1 of 34 · What the refusal suggests

The Officers Said No

A researcher set out to ask Florida police about the tools that watch the rest of us.

Trei McMullen, D.G.S., is the author of a 2025 doctoral dissertation submitted to American Public University System for the degree of Doctor of Global Security. The study is about the technologies law enforcement adopts and what they do to the privacy and security of citizens, with Florida as its case.

The officers he wanted to hear from were current and former officers in Florida. The original plan was to interview them in person.

Read the official record of the dissertation at ProQuest

Printed pages of the dissertation: pp. 1, 11, 28, 63, 121

2 of 34 · What the refusal suggests

The original plan

Sit down with Florida officers, face to face, and hear in their own words how they use the technologies the study is about.

The study was initially designed to include in-person interviews.

Printed pages of the dissertation: pp. 11, 28, 63, 121

3 of 34 · What the refusal suggests

They declined

Not all of them. But enough of them that the interviews could not go forward.

The dissertation records the difficulty in its own limitations section: participation rates among law enforcement personnel were low, and it calls this one of the primary limitations of the research.

Printed pages of the dissertation: pp. 121–122

4 of 34 · What the refusal suggests

The reasons they gave

The dissertation records why officers were reluctant to take part.

  • Officers doubted that their identities could be protected.
  • They cited a concern that even major corporations struggle to safeguard sensitive information from cyber threats.
  • They feared potential professional consequences.

These are the concerns the study reports officers expressed. They are the reasons officers gave, not a finding that any particular protection had failed.

Printed pages of the dissertation: pp. 121–122

5 of 34 · What the refusal suggests

What replaced it

An anonymous online survey went ahead instead.

The research approach was modified to an online survey format. The dissertation says the change improved response rates, and that it limited the depth of qualitative insight that direct interviews could have produced.

To safeguard participants, the survey did not collect identifiable information such as names.

Printed pages of the dissertation: pp. 44, 51, 122

6 of 34 · What the refusal suggests

What the refusal suggests

Explainer commentary

The officers the study most needed to hear from did not trust that their own words could be kept private.

This reading rests on the reasons the study records at pp. 121–122. It is the explainer author's characterisation of what the refusal means, not a finding or a conclusion of the dissertation.

Printed pages of the dissertation: pp. 121–122

7 of 34 · What was actually asked

The question

How does unconscious bias affect the technologies law enforcement adopts, and how does that affect the security and privacy of citizens?

That is the problem the dissertation sets out to address, using Florida as a case study.

Explainer commentary

The dissertation poses four research questions. As printed, the survey it ended up with reaches only one of them directly.

That comparison is the explainer author's reading of the published research questions against the printed survey instrument at pp. 27–28. It is not the dissertation's own statement. The same point is set out in full later in this deck.

Printed pages of the dissertation: pp. 11, 27–28

8 of 34 · What was actually asked

The limit the study keeps

Unconscious bias is a pattern across institutions, not a charge against every officer.

The study takes its definition from a Harvard Business Review article: preconceived stereotypes an individual holds towards people of a particular community, shaped by that person's social world and their own experiences. The difficulty is in the word itself — the unconscious part is what makes it hard for anyone to recognise in themselves.

The dissertation says the problem does not lie with individuals alone. It says existing evidence demonstrates unconscious bias including among well-meaning officers, and that the problem lies with technology too. A tool can carry the bias when the person holding it does not.

Printed pages of the dissertation: pp. 11–14, 22, 85

9 of 34 · What the tools are

The tools you already know

Body-worn cameras, dashboard cameras, drones, and artificial-intelligence software.

The dissertation lists the common uses of that software as identifying car nameplates, predicting and analysing crime rates, and identifying people through facial recognition. It sorts police-adopted technologies into software, such as face-detection algorithms, and hardware, such as body cameras.

Printed pages of the dissertation: p. 14

10 of 34 · What the tools are

Two kinds of face matching

Verification confirms you are who you claim. Identification searches a crowd for a match. Policing uses the second.

In plain language

There are two kinds, and the difference matters more than it sounds. Verification compares your face against the one face it already expects, which is what happens when your phone unlocks. Identification takes a face from anywhere and searches it against a database of many faces. Policing uses the second kind.

The dissertation describes verification as one-to-one matching, widely used for cell phones and secure access, and identification as one-to-many matching, used in surveillance and investigative systems. It says the range of error in identification systems is significant, citing numerous wrongful arrests based on misidentification.

Printed pages of the dissertation: pp. 9, 29–30

11 of 34 · What the tools are

Scoring a person

COMPAS scores how likely a person is to commit another crime.

In plain language

A judge sees that score before deciding a sentence. Because the way the score is worked out is not published, a person can receive a sentence different from someone else's and never learn the reason for it.

The dissertation reports that judges do not use COMPAS alone but consider its reported scores, that people assessed at high risk of re-offending receive longer sentences, and that the workings of the algorithm are not publicly available. It cites reporting that COMPAS-assisted judgments lead to higher conviction rates and lengthier judgments for young people.

Printed pages of the dissertation: pp. 9, 11, 22

12 of 34 · What the tools are

Scoring a place

Risk Terrain Modeling maps where a neighbourhood's risk features converge, so an agency can focus its resources there.

In plain language

This one looks at the features of a place rather than at a person. It marks where those features cluster together in ways associated with crime, so that an agency can direct its attention there. It names locations, not individuals. The catch is simple: people live in those locations.

The dissertation describes the model as treating the physical environment of a city as a landscape of interwoven risks, and notes that the implications for citizens living in these high-risk areas cannot be overlooked.

Printed pages of the dissertation: pp. 9, 37, 38

13 of 34 · How bias gets in

The mechanism

Predictive policing systems often rely on historical crime data.

Police records are a record of policing. The dissertation says that data is itself shaped by biased practices, and describes what can follow from training a tool on it.

A risk the dissertation describes — not a result it measured
  1. Historical crime records, which the dissertation says are themselves shaped by biased practices.
  2. A system trained on those records directs more police presence to the same neighbourhoods.
  3. More records are generated in those same neighbourhoods, and the cycle returns to step one.

The dissertation sets this out at p. 84 as a feedback loop that exacerbates disparities. It is a risk the dissertation describes, not a result this study measured.

In plain language

Police records are a record of where police have already been. The dissertation describes what can follow from that. If a program learns from those records, it may send officers back to the same streets, which produces more records from those same streets. The study presents this as a risk it describes, not as a result it measured.

Printed pages of the dissertation: pp. 24, 84

14 of 34 · How bias gets in

Four kinds of bias

The study names four: deployment bias, algorithmic bias, surveillance bias, and institutional bias.

  • Deployment bias — when predictive policing tools and similar technologies disproportionately allocate resources to certain communities (p. 84).
  • Algorithmic bias — the replication of historical biases within automated systems and AI tools (p. 85).
  • Surveillance bias — when surveillance technologies disproportionately target minority populations (p. 86).
  • Institutional bias — systemic discrimination embedded in organisational policies and practices (p. 86).

Cameras and drones raise the question of who they are aimed at. The dissertation raises a separate concern about who ends up holding what they record: it points to data gathered on citizens' private information by private companies contracted for surveillance, and says there is a need for clarity about how the government intends to safeguard it.

In plain language

The study sorts the problem into four kinds: where the tools get sent, what the software itself does, who the cameras are pointed at, and what an agency's own policy allows. You do not need to keep the labels. You do need the idea underneath them, which is that nobody has to intend any of it.

Printed pages of the dissertation: pp. 15, 81, 84–87

15 of 34 · Florida, concretely

Florida widens the law

In 2021, Florida Senate Bill 44 broadened the permissible use of drones by law enforcement and other governmental agencies.

  • Traffic management.
  • Evidence collection at crime scenes.
  • Search and rescue operations.

The dissertation reports the bill as an example of how quickly this technology has been integrated into daily practice.

Printed pages of the dissertation: p. 16

16 of 34 · Florida, concretely

Two voices, same drone

Two people quoted in the dissertation, disagreeing about the same tool.

“I would like to see us get to the point in Volusia County where every supervisor has a drone mounted on their car. The more information in the hands of the deputies, they can make better decisions.”

Volusia County Sheriff Mike Chitwood, quoted in the dissertation, p. 16

“we’ve learned from history that when you give police the ability to use technology in extenuating circumstances, those emergency circumstance parameters tend to grow and grow in authority and scope”

Matthew Guariglia, policy analyst at the Electronic Frontier Foundation, quoted in the dissertation, p. 19

Printed pages of the dissertation: pp. 16, 19

17 of 34 · Florida, concretely

Limits, and what went wrong

Florida then restricted which manufacturers’ drones government agencies could buy and use.

The law required the state to publish a list of approved drone manufacturers meeting strict security standards by January 2022, and required governmental agencies to phase out drones not produced by an approved manufacturer by January 2023. In April 2023 Florida banned drones manufactured by foreign entities, citing the potential for espionage and data breaches.

The dissertation also cites reports of what can go wrong. In Pasco County, a programme that flagged individuals as potential future offenders based on past interactions with the justice system drew complaints from residents who reported feeling harassed by excessive police visits. In Volusia County, drones praised for improving response times in emergencies were criticised for being used to monitor protests.

Explainer commentary

These are reports the study cites, not incidents it investigated. Its own method was a survey of volunteers and a reading of published statements.

Printed pages of the dissertation: pp. 16–17, 42–52, 84, 88

18 of 34 · How small this study is

The size, before any number

Two anonymous online surveys. Thirty people in total.

  • 20civiliansfifteen of them living in Florida, five elsewhere
  • 10officersall current or past Florida law enforcement officers
  • Participants were screened in by the survey’s first questions. Nobody was chosen at random.
  • People volunteered. The dissertation notes that groups less familiar with, or hesitant to engage with, these technologies might not be represented.
  • Everyone reported on themselves. The dissertation notes the possibility of response bias, and that participants may have given socially desirable answers.
  • The civilian sample was intentionally capped at twenty, which the dissertation calls appropriate for qualitative research and a limit on generalising the findings.

These are thirty people’s answers. They are not Florida’s numbers, and the dissertation does not present them as such.

Printed pages of the dissertation: pp. 44, 51, 52–53, 63, 75, 122

19 of 34 · What those people said

Both sides split

Asked about police technology and privacy, each group divided down the middle.

Civilians, asked whether they perceive the use of technology in law enforcement in terms of privacy and security to be too much

20 civilians answered

  • 10 Said yes
  • 10 Said no

pp. 28, 57, 77

Officers, asked how they perceive the use of technology in law enforcement in terms of privacy and security

10 officers answered

  • 5 Agreed it is too much
  • 5 Disagreed

pp. 28, 67, 79

The two groups were asked about the same subject in different words: the printed civilian item asks whether that use is ‘too much’, and the printed officer item asks how the officer perceives it. The dissertation reports the same even split for both groups.

Printed pages of the dissertation: pp. 28, 57, 67, 77, 79

20 of 34 · What those people said

Voice, measured two ways

Both groups were asked about having a voice. They were not asked the same question.

Civilians, asked whether having a voice in the laws that govern them would make them more likely to follow those laws

20 civilians answered

  • 14 Said yes
  • 6 Said no

pp. 28, 58

Officers, asked — as the item is printed — whether they feel citizens are more likely to follow laws when they feel heard

10 officers answered

  • 7 Said yes
  • 3 Said no

pp. 28, 68

A discrepancy inside the source. The officer item as printed at p. 28 asks about citizens. The results chapter at p. 68 describes that same item as if it asked whether having a voice in the laws governing them increases the officers’ own likelihood of following those laws. This deck reports both and does not reconcile them.

What was measured on both sides is a belief about voice. Neither item measured anyone’s behaviour.

Printed pages of the dissertation: pp. 28, 58, 68

21 of 34 · What those people said

What the officers reported

Four answers from the ten officers.

Had an experience where their own or a fellow officer’s technology, such as body cameras, failed

10 officers answered

  • 6 Said yes
  • 4 Said no

pp. 28, 65

Had an experience where an officer’s technology, such as body cameras, assisted in a situation

10 officers answered

  • 8 Said yes
  • 2 Said no

pp. 28, 66

Knew what happens to the data captured by the software and hardware officers use

10 officers answered

  • 5 Said yes
  • 5 Said no

pp. 28, 68

Training, reported without a denominator. The dissertation gives its training figures as percentages and does not say how many officers answered that item. Of those who did answer, it reports that 44 percent rated their training mediocre, 22 percent described it as a crash course, and 33 percent rated it good or great. Because the response count is unstated, those figures are stated here in words and are not charted (pp. 67–68).

Printed pages of the dissertation: pp. 28, 65, 66, 67–68

22 of 34 · What those people said

What civilians reported

Seven of the twenty civilians reported having been wrongly identified by police technology or an identification system.

Civilians, asked whether they had ever had an experience where they were wrongfully identified by any form of police technology or identification system

20 civilians answered

  • 7 Said yes
  • 13 Said no

pp. 28, 56

The printed survey item asked only whether the experience had happened. It did not ask what happened. The study does not establish what those seven experiences were.

Printed pages of the dissertation: pp. 28, 56

23 of 34 · What the executives said

What the executives said

A third strand: a reading of twenty published statements by law-enforcement executives, drawn from trade magazines and reports.

Each was coded for whether it acknowledged bias in these technologies, which kind of bias it named, whether it proposed solutions, and how it characterised the impact on public trust.

The study reports that most of the executives reviewed acknowledged the presence of bias, and that about a third of the reports provided no concrete solutions.

The figures used here are the ones in the study’s narrative at pp. 89–90. The study’s own summary table at pp. 82–83 does not agree with that narrative. This deck reports the narrative figures, as the study states them, and does not attempt to reconcile the two.

Printed pages of the dissertation: pp. 82–84, 89–90

24 of 34 · Where the study is thin

Where the study is thin

Better to say it plainly than to let a reader find it.

The dissertation’s four research questions include how agencies decide which emerging technologies to adopt, what methods detect and measure unconscious bias during recruitment, training and tenure, and what training programmes address it.

Explainer commentary

As far as the printed survey shows, those questions were not asked. The items reach only the second research question, and only through what people perceive.

The methods chapter describes the survey design as including both multiple-choice questions and open-ended prompts in one place, and describes the civilian survey as consisting exclusively of multiple-choice questions in another.

The dissertation says that moving to an online format limited the depth of qualitative insight that direct interviews could have produced, and that the small sample constrained its ability to generalise. It also reports that some participants viewed research on predictive policing as tied to disputes over diversity, equity and inclusion programmes, making them hesitant to engage.

Printed pages of the dissertation: pp. 27–28, 44, 51, 53–68, 121–123

25 of 34 · Where the study is thin

Take the pattern, with its limits

Explainer commentary

Do not take the percentages to the bank. Take the pattern.

Explainer commentary
  • On privacy and on voice, the two groups gave similar answers, allowing for the different wording of the questions put to them.
  • Officers reported that the tools fail.
  • Training was thin for most of the officers who answered that item, and the study does not say how many answered.
  • Five of the ten officers said they did not know what happens to the data these tools capture.

This is the explainer author’s summary of the results on the preceding slides. Its factual premises are those results, at the pages listed below.

Printed pages of the dissertation: pp. 28, 57–58, 65, 67–68

26 of 34 · The idea underneath

Why anyone adopts a tool

The study’s first frame is the Technology Acceptance Model 2.

The dissertation describes it as a framework for the micro-level factors in technology adoption, emphasising perceived usefulness, ease of use, subjective norms and image. Its concern is that the urgency agencies feel to remain technologically current can result in hastened implementation, to meet internal expectations or to match perceived advances at a counterpart agency.

In plain language

People take up a tool when it looks useful, when it seems easy, and when the people around them expect it. An agency can also adopt one in order to look current. The study’s concern is that an agency moving in a hurry may never ask what the tool does to the people it is pointed at.

Printed pages of the dissertation: pp. 10, 25–26

27 of 34 · The idea underneath

Why anyone follows a rule

The study’s second frame is Procedural Justice Theory.

The dissertation describes it as built on four principles — voice, neutrality, respect and trustworthiness — and cites Tyler’s argument that when people feel they have participated in establishing the processes and rules that affect them, their compliance and support for those processes increase.

In plain language

People tend to follow rules they believe are fair. The study uses a theory that says four things make a rule feel fair: having a voice in it, an even-handed process, being treated with respect, and being able to trust the people in charge.

This is the theory the study reasons with. It is not a compliance effect that this study’s survey measured.

Printed pages of the dissertation: pp. 9, 26, 110

28 of 34 · The idea underneath

The refusal beside the theory

The officers who held back gave the reason the dissertation records: they were sceptical that their identities could be protected and that sensitive information could be kept safe.

The study’s own theory says individuals are more likely to comply with laws and support law enforcement when they perceive the processes as fair and inclusive, and when they have had a hand in creating the rules.

Explainer commentary

Reading the officers’ refusal through the study’s own theory is the explainer author’s interpretation. The study did not test it.

Separately, seven of the ten officers who did answer the survey agreed with the printed item that citizens are more likely to follow laws when they feel heard.

Explainer commentary

Side by side: the officers who held back wanted their own information protected, and the officers who answered believed people cooperate when they have a voice. In this small study, that is what both sides said.

In plain language

The officers who held back said they did not trust that their information would be kept safe. The study’s own theory says people cooperate with a process they trust and had a voice in. Setting those two things beside each other is the explainer author’s reading. The study did not test it.

Printed pages of the dissertation: pp. 21, 26, 28, 68, 110, 121–122

29 of 34 · What you can do

What the study recommends

The dissertation closes with recommendations. These are the study’s own.

  • Routine independent audits of predictive tools, involving independent oversight bodies.
  • Public transparency reports detailing the deployment, outcomes and limitations of these tools.
  • Open-source or independently audited algorithms for tools such as COMPAS.
  • Training that covers ethical considerations and data security, not only technical proficiency, with consistent maintenance protocols.
  • Civilian advisory boards and community forums with a real part in technology policy.
  • Feedback gathered from officers and from residents.
  • Publicly accessible policies on data retention and usage, limits on retention periods, and robust cybersecurity protocols.
Explainer commentary

The explainer author endorses these recommendations. The recommendations are the study’s; the endorsement is not.

Printed pages of the dissertation: pp. 57, 61–62, 72–74, 106–107, 113–119

30 of 34 · What you can do

If you are a citizen

Explainer commentary

Ask what happens to the data your city’s police tools collect. When your city or county holds a meeting about a new camera system, a new drone programme, or a new piece of software, go to it. Help make the rules.

The advice above is the explainer author’s. Its factual premises are:
  • Five of the ten officers in this study said they did not know what happens to the data captured by the tools they use (pp. 28, 68).
  • Fourteen of twenty civilians said having a voice in the laws that govern them would make them more likely to follow those laws; seven of ten officers agreed with the printed item that citizens follow laws more readily when they feel heard. The two questions are worded differently (pp. 28, 58, 68).

Printed pages of the dissertation: pp. 28, 58, 68

31 of 34 · What you can do

If you are an officer

Explainer commentary

If the training you got on these tools was a crash course, ask for more. Ask what happens to the footage your equipment records.

The advice above is the explainer author’s. Its factual premises are:
  • Of the officers who answered the training item, the study reports that 22 percent described their training as a crash course and 44 percent rated it mediocre. The study does not say how many officers answered (pp. 67–68).
  • The dissertation says concerns about how surveillance footage is stored, accessed and reviewed may contribute to hesitation among some officers (p. 66).

Printed pages of the dissertation: pp. 66, 67–68

32 of 34 · What you can do

If you lead an agency or council

Explainer commentary

Publish the policy before you buy the tool, not after. Put someone from outside the department on the body that reviews it. Get the audit, and read it.

The advice above is the explainer author’s. Its factual premises are:
  • The dissertation says agencies’ eagerness to adopt technology often overlooks critical components of implementation, and cites an investigation in which a department folded a new device into an existing policy, bypassing the public comment period that a required impact and use policy would have triggered (pp. 23–24).
  • It recommends independent oversight committees including community members, legal experts and data scientists (pp. 113–114), and routine audits involving independent oversight bodies (pp. 61, 73–74).

Printed pages of the dissertation: pp. 23–24, 61, 73–74, 113–114

33 of 34 · What you can do

Back to the beginning

The officers held back.

The dissertation cites research finding that persistent surveillance can, over time, diminish the likelihood that residents will report crimes or cooperate with law enforcement. Elsewhere it reports that individuals residing in heavily surveilled neighbourhoods reported increased distrust toward law enforcement.

Explainer commentary

Whether the officers’ reluctance and that reported reluctance are the same reflex is the explainer author’s reading, not a finding of the study.

The study cites a count that, as of October 2017, over 300 US agencies had adopted drones. A Florida sheriff said he would like every supervisor in his county to have one mounted on their car. And the dissertation cites the assertion that the adoption of these technologies happens more quickly than public discourse about them.

That is where the study leaves it, and where the reader comes in.

Read the official record of the dissertation at ProQuest

Printed pages of the dissertation: pp. 16, 32, 35–36, 39, 106, 121–122

Explore the research · The method behind the findings

Hear the perspectives.
Keep the limits in view.

A qualitative case study of technology, bias and policing in Florida.

  1. A focused question.

    Explore how unconscious bias may shape the adoption and use of policing technology, with security and privacy at stake.

  2. People who chose to answer.

    Anonymous surveys gathered civilian and officer perspectives. Eligibility screening shaped a voluntary sample.

  3. More than one perspective.

    The study compared the two survey groups and separately analyzed published executive statements about bias.

  4. Limits stay visible.

    Small samples, self-reported answers and difficulty recruiting officers constrain what the findings can establish.

Source: dissertation, pp. 42–52, 63, 75, 81–84, 121–123. Survey respondents and published statements are separate evidence sources.

Explore the research · Method and source

How the study was done

The dissertation brings together civilian surveys, officer surveys and published executive statements to explore technology and bias in Florida policing. How those sources were collected matters to how we read the findings.

← Back to the method overview

01 · The design

A Florida case, examined through people’s perspectives.

The dissertation describes a qualitative case study: an examination of how unconscious bias may affect the adoption and use of emerging law-enforcement technologies, and what that could mean for citizens’ security and privacy. Florida supplies the setting. Although the research sought perspectives from different parts of the state, it did not set out to compare geographic regions.

The design combines surveys and thematic analysis. Chapter Six also uses descriptive statistics to compare the civilian and officer responses. Counts show how these respondents answered; the sampling approach limits how far those patterns can be generalized.

Source: pp. 42–47, 52, 75.

02 · Recruitment and participation

A voluntary sample with eligibility checks.

The study used purposive sampling: it sought people whose roles or experiences were relevant to the research. The survey link was publicly available, with opening questions that screened for eligibility. The civilian recruitment description includes distribution through social media and excludes people directly affiliated with law enforcement. Participation was voluntary, rather than randomly assigned.

The civilian and officer surveys were administered through Qualtrics. The dissertation says the survey did not collect names or other identifying information. Its results chapters report the groups below; the executive material was a separate content analysis of published reports, speeches and interviews.

  • 20civilian respondents15 living in Florida; 5 elsewhere
  • 10officer respondentsAll current or past Florida law enforcement officers
  • 20executive statementsPublished material analyzed separately

Source: pp. 44, 46, 49–53, 63, 75, 82–84.

03 · The change in approach

Interviews were planned. Surveys became the way in.

The research initially called for in-person interviews. The dissertation reports that officers were reluctant to take part because of confidentiality and data-security concerns, including fears of professional consequences if their identities were exposed. The researcher shifted to an online survey format.

According to the dissertation, this improved participation but reduced the depth of insight that direct interviews might have provided. The format change is part of the study’s limitations, as well as part of its story.

Source: pp. 121–122.

04 · The analysis

From responses to codes, then themes.

Chapter Three describes a three-stage coding process. Coding means assigning labels to material so that patterns can be examined. The stages below summarize the process the dissertation reports.

  1. Initial coding

    Read civilian responses for emerging patterns, then compare them with officer responses to identify themes around the use of technology.

  2. Focused coding

    Narrow attention to material concerning unconscious bias and its possible influence on decisions about adopting and using technology.

  3. Thematic analysis

    Bring the codes together into broader themes, connect those themes to the research questions, and interpret the community implications.

Chapter Six compares the two survey groups using descriptive statistics and comparative measures. Chapter Seven separately examines published executive material, organizing it around acknowledgment of bias, types of bias, proposed solutions, public trust and perceptions of technology.

A qualification in the source. The methods account is not fully consistent about the survey format. Page 44 describes multiple-choice questions and open-ended prompts; page 51 describes the civilian survey as exclusively multiple choice. This summary preserves that difference. The coding stages above describe the dissertation’s stated method; they do not establish that every reported finding came from an open-ended answer.

Source: pp. 44–46, 51, 75, 81–84.

05 · The theoretical frames

Technology acceptance and public legitimacy.

Technology Acceptance Model 2

TAM2 helps frame why officers might accept a tool. The dissertation discusses perceived usefulness alongside social expectations, voluntariness and professional image. These factors help organize its account of technology adoption.

Procedural Justice Theory

This frame concerns community trust, citizen rights and participation in decisions. The dissertation uses it to connect choices about technology with the public’s view of whether those choices are legitimate.

Source: pp. 25–26, 43.

06 · Reading the limits

Who answered shapes what can be concluded.

The civilian sample was capped at twenty. Voluntary, purposive recruitment may leave out people who are unfamiliar with the technologies or hesitant to discuss them. Self-reported answers can also reflect what seems socially acceptable, or a person’s limited awareness of their own biases.

The dissertation acknowledges that its small sample and difficulty recruiting officers restrict generalization across agencies. It also identifies the loss of interview depth, rapidly changing technologies, limited literature, and some participants’ reluctance to engage because they associated predictive-policing research with disputes over diversity, equity and inclusion programs.

These limits matter when reading both agreement and disagreement between the groups. The dissertation calls for larger, more diverse samples and approaches that make participation feel more secure.

Source: pp. 52, 121–123.

The source

McMullen, Trei. 2025. Unconscious Bias on the Implementation and Utilization of Emerging Technologies by Law Enforcement Agencies, and Effects on the Security and Privacy of Citizens in Florida: A Case Study of Florida. Doctoral dissertation, American Public University System. ProQuest number 31848616.

Official public record at ProQuest

This presentation does not reproduce or serve the dissertation, its survey instrument, or its respondent tables. The official record above is the place to read the work itself.

About this presentation

This is attributed commentary on Dr. Trei McMullen’s research. The research, its design and its findings are his. The dissertation’s author was not involved in producing this presentation.

The passages marked Explainer commentary throughout the deck, and the article in the reader, are the work of Scott W. Waddell, D.S.I., who decided what this presentation says and is responsible for it. The article’s own byline and identifier appear inside the reader.

Every factual slide carries the printed pages of the dissertation that support it. The deck was built only from the accepted explainer article and the dissertation itself; nothing was added from any other source, and no statement here is stronger than the one the dissertation makes.

Workspace evidence for this section, rather than dissertation pages: the project brief §8 and the commentary scope and byline decisions recorded in the project’s decision register.

The Officers Said No · A 2025 doctoral study