Monthly Dinner Meeting
Topic: “Using Statistical Analysis to Assist with Writer Identification”
When: Thursday, September 17th, 2015 at 6:00 PM
Where: Temecula Library Click Here for Map30600 Pauba Road, Temecula, CA 92592
Dinner Cost:
$25.00 for non-members $20.00 for member
Make payment for Current Event
Thank you for making a payment. Please remember to bring a copy of the receipt as
proof of payment to the event.
.
Presentation Subject: Using Statistical Analysis to Assist with Writer Identification.
Abstract
Identification of handwriting relies on the document examiner’s ability to discover the variability of the subject’s known writing. Although qualitative assessment is prone to judgment error by the document examiner, variability in forensic document examination is often presented in qualitative rather than quantitative terms. Qualitative analysis is typically neither repeatable nor reproducible. A quantitative approach to writer identification can and should be used to reduce bias and errors in document examiners’ opinions.
Literature shows that variability of the proportions of height of a person’s written letters remains stable within common cause variability across writing sessions. This presentation applies aspects of statistical process control to determine whether a writer of a known document is a potential author of a questioned document, based on such variability.
For example: The ratio of the height of lower case letters extending into the upper zone and height of lower case letters remaining in the middle zone is calculated. The ratios of the length of lower case descenders to the height of middle zone letters are calculated. For each calculated ratio a run chart is created for the ratios of the known writing. The same is performed for the slant angle of a given letter.
The standard deviation of the ratios or angles are computed to one, two, and three sigma. These measurements are plotted to create a control chart.
The measurement of the angle or ratio of the questioned writing is plotted on the control chart to determine how the questioned writing compares to the known writing with respect to the variability of the known writing. The number of standard deviations from the mean of the known writing is explored to determine writer identification.
The null hypothesis is if the ratio of the given letters in the questioned writing falls beyond 2.5 standard deviations from the mean the questioned writing may have been written by a person other than the known writer. The result is an indicator of authorship rather than a conclusive determination.
Speaker Bio: Mike Wakshull, MSc, CQE is a practicing forensic document examiner, author and speaker. He authenticates documents and handwriting for legal purposes. Mike managed Corporate Quality Risk Management at Amgen. He moved to Temecula to lead writing and implementation of global Quality Risk Management procedures at Abbott Vascular.
Mike has been qualified as an expert in document examination in California Superior Court in Riverside, Orange, San Diego and Los Angeles Counties for civil and criminal cases. He has worked on federal civil and criminal cases.
Join us for a wonderful presentation of Handwriting Identification using Statistical Analysis by Mike Wakshull, September 17th, 2015!!!