Competition - post a photo you've taken of nature >>

What does NULL HYPOTHESIS mean?! :/

A student observed that the yield of oranges increased as the abundance of honeybee hives in the orange hives increased. She counted the number of hives in different groves and then measured the yield in each grove. She decided to determine whether there was any association between the two variables.

Grove no |No. of honeybee hives/ grove |Yield of Oranges in tonnes/ grove
1 | 122 | 107
2 | 66 | 32
3 | 79 | 61
4 | 94 | 111
5 | 43 | 48
6 | 64 | 72
7 | 30 | 34
8 | 91 | 76
9 | 110 | 119
10 | 76 | 90
11 | 61 | 90
12 | 104 | 59
13 | 67 | 64
14 | 30 | 46
15 | 92 | 90

So I used the Correlation coefficient to get the calculated value which is 0.68125. and the critical value is 0.5385.

Now the question says:
State a suitable null hypothesis for this investigation.

Reply 1

satanslilsis

The null hypothesis is a statement where you'd say there will be no relationship between the two variables and any changes would be due to chance.
So, There will be no affect of the number of honey bees on the yield of oranges, any change in this yield will be due to chance :smile:

You'd only accept this hypothesis if the result was insignificant xx

Reply 2

cheshiremum

The null hypothesis is what you are trying to disprove, so as the poster above says, in your case, no relationship.

Reply 3

the bear

H_o : there is no correlation between hive numbers & orange yield

Reply 4

Adorable98

Original post by satanslilsis

You'd only accept this hypothesis if the result was insignificant xx

Original post by cheshiremum

The null hypothesis is what you are trying to disprove, so as the poster above says, in your case, no relationship.

Original post by the bear

H_o : there is no correlation between hive numbers & orange yield

Oh, I see! Cheers!! :h:

Reply 5

Adorable98

So one more thing, if the null hypothesis means that there will be no relationship between the two variables, how can it also mean that there's no significant difference between two variables :s-smilie:

Spoiler

BUMP..

Original post by Adorable98

BUMP..

General forum etiquette is not to bump unless it's past one day.

As for your question, generally if there is no relationship, there will be no significant difference between variables.

Here's an example:
There's a jar filled with black and white marbles in an equal proportion. You are calculating the probability of picking out a white marble when using a right hand versus the probability of picking out a marble with a left hand.

Since there already is a 50% probability of picking a white marble, there should be no difference between using a right hand or left hand with a high sample volume.

[EDIT] If there was a relationship between the two, there should be a significant difference.

An example of this would be a colour change reaction between 2 compounds. If compound 2 is the limiting factor, there should be a degree of colour change dependent upon the concentration of compound 2. Thus, a significant difference. However, if Compound 2 does not react with Compound 1, there is no reaction, thus a value of 0 for colour change.

(edited 8 years ago)

Reply 8

Eloades11

Original post by Adorable98

Spoiler

Those two statements are different. Usually if you're testing a hypothesis you are either looking for a relationship between two variables or you can be looking for a significant difference between two variables. Depending on which one your main hypothesis is, the other one respectively will be the null hypothesis.

No relationship between two variables is not the same as there being no significant difference between two variables. You will have to choose your null hypothesis carefully, making sure it counters your main hypothesis.