Most of our biological work involves whether or not there exist differences among different groups. Whenever we collect two groups of scores by random methods we are likely to find that the scores (and their means) differ to some extent. Some differences occur simply because of chance. When looking at two groups we are then faced with the problem of whether the observed differences are merely due to chance or not. Saying that one mean of one sample is larger than another mean of another sample is just not enough evidence that a real difference exists between the two larger populations.
The procedures of statistical inference enable us to determine, in terms of probability, whether the observed difference is within the range which could easily occur by chance or whether it is so large that it signifies that the two samples are indeed probably from two different populations with real differences in characteristics. One of the first steps in the decision-making procedure is to state the null hypothesis. The Null Hypothesis is an hypothesis of no differences, i.e., an hypothesis that there is no difference in some characteristic between test groups. (Ho:m1=m2) The actual hypothesis that the researcher is interested in is usually the alternative hypothesis (H1:m1=m2) if the null hypothesis fails, i.e., that there probably is a real difference. Statistics can not prove there is a difference between two groups. The best that it can do is to tell you that an observed difference between samples is substantially different (i.e., significantly different) from what would be expected by chance if the samples came from two groups which were the same (or had the same characteristics).The researcher must choose the point at which the probability of observing such a big difference between two groups by chance is so low that the researcher can reasonably reject the null hypothesis of no difference and conclude that there probably is a real difference. The point chosen is called the
alpha level (a ), usually at 5% or 1%. This means that such a large observed difference between groups with the same characteristics would occur only 5% of the time by chance. This also means that except for a (perhaps important) 5% chance of being wrong the researcher can reasonably reject the hypothesis that the two groups are the same. The null hypothesis is assumed to be true unless the data cause the researcher to reject it.The term "p value" means that under the null hypothesis of no difference between groups the observed difference between groups (or a larger difference) will occur with the probability "p" with random sampling. When the null hypothesis is rejected you would write p < .05 (when
a = .05). When the null hypothesis is accepted you would write p > .05 (when a = .05).Thinking in terms of a null hypothesis is part of good research procedure. It makes you stay objective and avoid bias. You don't just assume that your pet hypothesis is true and try to gather data to support it. You might ignore data subconsciously that refutes it. You should assume your pet theory is not true until the evidence you collect overwhelmingly compels you to conclude that it, indeed, must be true.
| Current Percentage Metric Time is |
|
O'Clock |
| Front Page | Contact Info | Office / Teaching | Acad. Background | IUP Links |
| General Biology I | General Biology II | Biometry | Ornithology | WWW Links |
Dr. Winstead's Biology Careers and Job Searches Starting Point
Dr. Winstead's Current Local and World Standard
Percentage Metric Time Clock
Dr. Winstead's Nintendo DS Battery Indicator Light Solution
for Color Blind (Color Deficient) Users