, and multicultural contexts (pp. 269-298). Hoboken, NJ: John Wiley.

32.

Pennell

B. E.

Mneimneh

Z. N.

Bowers

Chardoul

Wells

J. E.

Carmen

V. M.

. . .Gemma

G. V.

(2008). Implementation of the World Mental Health surveys. In Kessler

R. C.

Üstün

T. B.

(Eds.), The WHO world mental health surveys: Global perspectives on the epidemiology of mental disorders (pp. 33-57). New York, NY: Cambridge University Press.

33.

Podmore

Chaney

Golder

(1975). Third parties in the interview situation: Evidence from Hong Kong. The Journal of Social Psychology, 95, 227-231.

34.

Pollner

Adams

R. E.

(1994). The interpersonal context of mental health interviews. Journal of Health and Social Behavior, 35, 283-290.

35.

Pollner

Adams

R. E.

(1997). The effect of spouse presence on appraisals of emotional support and household strain. Public Opinion Quarterly, 61, 615-626.

36.

Reuband

(1992). On third persons in the interview situation and their impact on responses. International Journal of Public Opinion Research, 4, 269-274.

37.

Schwartz

S. H.

Bardi

Bianchi

(2000). Value adaptation to the imposition and collapse of communist regimes in East-central Europe. In Renshon

S. A.

Duckitt

(Eds.), Political psychology: Cultural and cross-cultural foundations (pp. 217-237). New York, NY: Palgrave Macmillan.

38.

Shane

S. A.

(1995). Uncertainty avoidance and the preference for innovation championing roles. Journal of International Business Studies, 26, 47-68.

39.

Silver

B. D.

Abramson

P. R.

Anderson

(1986). The presence of others and overreporting of voting in American national elections. Public Opinion Quarterly, 50, 228-239.

40.

Smith

P. K.

Bond

M. H.

Kağıtçıbaşı

. (2006). Understanding social psychology across cultures: Living and working in a changing world. London, England: Sage.

41.

Smith

T. W.

(1997). The impact of the presence of others on a respondent’s answers to questions. International Journal of Public Opinion Research, 9, 33-47.

42.

Streib

G. F.

(1973). Privacy in the research interview: Issues related to design, field tactics and analysis. Journal of Comparative Family Studies, 4, 276-285.

43.

Taietz

(1962). Conflicting group norms and the “third” person in the interview. American Journal of Sociology, 68, 97-104.

44.

Tourangeau

Smith

T. W.

(1996). Asking sensitive questions: The impact of data collection mode, question format, and questions context. Public Opinion Quarterly, 60, 275-304.

45.

Tourangeau

Yan

(2007). Sensitive questions in surveys. Psychological Bulletin, 133, 859-883.

46.

Triandis

H. C.

(1972). The analysis of subjective culture. New York, NY: John Wiley.

47.

Triandis

H. C.

(1989). The self and social behavior in differing cultural contexts. Psychological Review, 96, 506-520.

48.

Triandis

H. C.

(1995). Individualism and collectivism. Boulder, CO: Westview Press.

49.

Van Nimwegen

. (2002). Global banking, global values: The in-house reception of the corporate values of ABN AMRO. Delft, The Netherlands: Ebuorn.

50.

Welkenhuysen-Gybels

Billiet

(2001). The impact of third party presence in survey interviews on the measurement of political knowledge. Acta Politica, 36, 287-306.

51.

Zipp

J. F.

Toth

(2002). She said, he said, they said: The impact of spousal presence in survey. Public Opinion Quarterly, 66, 177-208.