Introduction

This chapter examines whether there are different results due to the ethnic background of the interviewer in a survey on ethnicity among second-generation Surinamese carried out in Amsterdam.

The literature on interviewer effects is part of a larger body of work on response effects. Section 5.2 first reviews response effects in general and then looks at research on the influence of the ethnic background of the interviewer in survey research (in the Anglo-Saxon literature simplified to race of interviewer effect). Attention is given to the respondents, the subject matter of the questions, methods and explanations for this phenomenon. Section 5.3 describes the research design of the survey of 300 secondgeneration Surinamese. Section 5.4 looks at the outcomes of this study using first a method based on individual scores and then multi-level analysis. Finally, section 5.5, the conclusion, discusses the relevance of the outcomes in terms of the theories that are found in the literature.

Existing research on response effects and race of interviewer effects

Response effects

Van der Zouwen (1989: 9) defines ‘response effects’ as those particular effects on the response that are caused by variables other than the ones which the researcher wants to measure with the stated question. Dijkstra & Van der Zouwen (1982) refer to four types of variables that can cause response effects. First, they distinguish variables that are related to the characteristics of the interviewer, which can include role dependent and role independent characteristics. Role dependent characteristics are, for example, motivation and skilfulness in interviewing in the intended manner. Role independent characteristics include, among others, ethnic background, sex, age and other personal characteristics of the interviewer, but in addition, they also include the interviewer's own expectations and opinions regarding the subject matter of the survey and respondents’ answers.

The second type of variables relates to characteristics of the respondents. These include role dependent characteristics, such as the motivation to answer honestly, and role independent ones, for example, ethnic background, sex and age, but also psychological variables such as the tendency to answer in a socially desirable way.

Third are variables related to the interview situation, like the extent of anonymity, the place and time of the interview, and also whether third persons are present, in front of whom the respondent may not feel free to say what he or she wants or may be distracted.