jon rogers wrote:
>
> Hi
>
> I have a statistics problem which I cannot figure out on my own.
>
> I have measured picking time (g/h) of four different varieties of
> strawberries (the four varieties differing in the size of their fruits). My
> initial hypothesis was that the picking speed would differ amongt the four
> varieties (due to the different sizes of the fruits).
>
> Having preliminarily inspected my data, it appeas that the main influential
> factor is not strawberry variety, but rather the total amount of
> strawberries of each variety ("total yield") - people thus tend to pick
> faster when there are many berries to pick. For each variety, all
> strawberries were picked.
>
> I have measured picking time (h), picking speed (g/h), and total yield (g,
> all the strawberries there were for each variety). Is there still a way for
> me to test my initial hypothesis? If so, how should I do it? How do
> separate / account for the influence of total yield on the picking speed?
Looks like multiple regression might be appropriate - though I'd want to look
at descriptive stats (means, s.d.s and some plots).
A lot depends on details of the study such as whether people only picked from
one variety or several.
In essence you want to see whether fruit size or yield accounts for more
variance in the picking speeds (g/h).
Thom
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
. http://jse.stat.ncsu.edu/ .
=================================================================
