Confidence Flashcards
What is significance?
- Probability of H0 being true
- OR if H0 is true, how likely would the observed outcome be?
What is a type I error?
A ‘false positive’ - concluding there is a relationship/pattern/presence when there’s not
What is a type II error?
A ‘false negative’ - concluding there is no relationship when there is one
Which type of error is preferable in geography?
Type II
How do we minimise error?
Low significance thresholds
Describe the love canal error
- Love Canal, NY state
- Former landfill site - 22,000 tonnes of waste
- Sold at auction with no mention of contamination
- 1977 rain caused chemicals to come to surface
- Birth defects, skin disorders, leukaemia
Describe the Andrew Wakefield error
- Falsely linked MMR vaccine to autism in children aged 3-10 years
- Parents didn’t vaccinate their children
- Immunisation dropped from 92% to 73% in the UK
- Data had been misinterpreted
How to we avoid error?
- Challenge statistical tests for robustness
- Think about how big a sample you need before collecting data
Why do we need a large sample?
- Increasing sample size increases the statistical power of the test
- Large samples allow small changes to be detected
What is standard error of the mean?
- Standard deviation of the mean of the means of a number of samples from the same population
- Can be calculated from a single set of measurements
What is the central limit theorem?
Given a sufficiently large sample size from a population with a finite level of variance, the mean of all samples from the same population will be approximately equal to the mean of the population
How is standard error calculated?
Std. Error = (std.dev of population)/square root of n
For a large population the mean distribution is approximately….
Normal, even if the population distribution is not
What is standard deviation?
How individuals differ from the sample mean
How do we estimate standard error if we don’t know the population mean?
Estimated std.e = sample s/√(n-1)