Quantitative Flashcards
Variable
characteristic that can be recorded about object/subject in study
Types of variables
Numeric - continuous (come from measurements) - discrete (counting) Categorical - nominal (no natural order) - ordinal (has a natural order)
Population
The larger group to which results are generalised
Sample
Representative group for drawing conclusions about the population
Advantages of sampling
More time efficient, more economical, can be more accurate
Sampling bias
When members of a sample over or under represent attributes of the population related to areas being studied
Probability sampling
Probability of each individual being selected can be calculated and is usually made to be equal chance
Types of probability sampling
- random sampling
- systematic sampling
- stratified random sampling
- disproportionate sampling
- cluster sampling
Simple random sampling
each member has equal chance of selection
Systematic sampling
every nth subject chosen (eg. every 9th person is chosen)
Stratified random sampling
Population split into subgroups from which random sample is drawn
Cluster sampling
successive random sampling of a series of units in a population, convenient but may compound effects of sampling bias
Disproportionate sampling
eg. If a population is 90% females and 10% males but you want even numbers of females and males (females have less chance of being chosen) so disproportionate
Types of non-probability sampling
- convenience sampling
- quota sampling
- purposive sampling
- snowball sampling
Convenience sampling
Subjects are chosen on the basis of their availability
Quota sampling
Researcher guides sampling process until a quota is met
Purposive sampling
Subjects ‘hand picked’ based on certain criteria
Snowball sampling
Relies on original participants identifying/referring people with similar characteristics
Experimental studies
Control over IV, experimental conditions and construction of groups for comparison. Intervention or treatment. Aim to provide evidence that IV is cause of change in DV
Observational studies
Subjects observed in natural state, groups are self selected, may be measured/tested but no intervention, can be retrospective or prospective
Randomised Controlled Trial (RCT)
2 or more groups allocated using randomisation (one is control, other receives experimental variable), pre and post treatment measures, changes that occur in experimental but not control group can be attributed to treatment
Double blinding
Neither pt nor assessor knows which treatment the pt is receiving
Crossover design
pt acts as own control, order of treatment is randomised
Factorial design
Several factors compared at the same time, each subject receives combination of all factors, suitable for investigating interactions
Quasi-experimental designs
- one group post test only
- one group pretest - posttest
- non - equivalent control group
- non - equivalent control group pretest-posttest
- single subject
Single subject design
- single pt or institution followed over time
- differs from case study as it is a structured experiment
- outcome measured before and after intervention
- external validity and ability to generalise is weak
Retrospective studies
Look at past data for potentially influential factors
Prospective studies
Follow participants, observe exposure and track outcome
Attrition
Loss of subjects in a study (could be due to mortality, pt moving, declining to continued participation etc)
Case control studies
retrospective, take and compare histories of groups, typically examine effects related to cause of condition
Cross sectional studies
provide a ‘snapshot’ at a certain point in time, often used to determine prevalence of disease/disability
Case report
clinical history of single pt, not a strong design for determining cause effect, can present clinically important info for future studies