showing results?

Question

Im doing a science fair project on different concentrations of a growth hormone on plants.
I have 15 plants in one group and 15 in another.
I measure the height and number of leaves.
Now the problem is im not certain how to compare the two groups to another. Can you average the height and leaves of one group and compare it to the other? but the problem with that is that all plants genetically differentiate in the height and number of leaves,right? so what can i do?

Answer 1

Ah, statistics!

Yes, each plant will have different genetics, so if you were doing this as an important study, you'd want to reduce the chances of that being a variable. You could do that to a degree by choosing seeds/young plants with the same characteristics and from the same parent plants, so there would be less genetic variation. At the very least, you'd want to use plants of the same species that were all about the same size and with the same number of leaves when you started.

The best test to use to compare these are a two-sample t-test: http://www.socialresearchmethods.net/kb/stat_t.htm Give me a few minutes and I'll write you a more detailed answer with a table you can use to analyze your results.

To calculate the t value, you'll need to find the average and variance for each sample (the treated group and the control are the two samples) for both the height and number of leaves. To calculate the varance follow these steps:

1) Make a list of each value for a characteristic (height, for example) and add these together. Your answer______

2) Square the result of #1. your answer_______

3) Next to the first column, square each value (multiply it by itself), then add these together. Your answer_________

4) Divide the result of #2 by the number of samples (in your case, 15) Your answer_________

5) Subtract your results #3 - #4 = ____________ This is your variance.

Do this with the other 3 samples (you'll have results for height: control, height: treatment, # leaves: control, # leaves: treatment.

To do the t-test for a characteristic (do height to start), follow these steps:

5) Add the variences for the control and the treatment group together. Your result____________

6) Add the number in the control and the treatment group together (for you, this would be 30)

7) Divide #5 by #6. Your result________________

8) Calculate the following: (result from #7/number in control) + (result from #7/number in sample) your result___________

9) Take the square root of #8. Your result_______________

10) Figure out the average height of your control and average height of the treated group. ______ & _______

11) Subtract the averages so you get a positive number. Your result____________________

12) Divide. #11/ #9 ______________ This is your t value!

Now to interpret your result on this table:
t table: http://www.statsoft.com/textbook/sttable.html#t

13) Find your d/f (degrees of freedom) on the left column - this is your total sample size minus 1 (29 for you).

14) Across the top, find the 0.5 column.

15) Find the number where this row and column intersect (in your case, 1.699127).

16) If what you got for #12 is lower than 1.699127, your results would be called "significant" - the 0.05 means there was only a 5% probability your results were "by chance".

If you want to do this again using the 0.01 column and your #12 is lower, you have "highly significant" results - only a 1% probability your resultes were by chance.

Now try this again for # of leaves.

Even though it takes a few steps, the math isn't that hard. It also gives you professional-looking results. This type of statistics is generally something you won't encounter until college, where they use formulas like those shown in the link above.

Good luck at the science fair!

Answer 2

There could be several ways to do it....

simply graphing it.....I would probably average the data ( but make sure you comment on any "odd" ones eg that was extrememly tall/short had lots of leaves etc.

You could do a bar/histogram graph....on say the left y axis have height and the right y axis could be number of leaves.
Use different colours for height and number of leaves and have group 1 info together ( so you have 2 bars touching, one with height and 1 with number of laves) and leave a gap and have group 2 ( writing group 1 and 2 on the x axis below)
This way you would compare all the info in one go.

In answer to the question that plants differ genetically, then technically yes. You should comment on this in your discussion, especially if you have " odd" or outlyer results.
I'm not sure which hormone you used....I forget which one does what these days...maybe auxin..whatever, but learn more about how the hormone works too so you can comment about your results in regard to not only the plants ( I assume they are all the same type) and the hormone.