Really need some help quick plz!!?

Question

I'm doing a biology project as part of my A2 course...

Im looking at the relationship between Leaf size and Light Intensity. I found that length and width of leaves are greater when light intensity is less (Leaves found in the shade).

So what kind of Statistical test should i use and why

1) Students T-Test

2) Mann Whitney U test

3) Spearman's Rank Correlation Coefficient

and what kind of graph should i do aswell? Scatter diagram or the Box and whisker diagram?

Plz help!! Thanks in advance!

Answer 1

It depends on how you treat your data.  In the best case scenario...

Since leaf size and light intensity can both be measured as continuous data, the best statistical test should be a correlation.  Since you don't have ordinal data unless you are ranking the light intensity instead of measuring it directly, a normal correlation resulting in an r^2 value is the most appropriate.

However, if you have classified light intensity simply as light or shade, then that data is considered nominal and, if your life size measurements are distributed normally, a standard t-test is most appropriate.

In other words, your leaf size data is continuous.  If your light intensity data is

1)  nominal (light or shade) = t-test

2)  ordinal (dark, shade, partial shade, etc.) then you need a rank test

3)  continuous (measurements in lumens, etc.) then you use a correlation.

The appropriate respecitve graphs would be:

1)  Bar graph (box and whisker?  I haven't heard this)

2)  Bar graph

3)  Scatter plot with regression for line fit, r^2, and P value.