[MORPHMET] Procrustes ANOVA

[Tarek Ismail]

Dear All 

I preparing a geometric morphometric work on an isopod species in Egypt. In 
morphoJ, the Procrustes ANOVA results give 0 values with one of the 
parameters (i.e. sex). The article deals with shape changes in two populations. 
in the analysis, the population is the main effect (Extra 1), sex (Extra 2) and 
individuals is the random effect.  I get the SS and MS for sex 0. So, what 
is the problem? I need help in this point. the results are:

Classifiers used for the Procrustes ANOVA:
Extra main effect(s): 
-- Extra 1: Population
-- Extra 2: Sex
Individuals: Individuals
Error 1: Image

Centroid size:
Effect  SS  MS dfF  P (param.)
Extra 11795651.412017   359130.282403   5   7.46  <.0001
Extra 2  0.000.00   2   0.00  1.
Individual 9726394.53609948150.468000 2023709.53  <.0001
Error 1   2699.886705   12.980225 208   0.25  0.9812
Residual   105.123498   52.561749   2

Shape, Procrustes ANOVA:
Effect  SSMS   df F  P (param.)   
Pillai tr.   P (param.)
Extra 10.24092157  0.0005125991  470   5.94   <.00013.27  <.0001
Extra 20.  0.00  188   0.00   1.0.00  1.
Individual 1.63916376  0.863263  18988  25.13  <.0001   67.18 <.0001
Error 10.06715515  0.034347  19552  0.53   1.
Residual   0.00121531  0.064644   188


Thanks in advance

Tarek G. Ismail
t_gad_2...@sohag.edu.eg

-- 
MORPHMET may be accessed via its webpage at http://www.morphometrics.org
--- 
You received this message because you are subscribed to the Google Groups 
"MORPHMET" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to morphmet+unsubscr...@morphometrics.org.

RE: [MORPHMET] Procrustes ANOVA for FA Analysis

[imronnawawi]

Sent from Mail for Windows 10

From: M. Viviana Toro Ibacache
Sent: Friday, 3 February 2017 9:27 PM
To: Kate Rusk
Cc: MORPHMET
Subject: Re: [MORPHMET] Procrustes ANOVA for FA Analysis

Hi Kate,

If I understood correctly, I am not sure that you can do a Procrustes 
ANOVA-based FA test on a list of individuals where only few of them have been 
digitized a second time. If this is the case, then I would suggest this 
approach (might be other ones using what you already have but this is fast and 
works for me):

- have a separate landmark file that contains only the specimens you repeated 
(5 landmarks sets + 5 repeatded landmark sets = 10 landmark sets), whose names 
end on "-rep1" and "-rep2" or something alike but with the same amount of 
characters.
- create a classifier (eg "repeat) using the last characters of the name (in my 
example, first character would be -5 and the last, -1). Remember you have to 
create an "individual" classifier that covers all the remainig characters.
- run the procrustes ANOVA on this 10-individuals landmark set using as error 
term the "repeat" classifier. This way you will be able to quantify variation 
due to repeated landmarking against variation due to individual*side.

I hope this helps,
Viviana

2017-01-30 20:56 GMT+01:00 Kate Rusk <kmru...@gmail.com>:
Hello everyone,

I am a bit confused about how to properly import and classify data for a 
Procrustes ANOVA in MorphoJ. Here is some more information about my data:

15 3D landmarks on human faces (object symmetry)
479 individuals
15 landmarks from 5 individuals digitized a second time
Sex and age as additional main effects

I realize I need to include a classifier that indicates a repeat (to be added 
as an Error classifier), but I am unsure how to import these data. So far, I 
have imported 1) the original data text file which includes the identifier and 
the XYZ coordinates for the original 15 landmarks and 2) a classifier text file 
which includes sex and age as covariates. I have separate XYZ coordinates for 
the repeat landmarks, but I don't know how these should be imported to 
implement the "repeat" classifier.

Any assistance you could provide would be greatly appreciated!

Thanks very much,
Kate

--
MORPHMET may be accessed via its webpage at http://www.morphometrics.org
---
You received this message because you are subscribed to the Google Groups 
"MORPHMET" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to morphmet+unsubscr...@morphometrics.org.




-- 

Viviana Toro-Ibacache DDS MSc PhD
Profesora Asistente (Assistant Professor)
Centro de Análisis Cuantitativo en Antropología Dental (CA2)
Unidad de Anatomía, Instituto de Investigación en Ciencias Odontológicas
Facultad de Odontología Universidad de Chile

Dirección:  Sergio Livingstone Pohlhammer 943
(address)   Independencia, Región Metropolitana
                 Chile
Email:        mtoroibaca...@odontologia.uchile.cl
Web:         www.researchgate.net/profile/Viviana_Toro_Ibacache/
                 www.cienciaymemoria.cl/staff.html
-- 
MORPHMET may be accessed via its webpage at http://www.morphometrics.org
--- 
You received this message because you are subscribed to the Google Groups 
"MORPHMET" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to morphmet+unsubscr...@morphometrics.org.

-- 
MORPHMET may be accessed via its webpage at http://www.morphometrics.org
--- 
You received this message because you are subscribed to the Google Groups 
"MORPHMET" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to morphmet+unsubscr...@morphometrics.org.

[MORPHMET] Procrustes ANOVA for FA Analysis

[Kate Rusk]

Hello everyone,

I am a bit confused about how to properly import and classify data for a 
Procrustes ANOVA in MorphoJ. Here is some more information about my data:

15 3D landmarks on human faces (object symmetry)
479 individuals
15 landmarks from 5 individuals digitized a second time
Sex and age as additional main effects

I realize I need to include a classifier that indicates a repeat (to be added 
as an Error classifier), but I am unsure how to import these data. So far, I 
have imported 1) the original data text file which includes the identifier and 
the XYZ coordinates for the original 15 landmarks and 2) a classifier text file 
which includes sex and age as covariates. I have separate XYZ coordinates for 
the repeat landmarks, but I don't know how these should be imported to 
implement the "repeat" classifier. 

Any assistance you could provide would be greatly appreciated! 

Thanks very much,
Kate

-- 
MORPHMET may be accessed via its webpage at http://www.morphometrics.org
--- 
You received this message because you are subscribed to the Google Groups 
"MORPHMET" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to morphmet+unsubscr...@morphometrics.org.

Re: [MORPHMET] Procrustes ANOVA in MorphoJ

[Chris Klingenberg]

Dear Alison

The residual (sums of squares, mean squares etc.) gives you the statistics for 
the second error effect.

I hope this helps.

Best wishes
Chris

--

*
Christian Peter Klingenberg
Faculty of Life Sciences
University of Manchester
Michael Smith Building
Oxford Road
Manchester M13 9PT
United Kingdom

Web site: http://www.flywings.org.uk
E-mail: c...@manchester.ac.uk
Phone: +44 161 2753899
Skype: chris_klingenberg



On 7 Oct 2016, at 19:36, Alison Foster 
> wrote:


Hi everyone,

I'm trying to assess measurement error in a 2D file of 16 individual specimens 
with no object symmetry. For each specimen I have taken two different images 
and then copied each image twice, resulting in four images of each specimen. I 
have prepared ID strings as recommended on the MorphoJ online manual (for 
example phm1_1; phm1_2; phm2_1; phm2_2. The first three letters being the 
individual, the first digit being the image (Error 1) and the second digit the 
replica (Error 2)). Even though I've selected 'Ind', Error 1 and Error 2 in the 
Procrustes Anova text box, when I run it and get the results I don't have 
anything for Error 2. In the Reports tab, it says "The effect indicated for 
Error 2 has no degrees of freedom. The effect was dropped from the analysis"

I don't know why Error 2 isn't being calculated and would be very grateful for 
any suggestions or comments. I've appended the output from the Results below.

Many thanks,

Alison


Dataset: Humeri

Classifiers used for the Procrustes ANOVA:
Individuals: ind
Error 1: image

Centroid size:
EffectSSMS   dfF  P (param.)
Individual   67921.930157 4528.128677  1522690.43  <.0001
Error 1  3.1929780.199561  16   3.95  0.0005
Residual 1.6185850.050581  32

Shape, Procrustes ANOVA:
EffectSS  MSdfF  P (param.)   Pillai tr.   P (param.)
Individual 0.053860100.897668  600 123.74  <.0001
Error 10.000464280.007254  640  13.93  <.0001
Residual   0.66640.000521 1280





--
MORPHMET may be accessed via its webpage at http://www.morphometrics.org
---
You received this message because you are subscribed to the Google Groups 
"MORPHMET" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to 
morphmet+unsubscr...@morphometrics.org.

-- 
MORPHMET may be accessed via its webpage at http://www.morphometrics.org
--- 
You received this message because you are subscribed to the Google Groups 
"MORPHMET" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to morphmet+unsubscr...@morphometrics.org.

[MORPHMET] Procrustes ANOVA output in MorphoJ

[Natalia Siomava]

Hello everybody,

I try to quantify components of shape asymmetry in fly wings. For this, I 
analize the individual and side effects with Procrustes ANOVA in MorphoJ. I 
have three fly species, and for one of them I got a nice output with all 
the parameters and also Residuals. For the other two species, F and 
P(param) are missing for the individual-by-side interaction. Does anybody 
know why it happens and whether it's still possible to get these values? If 
it's not possible, what would be the best way to estimate fluctuating 
asymmetry (FA)?

*Example of an output:*
Shape, Procrustes ANOVA:
Effect  SS MS dfF   
 P (param.) Pillai tr. P (param.)
Individual 0.135055350.460311 2934   4.37  
<.000111.66 <.0001
Side   0.000432060.240033   18 2.28  
0.0016  0.18 0.0288
Ind * Side0.030914240.105366 2934

Many thanks in advance and best wishes,
Natalia 

-- 
MORPHMET may be accessed via its webpage at http://www.morphometrics.org
--- 
You received this message because you are subscribed to the Google Groups 
"MORPHMET" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to morphmet+unsubscr...@morphometrics.org.