week 10- inferential stats (tests of association)

Question 1

regression

Answer 1

• can make predictions and explore the relationship between two variables
• there are three types (simple, multiple and logistic)

Question 2

simple/linear regression

Answer 2

• identifies relationship between two continuous variables (numerical)
• perfect relationship = 1 unit of change in IV leads to 1 unit of change in DV
• regression line can be used to make predictions

Question 3

residual

Answer 3

distance between data points and regression line

Question 4

multiple regression

Answer 4

one DV (interval level) and several IV’s (numerical), see how several factors interact to impact the DV

Question 5

logistic regression

Answer 5

variables are binary/ordinal (categorical)

Question 6

interpreting multiple regression

Answer 6

a) correlation coefficient (r)
b) coefficient of determination (r2): proportion of variability in DV explained by IV’s, how good model is at determining IV
- closer to 1 means model predicts IV well
c) beta coefficient (B): degree of change in DV for every 1 unit change in IV
- quantifies magnitude of relationship

Question 7

types of inferential statistics

Answer 7

parametric and non-paramteric (decision on what to use is determined by nature of hypothesis being tested, level of measurement, distribution of the variable scores)

Question 8

parametric tests

Answer 8

• precise and most common inferential stats test used
• assumes that we’re using numerical data (interval/ratio), that DV is normally distributed and that we are estimating at least one population parameter
  ie. t-test, ANOVA, pearson’s R

Question 9

non-parametric tests

Answer 9

• variables can be nominal or ordinal (categorical)
• few if any assumptions
• does not estimate population parameters
  ie. chi square test

Question 10

t-test for independent groups

Answer 10

• tests differences in means between two groups
• each group consists of different people
• participants are only tested once
• DV must be numerical
• test statistic about 2 is desirable

Question 11

t-test for dependent groups (paired)

Answer 11

• tests differences in means between the same group (tested twice)
• assumptions for independent t-test apply

Question 12

ANOVA

Answer 12

• used to test differences in means between independent groups
• when 3+ groups are being compared on one DV
• participants are tested once
• test statistic is F statistic
• one and two-way (compares two DV) ANOVA

Question 13

repeated measures ANOVA

Answer 13

compares means at 3+ different points in time, same participants are tested more than once

Question 14

chi-square test

Answer 14

• most common non-parametric test
• allows you to determine difference in groups among nominal or ordinal data (compares frequency in each category)
• expresses a difference in proportions, not a difference in means

Question 15

assumptions of a chi-square test

Answer 15

1. categorical variable
2. mutually exclusive categories
3. expected frequencies of 5 or more
4. null assumes no relationship
5. large enough sample size (>6)