Statistics Flashcards
What is a type I error?
A type I error is a false positive and means that you claim that a unimpaired enviroment is impaired or that a measure had an effect when it did not.
What is a type II error?
Is a false zero result meaining that you claim that an impaired enviroment is unimpaired or that a meausre did not have any effect when it did.
What does alpha mean?
If our samples indicate a change:
Alpha is the highest risk we take that this is wrong (p-value gives the actual risk).
If our sample do NOT indicate a change:
1-alpha gives the probability that this is not correct.
What is beta?
If our samples indicate a change:
1-beta (power) give the probability that this is correct.
If our sample do NOT indicate a change:
The risk that this is wrong is given by beta.
Which factors contribute to the power?
Significance level
Effect size, distance between H1 and H0
The variance
Number of samples
What is statitistical power?
The chance that we are right when saying that there is a significant change.
What is a priori test?
A priori means to, based on some kind of previous data, to calculate how many samples thata re required to detect a effect size.
What is a post hoc analysis?
Post-hoc means to calculate the power afterwards. Somwtimes critizied.
What is the standard error of the mean (SEM)?
SEM is ameasure between the sample mean and the true mean. It is suitbale to use in controlled settings with a low deviation (lab experiments).
What is the standard deviation?
Is measure of the sample distubution around the mean. Suitable when the variation is large (ecological studies). The variance is the avergre sums of squares.
Precesion
How close samples are to each other
Accuracy
How cloes different samples are to the true value
Random independ selection
All objects are numerated and selected randomly. All have the same probability to be included in the sample
Systematic selection
Samples can be taken in either regular time intervals, at intersections in a grid or transect in regular intervalls on a line
Stratified sampling
Population is divided into smaller homogenous groups, (sometimes very deviating groups can be excluded), draw samples from each group using either systematic or random sampling.
