Questions, Hypotheses, and Predictions

• Reading assignment for Monday, Whitlock & Schluter, Chapter 6 (QUIZ = YES)

Definition: A hypothesisis a clear statement articulating a plausible candidate explanation for observations.

Q 2.1: Suggest some hypotheses that could explain the observation that people drive faster on the journey to work than on the way home.

A 2.1:

  (1) Time management - have a lot of work to do.

  (2) Energy difference - newly caffeinated vs. exhausted.

Question: Why does chimp activity vary during the day?

Hypothesis: Chimp activity pattern is affected by feeding regime.

Prediction: The fraction of time that a chimp spends moving around will be higher in the hour around feeding time than at other times of day.

Alternate hypothesis (statistical): \( p_{f} > p_{t} \), where \( p_{f} \) is the fraction of time that a chimp spends moving around in the hour around feeding time, with \( p_{t} \) the same metric for all other hours of the day.

Null hypothesis (statistical): \( p_{f} \leq p_{t} \).

Question: Why do whelks group?

Hypothesis #1: Whelks group for shelter from wave action.

Prediction #1: Whelks are more likely to be found in groups in areas sheltered from wave action.

Hypothesis #2: Whelks group for feeding.

Prediction #2: Whelks are more likely to be found in groups in areas of higher food density.

Note: Multiple hypotheses can explain the same prediction.

Hypothesis #1b: Whelks are more vulnerable to predators in sheltered areas, but grouping provides protection from predators.

“No matter how the study is organized, the important thing is that the best study will be the one that allows us to tease apart the influence of the different hypothesized unfluences on grouping behavior.”

- Ruxton & Colegrave

Experimental design example: Factorial experiment with predation and wave action.

• With wave action and without predation.
• Without wave action and with predation.
• With wave action and predation (interaction effect).
• Without wave action or predation (control).

“Perhaps the key to having really novel ideas is just to keep your eyes and ears open and try and question the things you see around you.”

- Ruxton & Colegrave

“We encourage you to design experiments that are interesting because of the question they ask more than because of the specific answer to the question that emerges from the data.”

- Ruxton & Colegrave

Three To-Dos:

• Ask yourself: Is it possible to explore this question in a way to obtain a valuable answer? (i.e. do you have the resources? Careful, though: What is valuable?)
• Ask yourself: Are you committed to obtaining an answer? (note: this is different than excited, which the book mentions - excitement comes and goes, but commitment sticks)
• Bounce ideas off others you respect, but don't rely solely on their responses!!!

“You should think of the Devil's advocate as a highly intelligent but sceptical person. If there is a weakness in your argument, then they will find it.”

- Ruxton & Colegrave

Reason #1: One aim of a pilot study is to allow you to become well acquainted with the system that you will be working on.

• Make sure observations are actually there!!
• Don't just trust the literature - observe for yourself

Reason #2: A second crucial role for a pilot study is that it gives you a chance to practice and validate the techniques that you will use in the full study.

• Practice protocols (minimize wasted time, money, and suffering)
  • e.g. Butterflies and plant choice for egg laying
• Try out different protocols
• Find biologically realistic values for variables
• Pilot data used for statistical power calculations

Causation vs correlation

https://xkcd.com/552/

“The only way to be certain of removing problems with third variables is to carry out experimental manipulations.”

- Ruxton & Colegrave

Difficulties:

• Sometimes not possible to manipulate, for practical or ethical reasons.
• Not technically feasible to carry out manipulative study.

Note: Correlational studies can be used as a first step towards a manipulative study. Also, if you are interested in natural variation, observational is a great way to go.