How should population sample characteristics be presented for a small sample size? I'm performing a clinical trial and have enrolled 16 patients, of which 3 had to be excluded. I suppose it would be relevant to present some characteristics of the included patients, e.g. age, weight, gender distributions.
Presenting gender is easy, but for age and weight: Does it make sense to present mean and standard deviation when n=13? Or should I present median and quartiles? Or mean and range?
 A: For presentation purposes, in most cases, it's preferable to show frequencies inside specific brackets rather than means, medians etc. That's true even for small samples.
Now, these brackets depend more on your study rather that the quantiles. Take "age" for instance. In my field of work I usually use the age brackets: 18-24, 25-29 etc because these are the different segments of the customer database we have. Even if 18-24 customers are much fewer than 25-29 customers. In your case you may choose groupings such as children < 9 months old, 9-12 months if your study is a pediatric study or something different if it's on elderly people and so on. I believe it's more a judgement call rather than anything else. 
Similarly for weight. If you want to assess under-weighted, normal and over-weighted patients then these should be your presentation brackets rather than some means, medians, quantiles etc
It's usually good to have some kind of an end goal in mind to guide you through your initial descriptive analysis but not fully dictate it or bias it if that makes sense.
