stat tests and research methods Flashcards
The t-distribution:
The t-test takes into account both he expected mean and a measure of the standard error of the mean based on the sample.
one sample design
Benefits=
- can be used to compare group data with known variables
Disadvantage:
we may not always no population values and we may want to compare two groups, or to investigate the change of behaviour over time.
IMD
Independent measures design:
- We have two groups and the values come from different people (ie. Each person provides one measure in one group only) group 1 mean =4.75 and group b mean= 9
Benefits: the measurements are independent, we don’t have to worry about learning effects due to the repeated exposure
Disadvantages=
- people in the different groups might be quite different in various ways: personality, motivation etc. we need large sample sizes to average out these effects, or we need to counterbalance factors that we know may influence the results
- We cannot study the behavioral over time.
RMD
here is a single group which provides data for both conditions, ie. The value s from each condition come from the same people.
Advantages:
- we don’t have to think about the differences in baseline factors such as personality because there will alwaus affect both conditions equally
- We can study changes in behaviour over time
- We can usually test fewer people
Disadvantages:
- Measurements are not independent we need to calculate the variance differently
- People know the treatment after the first condition and cant be naieve in the second round. This might not work for every experiment
- We need to carefully counterbalance the conditions to avoid unwanted order effects.
comparing means
One sample t-tests
- We have one group with values coming from different people. This is compared to a single value.
Non directional hypothesis:
- A lot of the time in psych, you will be interested in whether there is any difference between the two things being measured, for this we would use a non directional hypothesis.
- In our sample t-test example thios would be something like H0= sample drawn from populatioon. And-
- It was hypothesied that there would be a significant difference in MEP’s between performing a mental imagery task and at baseline.
directional hypothesis
- For a directional hypothesis, we would use a one tailored test, H1= our sample is drawn from the population., it was hypothesised that MEPS would be sig higher when performing a m=MIT compared to baseline.
- Use a one tale test can be in pos and neg test.
- Important to first evaLUATE WHETHER THEORETICALLY AN EFFECT IN THE OTHER DIRECTION MAY BE likley, if not use a two tale-tailored test.
When looking at efficacy of psych treatment is also important to know whether our treatment
independent measures t-test
- Two different groups participating in different conditions, and we want to know if there is a change with MEPs. Group A imagines throwing the ball Group B imagine it lying on the floor.
T-test to test Whether there is a difference between two means that is significant looking for the difference between the two means minus the population statistics (normal ttest)
pooled variance
average of the two sample variances, this only works if we have unequal sample sizes, if they are unequal use different approach.
degrees of freedom
need to add up the sample size for each group and then subtract two. Two groups estimating the mean for each group, and in both of the groups one value will be constrained and cant vary.
to tell the difference between groups.,
If we had hypothesized any difference between the groups, that would bean we are interested in both directions possibly changing so we look in the 0.5. empirical t was smaller than the critical t this time which means no significant result.
paired samples t-test
- One group of participants but two measurements from condition A and condition B.
- Interested in examining the difference between two conditions.
Calculate difference between each participant two scores, and we the calculate sums of squares by using the mean difference. D subscript= for paired samples we are interested in DIFFERENCE of scores not the raw scores.
- Interested in examining the difference between two conditions.
effect size measures
- We also need to work out how significant the effect is.
- Standard error (sm)= standard deviation/ sample size (n)
- Larger value of estimated standard error in the denominator will create a smaller value for D, this means a large sample variance will make it less likley to obtain a significant effect
- Larger value of estimated standard error in the denominator will produce a smaller value for t
This means a large sample size variance sz will make it less likley to obtain a significant effect, also think about how larger variances make it more difficult to see consistent patterns or trends in data.
effect size, cohen’s d.
- Cohens d gives an estimate of the effect size this is independent of the sample size: d is the mean difference divided by the standard deviation.
- We can compare these estimated d values, d= 1.325 and d=6.732 to conventional cut-off values
- D= 0.2 small effect (mean difference around 0.2 standard deviations
- D=0.5 medium effect (mean difference around 0.5 standard d)
D=0.8 large effect (mean difference around 0.8 standard deviation)
effect size- r2
- Another way to calculate an effect size is to estimate the percentage of variation explained by the treatmentr2= t2/ t2+ df
- We simply use out t-statistic and the df to calculate r2
r=.01= small effect
r2= .09= medium
r2= 0.25= large effect
- We simply use out t-statistic and the df to calculate r2