National open university of nigeria introduction to econometrics I eco 355

Example:
We return to the example concerning prenatal exposure to 
alcohol on birth weight in rats. Lets assume that the researcher's 
sample has n=16 rat pups. We continue to assume that population of 
normal rats has a mean of 18 grams with a standard deviation of 4.
There are four steps involved in hypothesis testing:
(I). State the Hypotheses:

Null hypothesis: No effect for alcohol 
consumption on birth weight. Their weight will 
be 18 grams. In symbols:

Alternative Hypothesis: ALcohol will effect birth 
weight. The weight will not be 18 grams. In symbols:
(II). 
Set the decision criteria:

Specify the significance level. We specify:

Determine the standard error of the mean (standard deviation of the 
distribution of sample means) for samples of size 16. The standard 
error is calculated by the formula:
The value is 4/sqrt(16) = 1.

To determine how unusual the mean of the sample we will get is, we 
will use the Z formula to calculate Z for our sample mean under the 
assumption that the null hypothesis is true. The Z formula is:
Note that the population mean is 18 under the null hypothesis, and 
the standard error is 1, as we just calculated. All we need to calculate 
Z is a sample mean. When we get the data we will calculate Z and 
then look it up in the Z table to see how unusual the obtained 
sample's mean is, if the null hypothesis Ho is true.
 
 

---

106 
 
 
 
 
 
 
 
(III). Gather Data
:
Lets say that two experimenters carry out the experiment, and they get these data:
Experiment 1 
Experiment 2