This preview shows pages 1–3. Sign up to view the full content.
ANCOVA LAB
STATISTICS 101A
PROFESSOR ESFANDIARI
The objective of this lab is to show you how to conduct and interpret the results of
•
ANCOVA (analysis of covariance)
•
The test of assumptions in ANCOVA, and
•
Posthocs in ACOVA.
As we discussed in lecture, the major objective of ANCOVA is to reduce error variance
through statistical control. In order to do this, we need to identify a covariate, or a
potential confounding factor or extraneous variable, that is correlated with the outcome.
The outcome scores are then adjusted for the effect of the covariate and ANOVA is
performed on the scores that have been adjusted for the effect of the covariate. So
basically ANCOVA is equivalent of ANOVA performed on scores that have been
adjusted for the covariate.
Another way to reduce error variance is through experimental control and that requires
making the variable that we think affects the outcome a factor in the model. For instance,
if we are examining the effect of two teaching methods on learning physics, but, we think
prior knowledge of physics affects the outcome, then we can make prior knowledge of
physics a factor in the experiment by dividing the participants into groups based on their
prior knowledge of physics (example above and below the mean or median) and making
it a factor in the experiment.
In a national mathematics study the following data was collected on eighth grade
students:
Pretest score on arithmetic in percentages
Posttest scores on arithmetic in percentages
The level of emphasis that teacher placed on “understanding the nature of proof”.
1 = little emphasis, 2 = some emphasis, 3 = a lot of emphasis
Research question:
If we control for the effect of the pretest score on arithmetic, is there
any relationship between the level of emphasis that the teacher plays on proof and score
on arithmetic test.
Let us first run a oneway anova without using pretest on arithmetic as a covariate and
see if we get significant results:
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document Table one: BetweenSubjects Factors
Value Label
N
emphasis on understanding
the nature of proof
1
less emaphasis
345
2
average
emphasis
93
3
more emphasis
This is the end of the preview. Sign up
to
access the rest of the document.
This note was uploaded on 06/22/2011 for the course STAT 101 taught by Professor Esfandi during the Spring '11 term at UCLA.
 Spring '11
 esfandi
 Statistics, Covariance, Variance

Click to edit the document details