Take our Survey Bookmark and Share
Examining the Evidence


Weighing the Evidence > Reliability
…lawyers and judges often have insufficient training and background in scientific methodology, and they often fail to fully comprehend the approaches employed by different forensic science disciplines and the reliability of forensic science evidence that is offered in trial.

Determining Reliability

Reliability is one of the most critical factors in ensuring that evidence presented in court is useful, true, and accurate—and that it will produce just results.

The concept of reliability pertains specifically to results or conclusions arrived at by an examiner. From test to test, lab to lab, and analyst to analyst, it’s important to ensure or challenge the reliability every single time.

Have all potential sources of error been accounted for? Is the result or conclusion reasonably free of error? How do we know? How can we know?

Factors in Reliability

A number of factors are considered to determine whether or not a result is reliable, such as:

  • Test error rates
  • Laboratory and examiner error rates
  • Statistical measures of uncertainty
  • Methods for reducing these variables through accreditation and certification.

The recommendation made in A Path Forward suggests that all measurements be reported with arithmetic indications of uncertainty, such as confidence intervals.

It’s worth noting here that some of the pattern/experiential forensic disciplines cannot adhere to certain aspects of reliability as closely as the hard sciences.

It’s not unusual to have psychiatrists or psychologists disagree as expert witnesses on the same question. Forensic pathologists often disagree on topics that require interpretation, such as manner of death, and some questions simply do not lend themselves to the rigorous repetitive testing that can be done in a forensic chemistry lab.

Increasing Reliability

Analysts can take steps to increase reliability and decrease errors. For example, they can conduct validation studies to help establish:

  • Reproducibility, which concerns the reliability of re-testing. Does re-testing of the same specimen under the same conditions return the same result consistently? (To some extent, reproducibility is established during validation studies or – less often – through proficiency testing.)
  • Inter-analyst or inter-rater reliability, which pertains to consistency and reproducibility among different analysts
  • Internal reliability in instrumental methods, which has to do with the measuring or separation system being used

Reliable, Valid, Accurate

A procedure can be valid, and based on valid science, but still not be reliable. You should also note that reliability does not necessarily establish the validity or accuracy of a result. A procedure or technique consistently giving an incorrect result can be reproducible. Techniques and protocols must be both reliable and valid.

In the following sections, consider the uncertainty present when something is measured in a scientific test, or how there are bound to be errors of varying degree whenever human beings are involved.

Learn how laboratories and analysts undergo periodic proficiency testing to ensure that techniques remain reliable, and how—after decades of rigorous scientific study and testing—DNA testing has become a gold standard of reliability that other forensic techniques strive toward.