Variation Flashcards
biased estimator
A random sample statistic that does not equal the value of it’s population parameter on average
degrees of freedom
the number of scores that are free to vary in a sample: the last score calculated is not free, therefore must be removed (n-1)
Chebyshev’s Theorem
For any distribution (including non-normal), 75% of the data will always fall within 2 SD of the mean, and 89% of the data will fall within 3 SD of the mean.
Population Variance
The average SQUARED distance population scores deviate from their mean. SS/N (not n-1)
Sample Variance
The SQUARED average variance of sample scores from their mean
SS/(n-1) (not N)
Sum of Squares (SS)
The differences of each score from their means squared and then added together. Used in SS/N and SS/(n-1)
Deviation
The difference of each score from their mean
Empirical Rule
In a normal distro of sufficient sample size, 68% of scores will always fall within 1 SD of mean, 95% fall within 2, and 99.7% fall within 3.
Interquartile Range
The middle 50% of scores, obtained by “chopping off” the top 25% quartile of scores and the bottom 25% quartile
Population Standard Dev
sqrt of SS/N; the square root of the population variance
Sample Standard Dev
Square root of Sample Variance:
sqrt of SS/(n-1)
Sum of Squares Computational
SS=(Sigma X^2)-(Sigma X)^2/n (The second sigma is the square of the sum of original scores, and the first is the sum of the squares of each original score)
Population Sum of Squares Definitional Formula
Sigma (x-M)^2/N
Sample Sum of Squares Definitional Formula
Sigma (x-M)^2/(n-1)
Quartiles
The quarters of a normal distribution, each with 25% of data.
