Unit 26: Data collection methods and measurement
Data collection method: Interviews
Example: problems in a home for elderly care
Interview
• Non-standardized questions are asked
• Interview topics and probes are used to study units of observation (respondents or informants)
To say something about a population of units of analysis (persons, couples, organizations)
• Method is obtrusive and verbal
When interviews?
• You want to know feelings or thoughts
• When variables are underdeveloped
• When people are unable to answer closed questions
• Or when you expect to find unexpected information
Eight phases
1. Formulate the research question
2. Select and contact respondents
3. Select variables or topics you’re interested in
4. Formulate interview questions & answering categories
5. Create a interview protocol: ordering an lay-out
6. Design interview instructions and interview training
7. Code the received data
8. Data storage and further data analysis
Respondents or informants?
‘to what extent do inhabitants of a home for elderly care face problems in care?’
➔ Interview care givers (informants) to say something about specific inhabitants with dementia.
Or to say something about a group.
➔ If a person is a unit of analysis himself: respondent
• Non-probability sampling about types of problems
‘which types of problems do people face?’
Select interview topics
• Think about aspects of elderly care that can be evaluated
• Make a ‘model’ of these aspects (sleep, washing, toilet, meals)
Questions and probes
Question: ‘what do you think of the availability of personnel after dinner?
Probe: ‘can you elaborate on that?’
➔ Give you more information
➔ Activity of the interviewer after an initial question to keep the conversation going and gather
more information
Interview protocol
Written instructions guiding an interview
• Instructions to the interviewer
• Key interview questions
• Probes to follow key questions
• Types of answers that are expected
, Coding
Outcome of interviews is written and needs to be ‘coded’ before analyzing
• When you want to know differences between units of analysis data can be coded in a data
matrix
When making an inventory of possibilities, output most often can’t be structured in a data matrix
➔ Use lists of possible outcomes
➔ Barriers in processes
➔ Diagrams of expected caused and consequences
Data collection method: Validity and Reliability threats in interviews
Validity and reliability in interviewing
Answers in one part of the interview should be similar to the answers in another part of the interview
(reliable answers)
• You need consistent answers
• Verify answers (repeat the answer that was given), be aware of unreliability and then report it
Information you get should be correct, should not be about something different (valid answers)
• Example: complaints about the house are actually indications of individual physical comfort
• Check completeness, relevance and clarity answers, ask the questions differently
Consistency between characteristics and the coding at the end
• Who is interviewing and where it takes place, affects what is in the mind of the respondent
• Be aware of socially desirable answers
• Record interview and not taking notes
• Be aware of interpreting answers
• Create relevant codes (valid codes)
• Inter coder reliability
➔ Let someone else code the same text and see if you come up with the same codes
Data collection method: Survey
Example: downloading copy-righted materials
Survey
A set of standerdized questions is asked to a sample of units of observations (respondents, informants)
to say something about a population of units of analysis (persons, couples, organisazions)
• This method is obstrusive and verbal
• You affect people which means its obstrusive
When?
• Clear research question
• Large number of units of observation
• Known and well defined variables
Seven steps in survey design
1. Select (the type of) survey
2. Designing and selecting survey questions
3. Design the questionnaire (item order and lay-out)
4. Questionnaire testing
5. Interviewer testing
6. Survey administration, checking response
7. Data cleaning and storage
Data collection method: Interviews
Example: problems in a home for elderly care
Interview
• Non-standardized questions are asked
• Interview topics and probes are used to study units of observation (respondents or informants)
To say something about a population of units of analysis (persons, couples, organizations)
• Method is obtrusive and verbal
When interviews?
• You want to know feelings or thoughts
• When variables are underdeveloped
• When people are unable to answer closed questions
• Or when you expect to find unexpected information
Eight phases
1. Formulate the research question
2. Select and contact respondents
3. Select variables or topics you’re interested in
4. Formulate interview questions & answering categories
5. Create a interview protocol: ordering an lay-out
6. Design interview instructions and interview training
7. Code the received data
8. Data storage and further data analysis
Respondents or informants?
‘to what extent do inhabitants of a home for elderly care face problems in care?’
➔ Interview care givers (informants) to say something about specific inhabitants with dementia.
Or to say something about a group.
➔ If a person is a unit of analysis himself: respondent
• Non-probability sampling about types of problems
‘which types of problems do people face?’
Select interview topics
• Think about aspects of elderly care that can be evaluated
• Make a ‘model’ of these aspects (sleep, washing, toilet, meals)
Questions and probes
Question: ‘what do you think of the availability of personnel after dinner?
Probe: ‘can you elaborate on that?’
➔ Give you more information
➔ Activity of the interviewer after an initial question to keep the conversation going and gather
more information
Interview protocol
Written instructions guiding an interview
• Instructions to the interviewer
• Key interview questions
• Probes to follow key questions
• Types of answers that are expected
, Coding
Outcome of interviews is written and needs to be ‘coded’ before analyzing
• When you want to know differences between units of analysis data can be coded in a data
matrix
When making an inventory of possibilities, output most often can’t be structured in a data matrix
➔ Use lists of possible outcomes
➔ Barriers in processes
➔ Diagrams of expected caused and consequences
Data collection method: Validity and Reliability threats in interviews
Validity and reliability in interviewing
Answers in one part of the interview should be similar to the answers in another part of the interview
(reliable answers)
• You need consistent answers
• Verify answers (repeat the answer that was given), be aware of unreliability and then report it
Information you get should be correct, should not be about something different (valid answers)
• Example: complaints about the house are actually indications of individual physical comfort
• Check completeness, relevance and clarity answers, ask the questions differently
Consistency between characteristics and the coding at the end
• Who is interviewing and where it takes place, affects what is in the mind of the respondent
• Be aware of socially desirable answers
• Record interview and not taking notes
• Be aware of interpreting answers
• Create relevant codes (valid codes)
• Inter coder reliability
➔ Let someone else code the same text and see if you come up with the same codes
Data collection method: Survey
Example: downloading copy-righted materials
Survey
A set of standerdized questions is asked to a sample of units of observations (respondents, informants)
to say something about a population of units of analysis (persons, couples, organisazions)
• This method is obstrusive and verbal
• You affect people which means its obstrusive
When?
• Clear research question
• Large number of units of observation
• Known and well defined variables
Seven steps in survey design
1. Select (the type of) survey
2. Designing and selecting survey questions
3. Design the questionnaire (item order and lay-out)
4. Questionnaire testing
5. Interviewer testing
6. Survey administration, checking response
7. Data cleaning and storage
