Log in
Register|
joystar ☆ Morocco, 2012-10-27 20:14 (4560 d 21:21 ago) Posting: # 9472 Views: 5,690 |
|
|
Dear All, can anyone please tell me free software for sample size calculation and randomization. Thanks in advance |
|
Helmut ★★★   Vienna, Austria, 2012-10-27 20:46 (4560 d 20:48 ago) @ joystar Posting: # 9473 Views: 4,960 |
|
|
Hi Joystar, I would suggest that you install R first (download, Windows FAQs including installation instructions). Then you can run all these nice programs posted here.  Two great packages are maintained by Detlew Labes:For a standalone randomization program by ElMaestro see this post. — Dif-tor heh smusma 🖖🏼 Довге життя Україна! ![[image]](https://static.bebac.at/pics/Blue_and_yellow_ribbon_UA.png) Helmut Schütz ![[image]](https://static.bebac.at/img/CC by.png) The quality of responses received is directly proportional to the quality of the question asked. 🚮 Science Quotes |
|
ElMaestro ★★★ Denmark, 2012-10-27 22:03 (4560 d 19:31 ago) @ Helmut Posting: # 9474 Views: 4,900 |
|
|
Hi Joystar, for sample size in simple designs there is also a program called EFG which isn't as advanced as all the others - an evil friend of mine called the user interface 'puristic' if I recall correctly. In contrast the aforementioned products that require installation of statistical packages, EFG is a standalone application on Windows machines, so it is just a single download, unzip and run. You can click here to download it. — Pass or fail! ElMaestro |
|
Helmut ★★★ Vienna, Austria, 2012-10-27 22:28 (4560 d 19:07 ago) @ ElMaestro Posting: # 9475 Views: 4,862 |
|
|
My Master, how could I – as an involuntary beta-tester and one of the few knowing what “EFG” stands for?  — Dif-tor heh smusma 🖖🏼 Довге життя Україна! Helmut Schütz The quality of responses received is directly proportional to the quality of the question asked. 🚮 Science Quotes |
|
d_labes ★★★ Berlin, Germany, 2012-10-29 18:02 (4558 d 22:32 ago) @ ElMaestro Posting: # 9480 Views: 4,809 |
|
|
My dear Ol Pirate ,there is one caveat in recommending ElF#... It was not under development since 2009. The flaw we have discussed in this and following posts regarding brute force power estimation is still present  . Ok, ok, this is nitpicking, I know, because the settings under which the flaw becomes obvious have nothing to do with 'real life settings' we usually deal with in power calculations in BE studies. But IMHO confidence in software we are using comes not at least from how they behave in extremal cases.Moreover: The normal power calculation give also power values for the designs "2x2x3" and "2x2x4". But they differ from the results of other software. Example: CV=30%, n=10, T/R=0.95 ElF#... power.TOST() Unfortunately ElF#... is not open source. Thus only you can check what's going on here. — Regards, Detlew |
|
ElMaestro ★★★ Denmark, 2012-10-29 18:15 (4558 d 22:20 ago) @ d_labes Posting: # 9481 Views: 4,743 |
|
|
Hi d_labes, ❝ Unfortunately ElF#... is not open source. Thus only you can check what's going on here. Would you have an interest in the C source?? Thanks for the positive criticism. Yes, probably there is a bug in the 223-calculation. Heck, I don't remember how it was written. It is an uncommon design anyway for the purpose of plain average BE. Nowadays it is almost exclusively used in relation with scaling which is another story altogether. Once I am back in my office I will see if I can locate the source code for EFG and get the bug corrected. Easiest option is prolly to just eliminate the 223-results and possibly even the 224-results, too. — Pass or fail! ElMaestro |
Ing. Helmut Schütz