Distributions and Sign Test Flashcards
What will a normal distribution curve look like for an exam that was the right level of difficulty?
- a symmetrical bell shaped frequency distribution
- most of the scores are clustered close to the mid-point
- the mode,median and mean are all at the mid point
Why is this? (in terms of task difficulty and normal distribution)
- if the task is the right level of challenging, this means few students will set a high score as this is difficult to achieve
- also few will get low scores as the challenge level is right
What are negative skewed distributions and what do they show?
Non symmetrical frequency distribution
Where is the mean, mode and median on negative skewed distributions?
- the mean skews the distribution to the left because it is affected by the lower scores. (greater range on left than right)
- the mode of the scores remains on the right
Why is this? (in terms of task difficult and left/negative skewed = mean<median<mode)
- task is too easy (low level of challenge)
- majority will achieve high scores
- since test is very easy, most students will do well - get a high score + avoid low score
What are positive skewed distributions and what do they show?
- non symmetrical frequency distribution
- the mean skews the distribution to the right bc it is affected by higher scores (greater range on the right than left)
- mode of the scores remains on the left
Where is the mean,median and mode on positive skewed distributions?
- the mean skews the distribution to the right bc it is affected by the higher scores (greater range on right than left)
- mode of the scores remains on the left
Why is this? (in terms of task difficult and right/positiveskewed = mean>median>mode)
- majority will get low scores that lead to a positive distribution
- mean is pulled to the right because it is affected by minority high score
What is the role of inferential statistics?
enables a researcher to judge whether the data are:
-> significant (there is an actual difference/similarity between data)
-> not-significant (there is no actual difference/similarity between data, result occured due to chance)
One tailed test:
indicated by a directional hypothesis, a specific outcome is tested (goes in one direction)
Two tailed test:
Indicated by a non-directional hypothesis - a non specific outcome is being tested (go in both directions)
What are p-values:
The probability that the findings have occured due to chance
It informs us if the difference between two conditions is menial enough to be a fluke or significant enough to be due to manipulation of IV. These are written as a percentage or a decimal. E.g. 5% or 0.05.
What does it mean if p<_ 0.05
=5%
= Less than 5% likelihood that difference/ correlation is due to chance
= 95% likelihood due to manipulation of IV
What is the significance level (p-value) that psychologists use?
-> a stricter, smaller p-value is used when there is no previous research, or the researcher needs to be confident of the result
-> a lenient, larger p value is used when the study is a replication of existing research or the researcher does not need to be confident of the result
What does a lower p-value show?
More confident that difference/correlation is genuine