expirimental design Flashcards
what is inductive reasoning + example
process of inferring a general law or principle from observation of a particular instance
cognitive dissonance
what is deductive reasoning + example
inference by reasoning from a general principle to a particular situation
describe the different stages in the scientific method
question->hypothesis->experiment->analyse data->draw conclusions-> impact of data: model may be strengthened, modified
experimental control?
all things the same EXCEPT the test variable
4 obstacles to overcome
perfectly definite change and do nothing else to influence result
biologic variability
chance; experimental errors, faults in methods, differences in reagents
bias in experimenter and experimental design
randomised…
subjects are assigned to either treatment or control groups in a random fashion
double-blind
subjects arent aware of the treatment they’re receiving and neither is the investigator
single blind
subjects arent aware of the treatment but the experimenter is
potential for bias
placebo controlled
a false treatment is necessary to measure the proportion of any observed effect accountable to the placebo effect
cross over
subjects are swapped over from the treatment to control group. assumes that the effects of the drug are completely reversible
matching or randomised block design
subjects matched for age, sex, weight, stage of life
extend observation to a number of groups. with replication and randomisation of treatments, the error from variation is made to fall equally on treated and untreated groups
what overcomes biological variability and chance
biologic variability and chance overcome with replication
why used statistics
overcome variation in natural systems by analysing whether a difference can be ascribed to chance events
data can be organised, analysed and compared
enable hypothesis to be testes
what do we need samples to be
representative of the entire population for valid conclusions to be drawn
too small; inaccurate
too large; costly and inefficient
what reduces chance of bias
randomisation, replication, blinding and placebo