Force Science Forum


A medium for the delivery of the research conducted or supported by the Force Science Research Center, Minnesota State University, Mankato.


Selective attention, commonly referred to in law enforcement as tunnel vision and tunnel hearing, plays a very significant role in an officer’s perception, performance, and memory in a high stress encounter. An aspect of this phenomenon that the Force Science Research Center (FSRC) at Minnesota State University, Mankato, is interested in researching is the officer’s attentional responses and the impact of that on the ability of an officer to effectively multi-task—particularly in a life and death encounter. Clinical investigation has informed us that the emotional response of an officer has a high degree of relevance on the officer’s attention and then on the ability of an officer to both engage in life-saving behavior and simultaneously give meaningful and relevant commands in an attempt to control a threatening subject. The observations have also led us to hypothesize that the more an officer perceives that he or she has control of a situation, the more he or she is capable of giving relevant, meaningful commands. The less control he or she perceives that he or she has over a situation and the more threatening the situation is, the less relevant and meaningful the officer’s commands are as his or her attention becomes focused on the need to engage in life-saving action to stop the threat. This article is the first in a series that FSRC will present as this phenomenon is explored and the most effective types of responses and the most effective commands for officers in high stress, life-threatening encounters are sought.


This article is the first in a series that FSRC will present as this phenomenon is explored and the most effective types of responses and the most effective commands for officers in high stress, life-threatening encounters are sought.


Command Types Used in Police Encounters



Police officers regularly encounter situations in which escalating emotion can lead to hostility or violence between law enforcement and potential suspects. These situations can quickly become dangerous for both the individuals involved as well as bystanders. In order to increase public safety and officer effectiveness, it is necessary to investigate and minimize the factors that contribute to antagonistic encounters between law enforcement and potential suspects. Certain communication styles and tactics may be more effective than others in eliciting immediate compliance and decreasing violence (Thompson, 1983). Likewise, a failure to communicate clearly and concisely has been shown in certain circumstances to exacerbate negative interactions (Forehand & McMahon, 1981). Although police behavior has been researched, an investigation of specific commands used by police has not yet been conducted (Bayley & Bittner, 1984; Bayley & Garofalo, 1989; Johnson, 2004; Reiss, 1971; Thompson, 1983).


In addition to command types, Johnson (2004) has noted that certain communication styles may be more appeasing to potential suspects than others. Johnson also states that citizens who felt treated fairly were more likely to comply with the law. Factors such as respect, tone, and demeanor of an officer’s verbal communication are important factors in reducing violence and increasing compliance (see also Reiss, 1971).


In studies evaluating compliance, researchers have consistently found that the clarity or feasibility of commands are important elements in eliciting compliance to demands (Bertsch, 1999). After a literature review of the studies looking at these components, Bertsch categorized commands into eight types. She further broke each type into alpha and beta command subtypes in order to distinguish between specific, feasible commands (i.e., alpha), and vague, unfeasible commands (i.e., beta) consistent with Peed, Roberts, and Forehand (1977). Bertsch studied the effects of these 16 command types and subtypes within the context of student compliance in a classroom setting. The command types included interrogation, question, regular, indirect, stop, don’t, negative, and other. This identification of command types has not yet extended into the law enforcement literature, which is surprising given the abundance of police forces in various countries and cultures, and the importance placed on compliance with police commands. Bertsch’s review and subsequent study showed clear and striking benefits to using concise and specific alpha commands. There has not been a study in the compliance literature to date which has shown any advantage to using nonspecific beta commands in an effort to elicit compliance.


Thompson (1983) investigated communication styles used by police officers. Thompson’s communication research focused on the use of an impartial perspective, which was achieved through evaluation of the facts (who, what, when, where, how) and an evaluation of the purpose of each encounter. Thompson also focused on using language appropriate to each individual and to appeal to the emotions, reasoning, and character of potential suspects. Johnson’s (2004) study showed significant public support for specific and general aspects of Thompson’s verbal judo during traffic stops. However, participant input was not obtained for additional and specific verbal commands. Additionally, Bayley and Bittner (1984) and Bayley and Garofalo (1989) described and evaluated police behavior, which involved both physical and verbal behavior. Specific command types were not investigated, however.


Mastrofski, Snipes, and Supina (1996) investigated a number of components related to compliance in officer-citizen interactions. No major differences in compliance rates were noted by Mastrofski et al. among requests issued by officers for citizens to leave another citizen alone, calm down and stop the disorder, and cease illegal behavior. They also studied the use of authoritativeness or force in both police entry and requests. Police entry refers to the style of the initial verbal approach with the suspect, including friendly/nonthreatening interrogation, command/ threat, and force categories. Experimenters found only the force entry tactic to be significantly different in eliciting compliance, actually producing less compliance. Investigators also found that officers exhibiting the most authoritative entry tactics were least likely to gain compliance. Additionally, Mastrofski et al. found no significant differences in compliance among the request categories, including suggestions and requests, persuasion and negotiation, and commands and threats. It was noted, however, that greater police experience was associated with a greater likelihood of making commands and threats and a reduced likelihood of making suggestions and requests.


One element that was investigated in Mastrofski et al.’s (1996) study was defined as coercive balance of power, which included elements such as the number of officers present, the use of a weapon, and the sex of the officer. The presence of male officers and higher numbers of officers were less likely to lead to compliance, although only the number of officers was significant. These findings are counterintuitive, and they make a clear case for further investigation.


Mastrofski et al. (1996) also evaluated the type of problem behavior categorized as traffic, minor offense, drugs, and serious. The more serious the offenses in this study, the lower the likelihood of compliance. Researchers also found race to be a factor in compliance with officer requests. Results indicated that White officers were more likely to produce compliance with minority citizens and minority officers were least likely to elicit compliance with White citizens. Additional results suggested that males were more likely to comply than females.


Mastrofski et al.’s (1996) research only included commands which were unambiguous, excluding indirect and beta commands altogether. Current literature lacks research on the use of specific command types based on form and feasibility. No information is yet available on how command type relates to violence, compliance, or latency of response.


The current study will expand upon the research of Mastrofski et al. (1996) by evaluating differences in Bertsch’s command types across suspect compliance, latency, violence, and type of crime in law enforcement/suspect exchanges. Due to the negative connotation associated with interrogation in law enforcement, for the purpose of this study, the interrogation command type will be re-termed interview.




Data Collection


Data from police interactions were accessed via prerecorded videos, direct observation, or live video recordings on ride-alongs with law enforcement.


Data collected while riding with law enforcement involved two different police departments and nine different officers over the course of 11 rides. A total of four observers participated in ride-alongs during the busiest shifts, between Thursday and Saturday evenings anytime from 3:00 PM to 4:00 AM. Riders observed the law enforcement interactions in person and through dash-mounted cameras. In addition, observers viewed six different recordings of police interactions. These included two dash camera videos, a Hard Copy video, a World’s Wildest Police Video, and two COPS videos. Officer commands were recorded as one of eight command types and one of two subtypes. Suspect compliance and latency were also recorded. Additional officer and department information along with circumstantial information were recorded as well. This included the type of crime, possession of a weapon, use of officer force, and the level of violence.



Independent and Dependent Variables


Independent variables included command type and type of crime. A command was defined as any verbal communication directed by law enforcement to non- emergency personnel in which a verbal or motor response was appropriate. Commands were divided into eight types and further classified into two subtypes. See Tables 1 and 2 for definitions and for examples of the eight command types and two subtypes.


IBoth the interview and question commands are phrased as a question. These commands are distinguished by the response, however. A verbal response would be appropriate for the interview command whereas it would be possible but inappropriate for the question command. The question command requires a motor response. For example, “What is your name?” requires a verbal response and is an interview command. A motor response is most appropriate to commands such as “Could you please sit down.”


The next two command types, regular and indirect, can often be confused. The regular command type is defined as an order that is stated directly. The indirect command type is a suggestion (allowing for nonresponse) to respond motorically or verbally and is not in question form. The indirect command does not state a specific command, but it is classified as a command because a specific response is desired by the issuing individual.


The next three commands—(1) don’t, (2) stop, and(3) negative—were combined to form an exclusionary command category. All of these commands are a request to terminate an ongoing behavior, and the don’t and stop commands can also be used to avert a future behavior. The differences between these commands lies in the use of the words “don’t” and “stop.” Don’t commands are defined as instructions to terminate an ongoing behavior or a future behavior generally proceeded with the word “don’t.” Stop commands consist of instruction to terminate an ongoing behavior generally proceeded by the word “stop.” Alternatively, negative commands are defined as instructions to terminate an ongoing behavior, which do not begin with the words “don’t” or “stop.”

The final command type, other, is defined as any command that cannot fit into only one of the above categories or a command that may fit in two or more of the categories at the same time. This command type is most often used when a command fits into more than one of the command types such as “Why don’t you stop it.”


These eight command types are further divided into two subtypes: (1) alpha and (2) beta. An alpha command is defined as a command in which a motoric or verbal response is appropriate and feasible. Contrary to this, a beta command is defined as a command in which compliance may be difficult due to vagueness, interruption, or indirectiveness.


Observers could also indicate up to two categories of crimes for each encounter. These crime categories included assault, threat, suicide, narcotics, theft, burglary, arrest warrant, disorderly conduct, and other. Due to limitations of the sample size, only the primary crime was used, and suicide, theft, and arrest warrant were combined.


Assault was defined as physical violence such as domestic assaults, bar fights, sexual assaults, etc. Threat included “terrorist,” verbal, or physical threats. Suicide was defined as a call when someone had committed suicide or was threatening to do so. Narcotics crimes included calls related to possessing, selling, making, or intending to sell drugs or drug paraphernalia or being under the influence of narcotics. Theft included taking property or merchandise in which there was no break-in and entry and no physical harm to others. This could occur during stealing and shoplifting calls if there was no break-in or harm to others. Burglary was defined as break-in and entry or physical harm to others while stealing.


Arrest warrant was categorized as a police call in which police were attempting to arrest someone because of a court order to do so. Arrest warrant calls did not include calls during which the officer decided to arrest an individual because of the circumstances of the situation rather than because of a court order. Disorderly conduct included any disturbance to others such as public indecency, peeping toms, public intoxication, disturbing the peace, etc. The other category included any calls that didn’t easily fit into any of these categories.


Dependent variables included level of compliance, latency of response, and level of violence. Level of compliance was divided into compliance, forced compliance, and noncompliance: Compliance was defined as an individual responding appropriately to an officer’s command by means of free will prior to another command by the officer; Forced compliance occurred when an individual responded appropriately to an officer’s command as the direct result of the officer using physical restraint, a Taser®, or shooting a gun; and noncompliance occurred when an individual did not respond appropriately to an officer’s command by free will or to the officer’s use of physical restraint, a Taser®, or firing of a gun. Latency of response was further divided into three levels: (1) immediate if compliance occurred within approximately 10 seconds, (2) delayed if between 10 and 30 seconds, or (3) none if more than 30 seconds.


A violent encounter was defined as an encounter between an officer and potential suspect in which the individual posed a threat to the officer in the form of a weapon, extreme agitation, substance intoxication, or physical force. Nonviolent was defined as an encounter between an officer and potential suspect in which the individual appeared to pose no threat to the officer.


Training and Interobserver Agreement


All observers received training on the operational definitions and use of coding sheets. Observers were allowed to practice independently on several videos. A reliability check was conducted on one of the five videos, which contained 22.9% of the video commands. Interobserver agreement was 93%. Prior to conducting ride-alongs, each observer was able to reach 100% agreement on the commands given in a video.



The primary investigator viewed and coded all six videos. For each new law enforcement encounter, the department and jurisdiction, the officer rank, and use of force were indicated if known. Additionally, the number of individuals giving commands, weapons possessed, the violence of the encounter, and the type of crime committed was recorded. Each command type, the level of compliance obtained, and the latency of any compliant response was noted for each police- suspect interaction.




Command Type and Subtype


Of the 1,801 commands given, a large portion were interview, n = 938, 52.1%; followed by regular, n = 563, 31.3%; indirect, n = 141, 7.8%; other, n = 99, 5.5%; question, n = 45, 2.5%; and the exclusionary commands, n = 15, .8%. Exclusionary commands were mostly stop commands, n = 7; followed by negative, n = 6; and don’t, n = 2. The majority of the command subtypes were alpha, with a total of 1,488 commands or 82.6%, whereas the total number of beta commands, n = 313, were less than 18%. See Figures 1 and 2 for frequencies of commands.





Results indicated that the level of compliance was significantly different across the six command types: x2(10) = 368.66, p < 0.001. The percentage of compliance per command was greatest for interview commands, 91%; followed by other, 85%; question, 78%; indirect, 68%; exclusionary, 50%; and regular, 48% (see Figure 3).


Results further indicate that there were significant differences in compliance and noncompliance across alpha and beta commands: x2(2) = 231.059, p < 0.001. The greatest percentage of compliance per command was found in alpha command subtypes: 82%, compared to beta commands, 41%. The percentage of alpha command subtypes that produced noncompliance was only 18% compared to 57% of the beta commands (see Figure 4).




Results indicated a significant difference in latency of compliance across the six command types: x2(10) = 54.604, p < 0.001. The exclusionary commands produced the smallest percentage of immediate compliance per command, with only 55%; followed by question commands, 87%; and regular, 89%. All other command types produced at least 90% immediate compliance. Exclusionary commands were also found to produce the highest percentage of delayed responses at 27%. All other command types produced 5% or less in delayed responses (see Figure 5).


Results further indicated that there were significant differences in compliance and noncompliance across alpha and beta commands: x2(2) = 14.02, p < 0.001. Alpha command types produced immediate compliance 94% of the time and delayed compliance 2% of the time. In comparison, beta commands produced immediate compliance 85% of the time and delayed compliance 5% of the time (see Figure 6).


Violent Versus Nonviolent


Of the 1,801 commands observed, 352 commands corresponded to a violent encounter and 1,334 to a nonviolent encounter. Results also indicate differences in the use of the six command types in violent versus nonviolent encounters: 2(5) = 213.398, p < 0.001. The greatest percentage of commands used in violent encounters was regular commands, 41%; followed by question, 35%; indirect, 31%; other, 14%; and interview and exclusionary, 9%. The greatest percentage of commands used in nonviolent encounters was exclusionary and interview, 91%; followed by other, 86%; indirect, 69%; question, 65%; and regular, 59% (see Figure 7).


Results further indicate that there were significant differences in the use of alpha and beta commands in violent and nonviolent encounters: X2(1) = 145.179, p < 0.001. The percentage of alpha commands used in violent encounters was 16% compared with 84% used in nonviolent encounters. Beta commands occurred 49.5% of the time in violent encounters and 50.5% of the time in nonviolent encounters (see Figure 8).