Hopkins vs Lobos writeup
2011
The goal of these field surveys was to provide you with the opportunity to evaluate sources of variability in the collection of data. There are three types of variability in data collected to address hypotheses.
1)Variability that can be used to directly test the hypothesis. For example if the hypothesis is that there are more kelp rockfish at Hopkins than at Pt Lobos, we would compare the some measure of abundance (like a mean) at the two sites. A difference between the sites is one source in the variability in the collected data.
2)Variability associated with a known source but one that negatively affects our ability to test the posed hypothesis or hypotheses.
3)Unexplained variability. Again this negatively affects our ability to test our posed hypotheses.
In this set of surveys we tested the hypothesis that there were differences in species abundances between Hopkins and Point Lobos. Therefore we predict that there will be differences in means between the two sites. Formally we state this in the null form:
Ho: There is NO difference in abundance between Hopkins and Pt Lobos (this could be posed for all species).
In addition to assessing differences between sites, we also surveyed on two days at each site. We don’t think there should be a real difference given the short time interval between samples, but we want to account for any such difference.
WHY?
We also think that the variability between dates might differ among the three taxonomic groups surveyed (algae, fish and inverts).
WHY?
There is also the third type of variability in our data (unexplained). This is usually controllable, typically through good survey design including sufficient replication.
What counts as replication in our surveys?
Questions to address in the write-up:
We are interested in determining if the species composition differs at Hopkins vs Pt.Lobos and if variability associated with sampling day might affect our results. The results of the surveys are shown in the figures (if you want other figures let me know). The primary questions (here stated as hypotheses):
H1: There is a difference in species composition between Hopkins and Point Lobos
H1a: Difference in species composition between Hopkins and Pt Lobos varies by taxa
H2: There is a difference in species composition between days
H2a: Difference in species composition between days varies by taxa
H3: Both Site and sampling day affect species composition (interaction effect)
H3a: the interaction between site and sampling day varies by taxa
Note that species composition is a multivariate term and that once differences have been established it is important to look to the species to see how they contribute to the differences.
When you consider these hypotheses it is important to also think about the adequacy of the sampling design – specifically:
1)Are the number of days a site is sampled important?
- Is this answer the same for all taxa?
2)What is the optimal number of transects that should be sampled in a given site to maximize statistical power?
- Is the answer the same for all taxa?
- Can you determine it for all taxa?
3)Are there differences among species with respect to the ability to sample them (adequately)?
- If so, what attributes of the species contribute to these differences?