Deck 13: Analysis of Covariance

Question 1

A covariate is?

A)A confined variable.
B)A variable which is always negative.
C)A variable which shares some of the variance of another variable in the study.
D)A pair of variables which have exactly the same error ratios.

Answer 1

C

Question 2

How many simple assumptions of ANCOVA exist?

A)None
B)1
C)3
D)4

Answer 2

D

Question 3

ANCOVA can be used?

A)To adjust for the influence on an uncontrolled variable in some research designs.
B)To analyse any data.
C)As an alternative to correlated measures ANOVAs.
D)To justify the standard deviation.

Answer 3

A

Question 4

Convenience sampling sometimes causes undercoverage bias.

Question 5

If observations are not randomly sampled and independent of one another,what happens in ANCOVA?

A)Exact results
B)Inaccurate results
C)Increased risk of error
D)The sample size must be increased.

Question 6

A research fellow is conducting an experiment to test a hypothesis.Doubling the sample size will increase the power of the hypothesis test.

Question 7

Increasing survey size does not affect survey bias.

Question 8

Convenience sampling protects against undercoverage bias.

Question 9

Which of the following is true?

A)ANCOVA is a way of partitioning two-way ANOVAs to create separate tests of significance.
B)ANCOVA is a non-parametric version of ANOVA.
C)ANCOVA is an extension of ANOVA.
D)ANCOVA is a way of testing the statistical significance of ANOVAs.

Question 10

Analysis of covariance can be used to adjust for pre-existing differences in intact groups.

Question 11

In hypothesis testing,the p-value is greater than the significance level.

Question 12

In experimentally designed research,each experimental factor has one treatment level.

Question 13

Doubling the sample size will increase the effect size of the hypothesis test.

Question 14

Doubling the sample size will increase the probability of making a Type II error.

Question 15

Postal surveys are susceptible to non-response bias.

Question 16

It is not possible to use ANCOVA to increase statistical power,only decrease it.

Question 17

By randomly assigning subjects to treatment levels,randomization spreads potential effects of lurking variables roughly evenly across treatment levels.

Question 18

With experimental designs,blinding ensures that subjects in control and treatment conditions experience the placebo effect,but does not guard against confounding.

Question 19

In ANCOVA there should be one covariate fewer than the number of independent variables.

Question 20

Random sampling is a good way to protect against voluntary response bias.

Question 21

In hypothesis testing,the p-value is computed from the significance level.

Question 22

In regression analysis,when the sum of the residuals is greater than zero,the model is non-linear.

Question 23

The p-value is a probability.

Question 24

What is invalidated if the distribution is skewed and sample sizes are small in ANCOVA?

A)p-values
B)z-scores
C)t-scores
D)Error ratio

Question 25

Influential points always reduce the correlation coefficient.