[R-sig-eco] Manova and DFA analysis

[Sarah Goslee]

Hi,

Those aren't really R questions, nor are they questions we can answer
without a great deal more information. Your very best course would be
to set up an appointment with a local statistician who can help you.

If that isn't an option for you, a statistics forum would be a better
place to ask.

Sarah

On Wed, Sep 3, 2014 at 7:10 AM, Mohammed Almalki <m11m116 at hotmail.com> wrote:
>  Dear all,
>
>
> I am new user for R program and I am
> looking for somebody to help me with Manova and discriminant function analysis DFA .
>
> I have four measurement traits for bird
> species (weight, wing length, tarsus length and bill length) and I would like
> to test for differences in body size between males and females of this species.
>
>
>
> FIRST, I applied MANOVA using (weight,
> wing length, tarsus length and bill length) as dependent variables and sex as
> an independent variable. In order to identify the significance of sex
> differences for each dependent variable using this form:
>
> rm
>
> data1<-read.csv("C:/Users/Desktop/CP/CP_NOMISS.csv")
>
> names (data1)
>
> attached(data1)
>
> head(data1)
>
> manova1 <- manova (cbind (Weight,
> Wing.Length, Tarsus.Length, Bill.Length)~ as.factor (Sex), data=data1)
>
> summary (manova1)
>
> summary.aov(manova1)
>
> After that I got four tables (one table
> for each variable)
>
> My questions are:
>
> 1. Is what I did correct and enough to
> get Manova results?
>
> 2. what is the most important result can
> describe the difference is it F value or Pr(>F) value
>
> 2. How I can describe the results in
> figure?
>
>
>
> SECOND, I applied discriminant function
> analysis (DFA) on the four morphological characters using the package MASS in
> order to identify the variable that differed most between males and females using
> this form:
>
> rm
>
> library (MASS)
>
> data1<-read.csv("C:/Users/Desktop/CP/CP_NOMISS.csv")
>
> head(data1)
>
> attach (data1)
>
> data1
>
> plot(data1[ ,c(2,3,4,5)], col=data1[ ,1])
>
> data1.lda <- lda(SEX~WG + WL + TL +
> BL, data=data1)
>
> data1.lda
>
> After that I got this result:
>
> Coefficients of linear discriminants:
>
>             LD1
>
> WG -0.001040297
>
> WL -0.011554912
>
> TL
> 0.030233583
>
> BL
> 0.498226667
>
> 1.Is what I did enough to say the
> variable that differed most between males and females is BL  0.498. And does this difference is reliable?
>
> OR there are other steps I have to do.
>
>
> Please excuse the long email.
>
> Thank you very much in advance for any
> help you can provide.
> Best regards,Mohammed


-- 
Sarah Goslee
http://www.functionaldiversity.org