START OF EXAM 2 chapter 8

Question 1

bivariate correlations

Answer 1

associations that involve exactly two variables

Question 2

A study is correlational if it has…

Answer 2

two measured variables

Question 3

how can data be displayed in a correlational study?

Answer 3

scatter plot or bar graph

Question 4

How will the reported result of a correlational study be shown?

Answer 4

correlation coefficient or a difference between means

Question 5

what are the most important validities for an association claim?

Answer 5

construct and statistical validity

Question 6

effect size

Answer 6

the strength of an association and the importance of a result between 2 or more variables (typically large effects are more important but a small effect size can also be very important in certain cases)

Question 7

statistical validity

Answer 7

the extent to whch statistical conclusions are precise, reasonable, and replicable

Question 8

very small or very weak effect size

Answer 8

0.05 or -0.05

Question 9

small or weak effect size

Answer 9

0.10 or -0.10

Question 10

moderate effect size

Answer 10

0.20 or -0.20

Question 11

farily powerful effect size

Answer 11

0.30 or -0.30

Question 12

what effect size is unusually large

Answer 12

0.40 or -0.40, possibly too good to be true (dont trust these)

Question 13

A larger sample size does what to the confidence interval?

Answer 13

narrows it and makes it more precise

Question 14

confidence interval

Answer 14

margin of error of the estimate, how precise is the estimate, a range designed to include the true population value 95% of the time

Question 15

what does it mean if the confidence interval does not include 0?

Answer 15

it is statistically significant; it is unlikely to come from a population where the association is 0

Question 16

what does it mean if the confidence interval does include 0?

Answer 16

the relationship is not statistically significant you cant rule out that the true association is 0

Question 17

what does a smaller sample size do to the confidence interval?

Answer 17

wider confidence interval and less precise

Question 18

if a result has been replicated what does that say about the association?

Answer 18

we can be more confident about the association

Question 19

what could be affecting the association?

Answer 19

outliers, restriction of range, curvilinear relationships

Question 20

when are outliers problematic?

Answer 20

when they have extreme values on both variables, and the study has a small sample size

Question 21

what does restriction of range do to an association

Answer 21

It can make it appear weaker than it really is

Question 22

Do r values describe the data as curvilinear associations?

Answer 22

No, it wont describe it well so dont use it

Question 23

If the relationship between two variables is curvilinear, what does that suggest about the correlation coefficient?

Answer 23

the correlation coefficient might be closer to 0 hiding the relationship between variables

Question 24

how can curvilinear relationships be detected?

Answer 24

using scatterplots

Question 25

is internal validity necessary for an association claim?

Answer 25

no but we need to protect ourselves from the temptation to making a causal inference

Question 26

spurious correlation

Answer 26

a bivariate correlation that exists only because of a third variable

Question 27

can a third variable potentially explain a bivariate association?

Answer 27

yes but the third variable must correlate with both variables in the association

Question 28

In an association between height and hair length reveals that taller people tend to have shorter hair what is a potential third variable?

Answer 28

gender: men tend to have shorter hair and are taller than women

Question 29

moderator

Answer 29

when the relationship between two variables changes depending on the level of another variable, that other variable is a moderator

Question 30

how do you measure associations?

Answer 30

measure the first variable and the second variable in the same group of people, then use graphs and statistics to describe the type of relationship

Question 31

what happens when one part of the association is categorical and one is quantitative?

Answer 31

use a bar graph, NOT a scatterplot; this allows you to estimate the magnitude of difference

Question 32

What does r measure?

Answer 32

the direction and strength

Question 33

What do you use when both variables are quantitative?

Answer 33

scatterplot

Question 34

When everything else is the same, what is considered more important, a large or a small effect size?

Answer 34

a larger effect size is considered more important than a small one

Question 35

When are small effect sizes important?

Answer 35

when a small effect size combines over many people or situations, it can have an important impact

Question 36

What do outliers do to medium and strong correlations?

Answer 36

It can make a medium-sized correlation appear stronger or a strong one appear weaker than it is

Question 37

how is restriction of range corrected?

Answer 37

using a statistical technique that estimates the full set of scores based on what we know about the restricted portion of scores OR adding more people on both ends of the spectrum

Question 38

When does restriction of range occur?

Answer 38

Occurs when one of the variables have very little variance, ex: income and school achievement is only measuring middle class parents, there would be a restriction of range correction to estimate the results of low income and high income families

Question 39

When do you ask about restriction of range?

Answer 39

Only ask about this when the correlation seems weak because it typically makes correlations weaker; maybe when you get a result you don't expect.

Question 40

When should you investigate a curvilinear association?

Answer 40

If the relationship between variables is 0,

Question 41

What is a cuvilinear association?

Answer 41

The relationship between two variables is not a straight line (it might be positive up to a point and then negative)

Question 42

Does a third variable always suggest an internal validity problem?

Answer 42

no ex: gender and its effect on height and weight. It was found that in both men and women separately, the taller someone is, the heavier they typically are; gender didnt affect the trend

Question 43

What matters about the population when it comes to external validity in association claims?

Answer 43

The way the sample was collected from the population of interest matters the size of the sample does not

Question 44

What kind of sampling increases external validity, and if this sampling is not used, are the results not valid?

Answer 44

Random sampling. If a study doesn't use random sampling, dont trash the results just use that study in a future study that can be generalized

Question 45

Do moderators impact external validity?

Answer 45

can inform external validity but it may not generalize