COMM 2513 Statistics Week 4
COMM 2513 Statistics Week 4 COMM 2513
Popular in Introduction to Statistics
Popular in Communication
This 2 page Class Notes was uploaded by Rachel Notetaker on Tuesday September 22, 2015. The Class Notes belongs to COMM 2513 at University of Oklahoma taught by Johnson in Summer 2015. Since its upload, it has received 48 views. For similar materials see Introduction to Statistics in Communication at University of Oklahoma.
Reviews for COMM 2513 Statistics Week 4
Report this Material
What is Karma?
Karma is the currency of StudySoup.
Date Created: 09/22/15
COMM 2513 Statistics Week 4 Correlation A A standard score that tells you the strength and direction of a relationship between two variables B Answers how two variables are linearly related C Correlation does not answer the question of which caused the other D Pearson Correlation Coef cient r sample correlation 1 Shows how a straight line can be t through the data points a What line can be drawn that ts through the data points best E Types of linear relationships 1 Positive 0 lt r 5 1 r should be between 0 and 1 but will not be equal to 0 but can be equal to 1 X and Y more in the same direction 2 Negative As X goes up Y goes down perfect negative will fall all on one line but the slope is negative 3 Zero A perfect zero means no relationship between our two variables When graphed it looks like a box No slope up or down Knowing about X doesn t tell us about Y F Calculating r r NEXY EXEY lNEX2 EX2NEYZ EY2 N number of pairs of scores EX add all of the X s EY add all of the Y s EXquot2 square each X and then add them all up EYquot2 square each Y and then add them all up EXquot2 add all the X s and then square the total EYquot2 add all the Y s and then square the total EXY multiply each X by each Y and sum products Characteristics of the correlation Powgt Each variable must be quantitative r always falls between 1 and 1 Linear only Rquot2 Coef cient of determination 1 The proportion of variability of Y that is explained by X 2 Allows you to compare the strength of two correlations P the population correlation coef cient a So correlations can be descriptive or inferential b r is the symbol for the sample correlation c use Greek symbols when talking about the population values lll Factors which may affect the correlation A Restriction of range always lowers the correlation B Outliers sometimes increase sometimes decrease Scatterplot make one to determine