data collection quad
DESCRIPTION
TRANSCRIPT
Data Collection Jacky Griffith/Anji Waring Faculty of Health and Social
Work
Methods of Data CollectionObjectives: To describe different methods of
data collection To discuss the strengths and
limitations of the different methods
Data CollectionExisting data
Already been generated by someone else. They are not creating the data themselve
New data Creating data that are new and not there bfore
Biophysiological testsLike
Height Weight BP
Ie from physiological data
Self report data What people say/tell/write. From interviews and questaire. Subjects have generated themselves Telling us things
Observation data Observations of the researcher What they see how they interprete Researcher led.
Data CollectionDimensions of quantitative data
Needs to be
Structured
Quantifiable
Need to turn it into numbers.
Objective Minimize bias No thoughts and feelings of researcher.
Biophysiologic and physical variables
In vivo – performed on or with people or living organismsIn life Open up to a number of potential extraneous data that make the results skewed.
In vitro (in glass) – data gathered and then subject to laboratory analysis◦ Gives absolute values◦ Tend to be accurate
◦ As long as the tools are accurate in the first place
◦ Easily standardisedTends to be used in randomised
trials – what a new drug does to someone. Does it work or is it just the same as doing nothing/or using another drug
Not used in surveys or qualitative work so much. Not looking for that type of material.
Looking at attitudes to weight
Self Report MethodsUsed in both qualitative and quantitative
Unstructured – qualitative approach
Structured - quantitative approach
Instrument: questionnaire or interview schedule
Research termilogy for the tool used.
May use tool such as GHQ general health questionaire, BDI beck depression inventory which are common to a lot of studies
May be used in a number of situation Have been already validated. Should always be referenced
May design new tool which needs validating
Needs to be tested to see if it is valid give you the results that can be applied to the area of interest.
The research should mention if they have invented or used old tools.
Need to road test an unvalidated tool.
Commones way Face validity Experts say that it does
what it purports to May design something specific
to this study eg survey questionnaire
Does not use scales
Can be done by interview with interviewer coding responses, questionnaire given out by researcher, by post
Writes down the answers in coded form.
Questionnaires/Scales Tools or instruments designed
to measure or explore specific issues or attitudes - not just a set of questions.
Need to go beyond and use question that will uncover things that are very specific
Types: Questions : - ◦ Closed questions◦ Yes/no ◦ A choice◦ Have to give an answer.◦ “ forced into an answer”◦ Open questions◦ Give your own answer and allow to
expand on what you really mean and give reasons.
◦ Say what you like, how you like. Response Categories:-
◦ Likert◦ 12345◦◦ Visual analogue◦ Yum –hmm-yuck
Questionnaire Design Aim:
◦ To measure what intend to measure
◦ Does what it says it does…..
◦ Validity – does it measure what it says it does
◦ External validity – generalisable to others
◦ Reliability – does it actually work
◦ To ensure respondent can complete the questionnaire.◦
Consider:◦ Wording◦ Are the words understandable for
the target sample?◦ Technical words◦ Level of language◦ Children/university graduates◦ Ambiguousness.
◦ DischargeCan it be interpreted in any other
way?You know what you mean, but
others may not understand…..
◦ information level required◦ descriptive – yes/no gives you
numbers◦ go beyond that with
scales/subsequent questions deeper level of information
◦ format of questions◦ nice and easy to get more difficult◦ hook them with something
interesting◦ need to know the demographics
◦ age◦ gender◦ educational level◦ job title
to make sense of the dataif you put that first, it can put
people off.
By the time you have gone through all the interesting stuff, then
people are more likely to fill the boring stuff as they finish
Leading questions They are let into something rather
than lead
◦ response categories◦ yes/no◦ choose from multiple choice◦ more than one choice◦ dichotomous choices◦ fixed choices
how are the responsing going to be managed
mix up positive and negative.
People re-write the questions
◦ length of questionnaire◦ people in pain…◦ time constraints
◦ method of administration◦ take away◦ face to face◦ sent◦ over the phone (almost like an
interview)
Strengths & LimitationsStrengths less costly
just print a lot don’t have to be there. But if you post it, it can be
dear Reach a lot of people
access large sample distance
allows a larger sample size anonymity
might be more willing to say what they feel with regard to what they want to say.
less threatening no power balance
discrepensies. data easy to analyse
Limitations response rate
can be quite low…. Need to put out more to get
suffient back Coverage
Can go back and ask “why”s No way of finding out where
attitudes originated. ability to complete
then these are excluded social desirability
people don’t want to admit, even if they are anonomous, that is not acceptable to the population as a whole.
Is saying what you think social acceptable
Is what they are saying ileagal? So they don’t want to even admit even if it is anonymous.
IInterviewsTypes:structuredsemi-structureddone in a very open way.
Not in the same oredeCan go backwards and forward
BoundriesCue questions “now can I ask you about…”
Unstructured In grounded theory Don’t know much about what you are studying.
Don’t have a schedule or cue questions.
“tell me about…..”No stearing at all
Could last one second or one millennium
Method face to facetelephone focus group
a group interviewconformity issues
Practical Considerations Design of interview schedule
What are the key areas that need to be covered?
Training of interviewers
If quantitive all be done in the same way
if qualitative:- beware of cultural difference and interpretive difference.
Place of interview Your teratory Their teratory
They have more power… Timing of interview
After the event … too soon, too far after.
20 mins to 90mins Recording the information
collected◦ Manually
◦ audiotape◦ videotape
Strengths & Limitations Strengths quick results clarification response rate additional information less superficial
Limitations practicalities expensive analysis bias social desirability Observation
What we see Actually observed rather
than what people say
When aim to describe what
happening in setting need overview of
activity/behaviour establish main issues as
prelude to further work.Overview
Leader on to something elseWhat are the main issues
Then look at something in more detail
Qualitative or quantitateive
Target Areas individual characteristics
non-verbal communication type of verbal communication behaviour
conformity
Types
Gold 1958
Covert Overt
Complete participation
Participant as observer
Complete observerUnethical
Used with babies and young children (but parents need to know
Participant as participantHawthorn effect
But if you are there long enough, people’s behaviour reverts
Practical considerations
Planning
◦ Structured (tick box) /unstructured (written down long hand)
◦ Field note can support other forms of data.
◦ what to sample◦ when to sample◦ how to record
◦ tool use (validated/unvalidated)
◦ recorded electronic◦ time available
Strengths and LimitationsStrengths first hand recording
you can see what is happening.
No self report bias
control over data collected
Limitations difficult to arrange observer bias only see what you want to see or expect to
see. time Hawthorne effect Training
Inter-rater reliability all interviews/observes are singing from the same song book.
Pilot StudiesFor new tools or doing
something new with an old toolTry it out first
Small scale version of the main study.
To smooth out future hiccups.
Try it out on similar types of group.
Some of the sample will do the pilot study.
Time & cost built into the proposal
May have to go back to basics and redesign the whole thing but it is better to find out now than after you have spent lots of money on the real thing. Carried out prior to main study
to:◦ test new methods ◦ check feasibility of design◦ ensure appropriateness of sample◦ identify potential pitfalls
DOES IT ACTUALLY WORK!!!!!!!!????!!!!
Modify aspects of main study according to results from pilot
Look to see how if there has been a pilot and does it work Did they need to change and how
Triangulation
Use of more than one method for collecting or interpreting data
Ie inteveiw as well as observation From surveyingFrom more than one set of data
Hoping that one will support the other
May show that there is a gap and may show if one or other is unreliable.Types: Methodological Theoretical Data Quantititve vs qualitative
Investigator2 people looking to see if they
find the same thing….To validate.