L9 - Probability & Significance: Type I & II Errors Flashcards
Level of significance defined
“The level at which the decision is made to reject the null hypothesis in favour of the experimental hypothesis. It states how sure we can be that the IV is having an effect on the DV and this is not due to chance.”
Chance defined
“Something has no real cause, it just happens, e.g. by chance you are feeling happy today, there is no real cause that you can identify.”
What are significance levels?
- From the results gained from our experiment (for both the control and the experimental conditions) we would look for whether a real difference exists between the two sets of data, and how certain we are that there is a real difference.
- If the two sets of data are very similar, then a statistical test might indicate that chocolate makes no real difference to mood and we might accept the null hypothesis.
- However, if there is a probability that there is a real difference between the two conditions (and this can be proved by conducting statistical tests) then we would accept the experimental hypothesis and reject the null hypothesis.
What is probability?
- Probability is a numerical measure that determines whether our results are due to chance or whether there is a real difference that exists between the experimental and control conditions (and therefore we can accept the experimental hypothesis).
- If a real difference exists (that can be calculated statistically) we can say that results are significant, and the null hypothesis can be rejected (and we would accept the experimental hypothesis)
What significance levels is usually used in psychology?
- Usually p<0.05 (5% level) is used as:
• A) It is not too strict or too lenient, but is a middle, fair value of significance
• B) It minimises the chances of making a Type 1 or a Type 2 error (see later on). - “p” stands for “probability” and the 0.05 (5%) value illustrates the level of significance that has been chosen (5% level of probability that results are due to chance/fluke, therefore 95% certainty that our results are showing a real difference between control and experimental conditions)
- This means that if the level of significance is achieved, then the probability of results being due to chance is 5% or less.
- 5% significance levels are usually used when there is a directional one tailed hypothesis that has been clearly stated in the research
What other significance level are used?
• Sometimes a 10% level of significance is selected, and this is expressed as; p<0.10 (10%), and this is often used when we allow a 10% margin of error, and we would be 90% certain that our results are really showing a significant difference
• Sometimes a very strict level of significance is selected at 1% which is expressed as: p<0.01 (which indicates there is a 1% probability that the results are due to chance). This is often used when research findings are critical and are very important e.g. when testing the effect of drugs on humans, we must make sure that results are not due to fluke but that a real difference occurs between the experimental and control conditions, and that is why we set a stricter significance level.
What is a type I error?
- error means that we reject the null hypothesis and accept the experimental hypothesis instead
- however the results are due to chance & not significant
- null hypothesis should of been accepted
What is a type II error?
- error means that we reject the experimental hypothesis and accept the null hypothesis instead
- however the results are statistically significant & not due to chance
- experimental hypothesis should of been accepted