Statistical experiments are carried out to

EXPLORE: gather data to study more about the process or the product.

ESTIMATE: use the data to estimate various effects.

CONFIRM: gather additional data to verify the hypotheses.

Hypothesis 1: A new bonding method to bond two films is expected to yield a higher bonding strength compared to the current method.

In order to verify the hypothesis, I started by gathering a list of key variables in consultation with the group involved in the project. The key variables identified are: bonding glue, temperature, density and thickness of the films, and pressure setting. I ran the following experiment first.

Experiment 1: Two films were bonded together by choosing bonding glue type A, temperature level to be 300^oC, the thickness of the two films to be 4 mils, and a pressure setting to be 200 psi.

Data 1: The bonding strength measured was lower than the current method.

Question 1: Why is data 1 not supportive of the hypothesis 1?

After thinking over this experiment I arrived at the following conclusion.

Induction 1: The temperature setting may be low causing the glue to perform at below optimum level.

Experiment 2: Three sets of two films were bonded together by choosing bonding glue type A, the thickness of the two films to be 4 mils, and a pressure setting to be 200 psi. The temperature settings for these three sets were taken to be 400^oC, 450^oC and 500^oC, respectively.

Data 2: The bonding strengths for the three specimens were as follows:

At 400^oC the strength was still lower than the current one;

At 450^oC the strength was higher than the current one;

At 500^oC the strength was lower than the current one;
 

Induction 2: The temperature setting at 450^oC seems to give a better bonding strength when all other variables are set at the above mentioned levels.

The above investigation in various steps illustrates the basic ideas in a statistical experiment conducted in a scientific way. The remaining series of steps, with possible modifications including varying the settings of the variables simultaneously, form the basis of an experimental design. This will be seen in great detail later.

The basic ideas of this experiment can be summarized as follows.

• Constraint: the films should not peel off under "normal" usage.

• Key variables: bonding glue, temperature, density and thickness of the films, and pressure setting.
• Goal: the effectiveness of such bonding method.

• Procedure: All possible configurations in actual production setup should be considered in the study.

ESTIMATION: Suppose our study shows that the bonding strength is affected by glue, temperature and setting, then we would like to estimate the strength.

CONFIRMATION: Once we find the optimal settings, we run additional experiments to verify that the settings are in fact "best".

(a) Continue with the production.

(b) Not to use the method.

(c) Suggest appropriate modification in the process.

However, if it is not scientifically done, the conclusion may be totally false.