Introduction and revision Flashcards
(23 cards)
What processes are involved in designing the basic features of an experiment?
- formulating research hypotheses
- transform into treatment conditions (IVs) + select design (between/within subjects)
- measure=DV, response measure manipulated on the basis of the IVs.
What are the different types of treatment variables?
- quantitative - variation in amount (best, minimises confounding/nuisance variables)
- qualitative - variations in kind or type
- classification - systematic variation of characteristics intrinsic to the subjects (serious confounding variables)
Define nuisance variables.
Potential IVs which left uncontrolled exert a systematic influence on the different treatment conditions (e.g. experimenter effect, time of day, subjects selected)
Define dependent variable.
- a measure that will capture the hypothesised differences
- somehow dependent on the nature of the independent variable (causal link)
Describe the idea of control in experimentation.
Statistical control through random allocation, e.g. of pts to conditions or treatment conditions to rooms, avoiding systematic variations and a confounding variable, e.g. temperature.
Statistical experimental control can be achieved through randomising participants to conditions. When can this not be done?
With classification variables.
What is another term for between groups design?
Completely randomised design.
What does completely randomised design involve?
Subjects are assigned to a group, therefore any differences in behaviour among the treatment conditions are due to pts.
What is randomised block design?
Matched subjects (or within subjects/repeated measures)
What is the difference between a research and a statistical hypothesis?
- research = fairly general statement about the assumed nature of the world that gets translated into and experiment
- statistical = a set of precise hypotheses about the parameters of the different treatment populations.
What must the null and alternative hypotheses be/do?
Mutually exclusive and cover all possibilities.
What is the function of a null hypothesis?
To specify the values of a particular population parameter in the different treatment populations = no treatment effects present in the population. E.g. means are all equal.
When is there support for the alternative hypothesis?
When parameter estimates are too deviant from values in the null hypothesis, suggesting that there is an effect.
When is the alternative hypothesis accepted?
NEVER. The null hypothesis is rejected, which IMPLIES acceptance of the alternative hypothesis.
When is the null hypothesis accepted?
NEVER - statistical tests already used assume it is true, so cannot be used.
What are the criteria for rejecting the null hypothesis?
- Calculating test statistics based on the properties of the F-distribution.
- Adopt a value (alpha) called the significance level.
- Chance that differences observed by chance assuming that null hypothesis was true = p.
- p<a = reject null hypothesis.
- P is not the significance level.
- Beta = p(alt hyp true)
- Alpha is also type I error rate.
What is a Type I error?
Reality - H0 is true, H1 is false.
Decision - reject H0, accept H1.
What is a Type II error?
Reality - H1 is true, H0 is false.
Decision - reject H1, accept H0.
What is the probability of a Type I error?
alpha
What is the probability of a Type II error?
beta
As alpha increases…?
Beta decreases and vice versa.
Under what circumstances might we be more willing to accept Type I errors?
If it is important to discover new facts.
Under what circumstances might we be more willing to accept Type II errors?
If it’s more important to not clog up the literature with false facts.
