Chapter 18

Question 1

Coarse Categorization

Answer 1

simplify variables measured on an interval or ratio scale – reduces the estimate of association

Question 2

Contingency Table

Answer 2

a table where you lay out all the possible pairs of values that states might have on two variables: (ex. high-high, high-low, low-high, and low-low)

Question 3

Dichotomous Variable

Answer 3

when there are only two possible values for the variable

Question 4

Dummy Coding

Answer 4

under this convention, we arbitrarily assign a value of 1 to one level of the variable and a value of 0 to the other level – used for coding a dichotomous variable

Question 5

Linear Association

Answer 5

for some data, when we look at a scatterplot, it might appear that a curved rather than a straight line best summarizes the association between the 2 variables. By convention, however, we normally use a straight line as the prediction function, and thus, we normally refer to associations between variables as linear associations.

Question 6

Marginal Frequencies

Answer 6

they tell us how many states have a given value on one of the variables, ignoring their value on the other. (p.456 for more info).

Question 7

Median Split

Answer 7

splitting variables into two categorical variables in relation to the median: low for below the median, high for above the median

Question 8

Null Hypothesis

Answer 8

a statement that specifies what we hope is NOT true in the population.

Question 9

Partial Association

Answer 9

the association that persists between 2 variables when a 3rd variable that might explain their association is held constant.

Question 10

Pearson Product Moment Correlation Coefficient (r)

Answer 10

the most widely used index of association between continuous variables, ranges between -1.0 and +1.0 with = meaning no associations and +- 1.0 meaning perfect association

Question 11

P-Value

Answer 11

a probability level that gives you grounds to reject or accept the null hypothesis

Question 12

R^2

Answer 12

square of the value r and indicates the proportion of variability in one variable accounted for by the other variable; aka tells us by what proportion our predictions of one variable improve when we base those predictions on knowledge of the other variable

Question 13

Scatterplot

Answer 13

We construct a table in which each case is represented by a single point on a two-dimensional graph, positioned according to its values on the 2 variables.

Question 14

Statistical Significance

Answer 14

an index of the degree of confidence we can have that an association we observe in a sample would emerge if we were to replicate the study using another sample from the same population.

Question 15

Spurious Association

Answer 15

an association in which two or more events or variables are not causally related to each other, yet it may be wrongly inferred that they are

Question 16

T-Statistic

Answer 16

test statistic associated with r and is computed by

t= r/√(1-r^2 ) × √(N-2)

Question 17

Type I Error

Answer 17

This outcome results when we reject the null hypothesis of no association in the population when in fact there is no association. In other words, our sample data led us to conclude that there is an association when in fact there is none.

Question 18

Type II Error

Answer 18

We failed to reject the null hypothesis based on our sample data when in fact that null hypothesis is false and should have been rejected. In other words, we failed to realize that X and Y are associated in the population when in fact they are. We failed to realize that the hypothesis that motivated the research in the first place could be correct.

Question 19

ϕ (phi)

Answer 19

a symbol to index / assess how much of an association there is between two dichotomous variables. Values range from 0 to +- 1.0 with 0 indicating no association at all while +1.= or -1.0 indicates a perfect association between the two variables

Question 20

X^2 (chi square)

Answer 20

way to compute statistical significance of the association / the significance test mostly associated with contingency tables