Definitions Flashcards
Null Hypotheses
There will be no difference between the control and intervention arms
This is assumed to be true at the start of the study and has to be DISPROVED.
Dependant variable
The outcome of interest (for example healing time of a wound)
Independent variable
The intervention factor (for example the dressing being used in the intervention)
PROBABILITY sampling
Designed to give an UNBIASED sample where everyone (who meets the criteria) has a chance of selection
This is to choose the SAMPLE of those entering the trial Four types: Simple random, stratified random, cluster and systematic random
NON-PROBABILITY sampling
Non-random and the chance of being selected cannot be estimated
Falsification
Hypothesis testing
Hypothesis
Statement of the relationship between 2 variables
Standardised
Can be repeated and verified
Reliability
Must be repeatable with consistent results, dependability
Validity
Must measure what it intended to measure, credibility
Stratified random sampling
Put in groups according to characteristics (like gender) and then randomly selected
Cluster sampling
Random selection of larger units (like hospitals) which participants are then randomly selected from
Systematic sampling
Random selection of predetermined intervals
Factors affecting sample size
Population, Design, measurement, practical factors
Single blind trial
One person knows which aim of the trial they are in, person assessing the outcome does not know
Double blind trial
Neither participant nor person assessing outcomes knows the aim
Internal validity
Study results legitimate because of the way the study was conducted
External validity/generalisability
Concerns whether results are transferable to other groups
Threats to internal validity
History: Events happening outside the study
Maturation: Changes that happen over time
Testing: Change due to experience of the test
Instrumentation: Changes in measurement rather than change in status
Mortality: Differences in study drop out
Selection bias: Participants different to non-participants
Threats to external validity
Selection effects: generalisability to other populations, when ideal sample population cannot be obtained.
Reactive effects: Response to just being in a study (HAWTHORNE EFFECT).
Measurement effects: Measurement and testing affects the generalisability
Descriptive statistics
A way of displaying and summarising quantitative (numerical) data in ways that are easily understood
Levels of measurement
Nominal (categories)
Ordinal (different categories that can be ranked)
Interval (different categories that are ranked with equal spaces in-between)
Ratio (different categories that can be ranked, with equal spaces in-between and a fixed zero)