Summer Fun! Statistics in Your Backyard

Minitab Blog Editor 17 July, 2012

There are some sounds that are quintessential summertime…the whir of the lawnmower, shrieks of children splashing in the pool, the crackle of a campfire. I’m sure we could think of a hundred more. For me, one sound that comes to mind in particular is the chirp of crickets in the evening. In this blog we'll recreate an old country trick using cricket chirps and, I hope, learn some new Minitab tricks along the way!

Our story starts with Amos Dolbear, an American physicist and inventor who lived in the late 1800s. Dolbear invented several elements of the telephone and was, therefore, attuned to sound of all kinds. The chirp of a cricket in his garden did not go unnoticed. In 1897, Dolbear published an article called “The Cricket as a Thermometer.”  In it, he described the correlation between the temperature of the air and the number of times a cricket chirps. The correlation formula became known as “Dolbear’s Law.

How Does Dolbear's Law Work?

Crickets are cold-blooded, like all insects. As such, their movements are affected by the temperature of their surroundings. This includes the movement of their wings as they make their mating call—the “chirp.”  There are over 900 species of cricket and, as you might suspect, not all crickets chirp at the same rate.  Dolbear developed his Law based on the chirpings of a snowy tree cricket.  Depending on the type of crickets in your backyard, you may get a slightly different result when comparing to the actual air temperature.  Let’s take a look at the results I got when I compared my cricket chirps and temperature in Minitab.

Chirps in Action

First I needed to decide which cricket chirp thermometer to use, since quite a few are available online. I went with the tried-and-true Farmer’s Almanac calculator, which very closely describes the correlation formula in Dolbear’s original article. The Farmer’s Almanac suggests counting the number of chirps the cricket makes in 14 seconds and then adding 40. This should closely approximate the current air temperature (in degrees Fahrenheit). 

With thermometer in hand, I trudged out into the backyard every day from May to June (*note—some numbers may have been collected by my husband, or neighbor, or even fudged a little…a Gage R&R might be apropos!) to listen to crickets falling in love. I also recorded the air temperature measurement and entered the data in Minitab.

Calculating the Chirps

Once the data was in Minitab, I used Minitab’s Calculator function to add 40 to my “Number of Chirps.”  The “Assign as a Formula” check box in the Calculator was very helpful because, once set, the column with the equation would automatically update whenever I entered data from a new day. 

I was also sure to add a comment to my “Number of Chirps” column to remind myself that these chirp counts were collected in a 14 second interval.  To do this, I simply right-clicked on the column and chose Column > Description…

Chirp Accuracy

The final step was to show how the cricket chirp calculation was correlated to the real temperature. I created a Scatterplot with Regression Line from the Graph menu and added a Text box describing the Farmer’s Almanac calculation that I used.  When I hover over the Regression Line, I can see the fitted equation.  That fit looks close enough to me!

One Chirp Further

This calculation really seems to work!  My next thought was that I could make a better calculation than Dolbear. I headed back out to the yard, but this time, instead of collecting chirps every 14 seconds, I collected chirps every 15 seconds. When I entered the data into Minitab, I modified the equation so that 37 was added to the chirp count. Finally, because I was comparing my method to Dolbear’s, I used Minitab’s Layout Tool to put the two graphs side by side. Much to my chagrin, the slope of my regression line is tilted ever so slightly more positive. A quick look at my R-Squared and coefficient values confirms what you might suspect: Amos Dolbear had it right!

Even though my Law didn't beat out Dolbear's, I did get to employ some neat Minitab tricks:

  • Using the Calculator to assign a formula to a column, and apply it to new data entered in that column
  • Adding a comment to a column to remind myself about an important facet of my data collection process
  • Creating a scatterplot with regression line, and hovering over the line to see my fitted regression equation
  • Using the Graph Layout tool to compare the results of my method and Dolbear's side-by-side