Lecture 4 Flashcards
When does data collection occur?
Quant: Specific step along the way. Occurs after sampling.
Qual: iterative cycle btwn sampling, data collection, data analyses.
Data collection principles: Qualitative
Investigator involvement considerable.
Collection and analyses are intertwined: inform each other, verify understanding
Time spent GAINING ACCESS and gathering data may be considerable.
Multiple methods for gathering data.
Data collection principles: Quant
Investigator involvement is minimal
Data analyses conducted when all data available
May take a long Time
Methods are prescribed, do not change after protocol is established
Investigator involvement in: phenomenology grounded theory participatory action research ethnography
p: active listener
g: fully participates
par: variable, depends on study
e: fully participates
Data collection approaches for both qual and quant (5)
- observe
- ask questions
- examine materials
- measure performance
- self report
How do you collect data in qual?
- fieldwork: notes, reflections
- observe
- ask questions: interviews or focus groups
- examine materials
- audio taping
- photo’s and videos
How do you collect data in quant?
- observe, count
- checklists
- interviews: structured (specific answers to count)
- examine materials
- questionnaires
- outcome measures
How do you gain access (people, context) to qualitative data?
- point of entry through the physical location and the experiences of the participants
- process depends on nature of the question
What do you consider while gaining access to qual data?
researcher position existing networks rapport previous research experiences known site vs. unknown
Why does previous research experience matter?
Shows credibility in researcher.
What happens after the researcher gains access? (“Learning the ropes”) - Qual
- ongoing fieldwork: learn personal stories of informants, gain familiarity w/ setting, become an insider
- Rich Point
What is a Rich Point?
- point at which the investigator realizes his/her assumptions are not sufficient to explain the nature of the context.
- Investigator must work at understanding what is really happening by asking questions that focus or refocus
How do you refer to individuals in studies?
- subjects - quant - passive, objective
- respondents - surveys
- informants- qual - subjective
- key informants -
- participants
- other terms - insiders, outsiders
What is a key informant?
Person particularly knowledgeable.
- useful for helping investigator understand what is happening
- used to obtain info about subgroups to whom the researcher has no access
How do you ensure data accuracy in qual research (7)? What are they important for?
- multiple gatherers
- triangulation (crystallization)
- saturation
- member checking
- reflexivity
- audit trail
- peer debriefing
**all important for criteria of merit
Multiple Gatherers
several investigators analyze independently, then compare results
triangulation (crystallization)
multiple approaches (observe, interview, collect materials, etc)
saturation
sufficient info, no new insights
member checking
validate assumptions w/ informants
Reflexivity
to elucidate investigator bias
audit trail
explain thinking and action leading to results
Peer debriefing
more than 1 investigator analyzing
How do you measure in quant studies?
- assign #’s to represent quantities of a characteristic, attribute or to classify objects: describes behaviours, allows comparisons across people/within person
- levels of measurement: nominal, ordinal, interval ratio
What are levels of measurement?
Nominal - naming
Ordinal
Interval and Ratio
Characteristics of experimental type levels of measurement
See Table.
Measurement scales - types
- Likert scale
- Visual Analogue Scale
Likert scale
- ordinal
- definitely satisfied, somewhat satisfied –> etc
Visual analogue scale
- interval or ratio
- can have numbers or categories
- no pain to worst pain
What are measurement instruments used for (5)?
- describe group
- to screen- screening test
- to determine risk - predict future, risk assessment
- to diagnose - diagnostic test
- to make comparisons, evaluate outcomes
Types of measures in quant?
- single items: weight, ROM
- composite scales: QoL, FIM
How can you be confident w/ the quality of your data?
- validity
- reliability
Validity
- extent to which an instrument evaluates what it is intended to measure.
- affected by both random and systematic error
Reliability
- repeatability
- sources of error - include instrument, rater
- due to random error
- measured by statistics
What are types of measurement error?
Random - occurs by chance
Systematic - consistent variation
What are types of validity in quant?
- traditional view: face validity, content validity, construct validity, criterion validity
- contemporary view: unitary concept w/ many facets; theory driven, assign meaningful interpretation to scores or outcomes
Face validity
Look up
Content validity
Look Up
Construct validity
- convergent - everyone is supposed to be the same
- divergent - everyone is different
Criterion validity
- concurrent: take new measure/old measure, see if you get the same result
- predictive: do above in the futre
What is the significance of contemporary view of validity?
If the research doesn’t affect the person, it doesn’t really matter in the end…. ??
What are assessments of reliability in quant?
- extent of internal measurement error: stability reliability (test-retest), internal consistency reliability
- extent of external measurement error (equivalence): inter and intra rater reliability
What are evaluative outcome measures in quant?
- “sensitivity to change”: ability to detect change, but the change may not be clinically meaningful
- “responsiveness” - detect clincially important change
What are psychometric properties in quant?
- properties of an ax instrument
- extent of validation for the pop. being tested
- term from psychologists, social scientists
What are clinimetric properties?
- properties of indexes, rating scales, etc used to describe or measure symptoms, physical signs, clinical phenomena in clinical medicine
- extend of validation of these clinical ax tools
What are evaluative assessments that are often used interchangeably?
- psychometric
- clinimetric
