Deck 9: Correlation Analyses

Question 1

A statistical test that can examine the relationship between two interval or ratio level variables while controlling for the influence of a third, potentially confounding one is

A) partial r.
B) Pearson's r.
C) factor analysis.
D) Kendall's tau.

Answer 1

partial r.

Question 2

Two variables reflected a strong, statistically significant negative correlation within a research sample. How should this be interpreted?

A) The values of one variable cannot have caused the values of the other variable.
B) The values of the predictor variable caused the values of the outcome variable.
C) The values of the outcome variable caused the values of the predictor variables.
D) The values of the predictor value may have caused the values of the outcome variable.

Answer 2

The values of the predictor value may have caused the values of the outcome variable.

Question 3

A form of statistical analysis that uses correlation to identify redundancy in a newly constructed data collection instrument so that it can be eliminated or equivalent forms of the instrument can be developed is:

A) factor analysis
B) cluster analysis
C) Multiple R
D) factor rotation

Answer 3

factor analysis

Question 4

If, in the previous question, the correlation had been -1.0, what would happen after the students are asked to line up by weight?

A) About half of the students will move; others will stay in place.
B) Only the 13th student will stay in place; everyone else will move to the same position but in the opposite end of the line.
C) Everyone will have to move, but in no discernible pattern.
D) No one will have to move.

Question 5

If one or both variables are interval or ratio level but skewed, a correlation coefficient can still be produced by computing:

A) partial r.
B) Kendall's tau or Spearman's rho
C) Pearson's r
D) Multiple R

Question 6

A form of statistical analysis that uses correlation to form homogeneous groups of cases based upon there responses to several different questions is:

A) factor analysis
B) cluster analysis
C) Multiple R
D) factor rotation

Question 7

Which correlation coefficient below suggests the strongest relationship between two variables?

A) .65
B) .05
C) -.72
D) .50

Question 8

Correlation between variables can be displayed graphically using a:

A) frequency polygon.
B) stem and leaf plot
C) scattergram.
D) pie chart.

Question 9

In a social work class, there is a perfect positive correlation (1.0) between height and weight. All 25 students are asked to line up by height, shortest to tallest. What will happen if they are then asked to line up by weight, lightest to heaviest?

A) About half of the students will move; others will stay in place.
B) Only the 13th student will stay in place; everyone else will move to the same position but in the opposite end of the line.
C) Everyone will have to move, but in no discernible pattern.
D) No one will have to move.

Question 10

Which pair of variables below are probably correlated, but the correlation is not a linear one?

A) age and earned income for ages 16 to 90.
B) automobile weight and gasoline mileage
C) earned income and monthly house payment
D) alcohol consumption and severity of marital problems

Question 11

If two variables reflect a perfect (1.0 or -1.0) correlation,

A) the body of the graph of the correlation will contain a straight line.
B) there will be no exception to the overall pattern of the relationship between variables.
C) the two "variables" are probably just two different indicators of the same variable.
D) all of the above

Question 12

A correlation of .70 between two variables suggests that about percent of the variations in one variable is related to the variations in the other one.

A) 1
B) 70
C) 49
D) 100

Question 13

Correlation analyses are used for:

A) construction of data collection instruments such as scales.
B) describing the strength and direction of a relationship between variables.
C) determining if a relationship between variables is statistically significant.
D) all of the above

Question 14

A table that displays the bivariate correlations between several different pairs of variables is called a:

A) scattergram.
B) correlation matrix.
C) step-down graph
D) beta weight

Question 15

Which pair of variables below reflect a strong linear correlation, but the correlation is probably explained at least partially by one or more confounding variables?

A) education level and per capita income
B) number of siblings and mother's education level
C) soft drink consumption and drowning rates
D) all of the above