This applet is designed to teach students how bin widths (or the number of bins) affect a histogram. The histogram below is for the Old Faithful data set. The observations are the duration (in minutes) for eruptions of the Old Faithful geyser in Yellowstone National Park. Students should interactively change the bin width by dragging the arrow underneath the bin width scale.
This applet is designed to teach students how bin widths (or the number of bins) affect a histogram.
Target Student Population:
Students in an introductory statistics class.
Prerequisite Knowledge or Skills:
Basic understanding of how to read/interpret histograms. Other than that the applet is self-explanatory.
Type of Material:
An instructor may use this before, during or after a lesson on histograms. Could be assigned as homework or be shown in class with computer/projector. This could also be used for student self-exploration.
Java-enabled web browser.
Evaluation and Observation
Very clearly demonstrates that choice of bin width can dramatically alter the appearance of a histogram, possibly affecting our perception of key features such as number of modes and type of skew. The particular dataset used in the demo (eruption times of Old Faithful) appears to have only one mode when the bin width is large, but clearly has two modes for smaller bin widths. It is also nice that one can create weblinks to the applet to plot one's own data. The overall strength is that this applet is simple, clear and does a good job of demonstrating what it intends to.
* The concept that this applet is designed to teach, though important, is simple and fairly easy to grasp, which limits its impact.
* On the technical side, the applet provides a "click and drag" method for dynamically changing the bin width, which is convenient and intuitive, but lacks precision. It would be helpful if, in addition, the user could see, and to be able to input, numerical values of bin width.
* The use of the applet with one's own data seems to require entering all its numerical values directly within one's HTML file. This can be cumbersome for large datasets, as well as for plotting multiple variables from a single dataset.
Potential Effectiveness as a Teaching Tool
Should be very effective for teaching the importance of choosing appropriate scale and bin widths when plotting and interpreting a histogram. Should enhance a classroom discussion on this topic. It should also readily fit into an introductory statistics curriculum, regardless of specific textbook. Students learn on their own that the bimodal nature of the dataset is hidden, and for small bin widths the plot reduces to a spike at each data point.
There should be additional datasets (and associated histograms) and follow-up questions to get students to investigate other issues that emerge when adjusting bin widths. It would be helpful if the applet showed numerical value of bin widths as part of the demo.
Ease of Use for Both Students and Faculty
Very easy and intuitive to use the applet with the default dataset (eruption times of Old Faithful) that it comes with. Should be accessible to users with a broad range of technical backgrounds and hardware/software platforms. Moderately easy to link to the applet with one's own datasets.
If possible, it would be a key enhancement if the coding framework is extended to allow users to specify a datafile when linking to the applet. If the browser is not java enabled, there may be some issues downloading java on lab computers, which often have draconian security settings.