# Thread: It's posible to compare two means without considering standard deviations?

1. ## It's posible to compare two means without considering standard deviations?

My doubt is the following

'it's possible to compare means of two independent groups without considering standard deviations" ?
For example:

Control Group: mean +- standard deviation 50 +- 10 mm
Experimental Group: mean +- standard deviation 100 +- 30 mm

I am very surprised with the following conclusion made by a dentist friend mine

"Then we say there was an increase of mean 100% between groups !!"

(100 - 50) / 50 = 1 = 100%

However this conclusion seems odd to me, as it disregardes the variability of the data.
TIA
Ivan

2. ## Re: It's posible to compare two means without considering standard deviations?

Is treatment randomized?

Were groups the same size?

What does the continuous variable represent?

This seems wonky:

"(100 - 50) / 50 = 1 = 100%"

100 / 50 = 2 times or 200%

3. ## Re: It's posible to compare two means without considering standard deviations?

Yes. The treatment can be randomized. The same site sample, for example. The variable can tô be mm, for example.

Originally Posted by hlsmith
Is treatment randomized

Were groups the same?

What does the continuous variable represent?

This seems wonky:

"(100 - 50) / 50 = 1 = 100%"

100 / 50 = 2 times or 200%

4. ## Re: It's posible to compare two means without considering standard deviations?

The groups are differents
For example, ten subjects for Control and others ten subjects for Experimental.
.....................................
Originally Posted by balducci
Yes. The treatment can be randomized. The same site sample, for example. The variable can tô be mm, for example.

 Tweet