JamesB
Master Poster
- Joined
- May 27, 2006
- Messages
- 2,152
Now I don't have much of a science background, but I do have a basic understanding of statistics, and I was wondering if those of you with a better understanding could tell me if I was doing this right.
I was reading Jones new thermite paper and I noticed the graph on page 21 regarding the energy release of their samples, where they argue:
http://www.bentham-open.org/pages/content.php?TOCPJ/2009/00000002/00000001/7TOCPJ.SGM
But as they mention, only 2 of the 4 samples released energy above that of thermite. Looking at their bar chart (I don't see anywhere they give the exact number, so this is an estimate) one comes up with the following (plugged into Excel):
Sample Kj/g
1 1.2
2 2.5
3 7.2
4 5.8
Avg 4.175
StdDev 2.795681193
There is a huge variance and therfore a huge standard deviation. The thermite they are comparing it to has an energy release of 3.9 Kj/g. So when I had Excel do a z-test using 3.9 as the total population, it returned a .42, which if I recall correctly indicated there is a 42% chance that the population they are sampling actually has a lower energy release than thermite.
Now this type of sampling is normally used for things like marketing surveys, but isn't this well below the typical 2 standard deviations, 95%, normally used to determine statistical significance?
Note, this is also leaving aside the question as to whether having a higher energy release than thermite would actually mean anything to begin with.
I was reading Jones new thermite paper and I noticed the graph on page 21 regarding the energy release of their samples, where they argue:
http://www.bentham-open.org/pages/content.php?TOCPJ/2009/00000002/00000001/7TOCPJ.SGM
It is striking that some of the red/gray chips release more energy in kJ/g than does ordinary thermite, as shown in the blue bar graphs above. The theoretical maximum for thermite is 3.9 kJ/g [27]. We suggest that the organic material in evidence in the red/gray chips is also highly energetic, most likely producing gas to provide explosive pressure. Again, conventional thermite is regarded as an incendiary whereas super-thermite, which may include organic ingredients for rapid gas generation, is considered a pyrotechnic or explosive [6, 24]. As this test was done in air it is possible that some of the enhancement of energy output may have come from air oxidation of the organic component.
But as they mention, only 2 of the 4 samples released energy above that of thermite. Looking at their bar chart (I don't see anywhere they give the exact number, so this is an estimate) one comes up with the following (plugged into Excel):
Sample Kj/g
1 1.2
2 2.5
3 7.2
4 5.8
Avg 4.175
StdDev 2.795681193
There is a huge variance and therfore a huge standard deviation. The thermite they are comparing it to has an energy release of 3.9 Kj/g. So when I had Excel do a z-test using 3.9 as the total population, it returned a .42, which if I recall correctly indicated there is a 42% chance that the population they are sampling actually has a lower energy release than thermite.
Now this type of sampling is normally used for things like marketing surveys, but isn't this well below the typical 2 standard deviations, 95%, normally used to determine statistical significance?
Note, this is also leaving aside the question as to whether having a higher energy release than thermite would actually mean anything to begin with.
Last edited: