For the natural scientist, statistics is primarily a set of mathematical procedures that allow him to search for laws in a world in which the element of chance is constantly in effect. Most frequently, the use of statistical methods consists in testing the validity of hypotheses. If, for example, we find that 20 students infected by the Toxoplasma protozoa are, on an average, taller than 72 uninfected students, the relevant statistical method allows us to estimate, on the basis of the heights of all the 92 students, the probability that the observed difference in the average height of the infected and uninfected students is only a matter of chance. In this case, the t-test told us that this probability equals only 2.6%, indicating that there is great probability that there is some dependence between the height of the students (men) and infection by Toxoplasma gondii. However, the results of statistical tests understandably cannot answer the question of whether the infection increases the growth of students or whether taller students have a greater probability of becoming infected by T. gondii or that the height of the students and the probability of infection are affected by a third factor. In this case, the suspicious joint factor that simultaneously affects the height of the students and the probability of infection is the level of testosterone.
- more or less403
- not at all791