Lecture 7, 21-11-2022
No levels in reality – naïve realism
- Empirical reality = reality
- (Naïve) realism and empiricism reality is unlayered (flat).
Levels of reality – critical realism
- Domain of empirical: reality as it appears to us our experiences.
- Domain of actual: reality as it occurs in events.
o Events: every time A happens, B happens.
- Domain of real: reality as it is structured and its mechanisms
Influences on the conduct of research:
- Epistemology
- Ontology
- Research question
- Practical considerations
- Values
- Traditions
Questionnaire: interview without interviewer, also digitally CASI.
Questionnaire advantages:
- Fast and cheap
- No interviewer and/or variation
- Respondent determines: speed, time slot, location etc.
Questionnaire disadvantages:
- Respondent is helpless
- Limited number and non-complex questions
- Easy to skip (not necessarily with web)
- Whole questionnaire is known in one go (not necessarily with web)
- Who answers? How is it answered?
- Must be able to read / use web
- Low response ratio (often max 20%)
- Lack of control / selection
Strategies to increase response ratio:
- Promise money (or send)/win/threatening
- Recruiting letter
- Post box number or stamps (mail survey)
- Reminders (mail surveys)
- Clear instructions (mail surveys)
- Attractive layout
Computer Assisted Personal Interviewing (CAPI)
Survey interviews :
- Used to be the most used collection method in social science research
- Structured questionnaire
- Structured questionnaire sequence
- Standardized question formulation
- Closed questions: multiple choice
- Interviews are structured and have closed questions to achieve
standardisation.
, - Objectives standardisation:
o Reliability measuring instrument
o Exclusion of interviewer effects
- Fast and relatively easy
Model of the survey interview:
1. Interpretation question
2. Look up the answer
3. Assessment response
4. Formulate answer
Hume’s train station
Paradigmatic sequence:
1. Do you agree with the following statement, totally agree; agree; do not
agree, do not disagree; disagree; totally disagree
Respondent said agree, interviewer said okay!
2. Most interviewers say uuuhm before saying answers. Then respondents
say: “if they live here for five years and speak Dutch, I think so! Right?”
Then interviewer said “so totally agree”. Respondent: “NO, only if they live
here for five years and speak Dutch!” Interviewer: “then ‘not agree, nor
disagree’.”
Fowler and Mangione (1990) suggest hardcore standardisation: only then
reliable instrument. So repetition of the question in case of mismatch, or
repetition of all response categories.
Consequence Fowler and Mangione standardisation abnormal
conversation
More flexible solution:
- Making interviewers aware of objectives of standardisation, goals of the
interview, and… let the interviewers solve it!
- Only some suggestions:
o I: I enjoy living in my neighbourhood. R: well, yeah.
o Interviewer can repeat some of the answer categories.
o I: are you totally agreeing, agreeing or neutral?
o Interviewer than can make meta-comment
If 40% of the questions led to a mismatch, is 40% unreliable? No, because
mismatch can lead to a more valid answer:
R: I do not understand the question
I: What we mean is…
Problems are articulated, frames become clear
Less acquiescence (saying yes or no, this is an unwanted form of response
behaviour) or satisficing.
Different between objectivism and positivism:
- objectivism: is there a reality out there? Is there something as an extravert?
- Positivism: We can measure it, it is there.
Designing questions, general rules:
- Start with your research question
- Use your operationalisation
, - Think about the measurement level, because this is harder with open
questions
- Use validated questions (or used elsewhere) for comparability
Question typology Bryman:
- Personal factual questions
- Factual questions about others
- Informant factual questions
- Knowledge
- Beliefs
- Attitudes
- Normative standards and values
- On question content
- Other methodologist opt for a more technique oriented approach
Open questions versus the closed questions:
- Advantages:
o Answers in own terms
o New unexpected answers
o As a preliminary investigation (pre-study)
- Disadvantages:
o Time-consuming
o More difficult to analyse
Open questions versus closed questions:
- Advantages:
o Faster and easier
o Better response rates
o Easy to analyse
- Disadvantages:
o Limited answer possibilities
o Forced choice
o Interpretation problems
Standardisation versus communication
Elmo Wilson versus Rensis Likert
- Closed versus partially open
- Lazersfeld mediated
- Open questions, but closed answer categories
- Picked up by Johan Galtung
Ask as an open question, but receive closed answers
Does social reality exists externally to social actors (constructionist
perspective)
Naturalism versus subtle realism:
- Naturalism: opposes positivism, and objectivism/realism because:
o People construct the social world (constructivism perspective)
o Naturalist say: “researchers can get to know how social actors
construct reality”
- Subtle realism critiques:
No levels in reality – naïve realism
- Empirical reality = reality
- (Naïve) realism and empiricism reality is unlayered (flat).
Levels of reality – critical realism
- Domain of empirical: reality as it appears to us our experiences.
- Domain of actual: reality as it occurs in events.
o Events: every time A happens, B happens.
- Domain of real: reality as it is structured and its mechanisms
Influences on the conduct of research:
- Epistemology
- Ontology
- Research question
- Practical considerations
- Values
- Traditions
Questionnaire: interview without interviewer, also digitally CASI.
Questionnaire advantages:
- Fast and cheap
- No interviewer and/or variation
- Respondent determines: speed, time slot, location etc.
Questionnaire disadvantages:
- Respondent is helpless
- Limited number and non-complex questions
- Easy to skip (not necessarily with web)
- Whole questionnaire is known in one go (not necessarily with web)
- Who answers? How is it answered?
- Must be able to read / use web
- Low response ratio (often max 20%)
- Lack of control / selection
Strategies to increase response ratio:
- Promise money (or send)/win/threatening
- Recruiting letter
- Post box number or stamps (mail survey)
- Reminders (mail surveys)
- Clear instructions (mail surveys)
- Attractive layout
Computer Assisted Personal Interviewing (CAPI)
Survey interviews :
- Used to be the most used collection method in social science research
- Structured questionnaire
- Structured questionnaire sequence
- Standardized question formulation
- Closed questions: multiple choice
- Interviews are structured and have closed questions to achieve
standardisation.
, - Objectives standardisation:
o Reliability measuring instrument
o Exclusion of interviewer effects
- Fast and relatively easy
Model of the survey interview:
1. Interpretation question
2. Look up the answer
3. Assessment response
4. Formulate answer
Hume’s train station
Paradigmatic sequence:
1. Do you agree with the following statement, totally agree; agree; do not
agree, do not disagree; disagree; totally disagree
Respondent said agree, interviewer said okay!
2. Most interviewers say uuuhm before saying answers. Then respondents
say: “if they live here for five years and speak Dutch, I think so! Right?”
Then interviewer said “so totally agree”. Respondent: “NO, only if they live
here for five years and speak Dutch!” Interviewer: “then ‘not agree, nor
disagree’.”
Fowler and Mangione (1990) suggest hardcore standardisation: only then
reliable instrument. So repetition of the question in case of mismatch, or
repetition of all response categories.
Consequence Fowler and Mangione standardisation abnormal
conversation
More flexible solution:
- Making interviewers aware of objectives of standardisation, goals of the
interview, and… let the interviewers solve it!
- Only some suggestions:
o I: I enjoy living in my neighbourhood. R: well, yeah.
o Interviewer can repeat some of the answer categories.
o I: are you totally agreeing, agreeing or neutral?
o Interviewer than can make meta-comment
If 40% of the questions led to a mismatch, is 40% unreliable? No, because
mismatch can lead to a more valid answer:
R: I do not understand the question
I: What we mean is…
Problems are articulated, frames become clear
Less acquiescence (saying yes or no, this is an unwanted form of response
behaviour) or satisficing.
Different between objectivism and positivism:
- objectivism: is there a reality out there? Is there something as an extravert?
- Positivism: We can measure it, it is there.
Designing questions, general rules:
- Start with your research question
- Use your operationalisation
, - Think about the measurement level, because this is harder with open
questions
- Use validated questions (or used elsewhere) for comparability
Question typology Bryman:
- Personal factual questions
- Factual questions about others
- Informant factual questions
- Knowledge
- Beliefs
- Attitudes
- Normative standards and values
- On question content
- Other methodologist opt for a more technique oriented approach
Open questions versus the closed questions:
- Advantages:
o Answers in own terms
o New unexpected answers
o As a preliminary investigation (pre-study)
- Disadvantages:
o Time-consuming
o More difficult to analyse
Open questions versus closed questions:
- Advantages:
o Faster and easier
o Better response rates
o Easy to analyse
- Disadvantages:
o Limited answer possibilities
o Forced choice
o Interpretation problems
Standardisation versus communication
Elmo Wilson versus Rensis Likert
- Closed versus partially open
- Lazersfeld mediated
- Open questions, but closed answer categories
- Picked up by Johan Galtung
Ask as an open question, but receive closed answers
Does social reality exists externally to social actors (constructionist
perspective)
Naturalism versus subtle realism:
- Naturalism: opposes positivism, and objectivism/realism because:
o People construct the social world (constructivism perspective)
o Naturalist say: “researchers can get to know how social actors
construct reality”
- Subtle realism critiques:
