Back to the Syllabus


The Bean Experiment

Introduction to Hypothesis testing

Barry Sinervo


Index

Procedure

Null versus Alternative Hypotheses

Analyzing the Data

The Lab Report


Procedure

Read Through the Following Procedure.

However, before going through with it Make up a Null Hypothesis and Alternative Hypotheses (see below)

Form groups of 4-5 and perform a selection experiment of sorts. First you might ask what the heck do beans have to do with behavior? Sinervo is loosing it. We shall soon see.

Count out a pile of 200 black beans.

Then designate one member of the group as the bean picker. The other members of the group can weigh out the piles of beans as they are collected, and then record and graph the data for the group.

The bean picker will pick approximately 5-10 between their thumb, index, and forefinger and gently roll around the beans and let some beans spill out until only one bean remains. Take this bean and set it aside in "lot 1". Repeat this procedure with the remaining pile of beans 9 more times so that you end up with 10 beans in lot 1.

Set aside lot 1 and have some of the members weigh out lot 1, while the bean pickers continues making lots 2, 3, ..., 20 by the same procedure. Each lot has 10 beans in it.

Index


Null versus Alternative Hypotheses

Let us first set up a null hypothesis. What pattern of bean size would you predict if you were randomly drawing beans from the pile as a function of size.

Recall the manner in which you collected data on beans.

Let us now set up an alternative hypothesis. What kinds of patterns might you predict if the drawing was in some way non-random. Describe mechanisms that might explain these patterns. You might in fact have one or more alternative hypotheses to test. You should try to describe at least two hypotheses, and two mechanisms. Discuss the null and alternative hypothesis amongst the members of the group.

Index


Analyzing the data

Graph the data that your group obtained on bean size as a function of lot# (e.g., 1 is the first lot, 2 is the second lot, etc.). If some member of the group knows how to use a simple graphing program it might be useful to plot the data on the computer (and get copies of the graphs to other members of the group). If some member of the group knows how to run a simple linear regression, even better! Run a regression line through the data.

What is the pattern of size change that is found in your samples of beans from the first draw to the last draw?

Is this pattern statistically significant? Do we reject or accept our null hypothesis?

How can you explain this pattern in light of your alternative hypotheses?

Everyone should obtain an regression (linear trendline) and an R2 value to report with your data. Talk to your TAs if you don't know how to do this. If you email your R2 value and sample size to your TA, they can provide you with a p-value, so that you can assess the significance of your data.


Index


The Lab Report

The Big Picture Question. What sort of implications does this simple experiment have for the origins of Agriculture and cultivated plants? Refer to the paper by C. Heiser (available on reserve in the library) in your discussion of the been experiment. Now make some brief observations on the behavior of a living creature during its feeding behavior. Make a simple hypothesis concerning either:

  1. how variation in the creatures movements might get it into trouble with some kind of predator (natural selection on it) or
  2. how the creatures own foraging might select for subtle differences in the food that it is feeding on (natural seletion on the prey).

Include your hypotheses concerning the creatures behaviors in the discussion. Make a simple statement about the "unconcious aspects of natural selection" in your lab report discussion.

Lab report (1 page with 1 page figure) is in section next week.

Include the answers to all of the above Big Picture questions in the form of a standard scientific report (see the Bean Lab Report Grading web page for more details, as what is listed below is only a brief summary):

  1. Introduction where you discuss your hypotheses. These hypotheses should be discussed near the end of the introduction. You should begin your intro with the big concepts and work down to the small concepts (e.g., your specific hypotheses). The big concepts would be: how does this simple-minded experiment relate to the real world.
  2. Methods where you discuss in very succinct terms what you did. Only include enough detail such that a protege following in your footsteps could repeat this monumental experiment.
  3. Results should describe the patterns. Words and graphics are integrated. You must report statistical tests in the Results. Describe any statistically-signficant patterns.
  4. Discussion includes a consideration of your hypotheses and proposed mechanisms. Also, discuss the big picture question as the last paragraph in the Discussion.

Be succinct and to the point when writing your lab report!

Index


Back to the Syllabus