How do I determine the class size for creating a histogram in statistics?

[Saladsamurai]

I have a quick question regarding a stats problem in my girlfriend's text. It is pretty easy I suppose, but I am not quite sure; the text does not appear to give a "general form" of how to obtain the info.

Feel free to correct me if I misuse words here, I am not familiar with the stats lingo yet.

It is asking to make a histogram. There is a table that gives us each state (the individual) and the number of Medical Doctors per 100,000 people is in each state (<--the variable, I presume).

Now I know that I find the Range by subtracting the Lowest Actual variable from the Highest actual Variable.

Here is where I am getting a little lost: I think now I am supposed to divide the actual range by the "class size" in order to find the number of classes so I can start to draw my histogram.

But how do I choose the size of each class? Is it arbitrary?

And why do I use the actuals to compute the range instead of the reasonable beginning/end?Thanks,
Casey

 

[EnumaElish]

I'd first plot the data by state to get a "feel."

Class size is arbitrary. It partly depends on whether you'd like the histogram to look like a "normal curve" or can live with "gaps in the middle."

You do not have to use the actual endpoints, but that's one way.

 

[Architeuthis Dux]

Hello Casey,

Thank you for reaching out with your question. Determining the class size for creating a histogram in statistics is an important step in accurately representing your data. The class size refers to the width of each bar in the histogram and can greatly impact the overall shape and interpretation of the graph.

To determine the class size, you first need to calculate the range of your data, as you mentioned. This is done by subtracting the lowest value from the highest value in your data set. However, instead of using the actual range, it is recommended to use a rounded version of the range. This will help ensure that your classes are evenly distributed and easier to interpret.

The size of each class can be chosen based on the number of classes you want in your histogram and the range of your data. A general rule of thumb is to aim for 5-15 classes, but it ultimately depends on the size of your data set and the level of detail you want to show. A smaller number of classes may result in a simpler, more general overview of the data, while a larger number of classes can provide a more detailed, nuanced representation.

It is important to note that the size of each class should be the same throughout the entire histogram. This helps maintain consistency and makes it easier to compare data points within each class.

In terms of why we use the actual range instead of a "reasonable beginning/end," it is because using the actual range ensures that all data points are included and accounted for in the histogram. If we were to use a smaller range, some data points may be left out and the overall picture of the data would not be accurate.

I hope this helps clarify the process of determining class size for creating a histogram. If you have any further questions or need additional assistance, please don't hesitate to reach out. Best of luck with your stats problem!

Sincerely,