Ch. 14 Flashcards
Goals of experiments?
- Eliminate bias
- Reduce sampling error (by increasing precision and power)
Design features that reduce bias
- Controls
- Random assignment to treatments
- Blinding
Controls
A group which is identical to the experimental treatment in all respects aside from the treatment itself.
Confounding variable
A variable that masks or distorts casual relationship between measured variables in a study
Experimental artifact
Experimental artifact is a bias in a measurement produced by unintended consequences of experimental procedures
Clinical Trial
Experimental study in which two or more treatments are applied to human participants
Methods of Reducing Bias?
- Simultaneous control group
- Randomization
- Blinding
Placebo Effect
An improvement in a medical condition that results from the psychological effects of medical treatment
Randomization
The random assignment of rtreatments to units in an experimental study
Advantage of randomization?
Random assignment averages out the effects of confounding variables
Blinding
Process of concealing information from participants (sometimes including researcers) about which individuals recieve which treatment
(Prevents behaviour change)
Single-blind experiment
Participants unaware of treatment recieved
Double-blind experiment
Like single-blind, except researchers administering the treatment and measuring the respons are unaware of which treatments the subjects are recieving
Methods of reducing the effects of sampling error?
- replication
- balance
- blocking
basically minimizing the “noise”
Replication
The application of every treatment to multiple, independent experimental units
How does replication reduce sampling error?
More units = more info = better estimates = more power
Balanced Experimental design
A experimental design in which all treatments have the same sample size
Blocking
The grouping of experimental units that have similar properties. Within each block, treatments are randomly assined experimental units
Randomized Block design
Like paired design, but for more than 2 treatmentsx
How does blocking work?
Group experimental units together, so that extraneous various between groups are accounted for
(ex. testing the treatment of a drug, but grouping patients by clinic because of potential extraneous variation between clinics)
Extreme treatments
Make treatments stronger to force an increase in the signal-to-noise ratio.
Help make detecting a response easier
Factor
A factor is a single treatment variable whose effects are of interest to the researcher
Factorial design
An experimental design in which all treatment combinations of two or more variables are investigated. A factorial design can measure interactiosn between treatment varialbes (ex. no effect w/ increased water or increased oxygen alone, but increasing both simultaneously have an effect)
Interaction
An interaction between two (or more) explanatory variables means that the effect of one variable depends on the state of the other variable.
Matching
A strategy in which every individual in the treament group is paired with a control individual having the same or closely similar values for the suspected confounding variables.
(cannot account for all confounds)