Identify a writer using statistics

Determine the variability of a writer

The underlying method for handwriting comparison is determining the variability of a person’s writing. Because people are not machines, every time we write the handwriting is different from other times. The concept is the same as a baseball pitcher throws the ball a a different speed with each pitch. A runner has a slightly different time in every race. These differences are called variability. The forensic document examiner determines the variability in the known writing of a writer. The handwriting analysis or signature analysis compares the questioned writing or questioned signature with the variability of the known writing.

Variability is a concept from statistics. Variability can be used to identify a writer using statistics.

Basis to identify a writer using statistics

Many established authorities such as Ordway Hilton, Wilson Harrison, Huber and Hedrick and others wrote about this topic. They state that proportions between tall letters (h, l, d, etc.) and small letter (a, c, e, i, etc.) remain consistent among writings of a person. In 2013 Italian forensic document examiner and researcher Fausto Brugnatelli published a peer reviewed article. Mr. Brugnatelli measured proportions to identify writer using statistics. His research validated using a quantitative method to identify a writer.

Potential source of error

In the case of signature forgery or forged writing, a person may trace a valid signature of another person. The forger may be a skilled artist who can copy the writing of another person. In these cases, the proportions of the tall and small letters may match those of the victim. The result may be a false identification. For this reason it is not appropriate to draw a conclusion whether writing is authentic solely based on statistics. Other factors must be examined.

Use a chart to identify a writer using statistics

This chart shows use of a quantitative method of scientific examination of questioned documents. This method discovers indicators the person who wrote the known signatures also wrote the questioned signature. In this example the height of the upper zone (tall) letter is compared with a middle zone (small) letter. The signatures were measured using the ruler tool in Photoshop to determine whether the questioned signature is statistically close to the variability of the known signatures. The ratio for each of the known signatures is shown by the blue undulating line. The horizontal brown line represents the ratio of the questioned signature.

In this chart the green line at ratio approximately 0.28 represents the average of the ratios of the known signatures. The questioned signature lies just below the average of the known signatures. This is an indicator the signature was written by the writer of the known signatures.