Title: The Interviewer as an Informant about the Interview Process
1The Interviewer as an Informant about the
Interview Process
- Geert Loosveldt
- Department of Sociology
- Catholic University of Leuven
- Belgium
2I. Introduction
- Two components of the interviewer role
- persuading people to cooperate
- obtaining information
- Tendency to enlarge the task
- Expert Witnesses
- they observe the respondent reaction
- Expert Knowledge
- useful in quality assessment research
- methods rating forms, focus groups, interviewer
debriefing questionnaires
3- Comments
- Interviewers are a source of error
- Data about the interview are not error free
- Question
- Whats the quality of the interviewers
information about the interview process - General research objective
- evaluation of the quality of the data collected
by means of interviewer reports
4II. Research objectives
- The interviewer report
- Debriefing questionnaire
- The interviewers evaluation of
- interview situation
- respondent behaviour
- cognitive and communicative skills
- motivation to answer
- easy to collect but not often used
5- Evaluation of the interviewer report
- First criterion association between the
information form the report and aspects of
response behaviour - strong association ---gt positive evaluation
- Second criterion Interviewer variance in
interviewers evaluation - same response behaviour --gt same evaluation ---gt
no interviewer variance - absence of interviewer variance ---gt positive
evaluation.
6- Expectations
- information from the report is related to
response behaviour - interviewers do not evaluate the same response
behaviour in the same way interviewer variance - evaluation is written down at the end of the
interview - interviewer report is not a major issue of the
training
7III. Data
- Belgian General Election Study 1995
- Flemish population aged 18-75 years
- 2099 face-to-face interviews in respondents
homes - Response rate 65
- Questionnaire voting behaviour, political
participation, political objectives and
interviewer report
8IV. Information from the interviewer reports
- Four general evaluation questions
- general not related to specific questions
9Clarification question Did the respondent ask
for clarification ?
- Indicator of difficulties experienced by the
respondent - More than 10 asked often or a lot for
clarification
10Resistance question Did you feel any resistance
on the part of the respondent in answering some
questions ?
Resistance ----------------------- Never
57.3Almost never 28.3Now and then
11.4A lot 2.0Often
1.1 -----------------------
2088 11
Sensitive questions income question, questions
about political preferences. Almost never or
never in 85 of the interviews
11Motivation questionHow do you judge the
motivation of the respondent
Motivation -------------------------------- Very
reserved 1.3Reserved
3.3Rather indifferent
15.3Moderately motivated 42.8Highly motivated
37.4 --------------------------------
2090
9
Less then 5 was considered reserved or very
reserved
12Ability questionHow did you judge the
capability of the respondent to understand the
questions ?
Ability ---------------------------- Totally
inadequate 0.5Very poor 3.2Poor
8.5Adequate
36.4High 31.7Very high
19.6 ----------------------------
2087
12
For most respondents the evaluation is
positive 12 of the respondents had problems
13- Observations
- Respondents were motivated and capable of
performing their role as respondent - All correlations between these four variables are
significant ---gt - Reliable scale interviewers general impression
of the respondent - high score positive evaluation
- low score negative evaluation
- four evaluative variables and general scale are
used as dependent variables
14V. Response behaviour and the interviewer
evaluation of respondents role performance
- An interviewer expects substantive and consistent
answers---gt - Item non-response and inconsistent answers
- can be considered as negative ---gt
- will influence interviewers evaluation
- examination of the effect of
- item non- response for difficult questions,
attitude questions and sensitive questions - inconsistent answers
15Description of the difficult questions
- A rating task six political parties on 8
different 11-point scales (6x8 questions). - A dont know filter in the introduction but not
on the show card. - 20 used DK answers at least 24 times out of 48.
- Answering these questions requires a lot of
cognitive effort and motivation. - Expectation
- an effect on interviewers evaluation of
respondent motivation and ability.
16Description of attitude questions
- An attitude scale on ethnocentrism with 13 items
(5-point scale) - A no opinion filter in the introduction but
not on the show card. - Fewer than 2 gave more than 6 no opinion
answers. - no opinion answer is an expression of limited
motivation. - Expectation
- an effect on the evaluation of the respondents
motivation
17Description of sensitive questions
- The income questions and the political
preference questions. - 10 of the respondents refused to answer both the
open and the closed income questions. - Question about the political party the
respondents had voted and the most important
reason for voting for this party. For 7 no
information. - Expectation
- an effect on the interviewers evaluation of the
respondent resistance to answering some
questions.
18Description of inconsistent answers
- Four pair of questions for which the answers can
be interpreted as being inconsistent or not. - 11.7 of the respondents gave an inconsistent
answer for one or more pairs of statement. --gt - inconsistent answers are not dominant
- Expectation
- A moderate effect on the interviewers evaluation
of respondents motivation and ability
19(No Transcript)
20VI. Analysis of the effects of response behaviour
on the interviewers evaluation
- A regression analysis for each interviewers
evaluation variable and the general scale. - Independent variables
- five aspects of response behaviour and
- two background characteristics education and
sex.
21- Background characteristics have sign. effects -
All expected effects are significant. More
effects are sign. - Interviewers evaluations are
more a reflection of a general picture of the
respondent
22VII. Analysis of interviewer variance
- Second evaluation criterion analysis of
interviewer variance for the evaluative variables
- General idea
- controlling for relevant respondent
characteristics, there should be no variation
between interviewers with respect to the
evaluative variables - Multi-level models
- respondents are level-1 units
- interviewers are level-2 units
23Model A Only the interviewer as independent
var. Differences between
interviewers are sign. High
intraclass correlations. Model B Model A and
response behaviour characteristics.
Five interviewer variances are still
significant Conclusion a negative evaluation
24Extension of model B random intercept and random
slopes
Only a few random slopes are significant. The
impact of item non-response for the income and
political preference questions on the evaluation
of resistance is not the same
25VII Discussion
- The interviewer is a rather reliable informant
- Question 1 How can we get better information ?
- More training to complete this part of the
questionnaire - differences between interviewers reflect
differences in the way interviewers react. - Question 2 How can we use the information ?
- Typology of respondents
- use the typology in a substantive analysis