Statistics Flashcards
What is the pearson correlation?
Extent to which the variability of a variable is related to another variable variability.
What is the difference between one-tailed and two-tailed hypotheses?
One-tailed explains the direction (> or <) and two-tailed just asks whether there is a difference without specifying what difference.
The determination coefficient (r squared) is
The shared variance in percentages
What does regression estimates?
The extent to which other variables (independent, predictor) explain the variance of a variable (dependent, criterion/outcome).
ax+b, which is the slope and which is the intercept?
a=slope
b=constant/intercept
What are the conditions for a linear regression?
-Dependent variable continous, normal distribution
-Independent variable any level (normal distribution for cont.)
-Medium correlation between predictor and outcome (higher or equal to .30)
-Predictors must be orthogonal (correlation between them lower than .30)
Research hypothesis is reject when p is lower or equal than alpha. True or false?
True
What does the independent T-test do?
Compares the means of 2 independent samples.
How are the variables in an independent T-test?
Independet variable=nominal level, dichotomous
Dependent variable=continuous (interval or ratio)
Homogeneity of variances is a condition for T-test that can be checked using:
Levene test or Hartley’s F Max
What are some types of dependent samples T-tests?
-Test-retest designs
-Natural pairs
-Analyse whether treatment affects investigated variable
What does an ANOVA one-way do?
Compares the scores obtained on at least 3 independent samples (analysis of variance)
ANOVA is an omnibus test. True or false?
True (it tests the overall effect without saying which group mean is higher or lower)
What test can be applied if one of the three conditions for ANOVA are not met (dependent variable continuous, normal distribution on each sample, homogeneity of variance)?
Kruskal-Wallis Test
What does post-hoc analysis do?
Computes the overall effect of independent variable on outcome (pairwise comparisons which control the error)