 |
|
Mixing control areas with experimental areas
Mixing several control areas with several
experimental areas solves this problem because we will be able to calculate the
average number for both species of mussel that exist in the experimental areas
compared to the control areas.
If variations between the areas are caused only by natural
variations that always occur along beaches, then there is no reason for
the averages to be different. When the experimental and control areas are chosen randomly,
there should not be a larger supply of food only to the experimental area from natural causes.
Since we have a mixture of experimental and control areas, every set
of areas (experimental and control), should on average be equal because
all of them have been exposed to the same natural variation.
If on the contrary it is as prescribed in the hypothesis,
we expect on average an increase in the number of mussels within the
experimental areas after competitors have been removed, independent of natural
variations between the different places. To be able to make a proper comparison
of these averages we must have several control and experimental areas.
Statistical analysis
This leads to additional complications. We now have the averages and variations between the different experimental and control
areas, which we have to manage. Therefore, we must statistically analyse our
findings. When we decide if competition has had any effect, it is not enough
to guess if the averages are different enough in relation to the variation. we need to analyze our findings with a statistical test.
We do not describe here how a statistical analysis is conducted, but to
be able to conduct the statistical test and to base a support of the hypothesis on falsification,
we must change the hypothesis to what is known as the null hypothesis.
 |
Page
9 of 13 |
 |
|
