TacticalBriefs-header.jpg (19377 bytes)

April 2001
Web Site Index and Navigation Center

Reality of the Street? A Practical Analysis of Offender Gunshot Wound Reaction for Law Enforcement

By Shawn Dodson



A uniform process to analyze the efficacy of law enforcement handgun ammunition, based on studying the relationship between offender gunshot wound reactions and commonly encountered random variables, currently does not exist. Such a study must examine an agency's entire population of shootings with a given cartridge to consider key factors that may have influenced offender reaction to being shot, and yet be straightforward enough that a professional statistician isn't required to process and analyze the data.

A valid methodology allows an agency to correctly assess wound trauma incapacitation efficacy. Standardization of the methodology among law enforcement agencies enables valid interagency comparisons of various handgun ammunition cartridge brands, calibers, bullet types and bullet weights.

Statistical Representations of Data

(While this particular part of the article is not an easy read, and some readers will struggle to grasp the information it presents, it does not characterize the effort required to comprehend the material presented in the rest of the article.)

In order to understand statistical validity one must have a basic understanding of statistical representations of data. There are two types of statistical analyses:

  1. Descriptive statistics are used to describe a situation involving a known population; it entails the collection, organization, summarization, and presentation of data. Examples of descriptive statistics include:

  • A baseball player’s performance statistics by career, season or season to date

  • The percentage of voters nationwide, or by state, which chose George W. Bush in the 2000 U.S. Presidential election

  • The average penetration depth of five bullets shot into a block of bare ordnance gelatin

  1. Inferential statistics are used to make inferences (estimations and predictions) from samples to populations, and involve statements and generalizations about the probability of their validity. Examples of inferential statistics include:

  • Earthquake predictions

  • Pre-election polling techniques used to predict election results

  • Estimating reliability of a weapon system based on performance data collected during operational tests

  • Evan Marshall's one-shot stop percentages

It is important in statistics to be able to distinguish between a population and a sample. Population is a statistics term used to describe the total of all subjects that possess some common characteristics that are under study. In the context of this paper, population refers to the whole number of criminal offenders who’ve been shot with a given ammunition cartridge.

Generally a population is very large and/or it is too difficult and expensive to study all subjects of the population, therefore a sample of the total population is selected. A sample is a collection of some, but not all, of the elements of a population.

A sample population may be selected by a researcher in a number of ways. In random sampling all items in a population have an equal chance of being chosen for the sample. In non-random sampling there’s bias toward selecting certain items in a population, and, more often than not, it leads to misrepresentations about a population and should be avoided.

Random sampling forms the basis of statistical inference, and there are four basic types of samples. The first type is called simple random sampling, in which the sample is selected in such a way that it allows every member of the population to have the same chance to be selected. The second type is called systematic sampling, in which each subject of a population is numbered and then every X th number is selected. The first number for the sample is selected randomly followed by selection of every X th number.

The third type of sampling method is called stratified sampling, in which the population is first divided into relatively homogeneous groups called strata. Two methods can be used for collecting the sample: one can select at random from each stratum a number of elements that corresponds to the proportion of that stratum in the whole population. Or one can select an equal number of elements from each stratum and weigh the results according to the stratum’s proportion to the population. In stratified sampling, every element has a chance of being selected. This method is appropriate when the population is already broken into groups. When designed well, stratified samples are the most accurate reflector of the characteristics of a population.

The fourth common type of sample is called a cluster sample, in which the population is divided into groups or clusters and then one of the groups or clusters is selected randomly. A properly crafted cluster sampling procedure can give a precise sample at less cost than simple random sampling.

Reality of the Street?

This paper is not intended as criticism of Evan Marshall's "one-shot stopping power" system of rating cartridge effectiveness, however examination of the defects inherent in his methodology is unavoidable. 

For the sake of discussion, let us assume, for the moment, that Marshall’s one-shot stopping power data are uncorrupted. Which type of inferential statistics sampling technique does Marshall use to estimate one-shot stops? You are correct if your answer is non-random sampling. Marshall obtains his data using non-random, opportunity sampling. This is a fatal flaw in his sampling methodology, which introduces a bias (sampling error) that substantially misrepresents the reality of the street.

A second critical error committed by Marshall is that he presents his non-random samples of unknown populations as though the data possess the absolute validity of a descriptive statistics analysis, which they obviously do not.

For these reasons, Marshall’s methodology is conceptually defective and there’s no need to delve any further. The fundamental errors in his method are significant enough that an independent review of his data is pointless.

Reality of the Street: A Composite Picture

A properly constructed, sample-based inferential statistics study of handgun stopping power, developed with the assistance of a professional statistician, is incapable of deducing the reality of the street with enough accuracy to produce a meaningful finding.1 Simply put, statistical uncertainty is so large it blurs any practical distinctions that may be found in wound trauma incapacitation effectiveness.

Therefore real-world performance is best examined using the methods of descriptive statistics, which offer two advantages over an inferential statistics study. Firstly, descriptive statistics are purely empirical. There is little error; the data are accurate because they represent a known population of shootings with a given cartridge. Secondly, descriptive statistics can be used and understood by anyone who comprehends basic arithmetic.

The disadvantage of a descriptive statistics study is there is little application of the methodology outside of the law enforcement community. Thus the variety of cartridges studied are limited to those commonly issued by large law enforcement agencies whose officers are involved in several shooting incidents each year. This disadvantage is far outweighed, however, by the quality of data.

The study’s scope must not be restricted to assessing one-shot stop stopping power because this produces a distorted view of wound trauma incapacitation efficacy. There are at least five and possibly six different types of shooting incidents, based entirely on an offender’s immediate reaction to being hit by gunfire. They are:

  1. Incidents involving one gunshot wound: One-hit stop

  2. Incidents involving one gunshot wound: One-hit failure to stop, not hit again

  3. Incidents involving multiple gunshot wounds: More than one-hit required to stop

  4. Incidents involving multiple gunshot wounds: Multiple-hit failure to stop

  5. Indeterminable

  6. Other

The first four types of shooting incidents (I through IV) are self-explanatory.

Shootings categorized as Indeterminable involve circumstances in which an offender’s immediate reaction to being shot cannot be established with certainty. These types of shooting incidents include, but are not limited to:

  • An incident in which an offender is incapacitated by multiple hits delivered in rapid succession, and it is impossible to determine if more than one hit was required to force him to stop (for example, multiple hits delivered in rapid succession by a single officer or multiple hits, fired by more than one officer, that arrive on target almost simultaneously).

  • An incident in which a barricaded offender is shot at, disappears from view, and his immediate reaction to being hit cannot be observed or ascertained. It might be impossible to determine precisely when he was hit, and what he did immediately afterward.

The first five shooting incident types cover all anticipated scenarios. However category VI, Other, is included to deal with unanticipated shooting incident outcomes that, for one reason or another, don’t fit any of the other categories. This is necessary because no shooting can be excluded from the study.

Random Variables Affecting Offender Gunshot Wound Reaction

It is not enough to merely examine offender gunshot wound reaction. To obtain an accurate, big-picture perspective of wound trauma incapacitation efficacy critical details of every shooting incident must be considered. This task involves identifying, recording, interpreting, evaluating and processing data about commonly encountered random variables that have the potential to influence an offender’s immediate reaction to being shot. A thorough study allows an assessment of the totality of circumstances involved, to possibly gain a deeper level awareness of why an offender reacted in the manner he did, which is of vital importance to understanding the situation of wound trauma incapacitation in a broader context.

Table 1 (click here to view Table 1) is an essential visual aid tool, which presents a coherent analysis of the relationship between commonly encountered random variables and offender gunshot wound reactions. It exemplifies a descriptive statistics analysis of a hypothetical law enforcement agency’s experience to date with the .40 S&W Remington 165 grain Golden Saber JHP cartridge, which the agency issues for general purpose use with department issued Glock handguns. The study spans the entire period of time from when the cartridge was first placed into service to present day, or until it was replaced by a different cartridge if a historical study is being performed.2 The example data presented in Table 1 are for illustrative purposes only, to help readers organize and make sense of the concepts presented in this paper.

There appear to be three major factors that influence an offender’s reaction to being shot:

  1. The offender's mental state, 

  2. Factors that affect a bullet's terminal performance in the intended target (penetration, expansion and energy transfer), and

  3. Wound trauma physiological disruption produced by the bullet(s).

Mental State

The most difficult factor to discern with certainty is an offender’s state of mind. It is an intangible element that cannot be quantitatively measured in terms of hard, physical evidence. Mental state is almost entirely dependent on eyewitness reports of offender behavior. For example: Do eyewitness reports suggest the offender acted in an irrational manner that indicates some kind of mental or emotional disorder? Did he have a history of mental or emotional problems? Was he unable to comprehend the immediate personal danger to himself when confronted at gunpoint? Determining an offender’s mental state is unavoidably subjective but it is an important factor to consider nonetheless, regardless of the chance for error.

Offenders who have consumed alcohol, who are high on drugs, who are psychotic, who show signs of intense emotional agitation, and/or who are demonstrating a firm determination to inflict as much harm as possible before being stopped seem less likely to break down mentally and be quickly incapacitated by non-vital gunshot wounds than offenders who do not exhibit these traits, according to anecdotal evidence.

An offender may act for reasons of panic and/or desperation, such as trying to flee, and his desperate actions, intentionally or unintentionally, put himself or others in danger. Therefore acts of desperation should not be confused with acts of determination, in which an offender is mentally resolved to stay in the fight or he is acting with sheer willpower, despite injury, to 'drive-on' in attempt to accomplish his ruthless personal mission of destruction.

If it is established that an offender was under the influence of alcohol or drugs or both, a uniform definition of under the influence must be established. Using Table 1 an an example database, an “X” should be entered when toxicology tests indicate chemical substance concentrations in the bloodstream were probably enough to impair judgment and/or cognition at the time he was shot. Qualified medical advice should be sought to determine standardized minimum value definitions for chemical substances that can cause mental impairment.

Law enforcement agencies using these database criteria are advised to record toxicology measurements, as well as the name and type of drug (stimulant, depressant or hallucinogen) on a separate worksheet.3

Adverse Conditions Affecting Projectile Terminal Performance

Knowing what the bullet(s) encountered from the time it exited the muzzle until it came to rest in the offender’s body or exited his body is important information that can be helpful in determining why he reacted the way he did. The bullet(s) may have encountered an obstacle that adversely affected its ability to penetrate, expand or disrupt tissues.

For example, heavy clothing and several layers of medium or light clothing can clog the cavity of a hollow-point bullet, hindering or preventing bullet expansion, and affect tissue disruption. Careful examination of the offender’s clothing can reveal whether or not the bullet(s) passed through multiple layers of fabric, such as through the seam of a shirt pocket flap, a rolled-up shirtsleeve, multiple layers of clothing and/or other soft barrier materials.

More often than not, laminated windshield glass severely damages a jacketed hollow-point handgun bullet, turning it into a hunk of misshapen metal, which prevents it from expanding as designed when it hits and penetrates a human target.

Automotive sheet metal can also interfere with a bullet’s performance. A hollow-point bullet that encounters sheet metal can plug with discs of metal punched out by the rim of the meplat as it passes through, in a manner similar to the way dough becomes caught in a cookie cutter. Internal door structures and seat frames can also damage the bullet before it reaches its mark, and prevent it from performing as designed.

Other obstacles are also important to consider. A bullet may strike and pass through soft tissues and bone in the offender’s hand, wrist or arm before it hits and penetrates his torso, which, if the bullet expands or becomes deformed, concomitantly decreases its sectional density, decreases its torso penetration velocity, decreases its ability to disrupt tissues in the torso, and decreases bullet penetration of the torso, all of which can affect his reaction. Additional examples of obstacles include, but are not limited to: a notebook in a breast pocket; a pack of cigarettes, the strap of a backpack, articles of clothing made from leather. Put an “X” in the Other Obstacle(s) column in Table 1 and identify the item(s) on a separate worksheet.3

Additionally, a bullet that hits an obstacle at an angle and passes through can cause the bullet to yaw and hit/penetrate an offender's body at an angle, affecting both terminal performance and offender reaction to being hit.

Ever wonder why a bullet recovered from a body sometimes doesn’t resemble a bullet recovered from a block of ordnance gelatin? If it appears severely distorted, deformed and gouged, as if it collided with something hard, it probably did. The question that must be answered is, did the damage occur before or after the bullet hit and penetrated the offender’s body? A bullet that initially hits an area where bone is directly beneath skin (hand, wrist, arm, shoulder, skull and spinal bones) can be damaged by the collision with the hard tissue and not expand.

Wound Trauma

Finally, what anatomy a penetrating bullet damages and disrupts is important to consider. If a bullet comes into direct contact with the spinal cord, spinal bones or brain, the offender usually has no choice in what his reaction will be, and collapse or incapacitation is almost always instantaneous.

If a bullet passes through a major blood distribution organ or a large blood-processing organ, rapidity of physiological incapacitation will be dependent on the rate and volume of blood loss, as well as the offender's individual tolerance for blood loss.

A bullet that hits an arm or leg only may be enough to force an offender to decide to voluntarily give-up. Hits to the head and neck that don’t produce damage to, or disruption of, central nervous system (CNS) structures or major blood vessels should be analyzed also. Rapid incapacitation resulting from an extremity wound(s) only is unequivocally psychological in nature, but it is important to keep track of such events because they might indicate how often psychological incapacitation occurs in the real world.

The factors listed in Table 1 encompass most situations that are likely to be encountered in a police shooting. If desired, additional data about projectile penetration depth and expanded diameter can be reported on a separate worksheet.3

In-depth Analysis Yields Big-picture Perspective 

The key performance indicator of this study is the offender’s immediate reaction to being shot. However data about random variables commonly encountered in shooting incidents must be processed and analyzed to provide meaning and value to the findings.

The example data in Table 1 are sorted into subsets based on offender gunshot wound reaction, which can be further sorted based on any one or combination of variables. Each row reports a single shooting incident and the factors involved. As shooting incidents are added, and data accumulate, a pattern of random variables may begin to emerge that builds a picture of common factors that directly affect an offender’s reaction to being shot.

For example, the data may show that offenders who are incapacitated by a single gunshot wound represent mostly best case conditions, in which an offender is sober, rational, wearing light clothing, is mentally unprepared to be shot and seriously injured, and the cartridge is fired from a full-size handgun. Shootings falling into this category could be represented by offenders who are psychologically predisposed to act as they did or the offenders had no choice in their reaction because of what organs the bullet damaged.

Additionally, the data may also show a trend that as the challenge increases more hits are required to force an offender to stop. For example, an offender is under the influence of drugs and/or alcohol, is highly agitated and aggressive, appears irrational and oblivious to the personal danger of being shot and seriously injured when a gun is pointed at him, is wearing heavy clothing, and the cartridge is fired from a short barrel (less than 4 inches) handgun. The limited example data in Table 1 have been fabricated to reflect the patterns discussed herein.

Provided uniform criteria and definitions are used, this descriptive statistics study can be used to facilitate interagency comparisons of the effects of commonly encountered random variables on offender gunshot wound reaction. Standardization can be implemented at the federal level (FBI Uniform Crime Reports), through state justice departments or state law enforcement firearms instructor associations. Be aware though, that just as the 2000 U.S. Presidential election produced 51 different results (among the states and the District of Columbia), a comparison of individual agency results will also probably vary considerably due to specific variables involved.

When formally shared with other law enforcement organizations, news media or presented as evidence in court, some elements of Table 1 should be modified. Incident column data should be expanded to report pertinent public record information such as date, location, the gunshot wound recipient's name and the shooter's name(s). These data permit independent review and validation by others.

If officers supply their own weapons, the Weapon(s) column should report exact weapon make and model. In addition, the Terminal Performance data column should also be expanded to include a column to report incidents in which the weapon involved had a barrel length of less than 4 inches, to identify situations in which velocity loss from a short barrel may have affected the bullet’s terminal performance. (Table 1 accomplishes this task in a different manner by listing the Glock 27, which has a barrel length of 3.5 inches.)


A review of statistical representations of data shows that an inferential statistics study based on a hypothesis of "...if I were able to collect enough actual shooting data, I might be able to predict with some degree of accuracy how a load might work in the future.", lacks statistical validity for two reasons.4 Firstly, because the study is comprised entirely of non-random samples of unknown populations. Secondly, because the study also misrepresents these non-random samples as a descriptive statistics analysis.

Likewise, a valid inferential statistics analysis -- using random samples of shooting incidents to estimate populations and subsets, and to estimate and weigh the application of random variables -- simply produces too much uncertainty to effectively discern and accurately quantify differences in stopping power.

The descriptive statistics analysis proposed in this paper is designed to permit interpretation of offender gunshot wound reaction by considering and evaluating common factors that can have an effect on wound trauma incapacitation effectiveness.

All material published in FirearmsTactical.com is copyrighted and cannot be reproduced without permission. Permission is hereby given to all law enforcement agencies to reproduce this article, Reality of the Street? A Practical Analysis of Offender Gunshot Wound Reaction for Law Enforcement, and Table 1, Example Analysis of Law Enforcement Agency Shooting Incidents, for any use associated with evaluating ammunition performance.

Interested agencies and individuals can obtain blank copies of Table 1, available in Microsoft Word 2000 format.  Two versions of Table 1 are available.  Click on the links below:

End Notes

  1. The methodology of any inferential statistics study must be reviewed and validated by more than one professional statistician to ensure it produces accurate results. 

  2. Historical analysis of previously issued ammunition, especially ammunition that developed a reputation for superior stopping power performance, permits performance comparison with current issue ammunition. 

  3. Table 1 merely presents a matrix-type list of commonly encountered random variables for analysis, in a concise format that facilitates ease of pattern recognition. Report any additional elaborating information on a separate worksheet (toxicology report data, specific obstacles encountered by a bullet, bullet penetration and expanded diameter, general hit location [shoulder, thoracic cavity, abdominal cavity and pelvic area], gunshot wound type [penetrating, tangential or grazing] and specific anatomy damaged by the penetrating bullet). The worksheet should serve as a reference document to Table 1.

  4. Marshall, Evan P.: "One Shot Stops - Latest Update!" Gun World 42(3), March 2001; p. 54

Shawn Dodson is a field engineer and former reserve police officer with the Bremerton (WA) police department. Click here to review his background.

Delivering you informative multimedia essays
about the "battlefield problem-solving" tactical aspects of armed self-defense.

Web Site Index and Navigation Center

Copyright © 1998 - 2001 Firearms Tactical Institute. All Rights Reserved.
Salus In Periculo, and logo are trademarks of Firearms Tactical Institute.