Basic
Concepts of Quantitative Research
Dr. R. Ouyang
Discussion
Interpretation of results
Hypothesized results
If
the results support hypotheses, it is relatively simple to deal with; if the
results do not support the hypotheses, if requires some explanation regarding
possible reasons. If the null
hypothesis can be rejected, that means the research hypothesis is supported but
it is not proven. If you do not reject
the null hypothesis, the research hypothesis is not supported. That does not mean the research is
wrong. The null hypothesis may be true.
Unhypothesized results
Often
during a study, an apparent relationship will be noticed which is not
hypothesized. To interpret
unhypothesized results should be careful.
You can collect and analyze data on these unforeseen relationships and
present your results. These findings
may then from the basis for a later study.
However, do not fall into the trap of searching frantically for some
thing that might be significant in your study does not appear to be going as
hypothesized.
Statistical vs. practical significance
Another
thing we need to realized in research is that the results are statistically
significant does not automatically mean that they are of any educational value.
For example, with a very larger sample, a very small correlation coefficient
may be significant but no real practical use to anybody. As the sample size increases the t ratio
tends to increase. Thus with very large samples, a very small mean difference
may yield a significant t. Thus, in a
way, the smaller sample sizes typically used in educational research studies
actually have a redeeming feature. Everything has its two sides. Smaller samples will increase the
possibility of rejecting null hypotheses; whereas larger samples will keep us
from finding some important relationships.
Replication of results
Replication
means that the study is done again. The
need for replication is especially great when an unusual or new relationship is
found in a study, or when the results have practical significance and the
treatment investigated might really make a difference.
Reference:
Gay, L. R. (1996). Educational research: Competencies for analysis and application. Upper Saddle River, NJ: Merrill.