0
$\begingroup$

When performing an experiment to report gain of body mass in an experiment, do any of you know of a paper whose authors have used a bootstrap to improve their confidence intervals?

I have a n = 10 per group for a total of 6 groups, measuring whole body mass on a weekly basis in small animal models of a disease. Experimental design is essentially treatment by the use of a drug, over a period of ten weeks, and analysis of the final gain of body weight

(essentially think of a google search term as bootstrap, body mass, mouse)

$\endgroup$
  • $\begingroup$ Can you add more details? For example number of samples, metric measured, basic statistical estimates (mean/variance) and the experimental design. For your question, as it is now, searching bootstrap+body mass in google scholar would do. $\endgroup$ – WYSIWYG Jul 31 '15 at 9:33
  • $\begingroup$ @WYSIWYG I am not able to find a study that has explicitly used a bootstrap for body mass in the context I am seeking (see updated post).. am I missing the obvious here? $\endgroup$ – Rover Eye Jul 31 '15 at 9:42
  • $\begingroup$ You should clarify why you want to use bootstrapping - what statistical tests you performed and what was the primary observation (which you say, has low significance). Is it 6 groups each for treated and control plus 10 individuals in a group? Anyways see this. $\endgroup$ – WYSIWYG Jul 31 '15 at 9:43
  • $\begingroup$ @WYSIWYG bootstrapping enables us to understand the true range of confidence intervals (and subsequently increase the accuracy) much more, than the usual way obtained by means and variances. $\endgroup$ – Rover Eye Jul 31 '15 at 9:48
  • $\begingroup$ Also have a look at this paper. Basically what I wanted to ask was, whether you need bootstrapping. Is there some variation between groups? $\endgroup$ – WYSIWYG Jul 31 '15 at 9:56

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.