Scholarship Experimental Design

Scholarship Statistics Study!

a statistical experiment investigates ________ - whether a/an t________ or i________ causes a change in the r________ or d________ variable

causality, intervention, response, dependent

the dependent variable is the ________ variable

response

the independent variable is the ________ variable

explanatory

a sample consists of ________ ________

experimental units

factors other than the treatment need to be ________ so that the ________ in the response variable can be solely attributed to the ________

controlled, variability, treatment

________ ________ are when outcomes from one treatment are compared with the outcomes from another treatment / no treatment with the same group of participants

paired comparison

paired comparison results in ________ data

bivariate

________ can be drawn to show comparisons in a paired dot plot for bivariate data

arrows

participants can be separated into two groups using ________ ________

random allocation

random allocation ensures that the ________ of the two groups are similar, to prevent the influence of a c________ / l________ variable affecting response variable

characteristics, confounding, lurking

the group without the treatment is given a ________ or is the ________ group

placebo, control

a larger group ensures more similar ________ between the two groups, which should reduce ________ in results

characteristics, variability

________ may be appropriate, to analyse the ________ in the response variable between repeated measurements

replication, variability

________ is used to determine if the observed differences in the means or medians can be used to draw a conclusion about causality

re-randomisation

re-randomisation under ________ alone calculates and plots differences in means for many (about 1000) ________

chance, re-samples

the ________ ________ is the proportion of the re-samples which have an equal or greater difference in means/medians to the observed difference

tail proportion

if tail proportion __ 10%, then the experiment may likely be due to chance alone, or a combination of the treatment and chance acting

>

if tail proportion __ 10%, it is unlikely that the observed difference was due to chance alone, and that the treatment has an effect on the response variable

<

a causal inference can be extended to other ________ which share the same ________

populations, characteristics

the placebo effect is an example of ________ elements that come into play in an experiment, which are minimised by using single- or double-________

psychological, blinding

a report on an investigation using experimental design principles should use the ________ cycle

PPDAC