Data Analysis and Research Methods

Course Description and Syllabus

The Bathroom Scale Activity #6 - Part 1, Activity 6

This lesson is adapted from an activity in Engineering Data Analysis, Ford Motor Company

Key Purposes:

1. Students can construct a cause-and-effect chart (C-E Chart) to identify variability in a process. (In this lesson, there is a special emphasis on measurement variabiity.)
2. Students can compare two populations when the samples are dependent.

   Example: Blocked data (strength of two hands)

3. Students can evaluate significance graphically.

To see a Quick Time Movie of the Bathroom Scale Activity, Click the above image. File size may be large, please allow enough time for the download.

Description:

The Bathroom Scale Activity starts out with a short exercise called the Letter Test. It was originally called the F test, but this might have been misinterpreted as Fisher’s F test. The students have a brief opportunity to read a paragraph and record the number of F’s in the paragraph. They write down their observations and then the class results are recorded. The resulting variability (some students miss the f’s in the of’s.) can be used to emphasize the problem of witness reliability in courtroom cases. It can also motivate a discussion of measurement variability, which is pertinent to the Bathroom Scale Activity. In this activity, the hypothesis that the dominant hand is stronger than the non-dominant hand is explored by having students squeeze a scale with each hand and recording the strength in pounds. Before the activity is started, students use a C-E chart to discuss standardizing procedures. Special attention is brought to measurement issues: Who reads the scale? At what angle is the scale read? Is the scale zeroed after each squeeze? How is the scale held? Will the data be rounded? After comparing the dominant and non-dominant class data by dotplots or boxplots, it can be pointed out that this data consists of dependent samples, and that the set of differences obtained from all the students can be analyzed. "Between" and "within" variation are discussed in an informal manner. The advantage to analyzing differences is discussed.

The boxplot of differences can be compared to a boxplot with the same variability, but centered at 0. A scatterplot of the paired data can be explored, and its relationship to the y=x line, which would be the appropriate line if there were no difference in hand strength, can be studied.