I was teaching a Black Belt course in Washington, DC and i was asked to simplify the basic roadmap in a Hypothesis Test. In the example below, we are testing whether there is a correlation between two variables.
Example:
1. What is the practical question?
- Does an increase in tire pressure cause an increase in tread wear?
2. What is the X?
- Tire Pressure
3. What is the Y? (What am i measuring)
- Tread Wear
4. State the Null (Ho) and Alternative (Ha) Hypothesis
Note* in this example we are using Correlation: the Null Hypothesis is "r=0' (or there is no correlation)
- Null Hypothesis (Ho) = There is no relationship between tire pressure and tread wear
- Alt Hypothesis (Ha) = There is a relationship between tire pressure and tread wear
2. Gather data, Run the analysis and determine the P-Value
- Run a Corellation (the following metrics are an example of the outcome):
- r=.554, p-Value = .0228
3. Determine the Alpha Risk (100% - Confidence Interval)
- In this case the Confidence Interval was 95%, therefore the Alpha Risk is 5% (or 0.05)
5. What does the P-Value tell me ... (Reject or Accept the Null)
- Remember the saying: If P is low, then the "HO" must go! If P is high then keep the guy!
- (Remember that "P" is dependant on the Confidence Interval)
5. In this case we ....
- Reject the Null (because the P-Value (.0228) is lower than the Alpha Risk (0.05))
If you want to learn more, contact me at kclay@sixsigmadsi.com or call 479-739-4940. Check out my other blogs at http://www.sixsigmadsi.com/.
Subscribe to:
Post Comments (Atom)
0 comments:
Post a Comment