HELP on Statistics

#1
Hi All,
statistics is really not for me, so please help

I have 2 population of dogs one from the shelter n100 and one from a boarding kennel n100

I have 4 types of acoustic enrichment
no music
music 1
music 2
music 3

I am using a quality of life score of each dog after the animal have been listening to the song for a period of time

I am trying to calculate

1, is acoustic enrichment improving quality of life of the animal in the kennel
2) which type of music improve mostly
3) is there a difference in the improvement between the two population

What type of test should I do?

Thanks
 
Last edited:

Karabiner

TS Contributor
#2
It depends on whether all dogs have been in all 4 environments
or whethe each dog was allocated to only 1 condition. I suppose
the latter is true. If your quality of life score is interval scaled,
then a oneway analysis of variance (ANOVA) with post-hoc tests
would be an option for questions 1) and 2), and a two-factor
ANOVA for question 3) (where the interaction between dog type
and environment would be of interest).

With kind regards

K.
 
#3
HI K

all dogs have been in all 4 environments and each dog was allocated to only 1 condition.

Whats the best software to calculate the data attached?

Thanks
L.
 

Karabiner

TS Contributor
#4
If I understand you correctely, then you have a repeated-measures
ANOVA for question 1) and 2), and a "mixed" ANOVA wth a repeated
measures factor AND a grouping factor for question 3).

With kind regards

K.
 
#5
Thanks Karabiner, whats the best way to calculate the file attached? Sorry but I have been banging my head for few days as I really cant work this out with any of the software I have

Thank you so much!
 

Karabiner

TS Contributor
#7
That would cost you 45 Euros just for me having a look at it, and
32 Euros for each half an hour I'd spend on your data. So I guess
you might perhaps prefer to look for appropriate software.

Alternatively to repeated measures ANOVA, you could work on
questions 1) and 2) by using dependent samples t-tests for
6 pairwise comparisons between conditions (there are online
calculators for this, and I suppose Excel/open office can do
this, too). alpha level should account for multiple comparisons
then, e.g. by using a Bonferroni-corrected threshold 0,05/6.

But question 3 clearly concerns an interaction. If both both groups
did not differ with regard to the "no music" condition, you could
calculate difference scores (such as music 1 - no music) and compare
these between groups using independent samples t-tests.

With kind regards

K.