Statistical treatment of results in qigong experiments

Author: Lu Zuyin//Li Tipei
Institute of High Energy Physics, Academia Sinica, Beijing, China [1]
Conference/Journal: 1st World Conf Acad Exch Med Qigong
Date published: 1988
Other: Pages: 162 , Word Count: 387

From the statistical point of view, there are serious defects in most of the qigong experiments, viz, the defect arises from the small sample size of the experiment and the defect arises from the non-stable succeeding emission of qi through the whole experiment, so that it is not believed that valuable result can be obtained from such experiments.
But if we investigate the problem carefully, we will find that all the qigong experiments can be divided into two categories. Experiments designed for discovering new qigong effects belongs to the first category and the experiments designed for finding the laws of the effects belongs to the second. Up to now, most of the experiments are of the first category.
Two different methods of statistical treatment are related to these categories. The method of significance test is usually used for the treatment of the experimental result of the first category. According to the level of significance of the experimental event, one can easily discriminate the true qigong effect from the false appearance which arises from the background fluctuation.
Let µ be the average value measured of a physical quantity x. If x obeys the normal distribution N(x; µ, s2 ), where s is the standard error of measurement. Now we have measured a value of x, x*, which is significantly deviate from µ. If we denote the | x* -µ | / s as the 'significance' S, then we can say that the 'level of significance' of x is a
= N(å=x;µ,s2) The level of significance describes the probability of producing this unusual event by the background fluctuation only when there does not exist any condition to produce this unusual event. The lesser the level of significance, the higher the probability of existence a new phenomenon.
The significance test asks for a thorough knowledge of the background, but it does not ask for a big sample size of the events. A few events of sufficiently high significance (not necessary of the same highness) is enough to us to assert that the new phenomenon exists with high level confidence.
This is the case what we meet in the qigong experiment.
As an example, the experiment of the observation of the effect of qi on bovine thymus DNA is analysed.

The relation between S and a

S a
1 0.32
2 0.046
3 2.6x10-3
4 6.4 x 10-5
5 5.8 x 10-7
6 2.0 x 10-9