Observational Study
Study based on data in which no manipulation of factors has been employed
Retrospective Study
Observational study where subjects are selected and then previous conditions/behaviors are determined; need not be based on random samples and normally focus on estimating differences between groups/associations between variables
Prospective study
An observational study where subjects are followed to observe future outcomes; not an experiment due to lack of treatment, normally focuses on estimating differences among groups that can appear as groups are followed during study
Experiment
Manipulating factor levels to create treatments, randomly assigning subjects to these treatment levels, and then comparing the responses of subject groups across treatment levels
Random Assignment
To be valid, an experiment must assign experimental units to treatment groups at random
Factor
Variable whose levels are manipulated by experimenter
Response
A variable whose levels are compared across different treatments
Experimental Units
Individuals (subjects/participants when human) on whom and experiment is performed
Level
Specific values that experimenter chooses for a factor
Treatment
Process, intervention, or other controlled circumstance applied to randomly assigned experimental units
Principles of experimental design
Control, randomize, replicate, block
Control
This is done on aspects of the experiment that are known to have an effect on the response but aren't factors being studied
Randomize
This is done on subjects to treatments to even out uncontrollable effects
Replicate
If subjects of experiment aren't a representative sample from the population of interest, do this with a different group, preferably with a different population
Block
Do this to reduce effects of identifiable attributes of the subjects that cannot be controlled
Statistically Significant
When observed difference is too large to believe and may not have occurred naturally
Control group
Experimental units assigned to a baseline treatment level, normally the well-understood default treatment, or the placebo treatment.
Blinding
Any individual associated with an experiment who does not know how subjects were chosen is considered one.
Single-blind
When every individual in either of these cases (those who can influence results such as a subject, and those who can evaluate results such as a judge) is blinded
Double-blind
When every individual in both of these cases (those who can influence results such as a subject, and those who can evaluate results such as a judge) is blinded
Placebo
Treatment known to have no effect, administered so all groups experience the same conditions
Placebo Effect
Tendency of many human subjects to show a response even when administered a placebo (often 20% or more of subjects)
Blocking
When groups of experimental units are similar, it is often a good idea to gather them into what can isolate the variability attributable to the differences to see the treatment differences
Matching
In a retrospective/prospective study, subjects who are similar in ways not under study may have this done and be compared with each other on variables of interest
Designs
Either completely randomized or a randomized block
Completely Randomized Design
All experimental units have an equal chance of getting treatment
Randomized Block Design
Randomization occurs only within blocks
Confounding
When the levels of one factor are associated with the levels of another factor