Georgia's Tier Segregation System

The ladder: Tier I, Tier II, Tier III and the STEP step-downs

Georgia's Tier Segregation Management System was implemented in August 2013, according to the Department of Corrections' Tier Management Fact Sheet, as "a managed and measured stratification-oriented pathway" out of administrative segregation. The department's FY2014 Annual Fiscal Report records the only system-wide quantification it has ever published: "2,721 beds were converted to Tier beds" at all Level V facilities. That is a bed count from the system's construction, not a population, and it supplies the meaning of the phrase "Level 5 facility" that appears, undefined, in the Tier II procedure.

Tier I is governed by SOP 209.07, effective April 30, 2015, and applies by its own terms to "All State Prisons, County CIs, Private Prisons, Intensive Treatment Facilities, Transitional Centers, Probation Detention Centers, and Probation Boot Camp." It lists eight placement grounds, and only one — awaiting a disciplinary hearing — presupposes a charge; the rest include being "noted as a threat to the safe and secure operation of the Facility," being the subject of a criminal investigation, and simply having just arrived and not yet been classified. Confinement is capped at 30 days "to include investigation and disciplinary isolation time," with one exception: voluntary and involuntary protective custody, which carries no time limit at all. Exercise is guaranteed at five hours per week, at least one hour a day on five days.

Tier II is administrative segregation under SOP 209.08, effective April 11, 2016. The Fact Sheet quoted in at least five PREA audits describes it as "Administrative Segregation" with twelve eligibility criteria, an Administrative Segregation hearing within 96 hours, three phases, and 90-day Classification Committee reviews; audits use the two terms interchangeably. The procedure's own definition calls it a "Long Term Administrative Segregation stratification plan." Section IV.A.3 provides that people who "cannot return to general population because they pose a serious threat" remain in Phase 3 and are "reviewed every 90 days to determine assignment status." There is no maximum duration anywhere in the document.

Tier III is the Special Management Unit at Georgia Diagnostic and Classification Prison, governed by SOP 209.09 in a version effective June 18, 2026. The program is "a minimum 13 month (390 days)" — 60 days in E Wing, 60 in F Wing, 90 in D Wing, 90 in C Wing, 90 in B Wing, across five phases from most to least restrictive. Section IV.A.1.f states that people "shall not be held in the Tier III Program for more than 24 months," then reserves the right to retain them longer under six criteria including murder committed while incarcerated, escape, hostage taking, and a catch-all for someone who poses "such an exceptional, credible, and articulable risk" that no other facility will do. Every Tier III cell is single occupancy by policy.

The STEP programs are the step-downs. SOP 209.45 (Tier II STEP) and SOP 209.55 (Tier III STEP) were both written on April 25, 2019, both run twelve pages, and are textually near-identical: four hours out of cell per day seven days a week, one hour of exercise, 15-minute call limits, two-hour non-contact visits, 30-minute documented checks, general-population privileges. They diverge on one point. Tier II STEP says "Cells may be single or double occupancy." Tier III STEP says "All cells are single occupancy." Both state affirmatively that the STEP program "is not a restrictive housing unit" — a definitional carve-out that matters, because it will exclude the step-down population from any GDC production framed on the term "restrictive housing." GDC policy also ties placement to classification by definition: SOP 209.09 states that "All offenders assigned to the Tier III Program are classified as Close Security," and both STEP procedures say the same of their populations.

Placement, review and release: a one-way valve

The procedures build an asymmetry into the system. A facility can put a person into Tier II on its own authority — the Classification Committee recommends, the Warden approves within seven business days, and "in emergency situations, the Warden/Designee may authorize the immediate assignment" — but it cannot take a person out. Release under SOP 209.08 Section IV.F.5.c runs from the Tier II Unit Manager to the Warden and Regional Director, then to a Criminal Investigations Unit Security Threat Group check with ten business days to complete, then to the Director of Field Operations, "who will make the final decision," and finally to a statewide Tier Segregation Manager who reviews the file in SCRIBE and "determine[s] the appropriate placement." The same SOP names that statewide position, which necessarily holds or can produce a system-wide view of Tier placements.

The admission criteria are broad on their face. Criterion 1 in both the SOP and the Fact Sheet covers an incarcerated person "noted as a threat to the safe and secure operation of the Facility," and expressly includes someone who "ha[s] been threatened with bodily harm" — a victim, not an aggressor. Criterion 5 is "Failure in the Tier I program or refusal to participate," so declining a program is itself a route into indefinite segregation. Criterion 11 makes "Transfer from Georgia Diagnostic and Classification Prison" a standalone ground. None of criteria 1, 5 or 10 requires a disciplinary conviction. The system also has a documented intake from below: SOP 209.01, the discipline policy, caps disciplinary isolation at 30 consecutive days in prisons and 14 in detention and treatment facilities, and provides that where isolation beyond 30 days is "necessary to control and encourage behavioral change, a recommendation for placement of the offender in the Tier II program is warranted." Tier II is written into the discipline policy as the route around the sanction cap, in a section whose premise is a finding of guilt.

Review is where the asymmetry hardens. The appeal of an initial Tier II assignment goes to the Director of Field Operations. The appeal of the 90-day review — the decision that continues confinement — goes only to the Warden, who under Section IV.C.4 is the official who approved the placement in the first instance. Every subsequent 90-day decision to keep a person in indefinite administrative segregation is therefore reviewed by the same office that authorised it, with no route to Central Office. In Tier I the first 24-hour status decision "will not be subject to appeal by the offender"; the 96-hour hearing follows; and the Warden's decision on appeal "is final." At no point in either procedure is there a neutral or external decision-maker, a right to counsel, or a non-staff advocate.

Exit runs through a program that may not be available and back through Tier I. SOP 209.08 requires that the person "must be actively participating in the O.U.T. Program (Offenders Under Transition)" and, "if appropriate," must complete it before release, and then assigns them to Tier I for up to 30 days "as part of their reorientation to a less strict environment." The SOP contains no provision for what happens if the O.U.T. program is not offered at the facility. There is a further procedural trap: SOP 227.02 makes "involuntary assignments to Administrative Segregation" non-grievable, requires that a grievance "must not be placed in the Offender's file or referred to in SCRIBE case notes," and yet SCRIBE case notes are among the records SOP 209.08 requires the Unit Manager to review when deciding release. The one record that could show a grievance preceded a Tier placement is, by policy, absent from both the file and the release review.

Conditions: the arithmetic of a cell, and the question policy will not answer

The out-of-cell floors are stated plainly and they are low. SOP 209.08 Section IV.D.1.f offers "a minimum of five (5) hours per week of exercise," outside the cell unless security, safety or weather dictate otherwise, with no daily distribution specified — the five hours may lawfully arrive in any pattern. SOP 209.07 sets the same five hours for Tier I but distributes it across five days, leaving two days a week on which the policy guarantees nothing. Tier III (SOP 209.09) requires four hours out of cell per day, Monday through Friday, which "may consist of any of the following or any combination thereof" including table time, recreation, programming and visitation, and expressly excludes legal visits and weekend visitation from being counted against it. Both STEP procedures guarantee four hours a day, seven days a week — though the Tier II STEP text counts "shower, feeding ... etc." toward the four hours.

Reduced to a weekly arithmetic, the floors are 163 hours in cell per week, or 23.3 hours a day, for Tier I, Tier II, generic administrative segregation and the juvenile program; 148 hours, or 21.1 hours a day, for Tier III on weekdays, rising to 24 hours on Saturdays and Sundays; and 140 hours, or 20.0 hours a day, for both STEP programs. The consequence is counter-intuitive and documented: the step-downs guarantee 5.6 times more out-of-cell time than Tier II itself, and Tier III — the unit GDC labels most restrictive and the only one reformed by litigation — guarantees four times more than Tier II. A person in Tier II Phase 1 or 2 also has no tablet: SOP 204.10 confiscates the device on placement and returns it only at Phase III, for education and programming only, alongside a single monthly video visit.

Whether any of this is solitary confinement in the sense the UN Standard Minimum Rules use turns on cell occupancy, and GDC does not specify it for Tier II. Mandela Rule 44 requires two elements — 22 hours or more a day in cell, and the absence of meaningful human contact. The first is satisfied by arithmetic from the department's own floors. The second depends on whether the person has a cellmate, and SOP 209.08 is silent. SOP 209.06 goes further in the other direction: it lists single-celling in administrative segregation as an enumerated exception requiring assignment by medical, mental-health or classification staff plus evaluation within 24 or 48 hours, which implies that multiple occupancy is the ordinary condition. Two modes therefore run inside the same program under the same rules, and the policy distinguishes, guarantees and reports neither.

The audit record shows both modes running. Hancock State Prison's G-1 "Tier 2" unit houses "a maximum of 96 inmates, in double occupancy cells located on two tiers," in identical words across two audit cycles; a separate Hancock building holds 20 double-occupancy Tier 1 cells. Georgia State Prison's G Building held "up to 218 offenders in double bunk cells" while its K Building held 100 in "single cell lockdown cells" — the same program, both modes, one prison. At Hays, Tier I units were two-man cells and Tier II units were one-man. Smith State Prison's Cycle 4 audit describes 48 double-bunked cells housing 96 offenders per dorm. Telfair's E-1 housed 80 in two-man cells. Unit capacities across the audits run from 20 to 218, with no standard dorm size.

Supervision and clinical attention follow the same reduced pattern. SOP 209.08 requires health care rounds three times per week "excluding weekends and holidays," and excuses the Chief of Security, Unit Manager and Deputy Warden from daily visits on the same days; only the Institutional Duty Officer is required seven days a week. That 2016 standard conflicts with the Health Services SOP 507.04.33, effective 2022, which requires a daily visit and daily rounds in restrictive housing — and GDC has not harmonised the two, leaving the older, less protective standard written into the policy that governs the unit. SOP 219.01 requires that people in Tier I "should be asked" about suicidal thoughts once a month and those in Tier II during weekly rounds: the unit with the shorter maximum stay carries the less frequent screening, and the verb is permissive. Most sweeping, SOP 209.08 Section IV.D.1.l provides that "all of the above conditions or privileges are subject to modification due to extenuating operational circumstances" as deemed necessary by the Warden or designee — an undefined term with no time limit, no documentation requirement and no reporting obligation. In 2025 the department's own auditors found that people "assigned to the Tier II Program" and other lockdown units "are not being afforded a minimum of five (5) hours per week of exercise outside the cells," a finding verified against door sheets.

Scale: the one population GDC publishes, and the floor it does not

Georgia publishes a Tier III population almost by accident. The Special Management Unit at Jackson appears as a facility row in the monthly Inmate Statistical Profile series, and as of September 1, 2026 it held 154 people; across 75 monthly snapshots from July 2020 to September 2026 the figure ranged from 149 to 195. Against a design capacity of 192 single-bunked cells across six cellblocks — a figure that appears in GDC's FY2017 annual report and in its own facility description — the unit has run below capacity throughout the published series. Because SOP 209.09 requires single occupancy, cell count and bed count coincide there, which is true nowhere else in the Tier system. A separate October 2026 bed table produced under open-records request R031710-091426 gives the Special Management Unit 160 Tier III beds with 128 occupied.

That same table is the first per-facility count of the Tier II population GPS holds from the department, and it is the most important document in this record. It lists 884 Tier II beds across eight prisons, 804 occupied: Hancock 83, Hays 151, Macon 121, McRae Women's Facility 34, Smith 111, Telfair 117, Valdosta 129, Ware 138. Four were at or within one bed of full — Hays 151 of 151, Telfair 117 of 117, Ware 138 of 138, Hancock 82 of 83. It lists 119 STEP beds across four prisons (Georgia Diagnostic, Hays, Macon, Smith), 96 occupied, without splitting Tier II STEP from Tier III STEP. Combined, the three programs come to 1,163 beds and 1,028 occupied people on a single snapshot the agency did not date. Hays alone holds 151 Tier II beds plus 59 STEP beds, 210 in all, every one occupied — the largest restrictive-housing program outside the Special Management Unit. The table also settles a question the department's public fact sheet leaves open: GDC's Tier Segregation System fact sheet, in the version produced three days after the bed table in the same request, still names only seven men's prisons. McRae Women's Facility is not on it, and McRae has 34 Tier II beds. The Tier II program operates in a women's prison and the public description of the system omits it.

The table's own scope limit must travel with it: it covers Tier II, Tier III and STEP only, and GPS's request excluded Tier I, disciplinary isolation and protective custody. The 1,163 beds are therefore not a total of Georgia's segregation beds and must not be compared directly with the 2,721 beds converted in 2014 or with audit-reported cell counts, which include Tier I. Nor do they capture the whole system. In 2025 the Southern Center for Human Rights, citing a GDC Tier Management Bed Report obtained by open records, reported that more than 2,100 people were in Georgia's "tiers," including almost 1,000 in the long-term tiers — a figure GPS cannot reconcile with the October 2026 table.

The PREA corpus is the only per-facility dataset GDC publishes across the whole system. Scanning all 273 published facility audits, GPS found named Tier program housing in the audits of exactly nine facilities — Georgia State Prison, Smith, Hays, Valdosta, Telfair, Macon, Hancock, Georgia Diagnostic and Classification, and Ware — every one a close-security prison, with no exception in 273 documents. But the Tier label marks where the department concentrates long-term administrative segregation programming, not where segregation cells exist. Calhoun State Prison, whose own audit gives its level as "4/Medium," reports 197 segregation cells — more than close-security Baldwin (167), Ware (125) or Phillips (75). Wilcox, described in its own audits as housing medium-security general population, reports 168. Coffee, a private medium-security prison, reports 170. Dooly, medium security, reports 71 cells and its Cycle 4 audit describes a housing unit of "96 inmates in double bunks" designated for administrative segregation, isolation and mental health observation. Summing the most recent audit figure per facility gives 3,401 segregation cells across 30 facilities — a floor, not a total, and a cell count, not a bed count and not a population.

The historical Tier figures must be read as historical. Georgia State Prison's Cycle 2 audit itemises 13 Tier program dorms, 106 double-bunk Tier I cells at E-South, 218 Tier II beds at G Building, 100 single-cell lockdown Tier II beds at K Building and 49 "Tier II Mental Health Offenders" — 424 stated Tier beds, more than twice the entire Special Management Unit. Hays' Cycle 2 audit itemises 303 Tier beds across five buildings. Georgia State Prison closed on February 19, 2022. From inside it, the Southern Center for Human Rights alleged in a September 2021 class action that approximately 300 people were in its Tier II program, over 70 percent experiencing serious mental illness — an allegation in a complaint, not a finding, and the only facility-level Tier II population figure in any source GPS holds. The department's own annual reports give two incompatible versions of the same reduction: FY2018 reports a 41 percent, 671-person drop in extended restrictive housing since January 2017; FY2019 reports 40 percent and 675 people from the same baseline. Neither states the population before or after. After FY2020 the subject disappears from the annual reports entirely.

Two tracks: the litigation reformed the smaller program

Gumm v. Jacobs was filed pro se on February 12, 2015 by a man held in the Special Management Unit at Georgia Diagnostic and Classification Prison, and the Southern Center for Human Rights was appointed to represent him. The record that emerged was severe. The court adopted findings, based on Dr. Craig Haney's 2017 inspection, that residents were confined in cells "smaller than the average parking space" with as little as five hours a week out of cell, no outside light and no congregate programming; that 70 of the roughly 180 people then held were designated mentally ill; and that the named plaintiff had been held continuously for seven and a half years. Haney described the unit as "one of the harshest and most draconian" he had seen and "as chaotic and out-of-control as any such unit I have seen in decades." A 2017 survey figure puts the duration in perspective: 78 percent of the 182 people then in the unit — 141 of them — had been held more than two years, and the Southern Center reported an average of three to four years.

The 2019 settlement, approved May 7, 2019 with $425,000 in attorney fees, imposed a general 24-month limit on SMU confinement, mandatory 60-day and 90-day reviews with out-of-cell mental health evaluations by a licensed provider before assignment and at each review, a bar on housing anyone at mental health Level III or above in Tier III, minimum out-of-cell time, programming and reading material, and documentation and reporting requirements. GDC codified it into SOP 209.09 alone. That procedure now carries the 24-month ceiling, the mental-health exclusion, the 60-day and 90-day evaluations, and — added since — a quarterly review by a four-member statewide panel of the Deputy Director of Field Operations, the Statewide Mental Health Director, the Statewide Medical Director and a member of the Office of Legal Services, with personal approval by the Director, Field Operations or the Assistant Commissioner for Facilities required to hold anyone past 24 months. Compliance has been contested: a federal judge found in April 2024 that GDC violated the agreement, describing officials as "obstinately disobedient," appointing an independent monitor at the department's expense and imposing $2,500 a day in fines — $75,000 a month — until the terms were met. The court also found that the department had falsified records purporting to show that people were receiving regular reviews, citing a counselor who refused to sign forms for prisoners who were not present.

None of it reaches Tier II. SOP 209.08 has not been revised since April 11, 2016 — nearly three years before the settlement — and contains no duration cap, no mental-health exclusion, no panel, no mental-health evaluation requirement at placement and no out-of-cell minimum beyond five hours per week. The gap is written into the criteria. Tier III excludes both mental health Level III and Level IV; Tier II excludes only Level IV and admits Level III on mental-health recommendation. The effect is that people with the higher mental-health classification are routed away from the unit whose conditions were reformed and toward the one whose policy floor is five hours a week — and GDC's own clinical policy ranks Tier II among the highest-need settings it operates, grouping "SMHTUs/TIER II" with death row, the Special Management Unit and acute care, and assigning a 1:20 counselor ratio.

The legal consequence runs in the same direction. Wilkinson v. Austin, 545 U.S. 209 (2005), held that near-total isolation combined with indefinite duration and collateral consequences created a liberty interest requiring procedural protections. The factor Wilkinson treated as central — indefiniteness — is present in Tier II, where SOP 209.08 sets no outer limit, and absent from Tier III, where SOP 209.09 sets a 24-month ceiling with enumerated exceptions and a quarterly panel. On Wilkinson's own reasoning the due-process argument is stronger for the unreformed program than for the one that was litigated, though no court has applied it to Georgia's Tier II. The Department of Justice's October 2024 findings letter points at the broader failure: "GDC fails to control violence even in its segregated housing units and exposes incarcerated persons to an unreasonable risk of harm due to its inappropriate use of segregated housing," and the department "uses segregated housing as de facto protective custody, including for victims of sexual abuse."

The dead

The Tier system's harms are not only cumulative. Five people discussed in this record appear in GPS's death registry, and their registry entries are set out here as the registry states them.

• Sherry Elaine Joyce died April 27, 2024, at Arrendale State Prison; the registry lists the cause as Homicide. She was 61. She and Hallie Reed were strangled in the same mental health unit eight days apart, according to reporting in GPS's women's incarceration collection.
• Hallie Marie Reed died May 5, 2024, at Arrendale State Prison; the registry lists the cause as Homicide. She was 23. Reporting in the same collection states that she had requested protective custody after Joyce's death and was denied.
• Quintez Deshon Smith died August 29, 2022, at Hays State Prison; the registry lists the cause as Homicide. He was 25. Contemporary reporting in GPS's Hays collection records the cause as multiple sharp force injuries.
• Ricardo Castavio Daughtry died June 10, 2024, at the Special Management Unit; the registry lists the cause as Unknown/Pending, meaning GPS does not yet know the cause. He was 41. A court record identifies him as the lead plaintiff in the Special Management Unit class action.
• Carlito Omar Soldiew-Acosta died June 22, 2024, at Dooly State Prison; the registry lists the cause as Homicide. He was 38. Reporting in GPS's Dooly collection states that he was found in his isolation cell and had been dead at least 24 hours, with the coroner's office reporting that foul play was suspected.

These names belong in the Tier page because segregation is where several of them were held, and because the department publishes no segregation-specific death or self-harm data at all. GDC has not disclosed suicide and self-harm incidents by tier and facility, mental-health classification breakdowns of the Tier II and Tier III populations, or length-of-stay distributions for restrictive housing. Any count of deaths in Georgia's Tier system would be assembled by GPS from individual records, not read off an agency report.

Reading the record: what the audits can support, and what would settle it

The PREA corpus is the best per-facility segregation dataset Georgia publishes, and it is internally inconsistent. Any figure drawn from it must name the facility, the cycle and the field, and must never silently reconcile a questionnaire figure against a narrative one. Georgia State Prison's "Cycle 3" audit is its Cycle 2 audit republished: both documents are 181 pages, both carry a final report date of March 22, 2019, and their extracted text is identical at 546,523 characters with zero differences — so its segregation figure cannot be counted twice. Telfair's Cycle 4 report states both that "the facility does not have administrative segregation cells" and that "the facility head reported having 360 administrative segregation beds." The same "does not have administrative segregation cells" form response appears in later audits of prisons whose earlier audits reported substantial capacity. Pre-Audit Questionnaire fields contradict the same audit's narrative in at least three cases: Georgia State Prison's questionnaire says 12 segregation cells while its narrative itemises 106 Tier I cells, 218 Tier II beds and 100 Tier II single lockdown cells. Some labels leave the unit of measure undetermined — Washington's "Number of segregation/isolation units 48," Dooly's "Number of segregation/isolation 71" with the noun missing. Cycle 1 audits are two-to-four-page summaries with no physical plant table. And a single sentence of policy boilerplate about delayed PREA education appears in roughly 60 audits, including audits of facilities with no Tier housing, so a keyword search for "Tier Program" produces a badly inflated facility list. The word "tier" itself usually means the upper or lower gallery of a cellblock.

The records that would answer the scale question exist. GDC's FY2019 Annual Fiscal Report states that a "TIER 2 SCRIBE module went live March 2019 to effectively track offenders in Extended Restrictive Housing," and its FY2020 report states that the SCRIBE Warden's Dashboard was extended to show, for Tier II offenders, "the number of days an offender has spent in restrictive housing Tier II." The department has said in its own published reports that the data exists in an electronic system of record. SOP 209.07 and SOP 209.08 both require per-cell, per-bed accountability logs recording "cell number, bed number, offender name, offender number, race, in or out time" — by their own description a daily record of who is in each Tier cell and bed. Board Rule 125-3-2-.10(f) requires Division Director review and approval for any isolation confinement beyond 30 days, generating a countable approval for every long stay. SOP 209.09 requires the SMU Warden to notify the North Regional Director in writing whenever out-of-cell time is cancelled or shortened for three or more consecutive days — the only paper trail in the Tier corpus of actual out-of-cell denial, and one with no equivalent in Tier I or Tier II.

Why none of this becomes a published number is a question of policy design rather than data availability. A review of the 209 series found no procedure naming any Tier roster, periodic report or Central Office aggregate submission; the term "Central Office" does not appear in SOP 209.08 at all. The Statewide Classification Committee exists only for transgender and intersex housing decisions. Every Tier approval instead travels upward as a single case — committee to Warden to Regional Director to Director of Field Operations, with the Tier Segregation Manager placing individuals one at a time. GDC's reporting on the restrictive-housing population stopped after FY2020. And the structural reason the Tier population cannot be recovered from published statistics is that the monthly Inmate Statistical Profile reports security classification, not housing status: because every Tier III, Tier II STEP and Tier III STEP placement is classified Close Security by policy, the Tier population sits inside the published Close Security count and cannot be isolated from it. The Tier population is not merely unpublished; on the published structure it is arithmetically unrecoverable.

One caution applies to any request built on this record. SOP 219.01 asserts that "Offender records are confidential state secrets" under O.C.G.A. § 42-5-36(c) and routes all requests for offender records to the Office of Legal Services. A request framed around individual records will meet that exemption; a request framed around aggregate counts, unit rosters stripped of names, and the designation and capacity documents will not. The single most load-bearing missing document is the current designation list — which facilities are Tier I, Tier II, Tier II STEP, Tier III and Tier III STEP today. GPS holds the 2018 and 2019 fact sheets and the undated Close Security Facility Fact Sheet, but no current edition, and material changes have occurred since: Georgia State Prison closed in 2022, SOP 209.09 was revised in 2025 and again in 2026, and SOP 508.01 refers to "designated SMHTU/TIER II facilities" without naming them.