Experimental Design and Statisitcs Flashcards
Describe an Observational study
3 features
1) observe behaviour whilst it occurs naturally
2) descriptive method
3) can test hypotheses
What is a ‘variable’ ?
Any characteristic or factor that can vary
Issues with Observational studies (name 2)
OBSERVER BIAS - expectations of the observer & unconscious cues which influence the behaviour of P’s
REACTIVITY - change behaviour when we know we are being watched
How can we overcome issues with Observational studies?
OBSERVER BIAS
- code procedures
- make observers blind to hypothesis
- use several observers and rate their consistency
How can we overcome issues with Observational studies?
REACTIVITY
- use disguised observation for example two-way mirrors
- get P’s used to the observer
- use unobtrusive measures
What is a Correrlation study?
A relationship between two variables
Advantages of a Correlation study:
1) allows prediction
2) can study naturally occurring variables
Disadvantages of a Correlation study:
1) CANNOT establish a cause and effect relationship
2) third variable problem - unknown variable causing the change in the two variables rather than the variables themselves
What is an Experimental design?
- investigates the effect of an independent variable on a dependent variable
- variables are manipulated in a controlled setting
- all other conditions remain constant
What is the independent variable?
The variable that the experimenter manipulates
What is the dependent variable?
The variable that the experimenter measures
What is a ‘between subjects design’ ?
When participants are assigned to different experimental groups
INDEPENDENT MEASURES
eg one group of students is assigned to teaching method A whilst another group is assigned to teaching method B
Advantages of a between subjects design
- each P is naive to the experimental procedures
- essential when testing naturally occurring behaviours, eg gender
Disadvantages of a between subjects design
- large number of P’s needed for each experimental conditions
- individual differences eg personality traits
What is a “within subjects” design?
The same participants are tested under all the experimental conditions
REPEATED MEASURES
Advantages of a within subjects design
- fewer P’s needed
- no individual differences since the same P’s are used for each experimental condition
Disadvantages of a within subjects design
- order effects such as boredom or fatigue
* can stop this by using counterbalancing *
Define validity
Concerned with whether the method is measuring what it is supposed to be measuring
What is INTERNAL validity?
- deals with what is going on inside the study
- the extent to which an experiment supports a clear causal conclusion - is the IV causing the changes in the DV?
What factors limit internal validity?
1) confounding variables
2) expectancy effects
3) external validity
What are confounding variables?
An extraneous variable whose presence affects the variables being studied so the results you obtain do not actually reflect the actual relationship between the variables under investigation.
What are experimenter expectancy effects?
Subtle and unintentional cues which influence a P’s response
For example, an experimenter smiling when a participant behaves in the expected way
What are demand characteristics?
Cues that the participants pick up about the hypothesis which influence their behaviour
What is a placebo?
A substance with no pharmological effect
What is the placebo effect?
A change in behaviour/symptoms due to expectations
Influenced by the colour of the pill, packaging and knowledge of the practitioner
What are double blind studies?
- both the P’s and experimenters are kept blind to the experimental hypothesis
- this minimises the placebo effect as well as experimenter expectancy effects
What is external validity?
- the ability to generalise the findings of one experiment to other people and environments
- replication
What are descriptive statistics?
- describes a set of data using specific measures
- summaries and describes characteristics of data
What are inferential statistics?
- allow us to make inferences about a population based on findings from a sample
What is the mode?
The most common score in a set of data
How do you calculate the mean?
The result of adding all the values of data together and dividing the total by the number of data points.
What is the median?
The result when all the values in the data set are arranged in order and finding the middle value.
Define a ‘population’
The entire group that is of interest to the researcher
Define a ‘sample’
Subset of the population that the researcher is investigating
What is a normal distribution?
Bell shaped curve
Defined by the mean and the standard deviation
What is standard deviation?
Measures variation in terms of how far from the mean all the values in a sample fall.
What does a large standard deviation value mean?
The data varies a lot
What does a small standard deviation value mean?
The data varies a little
What does a zero standard deviation value mean?
All the values in the data set are identical
What does a t-test do?
Compares the differences between two samples or conditions
What is a t-value?
A ratio of:
difference between means / variability about the means
What does a large t-value mean?
The difference in means is large and the variability within the groups is low
What happens if the t-value decreases?
As the differences in means decreases, variability increases
What happens if the p value is < 0.05?
Reject the null hypothesis
Which t-test should you use for a within subjects design?
Paired
Which t-test should you use for a between subjects design?
Independent
What is an independent t-test?
A parametric test which compares means from two different separate groups
What do parametric tests do?
Make assumptions about our data
Eg - normal distribution, interval / ratio scale, homogeneity of variance
What is a paired t-test?
A parametric test which compares means from the same sample tested under two different conditions
When are non-parametric tests used?
When the data does not meet the assumptions of parametric tests - the data is not as distributed
Describe the Mann-Whitney U-test
Rank scores from the lowest to the highest
See if a group has a consistently lower ranking
Calculate a U value and the p value
A significant p value is <0.05
Describe the Wilcoxon Signed Ranks test
Calculate the difference between each pair of scores
Rank the difference scores are ranked from lowest to highest
Add together the ranks of the positive scores and the negative scores
The smallest value of the summed ranks = T value
A significant p value is <0.05