Psyc 317 Test 4

measure reliability -factors that distort correlation coefficients

less reliable a measure, the lower its correlations with other measures - a correlation between two scales can never been high than either scale's cronbach's alpha

correlational research

used to describe the relationship between two or more naturally occurring variables

statistical significance of r affected by 3 thing

1. sample size- more participants= more sig 2. correlation magnitude- larger r= more sig 3. move alpha from .05 to .01 or .001

correlation and causation

a correlation between 2 variables DOESNT imply causation - even a PERFECT correlation of 1.00 don't mean causality

correlational coefficient

a statistic indicating the degree to which two variables are related in a linear fashion - pearson correlational c. ( r ) is used to measure ranges from -1.00 - +1.00

covary

change ( the variables can change together

magnitude of correlation

expresses the STRENGTH of the relationship is unrelated to (-/+ signage) .10= small .30- medium .5- large

on-line outliers-factors that distort correlation coefficients

fall in the same pattern as the data and artificially deflatw

off line outliers -factors that distort correlation coefficients

fall outside the pattern of the data and artificially deflate

scatterplot

graph of participants scores on two variables straight line= perfect correlation

p value

index of how likely the true correlation in the population is different from .00 rule: if less then 5% probability the size of the sample correlation is due to chance ( p<.05) results are considered " statistically diff" and reflect the larger pop

direction of relationship

is determined by the sign of - or +

restricted range - factors that distort correlation coefficients

participants score are confined to a narrow range of the possible scores on a measure - artificially lower correlations below what they would be if the full range of scores was present - more serious distortion if curvilinear relationship

nondirectional hypothesis

predicts that two variables will be correlated but doesnt specify whether it will be positive or negative

directional hypothesis

predicts the direction of correlation - positive or negative, more powerful and can reach stat sig. with FEWER participants

coefficient of determination

r is NOT on a ratio scale r^2 : proportion of variance in one variable accounted for by the other variable (systematic) r^2 is on a ratio scale and is easily interpretable

curvilinear relationship

r=.00 indicates NO linear relationsip or curvilinear relationship

outleir -factors that distort correlation coefficients

score so deviant from the data that one can question whether it belongs in the data set typically more than 3 SD away from mean