Interpreting T-Tests: Two Examples From the Text
In this experiment, the mean increase in plasma volume for the patients
treated with albumin was significantly greater (t = 3.55, P <.001) than
for the patients treated with polygelatin.
The results indicate that gelatin is more effective than polygelatin
as a plasma expander.
In this study, there was no significant difference (t = 1.34, P >.10)
between mean serotonin concentration of the subjects dying from heart disease
and of the control group of subjects dying from other causes
The results show that mean serotonin concentrations do not differ between
heart disease patients and controls without heart disease. [incorrect]
The results do not support the hypothesis that serotonin concentrations
differ between heart disease patients and controls without heart disease.
What are the important parts of an interpretation?
Given a level of significance (alpha) the
correct decision regarding the null hypothesis must be made. This is the
most important step and if you get this wrong whatever follows is irrelevant.
Iím recommending two sentences. In a report, the first sentence would appear
in the Results section and the second sentence would appear in the Conclusion/Discussion
section. The first sentence must contain the t and P values and the words
"significant" or "significantly".
The Results sentence should show by tense and/or use of the definite article,
that a particular group of individuals (a sample of people, fish, petri
dishes, plants, lizards, etc.) is being addressed.
The Conclusion sentence should show by tense and/or the lack of the definite
article that a population is being addressed.
The Conclusion sentence should address the alternative (research) hypothesis,
not the null hypothesis.
We never say that we accept the null hypothesis. This point is made clear
on p. 245 of our text.