Chapter 4- Variability Flashcards
Variability
Having a range of values
It is the quantitative measure of the degree to which scores in a distribution are spread out or clustered together
Range
Highest score minus the lowest score
Interquartile Range
Range of scores that contains the middle 50% of a distribution
This is an important element of box plots
To find the IQR, you must find two different percentiles, the 25th and 75th. You subtract 75th- 25th
Ex. 0.25 x N, your answer is what place a score is. Like, if your answer was 10 then it would be the tenth number .
To find a semi- IQR- Take (75th-25th)/2
Standard Deviation
Descriptive measure of the dispersion of scores around the mean
It provides us with the width of the distribution
There are three different kinds of SD variability:
Lowercase sigma, S hat, and capital S
Standard deviation is the square root of variance
Deviation score formula
Shows clearly how each score contributes to the final result
Deviation score
Raw score minus the mean of is distribution
Tells us the number of points that a particular score deviates from, or differs from, the mean
Ex. If we got -5, we were -5 below the mean
Lowercase sigma
Measure of a population’s variability. A parameter. Describes variability when a population of data is available
Lowercase sigma and capital S have the same equation
S hat
Estimate of a population’s variability. A statistic. An estimate of lowercase sigma (in the same way that X bar is an estimate of mu). The variability statistic you will use most often in this book.
A hat on a symbol indicates that something is being estimated.
Note that the difference between S and lowercase sigma is that S hat has N-1 in its denominator because it IS AN ESTIMATE
Capital S
Measure of sample variability. Describes the variability of a sample when there is no interest in estimating lowercase sigma (which is the measure of a population’s variability)
Variance
Square of the standard deviation.
It is the standard deviation equation, before you take the square root of it. Variance is not very useful in descriptive statistics, but is in inferential statistics
What is considered an extreme score when talking about SDs and variability?
- More than 2sd above
- More than 2sd below
- Unrealistically large
- Unrealistically small
What factors can affect variability?
- Extreme scores (range is the most affected, IQR is the least)
- Sample size (range is the most affected, sd/variance/IQR are least)
- Sampling stability (range least reliable, sd/variance/IQR tend to be similar
- Open-ended distributions (sd/variance/range cannot be computed, only IQR can be estimated)