Central Tendency (Mean:
It can be affected by extreme…
It can be affected by extreme values so it may not be representative It uses all values so it is sensitive to central tendency.
Best used with interval level data.
Not affected by extreme values.
Useful if the data reflects categories.
Best used for nominal level data.
This is less sensitive than the mean as not all values in the set are representative.
It is not affected by extreme scores.
Best used with ordinal level data.
Spread of Data
It is easy to calculate
However it only uses twp values so its not sensitive or precise
It does not tell us how most values are clustered around the mean or whether they are evenly distributed.
The variance tells us about the spread of scores around the mean.
A small variance would imply that the scores are similar and close to the mean. A large variance would suggest that there are a variety of scores compared to the mean.
Standard deviation is the square root of the variance, so it tells us the average amount that a number differs from the mean.
A large standard deviation would mean a wider distribution of information.
Level of Measurements
When data is collected in discrete categories.
For example: agree or disagree.
Mode is the best measure of central tendency.
The data collected can be put into ranked positions or it can fall into natural order.
It is not measured on a standardised scale.
For example: degree of opinion- strongly agree, agree, disagree etc.
Median is the best measure of central tendency.
If you need equipment to measure, then it is interval level data.
For example: time, distance.
The mean is best used with interval level data.
This has a bell shaped curve and the mean, median and mode are all the same.
The data will be distributed symmetrically above and below the midpoint.
Examples: Blood pressure, heights of people, IQ scores.
Positive Skew: Mean, median and mode are not the same.
In a positive skew it is bunched below the mean and so bulges up the left.
It must be interval/ratio.
Must come from a population with a normal distribution.
They must have similar variances.
Statistical significance is the probability.
Psychologists use the rule that 5% or 0.05 is the cut off point to accepting or rejecting a null hypothesis.