Message Boards Message Boards

0
|
5836 Views
|
7 Replies
|
1 Total Likes
View groups...
Share
Share this post:

Perform an ANOVA with 9 vectors

Hi every one, i working with 9 different vectors, and i need compare the means, i think that i have to perform an ANOVA, but i don't know how to do it in mathematica, can some one explain me? Thanks

POSTED BY: Oscar Rodriguez
7 Replies

From your description, I think you need ANCOVA to find the dependence of each answer upon gender and age. Example Link: ANCOVA

For instance, I just want to see the first sheet in your xls:

data = Import["scale.xls", "Sheets" -> "Respuestas de formulario - Tabl"][[1]];

Pick out the categories (from the 2nd colume to the 12th). Use the Part/[[]] function and Span/;; function to extract the title list

In[11]:= title = data[[1,2;;12]]
Out[11]= {A1,A2,A3,A4,A5,A6,A7,A8,A9,AGE,Genero}

Use A1 in the perception table as an example. A1, the second colume: (Use rest to remove the head A1 it self)

In[14]:= a1 = Rest[data[[All,2]]]
Out[14]:= {6.,6.,3.,6.,7.,1.,6.,....,1.,2.}

age:

In[25]:= age = Rest[data[[All,11]]]
Out[25]= {21.,20.,18.,1...,21.}

gender:

In[78]:= gender = Rest[data[[All,12]]]/.{"Femenino "->"F","Masculino"->"M"}
Out[78]= {F,F,F,F,F,F,F,F,F,...,F,M}

Linear model fit for the A1

$A1 = \alpha + \beta D + \gamma x$

where D is the dummy variable/gender and x is a continuous variable/age. You need a ANCOVA/Analysis of Covariance. We need to make the input data in the following form: {gender, age, outcome/A1}

In[80]:= myinput=Transpose[{gender,age,a1}]
Out[80]= {{F,21.,6.},{F,20.,6.},...,{F,19.,1.},{M,21.,2.}}

Use LinearModelFit funciton to find the dependeces:

lm = LinearModelFit[myinput, {DGender, XAge}, {DGender, XAge},NominalVariables -> DGender];

The linear model symbolic expression contains lots of information:

lm["Properties"] // TableForm

including:

lm["ANOVATable"]

anova of lm

Lets find the data for male customers from 12 to 42.

In[90]:= maleData = Cases[myinput,{"M",__}][[All,2;;]]
Out[90]= {{20.,4.},{23.,5.},{22.,4.},{16.,4.},...,{42.,5.},{21.,2.}}

Use Show, Plot and ListPlot to merge the estimation and raw data points together:

Show[
 ListPlot[maleData],
 Plot[lm["M", t], {t, 12, 42}]
 ]

lm plot for male

Attachments:
POSTED BY: Shenghui Yang

Can you share the data? No need to be exactly the same numbers but the the dimensions should be correct.

POSTED BY: Shenghui Yang

i'm working from an spreadsheet, you need that sheet?

POSTED BY: Oscar Rodriguez

Sure, I can take a look. Please make come comment in the data so that I know which points you want to compare with each other.

POSTED BY: Shenghui Yang

thank you very much for your help, well first of all i going to try to write the most clear that i can (i'm not an anglo speaker) i'm learning english and mathematica.

I have this data base (this are not the real data) with answers of an perception scale (from A1 to A9), and i have another variables like gender, age etc. I would like to know how to compare all this vectors, for example all the means from A1 to A9, and if there more influenced by age or gender, i going to add the excel file.

I appreciate your help

Attachments:
POSTED BY: Oscar Rodriguez

thank you very much for your help, well first of all i going to try to write the most clear that i can (i'm not an anglo speaker) i'm learning english and mathematica.

I have this data base (this are not the real data) with answers of an perception scale (from A1 to A9), and i have another variables like gender, age etc. I would like to know how to compare all this vectors, for example all the means from A1 to A9, and if there more influenced by age or gender, i going to add the excel file.

I appreciate your help

POSTED BY: Oscar Rodriguez

you definitely put me on the right way, i going to study your indications, and i'll try to do it by my self. I am very exciting with mathematica, when i see the videos i think that is awesome what you can do with some code lines. I really appreciate your help

POSTED BY: Oscar Rodriguez
Reply to this discussion
Community posts can be styled and formatted using the Markdown syntax.
Reply Preview
Attachments
Remove
or Discard

Group Abstract Group Abstract