Null Hypothesis

[pan]

"You are testing to see if a new brand of fertilizer works better than the old brand. You have 20 plants and have divided them into two groups. The first will be grown with the old brand and the second will be grown with the new brand of fertilizer. You keep the water, temperature, light, etc. the same for both groups of plants and they are grown together in the same greenhouse. You measure the height in centimeters that the plants grow after a two week period."

 

What is your null hypothesis?

 

Would "The plants grown with the new brand of fertilizer would grow at the same or lower rate as the plants grown with the old brand." work?

[Axioms]

It will do, in my opinion anyways.

 

 

[pan]

Great, thanks!

 

How about:

"You have completed an experiment where you collected data on the growth of the fungus Aspergillus grown in media containing different concentrations of a particular vitamin. The concentration of vitamins was measured in mg/l and the fungal growth in milligrams."

 

 

Would "The fungus Aspergillus will grow at the same rate for all concentrations of vitamins." work?

Edited February 14, 2012 by pan

[tomgwyther]

Don't we need a control group, grown with no fertilizer or vitamins?

[pan]

According to my assignment, no. Is that odd?

[tomgwyther]

if it were my experiment, i would want to establish a sort of base line. That is: the rate at which plants grow unaided. Lets say that Fertilizer 'A' is less effective than fertilizer 'B'. It could be that 'A' performs no better than the control.

[CharonY]

From how the question is phrased it is of no relevance. Essentially you test whether A and B are from the same population. You have to be careful with the null, however. If your null is that B performs the same or is worse, it is a different test than A and B perform the same (which I think is the simpler and hence more likely intended question).

 

For the second question you are essentially a correlation analysis and you have to create the null differently. Think about the alternative hypothesis a bit more.