Quiz (Chapter 3)
Bivariate Correlation
The Key Concept Behind Correlation: Relationship
 (T/F) The concept of "relationship" cannot be investigated if we
measure each person in a group one time only on a single variable.
 What kind of relationship exists among the math and science quiz
scores shown below:
 highhigh, lowlow
 highlow, lowhigh
 little systematic tendency one way or the other

Person 
Math 
Science

Will 
7 
5 
Jill 
9 
3 
Bill 
8 
4 

 Add 10 points to each person's math score in the previous question
and then answer it again.
Scatter Diagrams
 If a scatter diagram is used to show how 20 people score on 2 variables,
how many axes will there be in the scatter diagram? How many dots
in the scatter diagram?
 If a "highlow, lowhigh" relationship exists between two variables,
the "path" of dots in the scatter showing this relationship will extend
from:
 upperright to lowerleft
 upperleft to lower right
 Where would you be located in a scatter diagram if you've got the
top scores and its a "highhigh, lowlow" relationship?
The Correlation Coefficient
 Which of the following correlations is "highest"?
 .35
 +.80
 .05
 +.65
 .95
 (T/F) Since a correlation of +1 is "perfect," a correlation of 1
is as "imperfect" as can be.
 "Positive" is to "negative" as "direct" is to "_______."
 What lowercase letter is generally used to represent a correlation
coefficient?
 (T/F) When used in reference to correlation coefficients, the adjectives
"modest" and "moderate" mean the same thing.
 (T/F) If r is +1.00, the means of the 2 variables will necessarily
be identical to each other.
The Correlation Matrix
 Which two variables had the lowest correlation in Excerpt 3.10?
 If all possible bivariate ("pairwise") correlations are computed
among 10 variables, there will be at least ___ (how many) correlation
coefficients in the correlation matrix.
 If a correlation matrix has 4 rows & 4 columns, then there were
___ or ___ variables to start with.
Different Kinds of Correlational Procedures
 What kind of measurement scale leads to rankordered data?
 Which termdichotomous or nominalcovers the other (and covers
other things too)?
 If people are called "tall" if they're over 5'6" (and "short" if
they're not), is this a true dichotomy?
 What symbols are used to designate Pearson's correlation and Spearman's
correlation?
 Which correlational procedure handles data sets having ties better
than does Spearman's?
 What do the letters bis stand for in the notation rbis?
What about the letters pb in the notation rpb?
 If we were to correlate hand preference (left vs. right) with gender
(male vs. female), which correlational procedure would be most appropriate:
phi or tetrachoric?
 What letter of the alphabet is used to denote Cramer's correlational
procedure?
Warnings About Correlation
 If the correlation is greater than .90, can we assume that a causal
relationship exists?
 In Excerpt 3.9, the coefficient of determination for ES and AG (for
the paperandpencil version) is equal to __ .
 In Excerpt 3.8, approximately what % of the variability in Spelling
scores was explained by variability among the Social Studies scores?
 (T/F) Depending on their location, outliers can either increase or
decrease the size of r.
 Does the term "linear" apply to these 8 pairs of math (M) and geography
(G) scores?
Name 
M 
G 
Joe

2 
1 
Sam 
1 
2 
Sue 
3 
2 
Ann 
2 
3 
Bob 
4 
3 
Dan 
3 
4 
Eve 
5 
4 
Pam 
4 
5 
 Two variables are considered to be "independent" if the correlation
of scores on the variables turns out to be close to __ .
 (T/F) The decision as to how an r should be labeled (e.g., as "high"
or "moderate" or "low") is quite subjective, and different researchers
might use different labels even though their correlation coefficients
are exactly the same size.
A Few Extra Questions that are Supposed to be More Challenging
 (T/F) In Excerpt 3.7, the correlation between Family Support and
Avoidance/Detachment suggests that these two factors were not at all
independent of each other.
 Look at Excerpt 3.1. If each county's "Votes Cast" percentage
had turned out exactly 10 percentage points higher than its "Voters
Registered" percentage, what two terms could we use to describe
the relationship between the two sets of percentages?
 (T/F) Given 8 totally different scores, it's possible to create a
"highhigh, lowlow" situation by (a) positioning the top 4 scores in
the X column & (b) positioning the lowest 4 scores in the Y column.
Click here for answers.
