Graunt's Table of Casualties

John Graunt's Table of Casualties

Keith Greiner

September 5, 2016

An analysis based on data from ...

John Graunt's 1662 publication of the "Natural and Political Observations Mentioned in a following Index and made upon the Bills of Mortality", provides helpful information to modern historians and analysts. It shows over 80 classifications of ailments that affected the lives, and deaths, of 17th century Londoners. It shows when the plague was present, and, demonstrated to 17th century leaders that plague events were not caused by changes in leadership. In that context it prompted something we, today, refer to as "data based decisions." For me, it also shows why it is so very important for data analysts to check their data. Analysts need to make sure rows and columns are in the correct order. They need to make sure sums of columns and rows add correctly. What we see in Graunt's data is that the original columns are out of order, and some of the additions are incorrect. Some columns appear to have been transposed. This essay discusses introductory information about Graunt's table, and shows an example of one of his summation errors. At the bottom of the page is a link to a comma-delimited file that may be imported into Excel for further analysis.

Data analysts need to always be aware of possible errors in any table. They should always examine the data in each cell and should calculate check-sums of columns and rows to determine if the original values are correct. I continue to be amazed at the number of tables published in journal articles and mass media that contain simple mathematical errors. It is more common than one might think. The data published by John Graunt concerning the causes of death in the 1600's provide a good example of the value of a critical examination of the data. Checking the data for basic addition is the first step in a long list of steps that require critical thinking skills.

John Graunt was described by statistician Karl Pearson as the Father of Statistics. He was a London haberdasher (men’s clothier, hat retailer). He was born in 1620 and died in 1674, at the age of 54. Since William Shakespeare died in 1615 at the age of 52, that would put Graunt at approximately a generation after Shakespeare. During Graunt’s life, there were three major plagues: in 1625, 1636, and 1665. Fortunately, for us, Graunt found a way to analyze the data on causes of death between 1629 and 1660, excluding the years 1637 through 1646. A portion of that table is shown below and a comma separated value (.csv) version of the table is available at the bottom of this page.

The 82 row by 30 column table is too large to present in this medium. Some authors have suggested that the very idea that a single-sheet 80x32 table cannot be presented in this medium is a way that our modern Windows (pun intended) on the world limit our ability to do high level critical thinking. The presentation becomes even more problematic when we reorganize it to account for Graunt's non-sequential presentation of years. However, an online image of the table, apparently in its original form, may be found at here (as of August 28, 2016). Following is the Table of Casualties as it appeared as a fold-out in a 1975 reprint. Click on the image to see a larger view.

Also see this site:(John Grount Table Image of Original) The entire document is located here as of September 8, 2016.

The table selection that is displayed, below, has been modified from the original in several ways:

1. This table has been changed to show years on the left and death categories at the top,

2. This table includes only three death categories,

3. This table includes two sections that show my check-sums of multi-year sums published by Graunt, and variations between my calculations and Graunt's.

Grant's table has several problems. There are sums that are incorrect, and some numbers appear to have been incorrectly copied or transposed. Grant did not have the benefit of a modern electronic spreadsheet, so the errors are understandable. Still, they should be considerations for anyone who might want to find meaning in his data.

The most notable problem is that Graunt’s combined total for the years 1629 through 1632, for Aged vs. Ague and Fever, have opposite errors. The Aged calculation shows an error of 148, while the Ague and Fever column shows an error or (148): that means it has an error of a negative 148. Without access to the original data, we are unable to find details of the error, but at least we know it is there. Apparently Graunt believed that a large proportion of the plague deaths were unreported. However, the increase in Ague and Fever deaths in 1635 suggests that perhaps some of the plague deaths were categorized as Ague and Fever. In today’s terminology, Ague, is malaria, a disease that has symptoms of fever and chills. Plague, could also have symptoms of fever and chills, so it is understandable that in 1635 there might be some mis-classification.

At the very bottom of this page there is an attached .txt file that contains two variations on the Graunt table. To download the file, click on the downward point arrow to the right of the line at the bottom of the page. To import the file into Excel, download the file, change the extension from .txt to .csv and import to Excel as a .csv file. That should work, but cannot be guaranteed. Carefully read the comments at the top of the file.

On September 10, 2016, Cable News Network (CNN) (See this link) reported that British researchers found DNA from a burial pit that was used between 1650 and 1670. Samples from teeth gathered at the site indicated that the people buried there did, indeed, die of the Black Plague. The article includes some numbers, that are not the same as those published by Graunt. So the numbers presented by CNN may not be traceable to a reliable quantitative source. On the other hand, we do have Graunt's quantitative table as a resource.

Table 1 shows Plague in the right-most column. There, we see that Plague deaths peaked at 10,400 in 1636; the last year before the gap. The time-series is also presented in Graph 1 and is shown below. Other sources such as this one at Harvard tell a different story and reference other “eyewitness” accounts that fill in some of the missing years. It is interesting that those references seen in the article on September 10, 2016 do not include or even mention Graunt’s data, if only for comparison. Reference to Graunt's table is appropriate because it reveals a longitudinal series that, we hope, used the same or similar methodologies each year.

There is an interesting essay about Graunt at this Xavier University site on September 10, 2016.

The .txt file is below. To use it as a .csv file, first download it, then change the .txt to .csv, and then open it in Excel. It should import without problems.

Another method is to open the .txt file and then use Select All --> Copy in the text editor followed by Edit --> Paste in the Excel spreadsheet. This should put all the lines of text into the left most column of the active sheet. Then use Data --> Text to Columns. That, or some variation of it, should put the data into the appropriate columns. If not, use the first method.