Assignment 6

Question 1

Analysis of variance

Answer 1

A method of hypothesis testing, normally used with more than two study groups, in which the likelihood of the outcome is estimated by a ratio of the variability between groups divided by the variability within groups.

Question 2

Variance between groups

Answer 2

The component of the total variability in ones data that measures the variability of group means around the grand mean. The between groups variance measures treatment effects if present. If there is no treatment effect the between variance component reflects only error variance.

Question 3

Variance within groups

Answer 3

The component of the total variability in ones data that measures error (i.e., natural) variability. This is measured by the variability of group measurements around the group mean.

Question 4

Statistical error

Answer 4

A conclusion error in hypothesis testing. This happens when one incorrectly rejects Ho or incorrectly does not reject Ho.

Question 5

Sum of squares

Answer 5

A component of variance that is formed by summing the squared differences of numbers around their mean.

Question 6

Mean square

Answer 6

Another term for variance. A mean square is formed by dividing a sum of squares term by its degrees of freedom.

Question 7

Factors

Answer 7

The systematic sources of variability in an ANOVA.

Question 8

Contrast coefficients

Answer 8

Positive and negative integers that are used in a sum of squares equation to test a particular pattern of means in a one-way ANOVA.

Question 9

Levels of a factor

Answer 9

The divisions of a factor. Each factor must have at least two levels.

Question 10

Type I error

Answer 10

A conclusion error in hypothesis testing when one incorrectly rejects Ho. This is analogous to a false positive.

Question 11

Planned comparisons

Answer 11

A method of hypothesis testing within ANOVA in which a specific pattern of group means is tested. A planned comparison is also called an a priori comparison because a global F does not need to be computed.

Question 12

Post hoc comparisons

Answer 12

The hypothesis testing that occurs among group means following a statistically significant global ANOVA. Post-hoc comparisons are also called a posteriori tests because they are conducted only if a global F test reaches statistical significance.

Question 13

F ratio

Answer 13

Used in the analysis of variance, the F ratio is the quotient of variance between groups divided by the variance within groups.

Question 14

Main effect

Answer 14

In a factorial design a main effect is the influence of one factor on the dependent variable while disregarding the other factors.

Question 15

Factorial design

Answer 15

An experimental design in which the influence of more than one factor is studied.

Question 16

Interaction

Answer 16

In a factorial design an interaction is the combined influence of more than one factor on the dependent variable.

Question 17

Test for trend

Answer 17

A specific type of a priori comparison in which one tests for a particular type of relationship (e.g., linear, quadratic) with an independent and dependent variable.

Question 18

Orthogonal components

Answer 18

Independent contributions to the total sum of squares in an ANOVA. In general, the term “orthogonal” means independent.

Question 19

Dunnett’s t-test

Answer 19

This is one of the post-hoc tests in which all groups are compared with a reference group. The results of the Dunnett Test should be reported only if the global ANOVA reaches statistical significance.