Skip to content

ANOVA from imported data has only 1 degree of freedom

7 messages · David Winsemius, shardman, Ben Bolker

shardman
# 
Hi,

I'm trying to analyse some data I have imported into R from a .csv file but
when I carry out the aov command the results show only one degree of freedom
when there should be 14. Does anyone know why? I'd really appreciate some
help, the data is pasted below.


/The imported table looks ike this this:/

        Order Transect Sample Abundance
1  Coleoptera        1      1        13
2  Coleoptera        1      2        12
3  Coleoptera        1      3        11
4  Coleoptera        1      4        13
5  Coleoptera        1      5         6
6  Coleoptera        2      1        18
7  Coleoptera        2      2        18
8  Coleoptera        2      3        16
9  Coleoptera        2      4        21
10 Coleoptera        2      5        11
11 Coleoptera        3      1        19
12 Coleoptera        3      2        16
13 Coleoptera        3      3         9
14 Coleoptera        3      4        32
15 Coleoptera        3      5        29

/The command I am using is this:/

anova2<-aov(Abundance~Transect,data=beetle)

/The results come out like this:/

Df Sum Sq Mean Sq F value  Pr(>F)  
Transect     1 250.00 250.000  7.2394 0.01852 *
Residuals   13 448.93  34.533                  
---
Signif. codes:  0 ?***? 0.001 ?**? 0.01 ?*? 0.05 ?.? 0.1 ? ? 1 

Cheers,
Sam


--
View this message in context: http://r.789695.n4.nabble.com/ANOVA-from-imported-data-has-only-1-degree-of-freedom-tp3887528p3887528.html
Sent from the R help mailing list archive at Nabble.com.
David Winsemius
#
On Oct 9, 2011, at 11:10 AM, shardman wrote:

            

      
      
      
    

        
To me this appears likely to be homework. If so then please read your  
college's rules regarding  academic honesty. In any case, please read  
the Posting Guide.

            

      
      
      
    

        

      
      
    

        
David Winsemius, MD
West Hartford, CT

  
  


  


      

    

    

      
      

        


  

        

      shardman
    

    

      
      #
    

  


  

      
Hi David,

Thanks for your message.

I can assure you this is not homework. I'm working on an ecology project and
am trying to analyse the results from the fieldwork. I don't want other
people to do the work for me I was just hoping someone might be able to spot
where I have made a mistake, I'm still teaching myself to use R after having
used SPSS for a long time.

I did read the posting guidelines but couldn't see any reason not to ask for
help, I'm sorry if I've got it wrong,
Yours,
Sam

--
View this message in context: http://r.789695.n4.nabble.com/ANOVA-from-imported-data-has-only-1-degree-of-freedom-tp3887528p3887979.html
Sent from the R help mailing list archive at Nabble.com.

  
  


  


      

    

    

      
      

        


  

        

      David Winsemius
    

    

      
      #
    

  


  

      
On Oct 9, 2011, at 2:08 PM, shardman wrote:

        

            

      
      
      
    

        
Please read the Posting Guide again. Nabble is not the standard venue  
where people read this list and you seem to have missed the part where  
the PG asked you to include relevant context in replies. It also also  
asks you to describe your scientific domain (which you have now done)  
and academic position (at least in part so the list can remain non- 
homework oriented).

My memory from your initial posting was that you had numeric  
identifiers for you groups and did not wrap the group variable name in  
factor(), so it was treated as a continuous variable. Doesn't SPSS  
have some mechanism to flag a variable as discrete?

Perhaps(from memory) something along these lines:
aov(outcome ~ factor(group), data=dat)

(Further comments from memory of earlier posting.)
You also were asking why there were not 14 df in the output but there  
were since 1+13 = 14. Surely you were not expecting 14 df to be  
associated with the grouping variable when there were only 3 groups?

            

      
      
      
    

        

      
      
    

        
David Winsemius, MD
West Hartford, CT

  
  


  


      

    

    

      
      

        


  

        

      shardman
    

    

      
      #
    

  


  

      
Hi David,

Apologies again and thankyou for your help, I've edited my original post to
clarify what I was asking. What I meant was that the factor had only 1
degrees of freedom when it should have had 2 (14 in total), so you're right
there were 14 but not in the right place.

In SPSS you select one column as a factor and another as a dependent
variable so this wouldn't happen, it's easy to use but not that versatile
and very expensive. I've been told good things about R so I'm trying to
teach myself.

I followed your suggestion and I now have the results I need,
I'll take more care with posting in future,

Yours,
Sam

--
View this message in context: http://r.789695.n4.nabble.com/ANOVA-from-imported-data-has-only-1-degree-of-freedom-tp3887528p3888322.html
Sent from the R help mailing list archive at Nabble.com.

  
  


  


      

    

    

      
      

        


  

        

      Ben Bolker
    

    

      
      #
    

  


  

      
shardman <samuelhardman <at> hotmail.com> writes:

            

      
      
      
    

        
[snip]

            

      
      
      
    

        
I was going to tell you to read up on the distinction between
factors and numeric variables in statistical models, but I can't
immediately find a reference on line (this is bound to be
in Peter Dalgaard's book or any other basic-stats-with-R
text).

str() will tell you whether the columns are numeric or factors.

Try

beetle <- transform(beetle,Transect=factor(Transect))

[or ?read.csv and use colClasses explicitly to specify
the classes of the columns]

before running your anova.