Multiple Egg Sacs

In breeding the females, I have noticed that they put out multiple egg sacs without having remated. They seem to be able to store sperm for long periods of time after mating with a single male. I am interested to see what sort of role this multiple egg sac laying will play in my manipulation. Will one of the male's sperm be used preferentially over another's? After that male's sperm is used up, will she then use the other's? Is the female able to influence the sperm used to fertilize eggs or does she use whatever is in her storage sac? I should be able to tell when I am able to see whether the percent of eggs hatched changes with each egg sac for the females mated with a sterile male and a fertile male. Maybe the mate order will play a role. This could be pretty interesting! I think I would use MANOVA here?

ANOVA and spiders

When I was going over Chapter 7 in G&E, I had a hard time getting through it because I kept trying to apply everything to my spider study. I even drew in my own little chart next to the one in the book showing a two-way layout in ANOVA. My two factors in this case will be mate order (whether the male is the 1st or 2nd to be mated with the female) and % of intromissions (20,50,80). These are my two categorical predictor variables. My response variable is the percent of eggs hatched. I hope to see a significant response from the % of intromissions, which supports the hypothesis that sperm number determines paternal advantage. I do not expect to see a main effect of mate order. If there was, this would support the hypothesis that the priority pattern is determined by sperm placement, in that the first male is able to place his sperm closer to the fertilization duct and therefore fathers more offspring. I think this will work. If you see any flaws, please feel free to point them out. Also, if I follow the Rule of 10, I will need 10 trial runs or replicates. There are 16 males per replicate (two pairs for sterile, sterile; two pairs for fertile, fertile; two pairs for fertile, sterile; and two pairs for sterile;fertile). The number of intromissions will be varied among the pairs with one being 50/50 and the other 20/80. Total this means I need 160 males!!!

Regression

I really liked working through some sample data sets with R on Tuesday. It helped me learned how varying some of the terms could change the coefficent of determination and the p-value. Here is the summary of what I learned:
Regression is used for two continuous variables (both independent and dependent). A regression assumes that there is a cause and effect relationship between x and y. Also, this technique asumes that the x variable is measured without error (can this be done?!?) As far as a linear relationship between two variables, the null hypothesis proposes that there is none and that the slope is 0. The alternative hypothesis depends on slope. Also, the slope, the coefficient of variation, and the sample size all affect the p-value. Remember the p-value is the probability that the slope of the line equals zero and that there is no relationship between variables (null hypothesis).

ANOVA

Heidi's presentation on Tuesday was very helpful. I especially liked how she covered the assumptions of ANOVA. I also thought that having Thursday's lab also be about ANOVA was well-planned. Though I think that I am getting a good grasp of this test, I think that it is pretty important for us ecology people to truly understand this well. I think that actually doing an ANOVA by hand with a real data set would help to make it more concrete in our minds. Back in ecology lab, this old-fashioned "work out the problem" method helped me to understand t-tests and chi-square tests. I know that it is old school but sometimes using computers does not equal understanding. Anyone of the same opinion?

Spider Update

Sorry it has been so long since my last posting, but I am going to try to make up for it. The good news is that I am finally ready to do a run for my project. I have 16 mature males and 9 mature females ready to go. I am mating them tomorrow! I also will be meeting with my advisor. So I plan to talk about statistical design. As of now, it looks like I will be using ANOVA, ANCOVA, and some t-tests. I am sure this may change as I go along and get a better idea of what I am doing. Also, the males that I am planning to sterilize will have to be placed in a straw using cotton as the stoppers. I took from this class that I should also do this straw cotton treatment for the fertile males as well minus the irradiation of course. Something I am also planning to avoid is confounding my study with time/seasonal issues. Since I can't run all of my runs at once, I plan to have all of my treatment and control groups represented in each run!

Tuesday's Class

I really enjoyed class this Tuesday. Rebecca did a really great job of putting it all together. It was very useful to define a study, hypothesis, and design according to whether the dependent and independent variables were continuous or categorical. The exercise actually made us think about relevant biological variables and how we should test them. I learned a lot about when to use different tests. Ultimately, this will be very helpful for my independent project/honors thesis. Now if we could just do this with a Thursday lab, I feel like I would get so much out of R. By the way, based on the biostats test, I found R to be useful in running analyses but inputting data was a nightmare. Anyone have any good tips or is it just hard to do? It is just not as friendly as a spreadsheet like Excel.

JV Ch. 6

I found this chapter to be a little more enigmatic than previous ones. There seemed to be a lot of emphasis on trying to do exercises that supported the Central Limit Theorem. This exercise was less practical and a little too redundantly didactic. What can we actually use from this chapter? I guess the most useful thing to me was learning to do the loop. I figure that this will come in handy in the future. I guess I would like to see this course take a step towards practically applying the information in JV to a biological study. Is there anyway that we could use a different biological data set and manipulate it with what we learned in each JV chapter? I think that I would learn a lot more from this and be more motivated in doing the exercises. After all, the purpose of this course is to learn how to use statistics in our projects and studies, and as of now I don't think I am anywhere close. Let's put JV in a different context!! I don't want to learn about R for the sake of doing just that. I want to use R as a tool in my own studies, and the examples in the book have so far been too theoretical.