Introduction (Recap) Flashcards
What is p the probability of?
Observing results as extreme as observed if the null hypothesis is true.
How did Hume contribute to statistics?
He believed cause and effect should occur close in time (contiguity).
That the cause must come first (temporal precedence).
Effect shouldn’t occur without cause.
What is contiguity?
Cause and effect should occur close in time.
What is temporal precedence?
Cause must come before the effect.
How did Mill contribute to statistics?
He wrote about confounding variables.
Give two advantages and two disadvantages of experimental methods.
Advantages: randomisation and active control groups.
Disadvantages: Hard to execute in real-world settings. Hard to apply to real-world situations.
Give two advantages and two disadvantages of observational methods.
Advantages: High ecological validity. Sometimes the only option for some IVs.
Disadvantages: Sometimes difficult to eliminate the other confounds. Experimenter bias when recording data.
What is a dummy variable?
Dummy coding is a way of entering categorical data into the straight line equation.
Use it with two groups. One reference group.
Why can dummy variables be useful?
They allow for the straight line equation to be used for t-tests and ANOVAs.
What is a census?
When your sample consists of the whole of the population you are interested in.
Explain the significance of the p-value when you have a census.
There is no significance as there’s no need to generalise results.
Name 4 characteristics of a normal distribution.
Area under the curve is SDs.
Bell-shaped.
Symmetrical.
Infinite.
Define standard error.
The standard deviation of the sampling distribution of means.
What is the advantage to having lots of sampling means?
The closer to the population mean you will be.
What do standard errors tell us?
How widely spread sample means are around the population mean.
What is the equation for standard error?
standard deviation divided by the square root of the number of participants. s/root n.
Define confidence intervals.
CIs are a range of scores within which the population mean will fall for 95% of samples.
This does not necessarily mean that our own 95% CI contains the population mean.
How do you calculate the lower and higher boundaries of CIs?
Lower: mean - (1.96 x SE).
Higher: mean + (1.96 x SE).
What are the three Cohen effect sizes for correlations (small, moderate and large)?
Small: .1
Moderate: .3
Large: .5
What is the correlation between a variable and itself?
1.
Correlations aren’t transitive to other variables.
What does this mean?
If X and Y have a positive correlation and Y and Z have a positive correlation.
This does not necessarily mean that X and Z have a positive correlation!
How many DPs should you report all numbers to?
2DPs (except p-values and effect sizes).
When reporting correlations what 3 things must you describe?
Their direction (positive and negative), magnitude (small, moderate or large) and significance.
How do you calculate degrees of freedom for Pearson’s correlation?
N-2.
How do you use SPSS to conduct a paired t-test?
Analyse - compare means - paired-samples t-test.
Put both variables into ‘paired variables’.
How do you calculate Cohen’s d?
d = M1 - M2/ SDpooled.
SDpooled = root (SD1 squared + SD2 squared / 2).
What should you report for paired t-tests?
Mean + SD.
t(df) = , p = , d = .
How do you use SPSS to conduct an independent-samples t-test?
DVs into ‘test variable’ box.
IV into ‘grouping variable’ box.
Define groups.
What should you report for independent samples t-tests?
Mean + SD.
Levene’s test for both DVs.
DV1: t(df) = , p = , d = .
DV2: t(df) = , p = , d = .
How do you use SPSS to conduct a bivariate correlation?
Analyse - correlate - bivariate.
Put all variables into the ‘variable’ box.
What should you report for correlations?
Direction, significance, r, p and magnitude.