Comparing Data Sets Using Statistical Analysis in Excel - A Basic Statistics Lesson
(Page 2 of 4 )
In order to maximize what you learn from this tutorial, I will provide a short background regarding statistical analysis.
Inferential statistics is the most underestimated (and underutilized) statistical tool in the engineering and business industry. Some managers or even engineers resort to hunches or guesses based on the numerical data available without using valid statistical techniques to test it.
Commonly, there are two major ways to compare data:
- Compare the means of the two or more data sets. For example, you can run a paired experiment. This is an experiment where the subject is the same but is subjected to different test conditions. A classic example is a single rat subjected to two different medications in an experiment designed to find out the effect of each medication.
- Compare the variation of two or more data sets. In most comparative studies, the mean is the most important piece of data. But there are also times when the variation is also studied. This is particularly useful in studying the stability of manufacturing processes, for example.
In statistical literature, a T-test is used to compare the means of two datasets. ANOVA (Analysis of variance) is used to compare the means of two or more data sets. An F-test is used to compare the variation of two data sets.
In MS Excel, the following are the available tools for both built-in and data analysis toolpak features:
WARNING: In any scientific research, the accuracy of the measurement system is extremely important. So make sure you have validated the accuracy of your measurement system before proceeding with the gathering of data and doing Excel analysis.
So if you are ready, let us proceed to case studies (T-test and F-test). Advanced statistical techniques for comparison like ANOVA will be covered in a separate article because they are broad and have a variety of applications.
Next: How to conduct a T-test in MS Excel >>
More BrainDump Articles
More By Codex-M