Parameter: describes a population.
Statistic: describes a sample.
Correlational Method: 2 different variables are observed.
Experimental Method: 1 variable is manipulated and another observed.
Ind. Variable: manipulated by researcher.
Dep. Variable: observed and measured.
Continuous Variable: infinite number of possible values that fall between any two observed values.
Ordinal Scale: organized, ordered values.
Interval Scale: ordered categories, but that must form intervals of identical size.
Discrete Variable: Separate, distinct categories.
Interpolation: Take the distance from top of interval and divide it by interval width. Subtract the distance you found
from the top of the interval of whatever is being measured (i.e., percent). Each cumulative percentage is associated w/
the real limits.Scores: width = 1 pt. (half = 0.5 pts)
Percentages: width = 24% (half = 12%)
44%-12% = 32%
•Changing the value of any score will change the mean.
Mean for population: µ
Mean for sample: M
Weighted mean: M= Σ X 1 + Σ X 2
N1 + N2
•Adding/removing a score will change mean, except if score is = to mean.
•The height of a graph should be approx 2/3 to ¾ its length.
•Variability is defined in terms of distance.
Interquartile Range: Q3 – Q1
Deviation: distance from mean. X - µ
Pop. Variance: mean squared deviation.
Standard Deviation: √variance.
4.Sum of squares
5.Average of variance
6.Square root of variance
Pop. Standard Deviation: σ2 = √ SS N
Sample standard deviation:
S = (√SS/df)
Sample Variance: s2 = SS
n – 1
•Adding a constant to each score does not change SD
•Multiplying each score by a constant makes SD multiplied by the same constant.
Z-scores: determines the exact location of each X value within a distribution.
Z = X - µ X = µ + zσ
Variance for sample of z-scores: use n-1 when calculating z-scores for a sample!
Chebyshev’s Theorem: at least (1 – 1) of the data values must lie within z standard deviations of the mean, where z > 1.