SES - Inferential Stats Flashcards
Left skewed?
Mean < Median.
Right skewed?
Mean > Median.
Descriptive stats?
Provide a concise summary of data either numerically or graphically.
What does sampling involve?
Randomly selecting a subset of people from a population who are representative of it, perhaps in stratified groups e.g. gender, sport etc.
Why do we use sampling? What is this known as?
To infer things about the population = Statistical inference.
2 types of hypothesis? What do they state?
- ) Experimental = States effect will be present.
2. ) Null = States effect will be absent.
What does testing hypothesis involve?
Applying or building statistical models to or of the data collected.
Why is mean a statistical model?
It is a hypothetical value and not one which is observed.
What does null hypothesis significance testing involve?
Computing probabilities to evaluate evidence for the competing hypothesis.
What does null hypothesis significance testing provide?
A rule-based framework for deciding whether to accept a hypothesis.
How many decimal places is probablility reported to?
2
4 basic principles of null hypothesis significance testing?
- ) We assume the null hypothesis is correct.
- ) We fit a statistical model to our data to determine how likely it is to get a result like this if the null hypothesis was correct.
- ) To determine how well the model fits the data, the probability of that data fitting the model if the null hypothesis if correct is calculated.
- ) If probability is small, we accept the experimental hypothesis.
P value? What is it known as?
A probability number showing the strength of the evidence against the null hypothesis.
Known as alpha level.
When do we accept the null hypothesis in relation to p-value?
2 data sets are compared and if they are not really different the p-value is closer to 1 and we accept the null hypothesis.
When do we reject the null hypothesis in relation to p-value?
2 data sets are compared and if they are very different the p value is closer to 0 and we reject the null hypothesis.
What is the standard across science when using p-values in relation to the null hypothesis? Example?
To use the p-values at or less than 0.05* to reject the null hypothesis and accept the experimental.
E.g. 95% due to systematic effect and 5% chance of error.
What is the term ‘significant’ used to describe in relation to p-values?
To describe the differences/relationships for which a p-value under 0.05 is found.
Type 1 error? What is it known as?
Occur where we erroneously find a p-value to reject the null hypothesis, when in fact we should accept it.
Known as a false positive.
Type 2 error? What is it known as?
Occur where we erroneously find a p-value to accept the null hypothesis, when in fact we should reject it.
Known as a false negative.
When do inflated error rates occur?
If we run multiple statistical tests on the same data set.