National open university of nigeria introduction to econometrics I eco 355

the treatment has 
no effect
. For our example, we formally state:
o
The null hypothesis (Ho) is that prenatal exposure to alcohol 
has no effect
on the birth weight for the population of lab rats. The birth weight will be 
equal to 18 grams. This is denoted
o
The Alternative Hypothesis (H1).
This hypothesis states that 
the 
treatment does have an effect
. For our example, we formally state:
The alternative hypothesis (H1) is that prenatal exposure to alcohol 
has an effect
on the 
birth weight for the population of lab rats. The birthweight will be different than 18 
grams. This is denoted
4.
Second Step: Set the Criteria for a decision.
The researcher will be gathering data from a sample taken from the population to 
evaluate the credibility of the null hypothesis.
A criterion must be set to decide whether the kind of data we get is different from what 
we would expect under the null hypothesis.
Specifically, we must set a criterion about wether the sample mean is different from the 
hypothesized population mean. The criterion will let us conclude whether (reject null 
hypothesis) or not (accept null hypothesis) the treatment (prenatal alcohol) has an effect 
(on birth weight).
5.
Third Step: Collect Sample Data.
Now we gather data. We do this by obtaining a random sample from 
the population.
Example
: A random sample of rats receives daily doses of alcohol 
during pregnancy. At birth, we measure the weight of the sample of 
newborn rats. We calculate the mean birth weight.
6.
Fourth Step: Evaluate the Null Hypothesis
We compare the sample mean with the hypothesis about the 
population mean.

If the data are consistent with the hypothesis we 
conclude that the hypothesis is reasonable.

If there is a big discrepency between the data and 
the hypothesis we conclude that the hypothesis was wrong.

---

102 
Example
: We compare the observed mean birth weight with the hypothesized values of 
18 grams.

If a sample of rat pups which were exposed to prenatal alcohol has a 
birth weight very near 18 grams we conclude that the treatement 
does not have an effect. Formally we do not reject the null 
hypothesis.

If our sample of rat pups has a birth weight very different from 18 
grams we conclude that the treatement does have an effect. Formally 
we reject the null hypothesis.