PSYO 271 Final Exam Terms

analysis of variance (ANOVA)

allows us to test more than 2 group means

same purpose as t tests

considers within and between group variability

systematic variability:

variability between our groups

Random Error

without knowing what other factors account for variability

grouping variable:

predictor that explains values in the outcome variable, AKA independent variable

outcome variable:

dependent variable

grand mean (Mg)

the mean across all groups

between-groups variability:

variability arising from group differences

within-groups variability:

variability arising within each group

within groups-sum of squares (SSW)

looking for distance between groups and the mean of the group to which they belong

Bonferroni Test:

series of t tests performed on pairs of groups

factorial ANOVA:

multiple grouping variables

repeated measures ANOVA:

each person is measured 3+ times

Correlations:

relationships between two continuous variables

covariance:

variables differing together

inverse relationship

as one variable goes up, another goes down

linear relationships:

middle points through a scatterplot would be best represented by a straight line

curvilinear relationships:

line through middle of data will be curved

correlation coefficients

between -1.00 and 1.00

magnitude reports strength

.10 = weak

.30 = moderate

.50 = strong

Pearson’s r

• r acts as a descriptive statistic like M

• tells us about the linear relationship’s magnitude and direction

• r also acts as a test statistic like t because we can compare it to a r*

coefficient of determination

can ALSO calculate r2 as an effect size

spurious correlations:

variables related simply due to random chance

range restriction

if our data doesn’t have the full range of variability of a variable

outlier:

datapoint far away from rest of observations in a dataset

Spearman’s rho (ρ):

finds relationships with ordinal data

line of best fit:

central tendency of scatterplot. close as possible to all points

distance between line of best fit and each data point =

error = residual

least squares error solution:

equation of the line of best fit gives the smallest possible value of squared errors/residuals

Intercept and Slope

intercept: where line crosses on Y axis

slope: steepness, directionality of line

sum of squares error/residual:

distance from observed score to the line of best fit

sum of squares total:

distance from observed score to the mean

sum of squares model:

difference from prediction line to the mean (aka the observed effect/ability to explain variance)

average size of the residual =

standard error of the estimate

multiple regression:

multiple X variables as predictors for a single Y variable at the same time