Week 2 - Research Methods Flashcards
Correlation & the Scatter plot
Distributions that show the relationship between two variables (x & y) are called bivariate distributions typically displayed in a scatter plot.
Scatter plot –> indicates strength and direction of relationship
Correlational method
*Used in describing the relationships among naturally occurring variables ie. without imposing manipulations
*Commonly used in personality research to explain:
~ the relationships among various personality traits (eg are aggressive people more egocentric? –> co-vary)
~ the relationships between personality and some specific behaviour (eg are anxious people more likely to plan for retirement?)
*interrelated statistical analyses (Pearson correlation, partial correlation, regression analysis, structural equation modeling, & factor analysis)
Strength of relationship
Defined by degree of linearity
Positive (bottom L to top R)
Negative (top L to bottom R)
Correlation coefficient
Strength of the relationship between two variables is indicated by the size of the correlation coefficient (rxy).
Cohen & Cohen: (r .10 = sml, r .30 = med, r .50+ = large)
R2 = the proportion of variance in y scores that can be accounted for by variation in x scores (r2=.5x.5=.25 –> 25% of variance in y scores can be accounted for by variation in x scores ie common variance; 75% = unique or specific variance)
-1= perfect negative; +1= perfect positive; 0= no relationship
P = stat meaningful R= practically meaningful
Correlation limitations
Direction of causation –> with Correlational research it is not always clear to what extent personality traits can be said to have cause primacy.
Descriptive and incapable of isolating causal action
Third variable problem: the possibility exisits, pending further investigation, that some third, as yet unknown or unmeasured variable has causal primacy.
Experimental method
Hypothesised cause-effect relationships are put to a direct test.
Independent variable - cause, situational
Dependent variable - effect, behaviour
IV, believed to be the cause is manipulated to see weather it has the hypothesised effect of the DV.
Statistical tests - t-test, ANOVA, etc
Eg bandurra & bobo doll
Experimental limitations
Ecological validity –> to carry out manipulations of the IV, experiments have to be performed in labs or artificial settings therefore taking behaviour out of context & raises questions about legitimacy.
Sometimes experiments are simply not possible –> ethical reasons! technical reasons.
Experimental strengths
Has the ability to manipulate variables of interest - establish cause-effect relationships.
Independent replications
Mixed methods
Non-manipulated IV (ie participant variables) –> any time the IV groupings reflect naturally occurring differences
