Popular in research methods
Test Prep (MCAT, SAT...)
verified elite notetaker
Popular in Psychlogy
This 4 page Study Guide was uploaded by Valerie Ho on Thursday October 8, 2015. The Study Guide belongs to 209 at University of Washington taught by Jacqueline Pickrell in Spring 2015. Since its upload, it has received 36 views. For similar materials see research methods in Psychlogy at University of Washington.
Reviews for Chapter Notes
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 10/08/15
Chapter 5 Revision Correlation and Correlational Research Correlation Association between two variables X and Y Correlational Research Positive and Negative Research To understand potential causeeffect relations Examine potential associations between naturally occurring variables lt s process 0 Measure X Measure Y NOT manipulate o Statistically analyze 0 Reduce confound variables ve X increase Y increase ve X high Y tend to be low Measuring Correlations The Correlation Coef cient Pearson39s r Interpreting the Strength of a correlation Measures direction and strength of variables measured on an intervalratio scale 0 Value Range from 100 to 100 0 Strength measured by absolute value Spearman s rho rankorder coefficient Measure relation of 2 quantitative variables when they re measured on an ordinal scale Cohen1998 Strength Absolute Value Weak 010029 Moderate 020049 Strong 050100 eg Correlation between TV watching and perceived crime risk 079 doesn t mean that the two are quot79 relatedquot correlation coefficient percentage of relation Coef cient of Determination r2 eg TVcrime r079 r2062 inY The proportion of variance in Y that is accounted for by the variance of X Thus 62 of the variance in the scores for perceived crime risk is statistically accounted for by the variance in the scores for TV watching l It informs us about the extent to which changes in X help predict changes Scatter Plots X axis horizontal are for variable X Y axis vertical are for variable Y Perfect correlation Points make a straight line on the graph Correlation does not establish Causation Reason Explanation The Bi Since the researcher measure variables instead of directionality manipulate temporal order between variables are Probem ambiguous We don t know whether X has caused Y or vice versa The Third Variable Probem When there s an alternative explanation for the correlation of X and Y ie there s no casually link between the two This quotalternative explanationquot confounding variable Z eg TV crime risk anxiety of people may be Z If it is actually peoples anxiety that is drawing the scores X and Y to increase then the relationcorrelation is not genuine namely spurious correlation Z does not have to be a cause of both X and Y for it to be an alternative explanation as long as it relates with one variable then the problem still exists Gaining and clearer casual picture Statistical Approaches Partial correlation 0 Deals with the third variable problem 0 Measure correlation of X and Z Y and Z so that they can be statistically be quotpartialedquot tered out of the analysis between X and Y 0 Thus the statistical relationship between variables X and Y are adjusted for Z Research Design Approaches Longitudinal Designs When data are gathered on same participants on 2 occasions overtime Two types Prospective Design CrossLagged Panel Studies When X is measured earlier Process thanY 1 MeasureXandYat Time 1 Thus reduce chances of the 2 Measure X and Y at Time 2 bidirectionality problem 3 Examine the pattern of correlations among X1 Y1 X2 Y2 Textbook example Leonard Eron and his colleagues 1960 Measured 3rCI grader s TVwatching preference and aggressive behavior Then again after 10 years to examine the correlation between the 4 variables Combining Stats and Design Approaches A combination of Partial correlation and longitudinal studies may be able to illuminate the bidirectionality and more of the thirdvariable problem thus gain a better picture for casual relationships 0 However it is impossible to illuminate all Z variables within a correlational study since there would always be factors that in uence a relationship within the real world Media Reports of Correlationa Findings Often give false impressions to readers that a casual relation was found instead of a correlational result Correlation and Prediction 0 Correlation enables prediction even without establishing a causal relationship eg if SAT scores correlate with college GPA then most likely higher SAT scores will predict higher college GPA The stronger the correlation between two variables will suggest a higher accuracy when predicting one variable from the other Using 1 predictor Using 2 Predictors Regression Analysis Multiple Regression aka simple linear regression Predicting scores of 1 variable from predict scores of 1 variable from another Explores quantitative linear relation between two variables Criterion Predictor Variable Variable Try to For estimation estimate Analysis of this sort is performed by a statistical software program 0 Usually generating a regression equation for X and Y Scatter plotting The regression line is a representation of the regression equation generated through computer software 2 or more factors eg predicting college GPA from SAT scores and high school grade 0 Using multiple variables for prediction doesn t mean that a combination leads to higher chances of A occurring 0 Also if two predictors do not correlate at all with each other then each predictor will contribute completely unique info to the prediction of for example college GPA Bene ts of Correlational Research Prediction in daily life Test Validation Venturing Where experiments cannot tread Hypothesis and Model Testing Convergence with Experiments Special Issues Nonlinear Relations Range Restriction Associations involving Qualitative Variables
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'