THINGS TO KNOW Flashcards
Define Null Hypothesis:
If 2 samples come from the same population, then ideally sample means should be identical.
Define Alternative Hypothesis:
Experimental manipulation has affected subjects. 2 samples are from diff population with different means.
What can a large difference between sample means suggest?
- 2 samples are actually from two different parent populations. Initial assumption that samples are from the same population are wrong.
- 2 samples are a poor reflection of the mean of the single population they are meant to represent.
The bigger the difference between the 2 samples…
the less likely one becomes and the more likely 2 become.
How many decimal places should you report too?
2 Decimal places.
What is homogeneity of variance?
The spread of scores roughly similar within each group, means both means are equally representative of their respective group.
At what point do results show inhomogeneity of variance?
When p is less than .05
What does Kurtosis mean?
The steepness and shallowness of a distribution curve.
What type of Kurtosis is Leptokurtic?
A steep curve, positive kurtosis value
What type of Kurtosis is Mesokurtic?
A normal curve, zero kurtosis value
What type of Kurtosis is Playkurtic?
A flat curve, negative kurtosis value
Define skewness:
Lack of symmetry in the distribution of scores.
Define negative skew:
Mean and median are smaller than the mode.
What is a type 1 error (Alpha error)
Rejecting the null hypothesis when it is intact true. Accepting the experimental hypothesis then finding it is actually due to chance.
How do you reduce a Type 1 Error?
Change the significance to .05 to .01
What is a Type 2 Error (Beta Error)
Accepting the null hypothesis when it is actually false. Researcher couldn’t find a significant result when there actually was one.
How do you reduce a Type 2 Error?
Increase the sample size. The larger the size, the easier to detect effects between groups.
What is a potential problem of trying to reduce a Type 1 Error?
You may increase the chances of creating another Type error.
What is a condition which distribution of a sample is normal?
The sample size is bigger than 30, the central limit theory.
What is z-distribution?
A normal distribution with mean of 0 and SD of 1.
What is a directional hypothesis? (One Tailed)
States the direction of the difference we expect to find.
What is a non-directional hypothesis? (Two Tailed)
Expect an effect, do not state the direction of the effect.
What one is most used in research?
Two tailed.
Why do we use 2 tailed?
- Its more easier to obtain a result
2. Rare to carry a one tailed test and then replicate the exact same results. Research tends to extend further research.