The problem is no longer whether warning signs exist. The problem is whether departments will adopt a noncompensatory, tiered, evidence-based screening architecture that forces those warning signs to matter.
Core Thesis
The next phase of police hiring reform cannot stop at proving that the current model fails. It must replace that model with a new architecture. The evidence already discussed shows that agencies possess predictive warning signs before appointment, yet use them weakly, inconsistently, and too often through subjective, late-stage review. The solution is therefore structural: police hiring must move from a discretionary, clinician-mediated suitability model to a noncompensatory, behavior-specific screening framework that begins earlier, uses validated indicators, separates common screening signals from severe red flags, constrains decision-makers through structured rules, and reserves clinical review for a bounded post-offer role rather than treating it as the system’s primary predictive engine. Pages 11 through 13 of the current deck frame that transition directly: from “current flawed architecture” to “proposed objective architecture,” from “generalized impression” to “validated prediction,” and from decentralized discretion to criterion-linked, actuarial decision rules.
The reason this follow-up matters is that reform often collapses after diagnosis. Institutions acknowledge that the existing process is weak, subjective, and too permissive, yet preserve its basic logic by adding more review, more language of professionalism, or more procedural complexity. That is not reform. That is continuity under new branding. A serious solution requires the opposite. It requires moving prediction upstream through standardized background questionnaires and empirical indicator screening; differentiating applicant risk through a tiered framework that separates population-level indicators from critical red flags; reducing discretionary latitude through structured cutoffs and auditable reasoning; and building a defensible record showing why a candidate was advanced or screened out. The point is not to eliminate judgment. It is to bind judgment to evidence. The point is not to create a harsher version of the same model. It is to create a different one.
This follow-up therefore advances a direct institutional claim: the real reform question is no longer whether warning signs exist, but whether departments will adopt a hiring system designed to force those warning signs to matter. A profession that delegates coercive public authority cannot continue treating validated risk as one more topic for discussion inside an unstructured “whole person” narrative. It needs predictive gates, not generalized impressions. It needs noncompensatory rules for severe indicators, systematic treatment of high-prevalence predictors, and a transparent audit trail that makes hiring logic visible to the institution and defensible to the public. Anything less leaves the current failure intact: warning signs gathered, warning signs softened, and warning signs rediscovered later as complaint, lawsuit, scandal, or termination.
Executive Summary
The first thought-piece established the diagnosis. Police departments do not primarily suffer from an absence of warning signs. They suffer from the institutional misuse of them. Predictive prehire indicators exist, many significantly relate to later misconduct, and agencies have often used those indicators only minimally in hiring decisions. The natural next question is therefore not whether the current model is inadequate. It is what should replace it. This follow-up answers that question directly. Its focus is the proposed solution: a transition from discretionary, subjective, clinician-mediated hiring to an objective, tiered, and behavior-specific screening system built around predictive discipline. The deck accompanying the original piece already sketches that solution in concrete terms. On page 11, it contrasts the “current flawed architecture,” which over-relies on a late, expensive clinical psychological evaluation, with a “proposed objective architecture” built from standardized background questionnaires, empirical indicator screening, background verification, and bounded clinical review. On page 13, it sharpens that contrast further by opposing the “clinical paradigm” of generalized impression and subjective interviews to an “actuarial paradigm” of validated prediction, actuarial decision rules, and criterion-linked screening.
The central premise of the follow-up is that no serious reform can remain compensatory. A system that allows severe red flags to be explained away by “whole person” balancing is the same system that created the current problem. That is why the new model must be noncompensatory. Some indicators should operate as systematic screening triggers because they are common and stable enough to support population-level exclusion pressure. Others should function as critical red flags because their severity and liability implications are too serious to be diluted by narrative context. Page 6 of the deck frames this point clearly in “The Predictive Risk Matrix,” distinguishing Tier 1 “Population Screening” from Tier 2 “Critical Red Flags,” and warning that a system which treats all adverse facts as merely “something to discuss” dissolves meaningful distinctions back into unstructured discretion. The follow-up should build directly on that insight. It should argue that the proposed solution is not simply more information or more caution. It is the adoption of a decision structure in which some facts must carry exclusionary force.
That new structure has several components, all already visible in the existing materials. First, prediction must move upstream. Page 11 makes the sequencing point: validated behavioral signals should be collected and assessed before the process reaches the high-cost, late-stage clinical screen. Second, the system must operate at the item level rather than through broad suitability narratives. Page 12’s “Blueprint for an Objective Model” identifies item-level specificity, tiering and calibration, and structured verification as core features. Third, discretion must be constrained rather than celebrated. The same page calls for reduced discretionary latitude, transparency in reasoning, and a defensible audit trail. Fourth, professional roles must be reallocated. The predictive work should be done by an evidence-based personnel-selection structure, while clinical evaluators retain only a bounded post-offer role. Finally, the system must become publicly legible. A screening decision should be capable of explanation in concrete terms: which indicator triggered concern, how that indicator is classified, and why the candidate advanced or was screened out. That is what makes the model objective rather than merely more elaborate.
The broader significance of the solution is civic, not merely administrative. The problem in police hiring has never been just an HR problem. Page 14 states this directly by framing the issue as “A Civic Crisis, Not Just an HR Issue,” and by emphasizing that delegating the state’s monopoly on force demands the highest level of predictive discipline, not discretionary leeway. That is the governing principle of the follow-up. The public does not bear the cost of weak police hiring in abstract organizational terms. It bears that cost through avoidable force, constitutional violations, sexual misconduct, false arrests, taxpayer-funded settlements, and degraded institutional legitimacy. A solution piece must therefore make clear that rebuilding police screening is not a matter of managerial optimization. It is a condition of lawful and responsible public power. The transition from impression to infrastructure is not optional modernization. It is the minimum reform required once the profession can no longer deny that the warning signs are visible, measurable, and predictive.
I. The Reform Mistake After Diagnosis
The first institutional mistake in police hiring is the one your earlier thought-piece already identified: departments continue to rely on a model that mistakes subjective clinical impression for predictive discipline. The second mistake is what typically happens after that critique lands. Institutions acknowledge that the current system is weak, imprecise, discretionary, or outdated, and then proceed to preserve its basic architecture with new language, new training materials, and new assurances that the process will now be administered more carefully. That is not a reform of structure. It is a reform of description. The underlying decision model remains the same: broad evaluative discretion, late-stage reliance on generalized suitability review, and an institutional preference for discussing warning signs rather than forcing them to carry predetermined consequences. If that architecture survives, the outcome will survive with it. The public may hear more disciplined terminology. It will not receive more disciplined screening. That is the central reform mistake after diagnosis, and it is already embedded in the logic of the slide deck itself. Pages 3 and 4, taken together, establish the precise problem. Page 3 identifies “The Evidence Deficit: The Data We Have vs. The Data We Use,” while page 4 labels the current model “The Category Error of Modern Police Hiring.” Those two slides do not merely criticize isolated practices. They expose a system that has information, understands in broad terms that the information matters, and nevertheless routes it through the wrong mechanism.
That distinction matters because institutions are often far more comfortable with critique than with redesign. A department can admit that hiring standards vary. It can admit that screening tools are imperfect. It can even admit that more evidence-based practices would be helpful. What it resists is the conclusion that the existing architecture must be displaced. Yet the deck makes clear that the current architecture is not failing because it lacks sincerity. It is failing because it relies on the wrong decision structure. Page 4 presents the contrast directly. On the left, “Clinical Impression” is described as the current state, a subjective approach that relies on late-stage clinical psychology to form generalized “whole person” narratives and flattens critical red flags into broad suitability judgments. On the right, “Actuarial Prediction” is presented as the required state, an objective approach that relies on personnel psychology to identify behavior-specific, empirically validated misconduct risks and bind human judgment to structured evidence. That is not a modest refinement. It is a change in governing logic. The institutional failure begins when departments read that contrast and assume they can keep the left-hand structure while borrowing some of the language from the right.
The slide deck does not support that softer reading. Page 3 shows why. The problem is not that agencies lack data about applicant risk. The slide states that 8,539 candidates were screened across more than 150 agencies, that 6,075 were hired and tracked over five years, and that 15 of 19 pre-hire misbehavior indicators significantly predicted later misconduct, with hazard ratios reaching up to 14.59. Yet the same slide identifies the institutional failure in stark terms: disclosing prior misconduct reduced a candidate’s hiring chances by only about 5%, and the graphic explicitly portrays that result as “95% Hired Despite Red Flags” and “5% Screened Out.” The bottom line on the slide is blunt: the profession does not suffer from an absence of information; it suffers from the institutional misuse of it. That sentence matters because it rules out the most common bureaucratic evasion. Departments can no longer plausibly say the problem is uncertainty, incomplete knowledge, or the inherent difficulty of prediction. The deck’s own diagnosis is that the information exists and is not being operationalized. Once that is established, the reform question changes. The issue is no longer how to improve the tone of the review process. It is how to rebuild the process so that known warning signs stop being administratively negotiable.
That is precisely where institutions often retreat. They accept the criticism in abstract but preserve the same discretionary format in practice. They leave red flags inside a qualitative “whole person” assessment. They continue to treat severe signals as something to be explained rather than something that should alter the candidate’s status. They keep the late-stage evaluative interview as the prestige center of the process. They promise better calibration while refusing to adopt structured exclusion rules. In effect, they convert diagnosis into improved rhetoric instead of improved design. The deck’s “category error” slide is especially useful here because it captures the difference between what institutions say they are doing and what they are actually doing. The left side of page 4 lists words like “maturity,” “judgment,” “concern,” and “fit” floating inside the cloud of clinical impression. That image is not accidental. It visualizes a model in which risks are translated into diffuse evaluative vocabulary. The problem with that vocabulary is not simply that it is soft. The problem is that it is structurally compensatory. Once every adverse fact is transformed into a generalized concern about overall suitability, every fact becomes easier to soften, contextualize, offset, or ignore. The very act of translating behavior-specific evidence into broad clinical language drains the predictive force from the warning sign.
This is why the follow-up cannot merely say that police hiring needs “better standards.” That phrase is too easy for institutions to absorb without real change. Better standards can mean more guidance layered onto the same discretionary process. Better standards can mean more careful documentation of the same clinical interview. Better standards can mean stronger admonitions to evaluators while leaving the candidate’s status dependent on generalized professional impression. None of that is enough if the architecture remains intact. The slide deck is stronger than that. It argues, by structure and by contrast, that the reform required is a transition from one paradigm to another. Page 13 makes this explicit by naming the shift. It contrasts “The Clinical Paradigm (Current Model)” with “The Actuarial Paradigm (Objective Model).” The clinical model’s core mechanism is “Generalized Impression,” its primary tool is “Subjective Clinical Interviews,” and its treatment of risk is “Flattened into ‘Overall Suitability.’” The objective model replaces those with “Validated Prediction,” “Actuarial Decision Rules,” and risk treatment “Tiered by Prevalence & Severity.” The significance of that slide is that it denies the idea that the current model can be rescued through incremental adjustment alone. It identifies two different governing systems. The reform mistake after diagnosis is failing to understand that difference and assuming that one can preserve the old system by speaking in the vocabulary of the new one.
The practical consequence of that mistake is predictable. If the process remains discretionary, late-stage, and compensatory, then severe indicators will continue to be absorbed into narrative balancing. If the process remains grounded in “overall suitability,” then departments will continue to treat evidence as advisory rather than binding. If clinical review remains the central prestige mechanism, then predictive information will continue to be subordinated to generalized professional judgment. The result is not mysterious. The slide deck already shows it. Red flags are gathered, discussed, and ultimately diluted. Candidates with meaningful warning signs continue to enter the institution. The department later responds to the downstream scandal as though the real failure began after the appointment. In reality, the architecture of that later failure was already preserved at the point of hire. The problem is not simply that institutions are slow to act. It is that they preserve a system designed to make early action easy to avoid.
That is why the first section of this follow-up must press a harder point than the original diagnosis. The original piece established that the current model is weak, subjective, and under-predictive. This follow-up must establish that critique alone is institutionally safe unless it is paired with structural displacement. A department can admit almost anything about the current model so long as it is not required to surrender discretion. It can tolerate the language of evidence. It resists the architecture of evidence. The slide deck, read honestly, leaves no room for that halfway position. Page 3 says the profession misuses the data it already has. Page 4 says the profession is using the wrong paradigm. Page 13 says the answer is not a more careful version of the current model, but a different one. The reform mistake after diagnosis, then, is preserving discretionary infrastructure after conceding that discretion is the mechanism through which predictive evidence keeps getting neutralized. Until that changes, every institutional promise of “improvement” will remain largely linguistic. The form of the process may look more sophisticated. The function of the process will remain the same.
II. Move Prediction Upstream
If the first reform mistake is preserving the old architecture after admitting its defects, the second is temporal: institutions continue to place the decisive predictive burden too late in the hiring sequence. The slide deck addresses this point most directly on page 11, “Moving Prediction Upstream,” and that page is one of the most important in the entire follow-up because it converts the critique from abstraction into process design. What it shows is that the current system over-relies on late-stage, high-cost clinical interviews to make generalized “suitability” decisions. The slide’s “Current Flawed Architecture” begins with a conditional offer, proceeds to background, and then places enormous weight on an “Expensive Clinical Psych Eval,” accompanied by a callout explicitly stating that the system over-relies on late-stage, high-cost clinical interviews to make generalized suitability decisions. The proposed architecture, by contrast, does not eliminate review; it reorders it. It begins with standardized background questionnaires, then empirical indicator screening using data-driven cutoffs, followed by background verification, and only then bounded clinical review. That sequence is not a cosmetic administrative preference. It is the heart of the proposed solution. The institution must stop asking the most expensive and least behavior-specific stage to do work that should already have been done earlier, more cheaply, and more objectively.
The sequencing question matters because timing shapes what kind of decision the institution is capable of making. A process that begins with standardized questionnaires and empirical indicator screening starts by asking what is known, what can be measured, and what the evidence says about the significance of that information. A process that reserves its main predictive energy for a late-stage clinical review starts by asking how the applicant presents, how a professional interprets the applicant’s narrative, and whether broad concerns can be reconciled with eventual clearance. Those are fundamentally different orientations. The first treats prediction as a screening task. The second treats prediction as an evaluative conversation. Page 11 sides decisively with the first. It explicitly states that the proposed objective architecture moves validated behavioral signals before the conditional offer and binds clinical review strictly to post-offer pathology checks. That sentence is critical because it tells us not only when the predictive work should happen, but also what clinical review should and should not be doing. Clinical review is not abolished. It is limited. The architecture reallocates its function. That is the difference between reforming the hiring process and merely adding new stages to the old one.
There is a practical reason for this upstream shift, and the deck does not hide it. The current model is expensive and inefficient because it carries too many candidates too far into the process before forcing the institution to confront what should already have been apparent. If validated indicators can be collected early, then high-risk candidates can be identified before the department invests in a stage that is both costlier and less well suited to item-level predictive discrimination. This is one of the places where the slide deck’s solution is especially elegant. It does not argue only from fairness or public safety. It also argues from institutional design. A process that puts prediction first is cheaper, clearer, and harder to manipulate. A process that puts generalized evaluation first is costlier, murkier, and more hospitable to discretionary drift. Departments often defend late-stage review as carefulness. Page 11 reveals that it is often the opposite: an inefficient architecture that postpones the hardest decisions until the point where subjective discretion is greatest.
Moving prediction upstream also changes the meaning of background investigation. In the current model, background can become merely preparatory—a file-building step before the more prestigious clinical or suitability review. In the proposed model, background becomes central because it is one of the earliest sites where validated indicators are collected and organized. The slide’s architecture makes that plain. Standardized background questionnaires come first. Then empirical indicator screening using data-driven cutoffs. Then background verification. The sequencing implies a hierarchy of function. Background is no longer just evidence gathering for a later holistic judgment. It becomes part of the predictive gate itself. That matters because the earlier thought-piece established that the profession does not lack warning signs; it underuses them. Moving prediction upstream is therefore not simply about speed. It is about ensuring that the warning signs matter before they are swallowed by narrative. The earlier the institution applies structured risk logic, the harder it becomes for later-stage professional impression to erase what the data already say.
The architecture on page 11 is strengthened further by page 12’s “Blueprint for an Objective Model.” The first principle listed there is “Temporal Sequencing & Early Collection,” which directs departments to move validated prehire screening upstream to gather predictive indicators early and at lower cost, before conditional offers. That principle is not ancillary. It is the opening feature of the objective model for a reason. Without proper sequencing, every other improvement remains vulnerable to dilution. Departments can have tiered indicators, structured verification, and transparency in reasoning, but if those features are introduced only after the institution has already committed itself to a candidate and made clinical review the prestige center of the process, the design remains unstable. The process must begin where the evidence is strongest and where the institution still has practical freedom to exclude without having to rationalize why a candidate already favored should now be removed. Moving prediction upstream therefore protects the integrity of the rest of the model. It places evidence at the point of maximum preventive utility.
This shift also answers one of the most persistent institutional objections: that hiring requires holistic assessment and that hard judgments cannot responsibly be made early. The deck’s architecture responds by distinguishing between predictive screening and bounded clinical review. It does not deny that some later-stage assessment may be warranted. It denies that such assessment should carry the main predictive burden. A process that starts with validated behavioral indicators is not prematurely rigid. It is properly sequenced. It recognizes that there are different kinds of questions in police hiring and that they belong in different stages. The question of whether an applicant presents behavior-specific indicators of future misconduct risk should come first because it can be asked through standardized instruments and verified through investigation. The question of whether an applicant presents a distinct clinical or pathological concern belongs later and in a narrower role. Page 11 makes that division explicit. The current model confuses the two. The proposed model disentangles them.
There is also a deeper institutional virtue in moving prediction upstream: it weakens the system’s dependence on performative rigor. Departments like processes that look serious. Late-stage clinical interviews, expensive evaluations, and extended review stages create the appearance of caution and professionalism. But the deck’s solution insists that appearance cannot substitute for design. A process can look intensive and still be structurally unserious if it postpones the actual predictive work until after discretion has had multiple opportunities to soften the evidence. By contrast, a process that begins with standardized background questionnaires and empirical screening may look less glamorous, but it is more faithful to the actual problem. The issue is not whether a candidate feels suitable after prolonged review. The issue is whether the institution can identify risk signals that should affect access to the badge before those signals are narratively reframed. Moving prediction upstream is therefore both a technical and symbolic correction. It reorders the process around substance instead of spectacle.
The same lesson appears again on page 13, where the “Actuarial Paradigm” is described as using validated prediction and actuarial decision rules, while the current clinical paradigm relies on subjective interviews and generalized impression. Although page 13 is more comparative than sequential, it reinforces the upstream logic of page 11. Actuarial systems work best when they are positioned early enough to serve as gates rather than late enough to become advisory overlays. If they come too late, they risk becoming one more input into the same generalized balancing process the current model already performs. If they come early, they structure the field on which all later review occurs. That is why the temporal issue is inseparable from the institutional one. Moving prediction upstream is not just about efficiency. It is about authority. It determines whether evidence governs the process or merely participates in it.
For that reason, the solution piece must be clear: prediction belongs at the front end. The current architecture waits too long, spends too much, and asks the wrong stage to do the wrong work. The proposed architecture begins with standardized collection, empirical indicator screening, and verification, and only afterward allows a bounded clinical role. That sequence reflects a more honest understanding of what police hiring actually is. It is not principally an exercise in professional impression. It is a screening function in which certain known signals must be identified and acted upon before the institution hands over authority. The moment departments understand that, the logic of page 11 becomes unavoidable. The task is not to make late-stage review more refined. The task is to stop relying on late-stage review to perform predictive work that should have already been completed upstream.
III. Make the Model Noncompensatory
If the predictive work must move earlier, the next requirement is even harder for institutions to accept: the model must become noncompensatory. This is where the follow-up becomes most forceful, because it is here that the deck’s proposed solution most directly collides with the culture of police hiring as it currently operates. A compensatory model is one in which a serious red flag can be offset by other favorable impressions or by an evaluator’s belief that the candidate is otherwise strong. A noncompensatory model is one in which certain indicators are so severe, so probative, or so institutionally dangerous that they cannot be washed away by “whole person” balancing. Page 6 of the deck, “The Predictive Risk Matrix,” is the clearest statement of this principle. It separates applicant facts into categories based on prevalence and severity and identifies two decisive zones: Tier 1 “Population Screening” and Tier 2 “Critical Red Flags.” The slide’s key teaching is that not all warning signs are equal and that evidence demands a calibrated, tiered response. But the most important line is its “Takeaway”: a system that treats all adverse facts as just “something to discuss” dissolves meaningful distinctions back into unstructured discretion. That sentence is the noncompensatory argument in concentrated form. If everything remains discussable, nothing remains decisive.
The culture of police hiring has long favored the opposite instinct. Departments are habituated to the language of judgment, maturity, context, rehabilitation, and overall suitability. Those concepts are not meaningless, but they become dangerous when used to absorb indicators that should operate as stopping points. Page 6 rejects that culture directly. It identifies Tier 2 “Critical Red Flags” as including domestic violence citations, unjustified use of force, and racially offensive behavior complaints. The action rule attached to that box is uncompromising: “Absolute red flags requiring hard exclusions despite smaller data counts.” That is the language of a noncompensatory system. It means the institution does not get to say that a domestic violence citation is concerning but outweighed by later performance, impressive demeanor, or a favorable interview. It means confirmed unjustified force does not become one factor among many. It means racially offensive conduct complaints are not simply contextualized into a broader character narrative. It means some indicators must stop the process.
This is probably the strongest part of the solution because it addresses the precise mechanism through which the current model fails. The earlier critique established that the profession underuses the warning signs it already has. The reason it underuses them is not only poor timing or inconsistent standards. It is also compensatory logic. Severe facts enter the system, but the system is designed to metabolize them into discussion. By the time the candidate exits the process, the red flag has been softened into an issue of “concern,” “fit,” or “overall judgment.” The predictive force of the fact is thereby neutralized. A noncompensatory model breaks that pattern by refusing to let certain indicators be converted into narrative. It preserves the indicator as an exclusionary event. That is why the “Tier 2” category matters so much. It creates a class of signals whose institutional meaning is fixed enough that they cannot be bargained away through discretionary interpretation.
There is a reason institutions resist this model. Noncompensatory rules reduce flexibility. They prevent departments from hiring a candidate they like despite a fact pattern they would rather minimize. They constrain the evaluator who believes he can see beyond the record. They limit the hiring manager who wants to solve a staffing problem by taking a calculated risk. And they weaken the prestige of the late-stage interview by making clear that some decisions have already been functionally determined by earlier evidence. All of this is precisely why the model is necessary. The problem in police hiring has never been the absence of discretion. It has been the overabundance of it. A system that gives institutions too many opportunities to reinterpret red flags will produce exactly the outcome the earlier thought-piece described: warning signs gathered, acknowledged, and nevertheless hired through. Noncompensatory design is the mechanism that interrupts that cycle.
The slide deck’s phrasing is especially helpful because it does not argue that every adverse fact deserves hard exclusion. Page 6 preserves nuance by distinguishing “Background Noise” from Tier 1 and Tier 2. Minor unrelated indiscretions are explicitly placed in their own zone. That distinction is important because it prevents institutions from caricaturing the noncompensatory argument as a demand for rigid exclusion based on any negative fact. The deck is more sophisticated than that. It says some facts are background noise. Some facts support population screening. Some facts require absolute exclusion. The problem is not that the model is too hard. The problem is that the current model refuses to distinguish among categories with enough discipline. It compensates across domains that should not be treated as morally or predictively interchangeable. Noncompensatory design, then, is not the abandonment of calibration. It is calibration taken seriously.
The force of the Tier 2 category is reinforced by page 5, “Dissecting Pre-Hire Risk,” which breaks applicant risk into four validated domains of behavioral signal: prior occupational trouble, law-enforcement trouble, temper and violence, and irresponsible behaviors. Page 5 ends with a synthesis note: prior behavior is the strongest signal of future behavior when the predictor closely resembles the institutional criterion. That principle strengthens the case for noncompensatory exclusion because some Tier 2 indicators are not just severe; they are highly commensurate with the risks the institution claims to be screening against. A domestic violence citation is not merely bad optics. It is an indicator tied to violence and abuse. Unjustified use of force in a prior law-enforcement setting is not merely adverse history. It is behavior sampled from the very context into which the institution is considering re-entry. A racially offensive behavior complaint is not simply awkward conduct. It is a warning sign about the candidate’s likely interaction with a public he will police under state authority. The closer the match between the indicator and the institutional risk, the stronger the argument that the signal cannot responsibly be compensated away.
This is where the follow-up should state the normative principle clearly: some facts must actually mean stop. That sounds obvious, but police hiring has not consistently behaved as if it were true. The slide deck is implicitly repudiating the institution’s longstanding habit of turning every severe fact into a “conversation.” The result of that culture is not mercy or nuance. It is loss of distinction. It is the collapse of different categories of risk into a single “whole person” pool where the evaluator’s subjective narrative becomes the ultimate authority. Page 6 expressly rejects that collapse. The hard-exclusion language for Tier 2 exists because the model recognizes that some indicators are too consequential to be entrusted to soft balancing. A candidate with such a signal is not simply higher risk in the abstract. The candidate presents a type of risk the institution cannot honestly say it failed to foresee if it proceeds with the hire anyway.
The civic dimension of the noncompensatory model should also be emphasized. This is not ordinary employment. The institution is deciding who will exercise force, restraint, discretionary command, and public credibility. In that context, the refusal to make some red flags decisive is not just an HR choice. It is an allocation of public exposure to foreseeable harm. Page 14 of the deck, though outside the pages the user highlighted for this section, states that delegating the state’s monopoly on force demands the highest level of predictive discipline, not discretionary leeway. That line captures why noncompensatory design is necessary in policing even if other sectors tolerate more compensatory balancing. The public is not protected by a model that says every dangerous fact can still be reconsidered through generalized narrative. The public is protected when the institution is willing to say that some predictors disqualify, period.
So Section III should be unambiguous. The proposed model must be noncompensatory because the current system’s biggest vice is its ability to launder severe risk through broad suitability language. Tier 2 indicators such as domestic violence citations, unjustified use of force, and racially offensive conduct complaints cannot be reduced to “something to discuss.” The system must preserve their force by treating them as hard stops. Without that feature, every other reform will remain vulnerable to the same old pattern: better forms, better interviews, better rhetoric—and the same institutional willingness to hire through what should have ended the candidacy.
IV. Separate Tier 1 from Tier 2
One of the most important reasons police hiring continues to fail is that it has not disciplined itself to distinguish among types of adverse information with sufficient rigor. It has treated applicant risk too often as a single undifferentiated field, to be summarized in a final judgment about whether the person seems suitable overall. Pages 5 and 6 of the slide deck provide the corrective. They show that the institution must separate Tier 1 from Tier 2 because not all warning signs do the same work, carry the same frequency, or require the same institutional response. Page 5 supplies the domains: prior occupational trouble, law-enforcement trouble, temper and violence, and irresponsible behaviors. Page 6 then translates those domains into a risk matrix based on prevalence and severity, distinguishing “Background Noise,” Tier 1 “Population Screening,” and Tier 2 “Critical Red Flags.” The point of reading these pages together is to show that the proposed objective model is not simply more data-driven. It is more discriminating. It insists that departments stop flattening qualitatively different signals into one vague narrative of overall suitability.
Tier 1 exists because some indicators are common enough, stable enough, and broad enough in their predictive value to justify systematic use at the population level. Page 6 identifies prior written reprimands, unfavorable terminations, and bad credit as Tier 1 indicators and pairs them with an action rule stating that they “must trigger systematic, population-level screening exclusions.” That is a powerful formulation because it shows the deck’s model is not limited to severe one-off facts. It also includes higher-frequency indicators that, taken seriously, should materially shape how agencies process applicant pools. These are not necessarily absolute disqualifiers in the same sense as Tier 2 red flags, but they are not casual discussion points either. They are the backbone of the institution’s screening architecture. They enable an agency to impose structured exclusion pressure across a broad applicant population rather than waiting only for spectacular red flags to emerge. In that sense, Tier 1 is how the model becomes administratively serious. It treats common predictive indicators as part of a repeatable and rule-bound system rather than leaving them to variable evaluator instinct.
Tier 2 performs a different function. It captures indicators that may appear less frequently but carry such severe liability implications that they cannot responsibly be handled as supplemental context. Page 6 places domestic violence citations, unjustified use of force, and racially offensive behavior complaints in this category and states that they require “hard exclusions despite smaller data counts.” The importance of this distinction cannot be overstated. A department that does not separate Tier 2 from Tier 1 will either overreact to ordinary indicators or underreact to serious ones. If everything is treated the same, then nothing is calibrated. If everything becomes “one more factor,” then high-severity indicators lose their force. The deck’s matrix prevents that collapse by making the institution specify what kind of problem it is dealing with. Is the indicator prevalent and useful for population-level screening? Or is it rarer but severe enough to require a hard stop? That is the kind of structured differentiation the current model lacks.
Page 5 helps explain why the distinction is necessary. It breaks the applicant’s history into validated domains rather than allowing the institution to process the file as one amorphous character narrative. Prior occupational trouble includes negligence warnings, unfavorable terminations, and repeated job-hopping. Law-enforcement trouble includes prior reprimands, demotions, and unjustified use of force. Temper and violence include documented physical altercations and domestic violence citations. Irresponsible behaviors include bad credit, moving violations, and support arrears. This structure matters because it reveals that adverse facts are not interchangeable. A negligence warning does not function the same way as a domestic violence citation. A prior written reprimand does not function the same way as an unjustified use of force. Yet the current hiring model too often treats them all as pieces of a generalized impression. The Tier 1/Tier 2 distinction is the corrective to that habit. It restores categorical differences that the “whole person” model tends to erase.
The mistake of flattening risk is not only analytical. It is institutional. Once all adverse facts are pooled together under overall suitability, the evaluator gains enormous power to decide how much any particular indicator should matter. That discretion is exactly what the slide deck is trying to discipline. Tier 1 reduces discretion by making some common indicators systematically relevant. Tier 2 reduces discretion even further by making some severe indicators effectively non-negotiable. The distinction therefore serves a dual function. It improves prediction, and it constrains the institution’s habitual tendency to turn structured evidence back into narrative judgment. Without this separation, departments can continue doing what they have long done: treating every file as unique in a way that obscures the existence of repeatable patterns. The more the institution insists every case is singular, the easier it becomes to deny that common decision rules should exist at all. The deck rejects that posture by insisting that prevalence and severity are knowable and should structure response.
This is also why Tier 1 and Tier 2 should not be confused with “minor” and “major” in a purely rhetorical sense. Tier 1 indicators are not trivial. They are the material from which serious screening architecture is built. Because they occur with enough frequency and show stable relations, they allow the institution to create population-level filters rather than improvising case by case. A department that ignores Tier 1 because the indicators are not sensational will miss the broader function of the model. Conversely, a department that treats Tier 2 as simply “more serious Tier 1” will fail to understand why the deck assigns hard exclusions to that category. The distinction is not only one of degree. It is one of institutional use. Tier 1 organizes screening across the applicant population. Tier 2 marks the points at which discretion must stop. Together, they create a model that is both scalable and morally serious.
The current model’s deepest flaw is that it dissolves both categories into a single vague story of overall suitability. That phrase sounds responsible, but it is analytically destructive because it treats all predictive information as ultimately commensurable. Under that logic, a candidate with repeated prior reprimands might be cleared because he interviews well. A candidate with a domestic violence citation might be retained in consideration because other aspects of the file seem favorable. A candidate with bad credit and an unfavorable termination might be viewed as complicated but promising. None of those outcomes is impossible under a subjective model because the model has no durable internal distinctions strong enough to resist discretionary smoothing. Pages 5 and 6 are an answer to that entire way of thinking. They tell the institution to stop asking only whether the person seems suitable overall and start asking what type of indicator is present, how often such indicators appear, how severe they are, and what category of institutional response they trigger.
There is also a broader legitimacy benefit to keeping Tier 1 and Tier 2 separate. A system that cannot explain why some facts generate systematic screening consequences while others generate hard exclusions is difficult to audit and easy to manipulate. By contrast, a system that explicitly distinguishes population-level indicators from critical red flags becomes more transparent both internally and publicly. Page 12’s “Blueprint for an Objective Model” later builds on this by calling for transparency in reasoning and a defensible audit trail, but those features depend on the Tier 1/Tier 2 distinction already being in place. Without it, every explanation collapses back into “the evaluator considered everything.” With it, the institution can say what kind of signal was present and why the signal required the response it did. That is the beginning of accountable hiring.
So Section IV should conclude with a clear proposition: the current model fails because it flattens distinct categories of risk into one vague suitability judgment. Pages 5 and 6 provide the alternative. Tier 1 supports population screening because some indicators are common and stable enough to justify systematic exclusions. Tier 2 supports critical red-flag exclusion because some indicators are so severe that they require hard stops. Not all adverse facts are equal, and a hiring model that refuses to act like that is not nuanced. It is undisciplined. The purpose of separating Tier 1 from Tier 2 is therefore not bureaucratic neatness. It is to force the institution to preserve distinctions that predictive evidence has already established and that public safety can no longer afford to lose.
V. Reduce Discretion by Design
The slide deck’s proposed solution does not merely recommend better judgment. It recommends less unchecked judgment. That distinction is essential because the institutional record has already answered the question of what departments do when broad discretion remains intact: they hire through the warning signs. The earlier parts of the deck established that departments often possess meaningful prehire misconduct indicators and yet give those indicators minimal practical effect at the gate. Once that finding is accepted, open-ended discretion can no longer be defended as a neutral feature of a thoughtful hiring process. It must be understood as one of the primary mechanisms through which predictive evidence gets softened, delayed, diluted, or ignored. That is why pages 12 and 13 are so important. They do not present structured rules, actuarial logic, and auditable reasoning as optional refinements for unusually disciplined departments. They present them as the core institutional safeguard against a system that has already shown its willingness to underuse the evidence in front of it.
Page 12, “Blueprint for an Objective Model,” states the point directly in several of its design principles. Among them are “Reduced Discretionary Latitude,” “Transparency in Reasoning,” and “Defensible Audit Trail.” These are not secondary administrative values. They are the instruments by which the proposed model protects itself from the same institutional behavior the original thought-piece exposed. The deck has already shown what happens when departments are allowed to process validated warning signs through broad, qualitative, and largely unstructured review. Those warning signs do not disappear, but they do lose force. They become concerns to be weighed rather than triggers to be acted upon. They become part of a generalized suitability file rather than the basis for a structured decision. The blueprint’s answer is therefore not to ask decision-makers to be more careful while leaving them the same latitude. It is to narrow the space in which they can neutralize evidence.
This is where page 13’s contrast between the “Clinical Paradigm” and the “Actuarial Paradigm” becomes more than conceptual. It becomes operational. The slide makes clear that the current model is governed by “Generalized Impression,” “Subjective Clinical Interviews,” and “Flattened into ‘Overall Suitability.’” The proposed model replaces those with “Validated Prediction,” “Actuarial Decision Rules,” and treatment of risk “Tiered by Prevalence & Severity.” That is not a simple difference in professional vocabulary. It is a shift in how authority is allocated inside the hiring system. Under the clinical paradigm, the evaluator’s broad impression remains the ultimate container into which the rest of the file is poured. Under the actuarial paradigm, the evaluator is constrained by structured decision logic that has already assigned institutional meaning to specific indicators. The significance of that change is enormous. It means the hiring system is no longer asking whether a red flag can be explained away by a persuasive narrative. It is asking what the decision rules require once the red flag is present.
Institutions often resist that move because they describe discretion as wisdom. They say hiring cannot be reduced to formulas, that experienced reviewers need space to interpret context, and that rigid structures fail to account for human complexity. The deck’s answer is not that context is irrelevant. It is that unrestricted context is what produced the current failure. Once an institution has demonstrated that open-ended discretion leads it to hire through serious warning signs, it loses the privilege of treating discretion as presumptively benign. In that setting, structure is not the enemy of good judgment. It is the condition that makes good judgment harder to counterfeit. A department that insists on preserving maximum latitude after acknowledging that it has repeatedly underused predictive information is not asking for nuance. It is asking to keep the same escape route that made the prior system permissive. Pages 12 and 13 reject that request by making structured reasoning an affirmative design principle rather than an optional discipline for unusually conscientious reviewers.
The deck’s logic here is especially powerful because it is empirical rather than moralistic. It does not say departments should reduce discretion because discretion is philosophically suspect. It says departments should reduce discretion because the institution has already shown what it does with it. That is the key institutional point this section must make. Open-ended review is not a neutral default. It is a known site of evidence dilution. It allows a hiring manager to preserve a favored candidate despite predictive concerns. It allows an evaluator to reframe a severe indicator as contextual noise. It allows departments to speak the language of reform while continuing to behave as though every file remains wholly open to personalized interpretation. The blueprint’s use of structured rules and actuarial logic is therefore not technocratic excess. It is a direct answer to a demonstrated failure pattern. The design must assume that, left unconstrained, the institution will continue to do what it has already done.
The phrase “actuarial decision logic” often triggers resistance because it sounds impersonal, mechanical, or bureaucratically cold. But the deck frames it differently. On page 13, the actuarial paradigm is not presented as anti-human. It is presented as anti-drift. It uses validated prediction, tiering, and rule-bound treatment of indicators to ensure that the same facts produce comparable institutional consequences rather than changing meaning from evaluator to evaluator. That consistency is one of the deepest benefits of reducing discretion by design. It protects not only public safety, but institutional coherence. A system that allows one reviewer to treat prior unjustified force as nearly disqualifying while another treats it as explainable has no stable screening philosophy at all. It has only local judgment with a common label attached to it. The deck’s solution requires more than goodwill; it requires repeatable logic.
There is also a defensive reason for structured reasoning. The more power a department retains to make highly individualized and opaque decisions, the easier it becomes to rationalize outcomes after the fact. If a problematic hire later becomes a complaint, lawsuit, or scandal, the institution can always say the file was carefully considered and that multiple factors pointed in favor of appointment. A structured system makes that kind of retrospective self-protection much harder. If the rules are clear, if the tiers are defined, and if the institutional meaning of validated indicators is documented, then the later question is no longer whether someone “used professional judgment.” It becomes whether the decision complied with the system. That is precisely why structured rules are a safeguard. They create institutional accountability at the moment of hiring rather than leaving accountability to be litigated only after harm occurs.
The blueprint on page 12 implicitly understands all of this. That is why it does not stop at advocating better inputs. It insists on a different decision environment. “Reduced Discretionary Latitude” means the evaluator’s role is narrowed. “Transparency in Reasoning” means the logic of the decision must be visible. “Defensible Audit Trail” means the institution must later be able to show how the signal was categorized and why the candidate advanced or was excluded. Taken together, these principles make clear that the problem is not simply collecting more predictive evidence. It is forcing the institution to live with the consequences of the evidence it already has. Without that design discipline, the current model will persist under a new label. With it, the profession begins to move from impression to infrastructure in a meaningful sense.
So the conclusion of this section should be direct. Structured rules, actuarial decision logic, and auditable reasoning are not optional add-ons for a reform-minded department. They are the heart of the reform because they address the mechanism through which predictive evidence keeps losing institutional force. Departments have already shown what they do with open-ended discretion: they hire through the warning signs. The answer is not to ask them to be wiser with that discretion. The answer is to redesign the process so that discretion can no longer function as a solvent for evidence.
VI. Reallocate Professional Roles
One of the most persistent sources of confusion in police hiring is the profession’s failure to distinguish between different kinds of professional work. The current model has tended to collapse too much of the predictive burden into the hands of clinical evaluators, especially at the late stage of the hiring process, as though the central problem were one of expert impression rather than validated risk architecture. Pages 11, 12, and 13 of the slide deck reject that arrangement. Read together, they support a different allocation of professional roles: the front end belongs to validated personnel-selection architecture, while clinical evaluators should occupy a bounded post-offer role focused on discrete pathology or functional concerns rather than acting as the main predictive engine of police misconduct risk. That is not a minor procedural correction. It is a core part of the reform model because the current system has been asking clinical review to do work it is not built to do.
Page 11, “Moving Prediction Upstream,” provides the clearest operational version of that argument. Its “Current Flawed Architecture” shows the system funneling applicants toward an expensive clinical psychological evaluation that is expected to make broad suitability determinations after the department has already invested time and attention in the candidate. The slide explicitly criticizes that design for over-relying on late-stage, high-cost clinical interviews to make generalized “suitability” decisions. By contrast, the “Proposed Objective Architecture” shifts the front-end work to standardized background questionnaires, empirical indicator screening, and background verification, with clinical review appearing only afterward as “Bounded Clinical Review.” That ordering matters because it signals a different professional hierarchy. The predictive decisions are no longer supposed to emerge primarily from the clinician’s generalized assessment. They are supposed to emerge earlier, from a structured system built to identify, verify, and classify known indicators of future misconduct risk.
Page 12 strengthens that shift by describing the architecture of the objective model. Although the slide uses design principles rather than profession titles, its logic is unmistakable. “Item-Level Specificity,” “Temporal Sequencing & Early Collection,” “Structured Verification,” and “Tiering & Calibration” are the features of a personnel-selection model, not a model centered on broad clinical impression. These are the tools by which a selection system is built, validated, and administered. They concern what information is collected, when it is collected, how it is verified, how it is sorted, and what consequences follow. Clinical review, by contrast, is not presented as the commanding logic of this architecture. It appears as a bounded component of a larger evidence-driven process. That is exactly as it should be if the institution is serious about separating predictive screening from generalized professional evaluation.
Page 13 then makes the role distinction explicit at the paradigm level. The “Clinical Paradigm (Current Model)” is described as relying on generalized impression, subjective clinical interviews, and risk flattened into “overall suitability.” The “Actuarial Paradigm (Objective Model)” is defined by validated prediction, actuarial decision rules, and risk tiered by prevalence and severity. That slide does not merely compare techniques. It compares professional orientations. The clinical paradigm assumes the primary task is interpretive: to synthesize impressions, context, and concerns into a final judgment. The actuarial paradigm assumes the primary task is structural: to use validated signals and predetermined decision rules to govern outcomes. Those are different forms of expertise. The follow-up must therefore be clear that the current system’s problem is not simply overreliance on psychology in the abstract. It is the misallocation of predictive authority to a mode of review that is inherently too generalized, too narrative-driven, and too late in the process to bear the institutional burden it has been asked to carry.
This matters because institutions often defend the current arrangement by invoking the prestige of professional judgment. They imply that clinical reviewers are uniquely qualified to “read” a candidate, understand nuance, and integrate complex factors that cannot be reduced to a checklist. But the slide deck’s solution is not hostile to expertise. It is hostile to the wrong expertise occupying the wrong place. A clinical evaluator may be well suited to identify certain pathology-based concerns, mental-health conditions, or functional issues that legitimately require post-offer assessment. What the clinician is not structurally well suited to do is operate as the principal translator of behavior-specific misconduct risk into front-end hiring consequences. That task belongs to a validated selection system, precisely because it depends on item-level evidence, tiered treatment, structured verification, and decision rules tied to known predictive relationships. The front end therefore belongs to the architecture, not to the impressionist.
There is an important institutional consequence here. Once the profession reallocates roles properly, it becomes harder for departments to use the clinical stage as the prestige center of the process while underdeveloping the actual predictive core. Under the current model, the late-stage clinical review often carries symbolic authority. It signals seriousness, caution, and professionalism. But the deck shows why that symbolism has been misleading. The predictive work should already have happened before the candidate reaches that stage. By assigning the central screening function to structured questionnaires, empirical indicator review, and verification, the proposed architecture demotes the clinical stage from institutional centerpiece to bounded specialist role. That is a profound but necessary demotion. It strips the system of one of its favorite illusions: that an expensive, expert-mediated late-stage review can substitute for disciplined front-end screening.
This reallocation also serves fairness. A system that expects clinicians to convert complex histories into one global suitability judgment risks asking too much of a single professional voice and then obscuring how the decision was actually made. A system that distributes the work more rationally is easier to understand and defend. The selection architecture identifies and classifies indicators. The verification process confirms them. The tiering structure assigns institutional significance. The decision rules constrain the outcome. Clinical review, where necessary, addresses the narrower question it is actually designed to answer. That is a cleaner, fairer, and more accountable division of labor because it prevents the system from hiding predictive decisions inside the opaque language of therapeutic or quasi-therapeutic assessment. The point is not that clinical professionals disappear. The point is that they stop being used as the all-purpose container for a problem that is, at bottom, about predictive gatekeeping.
The role shift is also essential because it aligns the profession with what the original diagnosis already made plain. The data problem is not on the back end of the process. It is not that the institution lacks evidence until the clinician speaks. The evidence begins earlier, in prior occupational trouble, law-enforcement trouble, violence indicators, and irresponsible behaviors. Those are the domains identified on page 5. If that is where the predictive material resides, then the front end must be designed around collecting and acting on that material. To continue treating clinical review as the primary gatekeeper after recognizing where the actual evidence lies is to preserve the category error under a more refined description. Pages 11 through 13 reject that. They relocate the institutional center of gravity away from generalized clinical judgment and toward validated selection design.
So the argument in this section should be framed without hesitation. The front end of police hiring belongs to validated personnel-selection architecture. Clinical evaluators should have a bounded, post-offer role directed at issues they are actually equipped to assess. The current model is defective because it asks clinical review to do predictive work it is not built to do and then treats that review as the profession’s principal claim to rigor. The proposed solution reallocates that authority. It restores predictive discipline to the selection system itself and leaves clinical review where it belongs: important, but not central.
VII. Build the Audit Trail
A hiring reform model that cannot explain its own decisions is not yet a reform model. It is only a more elaborate version of the status quo. That is why page 12’s insistence on a “Defensible Audit Trail” is one of the most important parts of the slide deck’s solution architecture. The point of the audit trail is not mere paperwork. It is institutional memory, accountability, and constraint. A department that cannot later show why a candidate was advanced or excluded, what indicators were present, how those indicators were classified, and what decision rules were applied is still operating inside the old world of drift, favoritism, dilution, and later denial. The audit trail is what converts predictive discipline from aspiration into durable organizational practice. Without it, even a well-designed model can be quietly undone by discretionary habit.
The need for an audit trail follows directly from the failure pattern already identified in the earlier slides. Departments have shown that they can gather meaningful warning signs and still fail to make those signs matter at the point of hiring. Once that is understood, the next question is not simply whether better rules exist, but whether the institution can later demonstrate that it followed them. The deck’s blueprint answers by requiring transparency in reasoning and a defensible record of how each decision was reached. This is the structural counterpart to the move away from generalized suitability. If a department continues to make decisions that can only be explained as “the evaluator had concerns” or “the candidate was deemed acceptable overall,” then the system remains vulnerable to the same weaknesses it is trying to cure. The audit trail forces specificity. It requires the institution to say which signals were present, whether they fell into Tier 1 or Tier 2, whether they were verified, and why the candidate either advanced or stopped.
This is indispensable because open-ended decision-making often survives by making itself difficult to reconstruct. When a problematic hire later becomes a source of complaint, the institution can hide behind generalities. It can say the file was considered carefully, multiple factors were weighed, professional judgment was exercised, and no single issue was dispositive. That language is almost impossible to audit because it does not reveal a decision path. A defensible audit trail changes that. It turns the hiring decision into a series of traceable institutional acts rather than a generalized impression sealed inside a final conclusion. That means later review—whether internal, judicial, legislative, or public—can ask the right questions. What indicators were present? What rules applied? Was the candidate subject to a Tier 1 screening trigger? Was a Tier 2 red flag present? Was the information verified? If the institution cannot answer those questions from its own record, then it has not truly left the old model behind.
The audit trail also protects the integrity of structured screening itself. Rules and tiers can be announced publicly, but if the internal recordkeeping does not preserve how they were actually applied, the system remains easy to manipulate. A department can proclaim that domestic violence citations are critical red flags or that prior law-enforcement reprimands trigger heightened screening, but if the case file does not record whether such indicators were found, how they were verified, and what institutional consequence followed, the announced model can be evaded without obvious trace. Page 12’s blueprint prevents that by linking reduced discretion, transparency in reasoning, and auditability as part of a single design philosophy. The audit trail is therefore not an afterthought. It is the enforcement mechanism for the architecture itself. It ensures that the model lives not only in the policy manual, but in the decision file.
There is also a fairness dimension. A transparent audit trail benefits not only the institution, but the candidate and the public. For the candidate, it creates a record that can be examined for consistency, defensibility, and compliance with stated standards. For the public, it creates the possibility of meaningful accountability when a later incident raises questions about why the candidate was cleared. In the absence of such a record, every dispute becomes a battle of narrative. The department says it exercised sound judgment; critics say it ignored obvious warning signs. But neither side can move beyond rhetoric if the file itself does not contain a structured account of how the decision was made. The audit trail supplies that account. It turns the hiring decision into something that can be reviewed as an act of governance rather than protected as an inscrutable act of expertise.
The institutional benefits are equally substantial. A defensible audit trail allows departments to monitor whether their own process is functioning as designed. Are Tier 1 indicators actually producing systematic screening consequences? Are Tier 2 indicators being treated as hard stops? Are certain decision-makers deviating from policy more often than others? Are verification failures weakening the screening architecture? None of these questions can be answered reliably if the department’s internal records preserve only a final suitability outcome without the path that produced it. In that sense, the audit trail is also a management instrument. It allows the institution to see whether the new model is being lived or merely recited. Without it, reform remains declarative. With it, reform becomes measurable.
The connection to pages 11 and 13 should also be made clear, even though page 12 is the anchor here. Page 11’s upstream architecture implies a sequence of institutional checkpoints: standardized background, empirical indicator screening, verification, bounded clinical review. Each checkpoint should leave its own trace. Page 13’s actuarial paradigm implies rule-bound treatment of validated indicators rather than generalized impression. That too should leave a trace. The audit trail is the record of those structural commitments in action. It shows that the process was not merely described as objective after the fact, but actually operated through objective stages while the decision was being made. This is why the audit trail belongs inside the core architecture rather than outside it as a compliance requirement. It is how the institution proves to itself, and later to others, that the system was not quietly pulled back into discretionary habit.
There is a final reason the audit trail matters: denial. Institutions confronted with later scandal often retreat into selective memory. They claim not to have known, not to have appreciated the significance of the warning signs, or not to have had enough information to do more. A robust audit trail makes those escape routes narrower. It records what the institution knew, how it classified it, and what it did with it. That does not guarantee a perfect system, but it does make institutional amnesia harder to sustain. It also changes the moral quality of later failure. If a department later faces litigation or public scrutiny over a candidate it cleared despite validated red flags, the audit trail will show whether the failure lay in the model, in its administration, or in a conscious departure from both. That is exactly the kind of accountability the current system has lacked.
So this section should make the point without dilution: a real reform model documents why a candidate was advanced or excluded. If there is no defensible audit trail, then the system is still vulnerable to the same old pathologies—drift, favoritism, dilution, and later denial. The audit trail is not clerical housekeeping. It is the backbone of institutional memory and the proof that predictive discipline actually governed the decision rather than serving as a slogan attached to it afterward.
VIII. Why This Is Not Just an HR Reform
The final mistake institutions often make when confronted with police hiring reform is to domesticate the issue. They treat it as an internal personnel question, a human-resources optimization problem, or a technical dispute about better assessment tools. Page 14 of the slide deck rejects that reduction outright. Titled “A Civic Crisis, Not Just an HR Issue,” it reframes the entire debate by locating police hiring where it belongs: inside the state’s delegation of force. The slide states that police officers do not enter ordinary workplaces and that the state’s monopoly on force demands the highest level of predictive discipline, not discretionary leeway. That line should govern the entire follow-up because it prevents the solution from being misunderstood as administrative modernization. The proposed model is not simply a better hiring process. It is a civic safeguard against preventable public harm.
That framing matters because ordinary employment mistakes and police hiring mistakes do not occupy the same moral or political register. A weak hiring process in a private office may produce inefficiency, poor morale, or financial loss. A weak hiring process in policing can produce bodily injury, coerced encounters, unlawful force, unconstitutional search and seizure, discriminatory enforcement, sexual abuse under color of law, false charges, taxpayer-funded settlements, and enduring erosion of institutional legitimacy. Page 14 captures this by moving the discussion out of the language of HR and into the language of public consequence. Once that shift is made, the tolerance for discretionary looseness should change as well. A state actor with authority to detain, arrest, use force, and shape constitutional reality on the street cannot be selected through a model that treats validated warning signs as merely one more topic for subjective discussion. The civic stakes are too high.
This is why the solution proposed in the earlier sections must be defended as public-law architecture, not just internal management reform. The transition to structured, tiered, evidence-based screening is not a matter of making hiring more efficient, although it may do that. It is a matter of making the state more disciplined in the way it delegates coercive power. That is the true significance of the move from impression to infrastructure. The profession is not simply choosing between two hiring philosophies. It is choosing between two different levels of seriousness about the risks it is willing to impose on the public. A discretionary, late-stage, compensatory model says, in practice, that a department may still gamble with warning signs so long as the file can be narratively defended. An actuarial, noncompensatory, auditable model says that some risks are too predictable, too severe, and too publicly consequential to be handled that way. Page 14 makes clear which of those positions the solution demands.
The civic dimension also helps explain why the proposed reforms cannot be left to local preference alone. When a department underuses predictive indicators in an ordinary workplace, the costs are borne largely inside the organization. When a police department does so, the costs spill outward—to the person stopped on the street, the community subjected to repeat misconduct, the family harmed by a foreseeable abuse of force, the city that later pays settlements, and the public that is told again that the incident was an unfortunate exception. That outward spill is what makes police hiring reform a matter of governance rather than simple administration. The state is not free to treat delegation of force the way a private employer treats hiring for a generic position. Page 14’s phrasing—“not just an HR issue”—should therefore be read as a warning against minimization. It insists that the institution not reduce constitutional and public-safety stakes to the language of internal process management.
This framing also clarifies why the original thought-piece’s critique of damage control matters. A department that acts decisively only after harm has occurred is not merely inefficient. It is abdicating a civic responsibility. If validated warning signs can be identified before appointment, then failing to act on them is not just a bad internal choice. It is a failure to protect the public from foreseeable misuse of state authority. That is why the slide’s final positioning is so strong. It forces the reader to see that predictive hiring discipline is not bureaucratic overreach; it is the minimum seriousness required when the consequence of error is not just a poor hire, but a dangerous public official. The proposed solution must therefore be defended in those terms. It is not about making departments look more modern. It is about whether the state will continue to hand force-bearing authority to candidates through a model it already knows is too subjective, too permissive, and too structurally forgiving of warning signs.
Page 14 also performs another crucial function: it answers the instinct to trivialize the architecture debate as a dispute among specialists. The earlier sections discussed temporal sequencing, tiering, actuarial logic, professional role allocation, and audit trails. Those concepts can sound technical if left standing alone. Page 14 translates them back into public meaning. The reason sequencing matters is that delay lets risk survive to appointment. The reason noncompensatory rules matter is that some red flags correspond to public harms that should not be risked. The reason role allocation matters is that the wrong decision-maker in the wrong place can convert predictive evidence into narrative reassurance. The reason audit trails matter is that the public needs to know whether the institution actually followed the model it claims to use. In that sense, the “civic crisis” frame does not replace the technical architecture. It explains why the architecture matters.
There is also a democratic accountability point embedded in the slide. When a state delegates force, it does so in the name of the public. That means the standards governing entry into that role cannot be treated as purely internal matters for professional guilds to manage without scrutiny. The objective model proposed in the deck is therefore not only more predictive; it is more democratically legible. It makes visible what kinds of warning signs matter, how they are treated, and why. A discretionary system, by contrast, hides behind professionalism. It tells the public to trust that the right people made the right judgment. Page 14 implicitly rejects that posture. It says the consequences are too profound for that level of deference. The public is entitled to a model that is structured enough to deserve trust rather than merely request it.
So the final section in this group should close on a clear point. This is not just an HR reform because police hiring is not just an HR decision. It is the state deciding who will carry its authority into the lives of civilians. The proposed solution—moving prediction upstream, making the model noncompensatory, separating Tier 1 from Tier 2, reducing discretion by design, reallocating professional roles, and building an audit trail—is therefore not internal process tinkering. It is a civic safeguard against foreseeable abuse of public power. Page 14 says the issue plainly: the delegation of force demands predictive discipline, not discretionary leeway. That is the line that should govern the whole follow-up, because it is the line that makes clear why the reform model cannot be treated as optional managerial improvement. It is the minimum architecture a constitutional democracy should require before it hands someone the badge.
IX. Conclusion: The State Must Stop Hiring Through Predictable Risk
By this point, the institutional choice is no longer difficult to describe. The evidence exists. The warning signs are not speculative. The current structure has already shown what it does with them. And the proposed replacement model has already been mapped. The state therefore stands at a point where continued reliance on the existing hiring architecture cannot honestly be framed as uncertainty, caution, or professional restraint. It is a choice to continue delegating force through a system that remains too discretionary, too late, too compensatory, and too structurally willing to convert predictive evidence into discussion rather than exclusion. That is the central conclusion of this follow-up. The problem is no longer simply that police hiring has been built around impression. The problem is that the state now knows enough to stop doing so and has not yet rebuilt the gate accordingly.
The slide deck leaves little room for institutional ambiguity on that point. Page 3 established the essential failure: departments possess data that significantly predict later misconduct, yet use those warning signs only weakly at the point of hiring. Page 4 identified the profession’s “category error,” showing that the current system continues to organize itself around clinical impression and generalized suitability rather than behavior-specific predictive evidence. Page 11 then moved from diagnosis to architecture, showing that prediction must be moved upstream through standardized background questionnaires, empirical indicator screening, and verification before any bounded clinical review occurs. Pages 12 and 13 refined that architecture into a full objective model: item-level specificity, tiering, calibration, reduced discretionary latitude, structured reasoning, and actuarial decision rules. Page 14 then stated the civic meaning of the entire exercise: this is not just an HR issue but a crisis in how the state delegates coercive authority. Taken together, those pages form a single argument. The existing model is not merely imperfect. It is no longer defensible as the primary gatekeeping structure for entry into policing.
That conclusion matters because states often preserve harmful institutions by refusing to admit that design itself is the problem. They prefer to frame failure as a matter of administration, compliance, training, or inconsistent execution. But the architecture presented in the deck does not support that evasive move. The problem is not simply that departments sometimes mishandle the existing system. The problem is that the existing system gives them too many ways to mishandle it. A discretionary, late-stage, compensatory model is built to absorb predictive warning signs into narrative balancing. It allows serious facts to become generalized concerns. It allows broad impression to outrank behavior-specific evidence. It allows hiring managers and evaluators to preserve flexibility where public safety requires structured discipline. So long as that design remains in place, the profession will continue to do what it has already done: gather warning signs, soften warning signs, and later confront the consequences as if they appeared only after the badge was issued.
That is why the title of this concluding section must be taken literally. The state must stop hiring through predictable risk. Not “become more careful around” predictable risk. Not “weigh” predictable risk more thoughtfully. Not “consider” predictive indicators more seriously while preserving the same discretionary logic. The force of the slide deck is that it makes those softer formulations inadequate. If a candidate presents validated indicators that place him inside Tier 1 population-screening territory, the institution must respond systematically. If a candidate presents Tier 2 critical red flags, the institution must stop the process. If a department cannot explain, through a defensible audit trail, why a candidate with meaningful indicators was nevertheless advanced, then the system has not solved the problem. It has only renamed it. The objective architecture is therefore not a recommendation to be selectively adopted. It is the minimum form of institutional seriousness that the evidence now demands.
The constitutional and civic stakes of that point cannot be overstated. Police hiring is not generic personnel management. The state is deciding who will carry firearms, exercise physical restraint, control movement, write reports with legal consequences, influence prosecutions, interpret compliance, escalate encounters, and embody public authority in some of the most coercive moments a civilian may ever face. In that context, preventable hiring error is not an internal management problem with internal consequences. It is an externalized public harm. It is borne by the person stopped, searched, struck, disbelieved, overcharged, injured, or killed. It is borne by neighborhoods that experience repeated abuse and are then told that each incident was an unfortunate breakdown rather than a foreseeable institutional choice. And it is borne by the public fisc when municipalities later pay to resolve claims that the state could have reduced through more disciplined front-end screening. Page 14 is correct to insist that this is a civic crisis, not just an HR issue. Once the state delegates force, every weakness in the hiring architecture becomes a weakness in constitutional governance itself.
This is also why the solution cannot be reduced to a single technique. The deck does not propose one magic screening instrument. It proposes a state of institutional discipline. Prediction must be moved upstream. The model must become noncompensatory. Tier 1 and Tier 2 must remain distinct. Discretion must be constrained by design. Professional roles must be properly allocated. The decision path must be recorded in an auditable way. These are not separate reforms that can be adopted piecemeal without consequence. They are interacting parts of a single gatekeeping philosophy. Remove one and the others become easier to evade. Leave prediction late and the clinical stage absorbs too much authority. Keep the model compensatory and Tier 2 red flags become narrative bargaining chips. Preserve broad discretion and the same warning signs will again be neutralized through impression. Omit the audit trail and the institution will later deny what it knew or how it acted. The deck’s proposed architecture works because it is cumulative. It understands that predictable risk survives where structure leaves room for it to survive.
A state that continues to rely on the current model after this diagnosis can no longer claim to be operating in good-faith uncertainty. It may still choose convenience. It may still choose staffing flexibility. It may still choose local discretion over standardized discipline. But it will be choosing those things with knowledge of what the design permits. That is what changes once a profession has both the evidence deficit diagnosis and the replacement architecture in hand. Before that point, the institution could plausibly say it was navigating a difficult problem with imperfect tools. After that point, the institution is deciding whether it prefers discretionary freedom to predictive responsibility. The slide deck makes clear what that preference means in practice. It means continuing to entrust force-bearing authority to a process that allows severe warning signs to be softened, postponed, or absorbed. It means continuing to act most decisively only after harm has already ripened into complaint, lawsuit, or scandal. In short, it means continuing to hire through predictable risk.
The profession therefore needs a sharper moral vocabulary than it usually permits itself. Weak screening is not just inefficiency. It is not just inconsistency. It is not just an unfortunate lag between research and practice. At this point it is a form of institutional indulgence toward risk the state has no right to indulge. A department does not become more humane by allowing a severe warning sign to survive inside a generalized “whole person” narrative. It becomes more permissive toward foreseeable public harm. A jurisdiction does not become more practical by preserving broad local discretion after knowing that predictive evidence is being underused. It becomes more willing to let variability in gatekeeping determine variability in public exposure to abuse. And a state does not become more thoughtful by leaving the primary predictive burden on late-stage professional impression. It becomes more willing to confuse procedural sophistication with preventive competence. The slide deck’s solution model exists precisely to strip those indulgences of their moral camouflage.
So the closing point is direct. The state must stop hiring through predictable risk because it no longer has the excuse that the risk is invisible, unknowable, or administratively intractable. The evidence is visible. The predictive categories are knowable. The architecture for action has been laid out. What remains is whether the state will accept the political and institutional consequence of acting on what it now knows. That means building the gate the evidence demands, not merely improving the language around the gate already in place. It means replacing broad suitability narratives with predictive discipline. It means using Tier 1 for systematic population screening and Tier 2 for hard exclusions. It means forcing decisions into a structure that can later be audited and defended. And it means recognizing, finally, that when the state delegates force through a process it already knows is too permissive, the resulting harm is not just the officer’s failure. It is the state’s own.
The ultimate reform question, then, is not whether police hiring can be improved. Of course it can. The real question is whether the state is prepared to stop treating predictable risk as administratively manageable and start treating it as disqualifying when the evidence says it should be. Until that happens, every later promise of accountability will remain compromised by the same prior fact: the warning signs were there, the structure was known to be weak, and the state handed over the badge anyway.
Read the Slide Deck
About the Author
Eric Sanders is the owner and president of The Sanders Firm, P.C., a New York-based law firm concentrating on civil rights and high-stakes litigation. A retired NYPD officer, Eric brings a unique, “inside-the-gate” perspective to the intersection of law enforcement and constitutional accountability.
Over a career spanning more than twenty years, he has counseled thousands of clients in complex matters involving police use of force, sexual harassment, and systemic discrimination. Eric graduated with high honors from Adelphi University before earning his Juris Doctor from St. John’s University School of Law. He is licensed to practice in New York State and the Federal Courts for the Eastern, Northern, and Southern Districts of New York.
A recipient of the NAACP—New York Branch Dr. Benjamin L. Hooks “Keeper of the Flame” Award and the St. John’s University School of Law BLSA Alumni Service Award, Eric is recognized as a leading voice in the fight for evidence-based policing and fiscal accountability in public institutions.