Download as pdf or txt
Download as pdf or txt
You are on page 1of 28

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/344138544

MANOVA/MANCOVA using SPSS

Presentation · July 2020


DOI: 10.13140/RG.2.2.19846.40004

CITATIONS READS

0 924

1 author:

Nasser Hasan
University College London
5 PUBLICATIONS   0 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Webinars in Psychometrics and Statistics View project

Belief-Importance Theory Questionnaire: Cross-Cultural Adaptation and A Psychometric Validation of the Arabic Version View project

All content following this page was uploaded by Nasser Hasan on 06 September 2020.

The user has requested enhancement of the downloaded file.


MANOVA/MANCOVA using SPSS

Presented by Nasser Hasan - Statistical Supporting Unit


7/23/2020

nasser.hasan@miami.edu
Overview
• Brief introduction of MANOVA/MANCOVA.

• Performing the Analysis Using SPSS.


MANOVA/MANCOVA using SPSS

Overview
MANOVA/MANCOVA

- MANOVA tests whether there are statistically significant mean


differences among groups on multiple DVs, [after controlling for
covariate(s) – MANCOVA].

- MANOVA/ MANCOVA is a straightforward extension of ANOVA/


ANCOVA, in which main effects and interaction effects of categorical
IV(s) are assessed on multiple DVs.
MANOVA/MANCOVA using SPSS

Overview
MANOVA/MANCOVA

- The unique aspect of MANOVA/ MANCOVA is that the variate


(supervariable, or a linear combination of dependent variables, Y*)
optimally combines multiple DVs into a single value that maximizes
difference across groups.

- In other words, a new DV (variate, supervariable, linear combination of


DVs) is created and then ANOVA is performed on the newly created DV
(Y*).
MANOVA/MANCOVA using SPSS

Overview
MANOVA Test Statistics

• Most MANOVA packages output many of the approximate multivariate tests. The
four most widely used measures for assessing statistical significance between
groups on the independent variables are:
o Roy’s Largest Root
o Wilk’s Lambda
o Pillai’s Criterion
o Hotelling’s Trace
MANOVA/MANCOVA using SPSS

Overview
MANOVA Test Statistics – What to Use?

- When there is only one factor with two levels, Wilks’ Lambda, Pillai’s trace,
Hotelling’s trace, and Roy’s largest root are the same. The associated F might be
slightly different, but the decision regarding whether effect is significant or not will be
the same.
- As sample size decreases, unequal n’s appear, and the assumption of homogeneity
of variance-covariance matrices is violated, Pillai’s criterion is more robust.
- In general, all four tests are relatively robust to violations of multivariate normality.
- Here are two suggestions:
- Roy’s root is not robust when the homogeneity of covariance matrix assumption is
untenable (Stevens, 1979)
- When sample sizes are equal, the Pillai’s trace is the most robust to violations of
assumptions (Bray & Maxwell, 1985).
Stevens, J. P. (1979). Comment on Olson: choosing a test statistic in multivariate analysis of variance. Psychological Bulletin, 86, 355-360.
Bray, J. H. & Maxwell, S. E. (1985). Multivariate analysis of variance. Sage university paper series on quantitative applications in the social sciences,
07-054. Newbury Park, CA: Sage.
MANOVA/MANCOVA using SPSS

Overview
MANOVA Assumptions

- Independence of observations
- Reliability of continuous variables
- Multivariate Normality (MVN) – MVN is assumed, but many times hard to
assess. Univariate normality does not guarantee multivariate normality, but if all
variables meet the univariate normality requirement then departures from
multivariate normality are inconsequential. As usual, with larger samples the
central limit theorem suggests normality.
- Linearity among all pairs of DVs – Departure from linearity reduces power as the
linear combinations of DVs do not maximize the difference between groups.
- Absence of multicollinearity and singularity among the dependent variables.
- Equality of variance-covariance matrices – variance-covariance matrices for all
groups (non-significant result from Box’s M test)
MANOVA/MANCOVA using SPSS

Performing the Analysis Using SPSS - MANOVA


Dataset

Please download the dataset using this link:

https://miami.box.com/s/tbt9oqs31certhwxr5l0z05a5h1q6uta

And open it in SPSS


MANOVA/MANCOVA using SPSS

Performing the Analysis Using SPSS - MANOVA


Variables in the Dataset

- Disability: Degree of disability à categorical


- Treat: Treatment (2 groups) à categorical
- WRAT_R: Scores on reading subtest of Wide Range Achievement Test
à continuous
- WRAT_A: Scores on arthimetic subtest of Wide Range Achievement
Test à continuous
- IQ à continuous
MANOVA/MANCOVA using SPSS

Performing the Analysis Using SPSS - MANOVA


Research Question

- Main effect of the degree of disability:


1. Disregarding the treatment, does degree of disability affect children’s
reading and arithmetic scores?
- Main effect of treatment:
2. Disregarding degree of disability, does treatment affect children’s
reading and arithmetic scores?
- Interaction between treatment and disability:
3. Does the effect of treatment on children’s reading and arithmetic
achievement differ as a function of degree of disability?
MANOVA/MANCOVA using SPSS

Performing the Analysis Using SPSS - MANOVA


Analyze à General Linear Model à Multivariate
MANOVA/MANCOVA using SPSS

Performing the Analysis Using SPSS - MANOVA


Move IVs to Fixed Factors and DVs to Dependent Variables
MANOVA/MANCOVA using SPSS

Performing the Analysis Using SPSS - MANOVA


Click on Plots à Move each IV to Horizontal Axis à Click on Add
MANOVA/MANCOVA using SPSS

Performing the Analysis Using SPSS - MANOVA


Click on Post Hoc à Move both IVs to Post Hoc Tests for
MANOVA/MANCOVA using SPSS

Performing the Analysis Using SPSS - MANOVA


Click on EM Means à Move everything from the left to the right
Click on Compare main effects and choose the appropriate method
MANOVA/MANCOVA using SPSS

Performing the Analysis Using SPSS - MANOVA


Click on Options and Choose needed descriptive
MANOVA/MANCOVA using SPSS

Performing the Analysis Using SPSS - MANOVA


SPSS output
The Box’s M of 6.86 indicates that the homogeneity of covariance
matrices across groups is assumed (F(15, 787.64) = .26, p = .99).
MANOVA/MANCOVA using SPSS

Performing the Analysis Using SPSS - MANOVA


SPSS output
• The Main effect of disability is statistically
significant (Wilks’ Λ = .26, F(4, 22) = 5.39, p =
.004, partial η2 = .50)
• The Main effect of treatment is statistically
significant (Wilks’ Λ = .14, F(2, 11) = 24.44, p <
.01, partial η2 = .86)
• The interaction effect is not statistically
significant (Wilks’ Λ = .91, F(4, 22) = .27, p =
.89, partial η2 = .05).
MANOVA/MANCOVA using SPSS

Performing the Analysis Using SPSS - MANOVA


SPSS output
MANOVA/MANCOVA using SPSS

Performing the Analysis Using SPSS - MANOVA


APA-write up:

A 3 (Disability) × 2 (Treatment) between-subjects multivariate analysis of variance was performed on two


dependent variables: WRAT-R and WRAT-A. Independent variables are levels of disability (Mild,
Moderate, Severe) and treatment group (Treatment, Control groups).
Results of evaluation assumptions of normality, homogeneity of variance-covariance matrices [The Box’s
M of 6.86 indicates that the homogeneity of covariance matrices across groups is assumed (F(15,
787.64) = .26, p = .99], linearity, and multicollinearity were satisfactory.
With the use of Wilks’ criterion, the combined DVs were significantly different by levels of disability
(Wilk’s Λ = .26, F(4, 22) = 5.39, p = .004, partial η2 = .50) and treatment group (Wilk’s Λ = .14, F(2, 11) =
24.44, p < .01, partial η2 = .86). No significant interaction was found (Wilk’s Λ = .91, F(4, 22) = .27, p =
.89, partial η2 = .05).
To investigate the impact of each effect on the individual DVs, a univariate F-test using an alpha level of
.05 was performed. Pair-wise comparison followed by a univariate F-test indicates that the significant
difference was found between children with mild disability and those with severe disabilities, only in
WRAT-A. The main effects of treatment were significant on both WRAT-R and WRAT-A, with
approximately equal effect. In particular, treatment group showed significantly higher means on both
WRAT-R and WRAT-A, when compared to control group.
MANOVA/MANCOVA using SPSS

Performing the Analysis Using SPSS - MANCOVA


Analyze à General Linear Model à Multivariate
MANOVA/MANCOVA using SPSS

Performing the Analysis Using SPSS - MANCOVA


Move IVs to Fixed Factors, DVs to Dependent Variables, and Covariates to
Covariates
MANOVA/MANCOVA using SPSS

Performing the Analysis Using SPSS - MANCOVA


SPSS output
The Box’s M of 6.86 indicates that the homogeneity of covariance
matrices across groups is assumed (F(15, 787.64) = .26, p = .99).
MANOVA/MANCOVA using SPSS

Performing the Analysis Using SPSS - MANCOVA


SPSS output
MANOVA/MANCOVA using SPSS

Performing the Analysis Using SPSS - MANCOVA


APA-write up:

A 3 (Disability) × 2 (Treatment) between-subjects multivariate analysis of covariance was performed on


two dependent variables: WRAT-R and WRAT-A, after controlling for IQ scores. Independent variables
are levels of disability (Mild, Moderate, Severe) and treatment group (Treatment, Control groups).
Results of evaluation assumptions of normality, homogeneity of variance-covariance matrices [The Box’s
M of 6.86 indicates that the homogeneity of covariance matrices across groups is assumed (F(15,
787.64) = .26, p = .99], linearity, and multicollinearity were satisfactory.
With the use of Wilks’ criterion, the combined DVs were significantly different by levels of disability
(Wilk’s Λ = .24, F(4, 20) = 5.15, p = .005, partial η2 = .51) and treatment group (Wilk’s Λ = .10, F(2, 10) =
45.13, p < .01, partial η2 = .90), after controlling for IQ. No significant interaction was found (Wilk’s Λ =
.95, F(4, 20) = .27, p = .96, partial η2 = .03).
To investigate the impact of each effect on the individual DVs, a univariate F-test using an alpha level of
.05 was performed. The main effects of disability was significant on only WRAT-A. The main effects of
treatment were significant on both WRAT-R and WRAT-A, showing a slightly higher effect on WRAT-A.
Pair-wise comparison followed by a univariate F-test indicates that the significant difference was found
between children with mild disability and those with severe disabilities, only in WRAT-A. The main
effects of treatment were significant on both WRAT-R and WRAT-A, with approximately equal effect. In
particular, treatment group showed significantly higher means on both WRAT-R and WRAT-A, when
compared to control group.
Thanks for Listening and Attending!
Multiple Regression Using SPSS
Any Questions?
Presented by Nasser Hasan - Statistical Supporting Unit
Can you please give us a minute to fill this survey as it will help
6/3/2020
us to evaluate our performance and take your feedback into
consideration for future webinars:
https://umiami.qualtrics.com/jfe/form/SV_a9N5Xta6OlybEeV

View publication stats

You might also like