Basics Flashcards
Define:
Independant Variable
The variable(s) that are manipulated
Define:
Dependant Variable
The measured outcome
Define:
Population & Sample
Population: The complete group we want to generalise results for
Sample: Subsection of population that’re are being tested
Data from samples can be used to make inferneces about the population
3 types of experimental design
Experimental design = the way ppts are allocated to groups
- Between Participants
- Within Participants
- Matched pairs
Define:
Between groups
AKA independant measures design
When different participants are used for each condition
define:
Within ppts
Every participant takes part in every condition
Define:
Matched pairs
different ppts take part in each condition, but they’re are matched for demographics such as age and gender
Advantages of:
Between Participants
Reduces order effects
order effects = effects that happen as a results of ppts repeating task, practice effects or fatigue effect
Disadvantages:
Between Ppts
Participant variables could effect results
i.e. if all older participants were randomly allocated to one group
Advantages:
Between ppts
Reduces risks of ppts varibles effecting results
How do you
combat order effects
Counterbalance
Alternating or changing the order each ppt completes each condition
Advantages
Matched pairs
Reduces ppt variables as ppts are matched for demographics
Disadvantages
Matched Pairs
- More time consuming
- Loose 2 sets of data if 1 ppt drops out
define:
Descriptive statistics
Stats used to organise and summerise a set of data
Name 3
What are measures of centeral tendancy
Mean, median & mode
Name 4
Measures of dispersal
- Range
- Standard Diviation
- Standard Error
- Confidence intervals
Define:
Catagorical Data
Data that has no numeric value i.e. male or female
(use mode as measure of central tendancy)
Define:
Ordinal data
presented in ranked order i.e place in a race
Differnces between scores are not consitent
Define:
p-value
weather it was statistically signnficant
usually under 0.05
what is type 1 error
sayig someting is significant when there is no effect
how do we combat type 1 errors
replication of study
Define:
Normal distribution
Scores are distrubuted symetrically either side of the mean
(Produces bell curve)
Define:
Skewness
Measure of distrbution where if skewness = 0 then data is normally distributed