Questionnaire Analysis & Literature Appraisal Flashcards
Factor analysis is where you are looking at
The structure of the items
Why do we screen the data?
To test wether it’s ready for parametric testing.
What is reverse scoring?
Change direction of phrasing in items to keep ppl thinking and paying attention while doing the questionnaire - both negative and positive framed items.
Cronbachs alpha is best used when you have a small or large questionnaire?
Small
Chronbachs alpha magic number is
.7
What are rogue items
Items that don’t hang well reliability wise with the other items
Why delete rogue items?
To help reliability.
Outcomes with good item internal reliability may need to check to see if they need to
Delete items that are redundant (they look likely they measure the same thing)
Alpha can be strong at .6 when
It’s measuring attitudes. You need to rationalise this.
Sample of convenience…
General population
Item analysis is where we compare
Top versus bottom scorers
Validity is a test that
Measures what is should measure
Validity needs to be
Reliable
The more valid a test is the more
We reduce measurement error
Construct validity is the
Adequacy of operational definition of variables
Does the measure that is employed to measure a behaviour actually measure the behaviour is an example of…
Construct validity
Examples of construct validity…
A measure of intelligence…through an IQ test
Nomological network is a network of
Evidence surrounding and supporting the validity of a construct
Nomological networks refers to a network of
Correlations which support NOT prove a construct
The larger the nomological network for a construct the greater
The validity
How many ways can we evaluate the viability of a scale in terms of construct validity?
Two
What do both ways of evaluating the viability of a scale in construct validity assume?
That there is already a good measure of the construct available
Two ways to evaluate the viability of a scale in terms of construct validity
Convergent validity
Discriminant validity
Convergent validity is where our
Questionnaire that converges with other questionnaires that measure the same/related constructs. Should correlate
Discriminant validity
A questionnaire should not correlate with questionnaires that measure different/unrelated constructs ie depression and anxiety
Face validity
What it looks like. Does it seem sensible. Ie items relating to your construct.
Content validity
Items on questionnaire judged to be appropriate by experts in the field
Population validity
Extent to which results found in a sample can be generalised to a population.
Criterion validity
Extent to which scores in a test are related to criterion validity
Predictive validity
Shown when a questionnaire predicts another variable ie high GPA predicts honours performance
Concurrent validity
Demonstrated by high correlation between 2 tests taken at the sane time. Assumes another questionnaire exists that us the gold standard for measuring that construct
2 purposes for literature review
Discover what has been done
What needs to be done
3 sources of literature are:
General
Secondary
Primary
2 types of info provided by general sources are
General intro to area
Clues to where more scientific info can be found
Two sources of info provided by secondary sources…
Scholarly summaries
Further references
Primary sources are
Original report of original work
Lit review contains mostly these
The 8 important areas to any journal article:
Preview of previous lot Problem and purpose Hypothesis Method Sample Results discussion References General comments
