+
Related Flashcards
Related Topics
Cards In This Set
| Front | Back |
|
A pilot is:
|
A small version of the studyconfirms the feasibility of the study
|
|
Methodological Studies
|
Studies for the purpose of designing instrumentsconduct tests for validity and reliability
|
|
Indications of a Good Instrument
|
Validity: the extent to which an instrument reflects the concept being examinedReliability: the extent to which the attribute is consistently measured
|
|
Types of Validity
|
Does the instrument measure what it is supposed to measure?3 types: ContentCriterion-RelatedConstruct
|
|
Content Validity
|
Is the content representative of the content domain under study? -Content Validity: rate each question or part of the survey to determine if it is valid or not. Question by question. More important type of content validity-Face Validity: ask others to evaluate if the instrument is measuring the right thing. Intuitive and could simply give the appearance of measuring the same thing.
|
|
Criterion-Related Validity
|
To what degree are the "observed score" and the "true score" related? (how much error is there?) Concurrent Validity: simultaneously administer 2 instruments- the new one and an already established one. We want a high correlationPredictive Validity: whether a current score is related to a score taken in the future
|
|
Construct Validity
|
To what extent does the instrument measure the theoretical concept or trait?
|
|
Construct Validity: Hypothesis Testing
|
Have hypothesis about result- then compare to your results
|
|
Construct Validity: Convergent
|
2 instruments in same person- same as concurrent
|
|
Construct Validity: Divergent
|
Give 2 instruments that measure different things. Negative correlations support construct validity
|
|
Construct Validity: Multi-trait/ Multi-method
|
Convergent and divergent at the same time- 3 instruments
|
|
Construct Validity: Known Groups
|
Pick groups of people that you expect to score a certain way.Find some that are high and are low on the same scale
|
|
Construct Validity: Factor Analysis
|
Categorize into domains
|
|
Reliability
|
Observers will receive the consistent results over timeExpressed as a coefficient from 0-10.70 or greater is consider acceptable 0= no correlation1= perfect correlation
|
|
Types of Reliability
|
Stability: obtain same results with repeated administrationEquivalence: consistency or agreement between different using tool, or agreement between alternate forms of the toolInternal Consistency: items within scale measure the same concept
|
