# What's the point of a histogram of baseline measurments?

1. Oct 13, 2009

### p3t3r1

Hey folks, I am trying to write a paper and my suprevisor told me to make baseline measurments of the two groups and plot a histogram for each group and include that in my results. Problem is, I don't know that purpose of this histogram is and I don't know what to write. I mean yeah histogram allows me to see outliers but what's the point other than that? For the rest of my results, I could say something about the t test value, p value etc.. but right now I am only saying that a histogram of the baseline measurment is shown in figure 1 & 2 without saying anything else and I feel like I am missing something here.

First time writing a research paper, some help here would be appreciated, thanks!

2. Oct 13, 2009

### Bob S

Hi p3t3r1-
A baseline histogram plot is useful to show that there is no background contribution to the parameter you are measuring from parameters you are not aware of, and may affect your measurement. Also if the parameter you are looking for sits on top of large backgrounds, the reader is assured that you have properly included a background subtraction in your analysis. So a histogram difference (subtraction) represents the parameter you are measuring, and nothing else. This method is used to verify the existance of a new fundamental partilcle, like the Higgs boson, for example.
Bob S

3. Oct 13, 2009

### p3t3r1

Thanks for the quick reply. So let me get this right, if I include a histogram and it looks normal for the baseline measurments, that means I have no background contribution? I thought that was excluded by doing a student t-test between the treatment group and the control group? Thanks!

4. Oct 13, 2009

### Bob S

If you do have a background contribution (even if it has structure), and you adequately subtract the background from the gross signal, your true signal is probably clean. If you were a pharma, you might be doing a double-blind experiment with your new drug and a placebo, dose rates, dose duration (time), body mass, and for different age groups, education, physical condition (illnesses), sexes, pregnant women, and [STRIKE]races[/STRIKE] ethnicity, income? et cet.
Does your drug work better for some groups than others? Histograms will quickly show these variables.

Bob S

Last edited: Oct 13, 2009
5. Oct 14, 2009

### vanesch

Staff Emeritus
A histogram is an empirical measurement of a probability density function. You want to say things about (the difference of several) probability density functions, so it is always a good idea to have a look at them, no ?
Yes, the student-t test will give you an answer to the question whether two populations come from distributions with the same mean value or not. But it can be interesting to look at the distributions themselves, no ?