Scientific Testimony: An Online Journal


(The articles will open in a new browser window.)

Challenging the Forensic DNA Evidence in People v. Marshall

This case study is presented in two parts. The first article, Examiner Bias in Forensic RFLP Analysis, discusses the scientific issues raised by the RFLP evidence in a rape case. It includes copies of the case autorads. The second article, People v. Marshall: The Legal Story, discusses the discovery battle and other legal machinations that brought to light the scientific problems in the case. It includes copies of key legal documents from the case. Readers are invited to comment on this case and on the analysis of it presented here by sending e-mail messages to the editor. All pertinent comments will be posted.

Update on Michael DeGuglielmo--The Marshall case is not the only case in which Mr. DeGuglielmo was accused of sloppy scientific work and unethical behavior.  See the Florida Supreme Court's opinion in Murray v. State (2002).

Unconventional Wisdom: The Lessons of the Oakland Hills Fire

Link to an article by JOHN J. LENTINI, C.F.I., DAVID M. SMITH, C.F.I. and Dr. RICHARD W. HENDERSON, C.F.I.

Editor's Comment:

How can you tell if a fire was caused by arson? For years, fire investigators were taught to look for key "indicators." Crazed glass, melted copper wiring, and melted steel were all said to indicate an unusually hot fire, consistent with the use of accelerants. Uneven burn patterns were said to reflect multiple ignition points, another indicator of arson. This conventional wisdom of fire investigation appears in textbooks and provided a "scientific" basis for expert testimony in thousands of cases.

In their provocative article, John Lentini and his colleagues argue that the conventional wisdom of fire investigators is simply wrong. Their analysis of 50 homes burned the 1991 Oakland Hills fire (a wild fire) showed a high frequency of traditional "arson indicators" where arson clearly had not occurred. Lentini and colleagues suggest that fire investigators have not realized the error in their conventional wisdom because there have been few careful, empirical studies of the results of "naturally occurring" fires.

When a report of the Oakland fire investigation was first published in 1992,1 it reportedly caused consternation among trainees preparing to take certification examinations in fire investigation. Should they answer the examination questions based on the conventional wisdom they had read in textbooks and learned in training courses or should they base their answers on the unconventional wisdom emerging from analysis of the Oakland Hills fire?

Readers should consider what this episode tells us about the nature of forensic science. How is it possible that conventional wisdom dating back at least 50 years could suddenly be found to be mistaken? In a field that claims to be scientific, why have so few studies been done to test fundamental assumptions? To answer these questions we must consider the institutional structure of forensic science and the nature of the "science" that it fosters. Look for much more on these institutional issues in future articles in this journal. Readers comments on these issues are welcome as well.

A link to a related article by John Lentini may be found below. Both articles are located on the web site maintained by Applied Technical Services, Inc., a private forensic testing organization. The ATSI website ( includes additional materials for those interested in fire investigation, including a study guide for the American Board of Criminalistics (ABC) certification exam in fire investigation.

The following materials, also by John Lentini, are also helpful in understanding arson evidence:

1. Lentini, J., Smith, D., & Henderson, R., Baseline Characteristics of Residential Structures Which Have Burned to Completion: The Oakland Experience. Fire Technology, Vol. 28, No 3, August 1992.

A Calculated Arson

Link to an article by JOHN LENTINI

A well-known arson investigator offers a critique of the investigation and scientific testimony in Commonwealth of Pennsylvania v. Han Tak Lee. The case is an interesting illustration of misleading expert testimony based on inconclusive data.