Statistical Significance Flashcards
What’s a sample?
We can’t test everyone from our population of interest
Rely on samples, drawn from our population to test our hypothesis
There may be variability, but it gives us an estimate of our population statistic
What is random sampling?
Ideal
Everyone in the population of interest has an equal chance of being selected
Many types:
simple random- randomly pick people from a list of names
systematic- every Nth person from a random starting point
Multi-stage- select institutions randomly, select people randomly from there
What is stratified sampling?
Help with representativeness
Random sampling but with contraints according to the characteristic of your population of interest
Many types:
Stratified sample- select randomly from people sorted by a particular characteristic (age, gender, culture)
Disproportionate stratified- select more from an unrepresented group
Cluster sample- random samples from specific geographic locations
What are non-random samples?
More common
Quota- people recruited because they fit researcher defined categories
Convenience- people selected because they are easy to recruit
Snowball- other ps suggest similar people
Purposive- ps selected because they’re of interest to the research
Theoretical sample- collecting additional ps as theory develops
How can samples create a normal distribution?
Testing many samples create a normal distribution around our true population statistic
Reduce in frequency above and below this value
In most cases, our sample mean is close to the population mean
Some samples have a weaker or stronger correlation than the ‘true’ correlation
Why are experimental designs useful for samples?
Good experimental design is needed to make sure samples are representative and of an adequate size
What’s statistical significance?
Distribution can be formed around the null hypothesis Statistical significance is testing whether data from the sample could have come from this null distribution
If it’s unlikely that the null hypothesis is true within the sample, then the data is significant (p value is .05 or lower), the alternative hypothesis can be accepted and the null rejected
What’s the null hypothesis?
Prediction of no effect
Hypothesis states there will be no correlation
What’s an alternative hypothesis?
Our prediction of what we think will happen
Hypothesis states what type of correlation there will be between the 2 variables
What are critical values?
For every sample size there is a critical value
Put againts the p value to test if their data is statistically significant
What’s the difference between a directional and non-directional hypothesis?
Directional- one tailed
Non directional- Two tailed
What an issue with stratified sampling?
Difficult to implement for lab based research
What’s a t distribution?
Compares the mean from both groups
Related to the sample size
Heavier tails
T value is outside of critical value then the data is significant
What’s a p value?
If the null hypothesis is true, it’s the probability of obtaining the observed results
IT DOESN’T indicate the probability of the null hypothesis being true
P value has to be .05 or less (5%)
Low p values can occur when there is a bigger sample size, as there is less variability
What happens if a p value is above .05?
Means researcher’s can’t tell the difference between the null hypothesis actually being true, or the researcher just not having a study designed well enough to detect an effect
