1 Session 20: Data Collection in the Field Ina F. Wallace RTI International IES/NCER Summer Research Training Institute, June 2007
Mar 29, 2015
1
Session 20: Data Collection in the Field
Ina F. Wallace
RTI International
IES/NCER Summer Research Training Institute, June 2007
2
Data Collection in the Field
Topics Covered
• Methods of Data Collection
• Use of State-of-the-Art Data Collection Procedures
• Ensuring Data Quality
• Conducting Small-Scale Validation Studies
• Ethical Issues in Data Collection
3
Methods of Data Collection• Interviewer Administered
– In Person• Surveys• Assessments
– Telephone
• Self-Administered– Paper and Pencil
• Mail-in• In-person
– Computer Assisted Self Interview– Web procedures
4
Data Collection Mode Characteristics
High Data Collector
Involvement
Low Data Collector Involvement
Paper Computer Paper Computer
Direct Contact with Respondent
Face-to-Face Paper and Pencil Interviewing/
Assessment (PAPI)
Face-to-Face
Computer Assisted Personal Interview/
Assessment (CAPI)
(Audio) Computer Assisted Self-Interview (CASI,
A-CASI)
Indirect Contact with Respondent
Telephone
PAPI
Telephone Computer Assisted Telephone Interview (CATI)
Mail, Fax, E-mail questionnare
Touch-tone Data Entry (TDE), Web, Electronic Mail Survey (EMS)
No Contact with Respondent
Direct Observation
Computer Assisted Data Entry (CADE)
Administrative
Records
Electronic Data Interchange (EDI)
Adapted From Biemer & Lyberg (2003). Introduction to Survey Quality
5
Face-to-Face InterviewingAdvantages• Maximum degree of communication and interaction
• Can utilize PAPI or CAI• Allows for long, complex interviews
• Variety of tactics can be used to gain cooperation
• Response rates usually high
Disadvantages•High cost due to travel•Can generate social desirability for sensitive questions
•Interviewers can affect respondents
•Can generate interviewer variance
•Presence of others can affect responses
•Falsification of interview
6
Telephone InterviewingAdvantages• Cost lower than face to face
• More easy to provide training, supervision, and feedback
• Social desirability bias less than with face to face
• Interviewer variance less
Disadvantages• Less flexible• No ability to use visual aids
• Must be shorter• Questions must be less complicated
• Less capacity for motivating the respondent
• No coverage of non-telephone units
• Response rates lower
7
Mail SurveysAdvantages• Lowest cost• Low social desirability • Can utilize visual aids • Allows respondent to take time and look up information (if necessary)
• No interviewer variance
Disadvantages• Little control over who completes the survey
• Item non-response high• Response rates lower• Require a long field period to obtain an acceptable response rate
• Respondents must have good literacy skills
• Limit to closed-format questions
8
Computerized Procedures• Computer Assisted Personal Interview
(CAPI)• Computer Assisted Telephone Interview
(CATI)• Computer Assisted Self Interview (CASI)• Audio Assisted Self Interview (ACASI)• Telephone Audio Assisted Self Interview
(T-ACASI)• Computer Assisted Data Entry (CADE)• Web Surveys
9
Computer Assisted ProceduresAdvantages• Complex skip patterns and
branching are easier to follow
• Previously acquired information can be programmed
• Ranges, inconsistencies, and illegitimate values can be checked
• Help screens can be included
• Data are available for analysis more quickly
Disadvantages• Time consuming to develop
and program computer • Must be adequately tested• Can’t change things easily
once fielded• Risk of computer crashing• Costly
10
Direct Assessment
• Used to obtain psychological and educational information about an individual
• Individually administered
• Research driven NOT for clinical or diagnostic purposes
• Follows a standardized protocol
11
Observational Techniques• Involve a researcher making observations
within a natural setting• Direct Observation - individuals know they
are being observed– Continuous Monitoring recording as much
behavior as possible– Time Allocation randomly visiting the sample
to observe at different times
• Unobtrusive Observation – individuals do not know they are being observed
12
Observational Techniques
Advantages• No or limited participation needed
• May be good for generating hypotheses
• Sensitive social issues can be obtained more accurately
• High in validity because of the depth of information collected
Disadvantages• Variance associated with
observers is similar to that found with interviewers
• Can be very costly as they are often labor intensive
• Time consuming• Reliability can be low• Generalizability can be
low
13
Summary: Issues in Choosing a Strategy
• Type of Population
• Nature of Research Issue
• Question Form
• Question Content
• Response Rates
• Costs
• Length of Data Collection
14
Methods of Administration
• Standardized Interviewing Approach– A protocol in which interviewers interact with
the respondent in a manner which is consistent across all interviewers
• Conversational Interviewing– A protocol in which interviewers modify and
adapt questions to the respondent’s situation
• Hybrid Interviewing – Combination of the two approaches
15
Factors that Can Explain Interviewer Variance
Interviewer•Characteristics•Appearance•Motives•Beliefs/attitudes•Perceptions•Expectations•Skills/Knowledge
Respondent•Characteristics•Knowledge•Interest/motivation•Confidence•Strength of convictions•Expectations
Interviewer EffectsQuestionnaire
•Definition clarity•Terminology/jargon•Question form•Instructions•Question wording•Question topic
Survey Conditions and Setting
•Mode of interview•Standardization •Interviewer training•Interviewer supervision•Monitoring/observation
From Biemer & Lyberg (2003). Introduction to Survey Quality
16
Factors that Can Explain Interviewer Variance
Interviewer•Characteristics•Appearance•Motives•Beliefs/attitudes•Perceptions•Expectations•Skills/Knowledge
Respondent•Characteristics•Knowledge•Interest/motivation•Confidence•Strength of convictions•Expectations
Interviewer EffectsQuestionnaire
•Definition clarity•Terminology/jargon•Question form•Instructions•Question wording•Question topic
Survey Conditions and Setting
•Mode of interview•Standardization •Interviewer training•Interviewer supervision•Monitoring/observation
From Biemer & Lyberg (2003). Introduction to Survey Quality
17
Role of the Interviewer/Assessor
• Gaining Cooperation– Converting reluctant participants
– Recognizing and dealing with refusals
• Training and Motivating Respondents– Interviewers implicitly and explicitly teach
respondents how to behave
• Obtaining Quality Data
• Ensuring Confidentiality
18
Assessing Children• Assessor Characteristics
– Flexibility– Vigilance– Self-awareness– Eye contact
• Establishing Rapport– Allaying apprehensions– Helping children feel at ease– Maintaining limits on permissible behavior
• Arranging the Physical Environment
19
Ensuring Quality Data
• Recruiting interviewers/assessors– Characteristics of all data collectors– Characteristics of interviewers– Characteristics of assessors
• Training on standardized procedures
• Certification as data collectors
• Monitoring performance
20
Training
Content of Training• Presenting the study • Contacting participants• Administering the
questionnaire/ assessment• Probing• Recording responses• Confidentiality• Interpersonal relations• Administrative procedures
Procedures for Training• Written materials• Lecture and presentation• Computer-based tutorials• Planned exercises• Role Playing• Observation• Certification
21
Interviewer Training Agenda• Introduction to the Study• Role of the Interviewer• Confidentiality Procedures• Review of Questionnaire and Protocols• Standardized Interviewing Techniques• Round Robin Interview• Gaining Cooperation• Presentation of Scenarios• Round Robin Demonstrations• Paired Mocks• Certification
22
Assessor Training Agenda
Day 1:
Large group • Introduction to Project• Role as an Assessor• Best practices with
children
Small groups • Pre-training quiz• Introduction to all
materials• Learn each
assessment–Observation–Round Robins
Day 2:
Small groups • Individual practice with
assessments• Paired practice with
assessments
Large group• Building rapport and
gaining cooperation• Administrative
responsibilities • Data security • Informed Consent
Day 3:
Small groups • Review of
assessments• Paired practice• Post-training quiz
Large group • Concluding remarks• Certification
23
Monitoring PerformanceWhat to Monitor• Detection and prevention
of falsified information• Compliance with rules
and guidelines about administering the questionnaire or assessment
• Performance on non-interview/assessment tasks
How to Monitor• Telephone verification
with a sample of respondents
• Direct observation – in-person or by telephone
• Audio-recording – CARI• Review of materials• Performance and
production measures
24
Human Subjects Issues• Institutional Review Boards • Informed Consent
– Who is carrying out research under whose sponsorship– Description of purposes and procedures– Duration of involvement– Risks and Benefits– Confidentiality– Voluntary nature of research including stopping early
and skipping questions– Compensation– Contact information
• Confidentiality Issues
25
Human Subjects Issues• Institutional Review Boards • Informed Consent
– Who is carrying out research under whose sponsorship– Description of purposes and procedures– Duration of involvement– Risks and Benefits– Confidentiality– Voluntary nature of research including stopping early
and skipping questions– Compensation– Contact information
• Confidentiality Issues
26
Data Collection EvaluationStage of Data
CollectionEvaluation Method Purpose
Design Expert review of instrument Identify problems with instrument
Pretest Behavior coding
Cognitive interviewing
Evaluate one or more stages of the response process
Pretest/Post-Data Collection
Respondent focus groups
Interviewer debriefings
Evaluate instrument and data collection procedures
Pretest/Data Collection
Supervisor observation
Telephone monitoring
CARI
Evaluate performance of data collector
Post-Data Collection
Experimentation
Internal Consistency
External validation
Re-interview/retest
Compare alternative methods of data collection
Evaluate components of MSE
Validate estimates
Adapted From Biemer & Lyberg (2003). Introduction to Survey Quality
27
Post-Data Collection Evaluations
• Experiments
• Observational Studies
• Internal Consistency Studies
• External Validation Studies
• Administrative Record Check Studies
• Re-interview/Re-Test Studies
28
Post-Data Collection Evaluations
• Experiments
• Observational Studies
• Internal Consistency Studies
• External Validation Studies
• Administrative Record Check Studies
• Re-interview/Re-Test Studies