# Sample size in proportional tests (prop.test/other suggestions that can include sample size)?

#### lucyd123

##### Member
Can I just say a big thank you to everyone on this forum for patiently contributing to my limited knowledge of statistics and holding up a light in the otherwise total darkness I experience with this subject.

This question relates to the weight of leaves on a sample of 90 trees measured prior cattle browsing, and the weight of leaves on a sample of 87 different trees taken post browsing.
Prior to browsing: 141.7 g of leaves <160 cm, 155.4 g of leaves >160 cm, (total leaves 297.1 g), sample size of 90
Post browsing: 34.9 g of leaves <160 cm, 169.6 g of leaves >160 cm, (total leaves 204.5 g), sample size of 87

So I followed the big R book on proportions, but the prop.test does not take into account sample size in the example in the book.
Does this matter? Can I include the sample size? Is this an appropriate approach?

For example:
I multiplied by 10 to not have decimal places. Then I simple entered
Prop.test(c(prior to browsing <160cm, post browsing <160 cm), (total leaves prior to browsing, total leaves post browsing)

So this my result:
> prop.test(c(1417,349), c(2971,2045))

2-sample test for equality of proportions with continuity correction
data: c(1417, 349) out of c(2971, 2045)
X-squared = 496.77, df = 1, p-value < 2.2e-16
alternative hypothesis: two.sided
95 percent confidence interval:
0.2816133 0.3309540
sample estimates:
prop 1 prop 2
0.4769438 0.1706601

Is this appropriate? Would I be able to refer columns instead of values? Any advice on best practice appreciated!

Best wishes,

Lucy

Lucy

Last edited:

#### lucyd123

##### Member
I've bumped this because I've changed the question!

#### GretaGarbo

##### Human
Prior to browsing: 141.7 g of leaves <160 cm, 155.4 g of leaves >160 cm, (total leaves 297.1 g), sample size of 90
Post browsing: 34.9 g of leaves <160 cm, 169.6 g of leaves >160 cm, (total leaves 204.5 g), sample size of 87
I did not understan this part. No! Wait! Is the 160 cm the hight above the ground?

That would just mean that you have two response variables: y1 = the weight of leaves below 160cm and y2 = the weight above 160 cm.

Then you can just do a t-test for each of the variables (or a WMW if you absolutely prefer that).

- - -

There is an important question about the trees (that would make the invetigation more objective). Did you randomize which trees to cut down before or after?

#### Dason

##### Ambassador to the humans
What makes you think prop.test doesn't take sample size into consideration?