Bivariate Correlations PSY 3213
Popular in Research Methods in Psychology
verified elite notetaker
Popular in Psychlogy
This 5 page Class Notes was uploaded by Emily Notetaker on Friday March 11, 2016. The Class Notes belongs to PSY 3213 at University of South Florida taught by Dr. Brannick in Summer 2015. Since its upload, it has received 13 views. For similar materials see Research Methods in Psychology in Psychlogy at University of South Florida.
Reviews for Bivariate Correlations
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: 03/11/16
Bivariate Correlation 2 main Psych Traditions Experimental o People are all alike; general laws of behavior Color perceptions, operant conditioning Correlational o People are all different; laws relating individual differences to outcomes Conscientious and job performance; cognitive ability ad grades in school o Cause and effect only suggested; statistical controls to help eliminate possible causes; predictions can be useful even if cause is unknown (e.g., grades) Experimental Designs People are assigned to condition Usually one or two categorical IVs; single continuous DV Ex: effects on type of therapy and type of drug on depression Correlational Designs People are measured, not assigned to condition Usually one or (may) more continuous variables; may also have categorical variables Ex: relations b/w level of anxiety, perceptions of parental involvement, and depression Morling covers o Correlational o Then experimental Correlational Research Assessing the association b/w tow variables o 2 or more variables are measured and a relationship est. IV can be categorical or continuous (scale) Is intelligence in children correlated with the number of books in the home? Are GPA’s higher in public or private colleges? Review of the Three Causal Rules Covariance o Things go together or are associated o Not which comes first or whether one causes the other o Storks’ nests and babies in Stockholm Temporal precedence o Cause must precede effect Internal validity o There can only be one explanation Can we say anything about causality in correlational research? Covariance o Correlations can establish covariance Temporal precedence? o Directionality of effect problem Internal validity o Third variable(s) problem Threats to Causal Conclusions in Correlational Designs Can we make a causal inference? o Unless you can do an experiments, it is impossible to rule out all third variables. What does plausible mean in this context? Additional threats to claims Measurement – construct validity o How did they measure their IV and DV? o Are your measures good measures of the constructs? Watching TV is associated w/ depression People who like spicy foods are risk takers External validity o How did they get their sample? o Will the results generalize? A study at a Bible college showed that degree of endorsement of fundamental beliefs was related to the amount of time spent volunteering College professors showed no preference for offering a Dean’s job to a male or female job applicant Statistical validity 1. How strong is the relationship? The correlation is an effect size that tells how closely related two variables are By conventions, a correlation is usually labeled small if it is less than .2, and large if it is .5 or bigger. But importance is not the same as effect size (consider aspirin) 2. Is it statistically significant? Statistical Significance 1 In statistics, something is said to be significant if it is it likely to be observed by chance For example, if we have N=100 people, then correlation as large as .40 is very unlikely if the truth is that the correlation is zero in the population We report p values that indicate the likelihood that the observed statistic is due chance. In our .40 example above, we should print “p < .50.” Statistical Significance 2 Large values become likely as N gets small, so p depends on sample size as well as the observed statistic Can’t tell significance just by looking at r Unless the association is really strong, it helps to have a large sample (at least N > 100) By convention, we usually declare a result to be statistically significant if p < .05. 3. Are there outliers? 8 7 6 5 4 3 2 1 0 0 1 2 3 4 5 6 7 8 9 4. Is it curvilinear? Y-Values 8 6 4 2 0 0 2 4 6 8 10
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'