Distributions Flashcards
What are the qualities of a normal distribution?
symmetrical, divided into deviations, each deviation has a known percentage of population.
Can we assume a normal distribution with random sampling design?
Yes but there is still room for error, so we calculate skew and kurtosis in our sample
Pearson’s First Coefficient of Skewness
(mean-mode)/st dev or
3(mean-median)/st dev (if no mode)
How to calculate skew
Pearson’s First Coefficient of Skewness:
(mean-mode)/st dev or
3(mean-median)/st dev (if no mode)
What are the values of mean, median, and mode in symmetrical distribution?
Similar or the same
What is the value of skew in a normal distribution?
Zero; the further from zero, the more skew
What is the acceptable range for skew?
between +/- 2
How to compute kurtosis
Divide range by 6- should be similar to your standard deviation
Is skew or kurtosis more important?
Skew- parametric stats can deal with a non-mesokurtic distribution if you have adequate skew
What is the relationship between skew, error, and power?
With a skewed distribution, we are more likely to make an error, and therefore have less statistical power.
What kind of information do z-scores give us?
individual scores in relation to distribution-instead of comparing to mean, as you do std. dev., it is a standalone score
What will z scores do in a normal distribution?
Follow the std. deviation
What will a z score equal to the mean be?
0
What will a z score 1 std. dev. above the mean be?
1
z-score formula
z=(x-m)/std dev
