Houston City Council: expanded roll calls

Technical appendix · 2000–2026 pooled fit and 2021–2026 check

Author

Boris Shor

Updated

October 12, 2026

This appendix documents the expanded Houston report. It presents 2000–2026 pooled positions, with tabs for each elected council on a common axis. A separate 2021–2026 fit provides a within-period check. The earlier 2021–2025 technical appendix describes the baseline report.

Source and analysis matrix

The input is Jacob’s October 12, 2026 package, houston_rollcalls_2000_2026_2026-10-12.zip, preserved locally under Jacob/2026-10-12-houston/original/ (SHA-256 4a905676a80a12b59942ca743338c153adf51868219f95c3fe3b0dc105f5d021, matching the author’s hash). It contains 39,764 actions and 631,078 member rows. Of these, 2,081 actions are contested, and 1,113 of those have a one-vote minority. The package no longer flags actions as scaling-eligible; retention is decided here. Code/61_prepare_houston_2026_10_12.r keeps every contested action with at least one yea and one nay among fitted members. It codes 1 for yea and 0 for nay, with absences, departures and omissions missing. Missing votes are never recoded as nays.

Compared with the September 25 package, this version adds 307 actions from September 2010 to July 2011 meetings, 59 of them contested. It also splits a merged member string, “Wanda Adams. Mike Sullivan”, into the two members it combined. Six uncontested procedural motions reappear under renumbered keys. All 31,909 vote cells on actions contested in both versions are unchanged, and the hold-out files are byte-identical. A bounded source check covered twelve seeded contested actions (eight newly added, four formerly merged) against the official minutes, and all twelve were supported. The audit record documents both checks.

Five actions previously verified against official minutes are already in this ZIP and are not appended twice. Three retain analysis-side corrections against saved page transcriptions: ORDINANCE 2024-0444 and MOTIONS 2025-0781 and 2025-0855. These three remain uncorrected in the October 12 package. Changed cells are in Results/Houston-2000/verified_overrides.csv; the ZIP is unchanged. Against the 2003–2020 hold-out, the current parser gets all 35 corrected yea totals and all 36 nay totals; the separately supplied 2000–2003 hold-out has 23 correct emitted full vectors and three explicit refusals rather than guessed results. See the earlier package review.

Following the minimum-observation rule in the State Ideal workflow, the matrix requires seven qualifying yea/nay votes per person. Only two one-vote name variants fall below it. All 2,081 contested actions still have a yea and a nay after the cutoff. The matrix has 77 members, 2,081 actions and 30,633 observed binary votes. The September 25 fit, which required two votes on each side, used 939 actions and 13,785 votes.

Excluded name	Qualifying votes
Carolyn EvansShabazz	5
M. J. Kahn	1

The separate modern subset covers 2021–2026 and contains 25 members, 140 actions, and 2,198 observed votes. It includes Mayors Sylvester Turner and John Whitmire and the ten council members who served across their terms.

Model and identification

The one-dimensional pscl::ideal() model is $\Pr(Y_{ij}=1)=\Phi(x_i\beta_j-\alpha_j)$ , where $x_i$ is a fixed member position, $\beta_j$ an item’s discrimination, and $\alpha_j$ its difficulty. Each person has one position within a fitted window. normalize=TRUE identifies location and scale. Display signs place Republican-affiliated members on the positive side. This is a sign convention, not an external ideological constraint.

The modern check uses two freshly initialized chains. The reported pooled fit (Code/62_fit_houston_party_start.r) uses exactly two chains, both started from researched party affiliation: Democratic −1, Republican +1, and 0 for Gabriel Vasquez, who switched parties while serving. This is the starting rule used for every city in the project. Labels set starting values only. They never enter the votes, likelihood, priors or post-fit positions, so the party comparisons in the report remain external checks. A start is mode selection, not proof of identification; see the earlier diagnostics below.

Each chain runs 500,000 iterations after 10,000 burn-in, thinned by 500, keeping 1,000 draws. A first run at the earlier length (50,000 iterations, thinned by 50) mixed too slowly at the two lone-dissent endpoints and among early Republican-affiliated members: maximum R-hat 1.135 and minimum ESS 21. Its estimates matched the longer run, which is the one reported. Chains, seeds and session information are saved under Results/Houston-2000/party-start-two-chain/.

Diagnostics

Fit	Members	Actions	Maximum R-hat	Minimum ESS
Modern, 2021⁠–⁠2026	25	140	1.010	191
Pooled reported fit, 2000⁠–⁠2026	77	2,081	1.040	192

These are classical univariate R-hat and coda effective sample sizes, not rank-normalized split R-hat or bulk/tail ESS. The two pooled chains’ member means correlate 0.9999 overall and at least 0.9999 within every elected council. The pooled scores correlate 0.954 with the independent 2021–2026 fit on its 25 members. These checks describe stability within the selected mode; they do not prove global identification.

Lone-dissent endpoints

Including one-vote minorities places Jolanda Jones (-5.05) and Helena Brown (5.87) far beyond every other member. They cast the sole minority vote on about 31% and 79% of their qualifying votes, respectively. Because normalize=TRUE fixes the overall spread, these endpoints compress everyone else toward zero relative to the September 25 fit. The ordering is little changed: rank correlation with the September 25 scores is 0.985, against a Pearson correlation of 0.757. The two endpoints also have the smallest effective sample sizes.

Earlier identification diagnostics (September 25 package)

Fits to the September 25 package showed that the 26-year scale is not globally identified by the votes alone. These diagnostics were not repeated on the October 12 matrix. A chain started from the leading eigenvector and a chain started from party labels reached the same solution. A chain from a dispersed random start reached a different mode, preserving the early councils’ ordering but reversing the modern end. Its chain means correlated with the eigen-start chain by period as follows:

Period	Members	Correlation between two chain means
overall	77	0.2614
2000⁠-⁠2001	15	0.9985
2002⁠-⁠2003	15	0.9988
2004⁠-⁠2005	15	0.9993
2006⁠-⁠2007	16	0.9987
2008⁠-⁠2009	16	0.9548
2010⁠-⁠2011	15	0.7962
2012⁠-⁠2013	18	0.9064
2014⁠-⁠2015	17	-0.9000
2016⁠-⁠2019	18	-0.9831
2020⁠-⁠2023	19	-0.9933
2024⁠-⁠2027	17	-0.9990

Pooling those eigen and random chains gave maximum R-hat 20.43. Two checks favored the party-consistent solution. It agreed with the independent 2021–2026 window, and it had slightly lower held-out log loss in three fixed-seed splits:

Split	Held-out votes	Party-consistent log loss	Alternative mode log loss
Split 1	2,156	0.4343	0.4445
Split 2	2,175	0.4211	0.4275
Split 3	2,158	0.4229	0.4283

The margins were thin, and neither check eliminated the alternative orientation. The current fit’s party starts select the party-consistent mode by design. The historical ordering should therefore be read as the solution corroborated by the modern window and by external affiliation, not as a uniquely identified scale.

Pooled member estimates

Comparison with the September 25 fit

The October 12 matrix adds one-vote-minority actions and the 2010–2011 meetings, and its party-start fit is renormalized around the two lone-dissent endpoints. Score changes therefore reflect a rescaled axis, not individual ideological movement. For example, Mike Sullivan moves from +1.97 to +0.77, and Addie Wiseman from +1.23 to +0.90. Jolanda Jones moves from -1.15 to -5.05.

These are the 77 reported estimates plotted in the report. The five mayors are identified by name because the source’s role field labels some earlier mayors as council members. Values are specific to this pooled Houston fit.

Official	Mean	95% lower	95% upper	Qualifying votes
Jolanda “Jo” Jones	-5.05	-6.27	-3.93	311
Amanda Edwards	-1.69	-2.70	-1.01	111
Edward Pollard	-0.90	-1.17	-0.67	286
Ada Edwards	-0.81	-1.05	-0.62	665
Letitia Plummer	-0.59	-0.77	-0.42	257
Jew Don Boney, Jr.	-0.51	-0.76	-0.31	218
Carol M. Galloway	-0.50	-0.64	-0.37	641
John E. Castillo	-0.46	-0.68	-0.30	234
Lee P. Brown	-0.44	-0.58	-0.30	537
Tarsha Jackson	-0.43	-0.61	-0.28	221
Abbie Kamin	-0.36	-0.52	-0.21	278
Tiffany D. Thomas	-0.36	-0.53	-0.20	276
Wanda Adams	-0.33	-0.49	-0.19	501
Ellen Cohen	-0.33	-0.48	-0.20	506
Edward Gonzalez	-0.30	-0.42	-0.20	619
Richard Nguyen	-0.30	-0.50	-0.15	92
Sylvester Turner	-0.26	-0.42	-0.12	330
Adrian Garcia	-0.26	-0.40	-0.13	390
James G. Rodriguez	-0.26	-0.38	-0.14	609
Peter Brown	-0.25	-0.43	-0.10	312
Annise Parker	-0.25	-0.34	-0.16	1,139
David W. Robinson	-0.25	-0.37	-0.14	413
Karla Cisneros	-0.25	-0.41	-0.10	335
Jarvis Johnson	-0.22	-0.35	-0.10	478
Melissa Noriega	-0.21	-0.34	-0.09	597
Larry V. Green	-0.21	-0.34	-0.09	460
Carolyn Evans-Shabazz	-0.20	-0.35	-0.04	289
Bill White	-0.19	-0.31	-0.08	463
Martha Castex-Tatum	-0.17	-0.32	-0.03	319
Carol Alvarado	-0.17	-0.26	-0.07	669
Ronald C. Green	-0.17	-0.29	-0.06	447
M. J. Khan, P. E.	-0.16	-0.28	-0.05	450
Jerry Davis	-0.16	-0.28	-0.05	523
Gordon Quan	-0.15	-0.24	-0.05	666
Sue Lovell	-0.14	-0.25	-0.04	513
Andrew C. Burks, Jr.	-0.14	-0.38	0.05	295
Stephen C. Costello	-0.14	-0.25	-0.03	561
Dwight Boykins	-0.12	-0.24	-0.01	196
Al Hoang	-0.12	-0.27	0.01	462
Mike Laster	-0.11	-0.22	0.00	507
Robert Gallegos	-0.07	-0.18	0.03	399
Mario Castillo	-0.06	-0.28	0.15	80
C. O. “Brad” Bradford	-0.06	-0.17	0.05	603
Joaquin Martinez	-0.05	-0.25	0.17	77
Jack Christie D.C.	-0.03	-0.15	0.09	511
Sallie Alcorn	-0.03	-0.18	0.10	291
Gabriel Vasquez	0.00	-0.09	0.09	504
Toni Lawrence	0.04	-0.06	0.12	460
Oliver Pennington	0.07	-0.02	0.15	584
Dave Martin	0.08	-0.01	0.17	486
Carroll Robinson	0.08	-0.01	0.17	490
Anne Clutterbuck	0.12	0.03	0.20	501
Brenda Stardig	0.12	0.02	0.21	408
John Whitmire	0.16	-0.06	0.37	82
Pam Holm	0.19	0.10	0.27	449
Mark Goldberg	0.22	0.14	0.31	656
Willie Davis	0.23	-0.02	0.48	74
Steve Le	0.24	0.09	0.41	108
Chris Bell	0.25	0.15	0.34	232
Twila Carter	0.26	0.02	0.51	71
Mark Ellis	0.39	0.31	0.47	677
Rob Todd	0.40	0.29	0.50	199
Bert Keller	0.40	0.31	0.48	534
Amy Peck	0.42	0.30	0.54	297
Orlando Sanchez	0.43	0.32	0.53	223
Michael Berry	0.43	0.35	0.51	594
Julian Ramirez	0.51	0.27	0.77	77
Shelley Sekula-Gibbs, M.D.	0.53	0.45	0.62	584
Michael Kubosh	0.54	0.43	0.68	394
Mike Knox	0.59	0.47	0.72	316
Mary Nan Huffman	0.65	0.50	0.81	174
Fred Flickinger	0.72	0.49	0.97	79
Mike Sullivan	0.77	0.65	0.89	490
Addie Wiseman	0.90	0.79	1.01	645
Greg Travis	0.95	0.75	1.18	224
Bruce Tatro	1.68	1.39	2.06	567
Helena Brown	5.87	4.79	7.03	317

Elected councils and bridging

Houston’s regular terms were two years through the council elected in 2013 and seated in 2014–2015, then four years beginning with the 2015 election and January 2016 seating (2003 city almanac; 2016 financial report). The city’s council page specifies January 2 term starts, and the City Secretary’s election index supplies the election years. The report assigns actions by those term boundaries, not by equal-sized analytical windows. A special election changes who serves within an existing council term.

Council	Elected in	Term years	Members	Actions	Observed votes	Continuing from prior
2000⁠–⁠2001	1999	2	15	247	3,374	—
2002⁠–⁠2003	2001	2	15	328	4,647	10
2004⁠–⁠2005	2003	2	15	138	1,933	9
2006⁠–⁠2007	2005	2	16	214	3,029	11
2008⁠–⁠2009	2007	2	16	118	1,628	11
2010⁠–⁠2011	2009	2	15	202	2,799	9
2012⁠–⁠2013	2011	2	18	317	5,061	10
2014⁠–⁠2015	2013	2	17	93	1,449	11
2016⁠–⁠2019	2015	4	18	122	1,914	11
2020⁠–⁠2023	2019	4	19	218	3,521	10
2024⁠–⁠2027	2023	4	17	84	1,278	10

The new source contains substantial 2000–2003 coverage; the 2024–2027 term is observed only through May 2026. Adjacent councils share 9–11 observed members, but roster overlap alone does not guarantee enough discriminating votes to align the full scale. The alternative random-start mode found for the September 25 package shows that this alignment is sensitive to initialization. Earlier ad hoc window fits were exploratory diagnostics and are not used as report periods or current estimates.

Validation and remaining limits

Party affiliation was researched separately from the roll-call package. Code/30_houston_pooled_party_evidence.r maintains evidence for the original 67 officials; Code/38_houston_2000_party_evidence.r adds ten source-backed records and produces Results/Houston-2000/party_evidence.csv and scores_with_party.csv. Each record has a source and one of four descriptive evidence types — party_record, party_organization, self_identified, reported — which say what kind of source the label rests on and are deliberately not a ranking. A documented partisan candidacy or party office is thereby distinguished from an endorsement or news characterization; no vote pattern or district was used to assign party. Gabriel Vasquez’s documented Democratic-to-Republican switch is carried as two interval rows in party_affiliation.csv, one per affiliation, and appears as Mixed only in the person-level analysis table, where his single pooled position cannot be assigned to one party. The final inventory is 45 Democratic, 31 Republican, and one Mixed; two-party comparisons use 76 officials. The 77-person labels set both chains’ starting values, but no party label enters the likelihood or prior. The separate modern-subset comparison and presidential-vote relationship use pooled positions for the officials to whom those external data apply. The 2024 district overlay allocates precinct votes to districts in proportion to intersecting area. The report places the median of the five 2023-elected at-large members’ pooled posterior-mean positions at the approximate citywide Trump share; that marker is excluded from the district-only correlation and fitted line. Unanimous actions are excluded from scaling, so the fitted dimension pertains to a selected subset of business; actions with a single dissent are included.

The 2023 mayoral check is built by Code/44_houston_mayoral_districts.r from the City’s official runoff canvass, with all Harris, Fort Bend, and Montgomery candidate totals reconciled to their county subtotals. The official October 2022 crosswalk identifies twelve Harris precincts marked as crossing council districts; the table includes one zero-vote precinct absent from its prose appendix. The primary series excludes all twelve, uses the predominant overlap with the official 2023 council polygons for remaining Harris precincts and 2022G Fort Bend precincts, and leaves 111 Montgomery candidate votes unmatched. It assigns 197,676 of 201,528 candidate votes (98.09%). This is near-exact, not certified district-level tabulation: precinct and polygon vintages differ, and a few minor secondary overlaps remain. mayoral_precinct_audit_2023.csv flags secondary area shares of at least 1%; among unmarked precincts, only one candidate vote lies in a precinct with a secondary share over that threshold. An area-weighted allocation of the matched precincts is reported as a sensitivity check and assumes uniform voter density. Both variants compare Whitmire’s two-candidate district share with the pooled scores of the eleven district members elected in 2023, not the council that was still serving on runoff day. The local Data/Houston/mayoral_history/readme.md retains hashes, coverage, and the split list.

Candidate endorsement indicators

Code/31_houston_endorsement_validity.r joins four separate organizational indicators to the 77 fitted people using a normalized first-and-last-name key, with one reviewed Ed/Edward Gonzalez alias. The source ledger preserves the original candidate name, office, election date, general/runoff distinction, URL, note, fitted local ID when matched, and all 31 unmatched candidate-election rows. Nonmatches include challengers and a controller, not just identity failures. The ledger has 82 positive rows. Duplicate general/runoff endorsements remain distinct; “ever endorsed” counts distinct fitted people. 2013, 2015, 2019, and 2023 Caucus general slates are complete; the federation’s 2023 general slate is complete. The 2015 list is a contemporaneous reproduction, not the official list itself. TOP’s 2023 announcement names three council candidates and references its mayoral choice, but is not assumed to exhaust all possible endorsements. No verifiable Our Revolution Harris County Houston municipal candidate slate was recovered. Other cycles have positive evidence only or no recovered slate; unlisted people are unknown, never coded zero across all 77.

Organization	Verified candidate-election rows	Distinct fitted officials ever listed
Texas Gulf Coast Area Labor Federation	18	9
Houston LGBTQ+ Political Caucus	56	24
Texas Organizing Project / TOP PAC	8	2
Our Revolution Harris County	0	0

The within-election comparison uses the 17 fitted officials with a recorded qualifying vote in 2024, the first year after the 2023 election. Only the November 2023 general slates define these flags. “Outside” means a member of this 17-person cohort not listed on that recovered slate, not someone from an undocumented historical cycle. The point-biserial column is the correlation of the binary listed flag with the pooled score; a negative value means listed officials lean toward the negative end. The ordered-pair share is the fraction of listed–not-listed pairs in which the listed official has the lower score. Group means and differences use posterior mean point estimates, not posterior-draw uncertainty. For TOP, the not-listed comparison is especially weak because the short announcement may not be a complete slate. The Our Revolution row is not estimable. None of these endorsements entered the fit.

Organization	Listed / outside	D / R listed	Mean listed	Mean outside	Difference	Point-biserial	Ordered-pair share	Within-D listed / outside	Within-D mean difference	Within-D ordered-pair share
Labor federation	7 / 10	7 / 0	-0.28	0.18	-0.46	-0.52	0.79	7 / 4	-0.02	0.46
LGBTQ+ Caucus	7 / 10	7 / 0	-0.24	0.14	-0.38	-0.43	0.77	7 / 4	0.1	0.43
TOP PAC	2 / 15	2 / 0	-0.51	0.05	-0.57	-0.42	0.93	2 / 9	-0.29	0.89
Our Revolution	0 / unknown	0 / 0	—	—	—	—	—	0 / unknown	—	—

All 2023-listed fitted officials in these three recovered slates are Democratic-affiliated; the 2024 cohort has eleven Democrats and six Republicans. This explains much of the all-party separation for labor and the Caucus. Among Democrats alone, their listed and unlisted members have nearly identical mean scores and pair orderings near one-half. TOP lists only two fitted winners, both Democrats; their lower scores among Democrats are suggestive, but the tiny group and partial-slate issue prevent a strong inference. The endorsement_2023_cohort.csv file records each member’s indicator status and coverage qualifier so the denominator can be inspected.

The Caucus’s complete 2013, 2015, and 2019 general slates also permit within-election checks against first-year observed council members, without pretending every unlisted 2000–2026 official was unendorsed. The 2019–2020 comparison excludes Jerry Davis, a holdover in District B while its 2019 election awaited the delayed December 2020 runoff; Tarsha Jackson entered only afterward. These older comparisons are descriptive: each uses a small within-election cohort.

Election → observed year	Cohort	Listed / outside	D / R listed	Mean difference	Point-biserial	Ordered-pair share	Within-D mean difference
2013 → 2014	17	10 / 7	8 / 2	-0.23	-0.54	0.81	-0.04
2015 → 2016	17	8 / 9	8 / 0	-0.62	-0.58	0.92	-0.20
2019 → 2020	16	6 / 10	6 / 0	-0.22	-0.23	0.60	0.27

The report’s aggregate density curves smooth the 76 fixed-party posterior means, giving each official equal weight; they do not integrate posterior uncertainty or service duration. Both groups use a Gaussian kernel with bandwidth 0.18 ideal-point units. Each aggregate curve begins at its group’s observed minimum, with a smooth taper to zero. Scores beyond +2.5 still enter the kernel estimate. The elected-council plot, starting in 2000–2001, computes each party’s median fixed pooled score among members who cast at least one qualifying vote in that council; a light dashed linear regression overlays each party’s observed-median line. The two-panel 2000–2001 versus 2020–2023 density comparison uses the same bandwidth and axes in both councils. Each party’s curve is restricted to its observed score range, tapered to zero at both endpoints, and renormalized to unit area; bottom ticks show the individual scores. Gabriel Vasquez is excluded because his affiliation changed during service. These are display conventions for small samples, not boundary-corrected population-density estimates. Members serving multiple councils recur at the same coordinate; changes in the party distributions and median lines reflect membership composition, not within-person movement or a time-varying ideal-point model. The current council is partially observed.

Individual party evidence

The source links below identify the particular record or report used for each label. “Reporting only” and “secondary guide/roster” are weaker than a partisan candidacy, party office, or self-identification. A few affiliations are inferred from a documented party process rather than an explicit self-description, as detailed in the basis column.

Official	Affiliation	Evidence type	Source	Basis
Abbie Kamin	Democratic	party_record	source	2026 Democratic candidate for Harris County Attorney.
Ada Edwards	Democratic	party_record	source	Texas House resolution identifies her as a Harris County Democratic Party field coordinator.
Addie Wiseman	Republican	party_record	source	Texas Secretary of State lists her in the 2010 Republican primary for state House District 127; council election was nonpartisan.
Adrian Garcia	Democratic	party_record	source	Ran in Democratic primary for Harris County sheriff in 2008.
Al Hoang	Republican	party_record	source	Texas Secretary of State lists him in the 2014 Republican state House primary; stronger than his earlier GOP council endorsement.
Amanda Edwards	Democratic	party_record	source	FEC identifies her 2020 U.S. Senate candidacy as Democratic.
Amy Peck	Republican	reported	source	Listed as Republican on the current council roster; described as the 'incumbent Republican District A Councilwoman' in local commentary.
Andrew C. Burks, Jr.	Democratic	reported	source	Chronicle calls him a lifelong Democrat and former county-party-chair candidate, despite GOP endorsement in nonpartisan 2011 council runoff.
Anne Clutterbuck	Republican	party_record	source	Candidate profile identifies service as a Republican state-convention delegate.
Annise Parker	Democratic	self_identified	source	Her campaign identifies her as a Democrat; she ran in the 2026 Democratic county-judge primary.
Bert Keller	Republican	party_record	source	2022 Republican state-house primary candidate profile identifies former Houston council member Bert Keller and his GOP party roles.
Bill White	Democratic	party_record	source	Democratic nominee for Texas governor in 2010.
Brenda Stardig	Republican	party_record	source	Filed as a Republican for Harris County commissioner in 2020.
Bruce Tatro	Republican	reported	source	Contemporaneous Houston Press identifies district council member Bruce Tatro as the principal GOP candidate for controller in 2003.
C. O. “Brad” Bradford	Democratic	party_record	source	Texas House journal records Clarence Bradford as Democratic precinct chair; he later was Democratic nominee for Harris County DA.
Carol Alvarado	Democratic	party_record	source	Texas Legislative Reference Library records her state legislative service as Democrat.
Carol M. Galloway	Democratic	self_identified	source	Archived Carol Mims Galloway biography lists her as a member of Harris County Black Democrats. Democratic affiliation is supported; no party candidacy or office is documented here. Do not confuse with Dallas Republican Carolyn K. Galloway.
Carolyn Evans-Shabazz	Democratic	reported	source	Listed as Democratic on the current council roster.
Carroll Robinson	Democratic	party_record	source	Roll Call reports a 2002 Democratic congressional primary runoff between council members Carroll Robinson and Chris Bell.
Chris Bell	Democratic	party_record	source	U.S. House official history: elected as a Democrat in 2002, after Houston council service.
Dave Martin	Republican	reported	source	Characterized as a Republican/conservative in local reporting; no party-record source located.
David W. Robinson	Democratic	reported	source	Houston Chronicle 2018 floodplain-vote report explicitly identifies Robinson among Democratic council members; no party-primary record located.
Dwight Boykins	Democratic	reported	source	Contemporaneous Houston Chronicle report calls him a Democratic council member; Republican primary voting and GOP donations are mixed signals.
Edward Gonzalez	Democratic	party_record	source	Ran for Harris County sheriff as a Democrat in 2016.
Edward Pollard	Democratic	party_record	source	2016 Democratic primary candidate, Texas House District 137.
Ellen Cohen	Democratic	party_record	source	Texas Legislative Reference Library records her state House service as Democrat.
Fred Flickinger	Republican	party_organization	source	Harris County GOP endorsement, District E, November 2023.
Gabriel Vasquez	Mixed	reported	source	Houston Press reports council member Gabe Vasquez changed his affiliation from Democratic to Republican in 2003. Both affiliations occurred within his service; exclude from fixed two-party comparison.
Gordon Quan	Democratic	party_record	source	Sought 2010 Democratic nomination for Harris County judge. Opponent Ahmad Hassan, not Quan, had run in both party primaries.
Greg Travis	Republican	party_record	source	Resigned to run in the Republican primary for Texas House District 133.
Helena Brown	Republican	party_record	source	Texas House resolution commends her service as Republican Party precinct chair.
Jack Christie D.C.	Republican	reported	source	Contemporaneous Houston Chronicle report calls him a Republican council member.
James G. Rodriguez	Democratic	reported	source	University of Houston 2009 nonpartisan election guide lists incumbent James Rodriguez as D.
Jarvis Johnson	Democratic	party_record	source	Texas Legislative Reference Library records his state House service as Democrat.
Jerry Davis	Democratic	party_record	source	Harris County official 2020 Democratic primary results list him as a candidate.
Jew Don Boney, Jr.	Democratic	party_organization	source	Boney's contemporary re-election advertisement lists endorsements by Harris County Democrats and Tejano Democrats; this is party-organization support, not a partisan ballot filing.
Joaquin Martinez	Democratic	reported	source	Listed as Democratic on the current council roster.
John E. Castillo	Democratic	party_record	source	Contemporaneous report of the March 2002 Democratic county-commissioner primary places former council member John Castillo against Sylvia Garcia.
John Whitmire	Democratic	party_record	source	Long-time Democratic state senator.
Jolanda “Jo” Jones	Democratic	party_record	source	Texas Legislative Reference Library records her state House service as Democrat.
Julian Ramirez	Republican	party_organization	source	Harris County GOP endorsement, At-Large 1, November and December 2023.
Karla Cisneros	Democratic	reported	source	Houston Chronicle 2018 floodplain-vote report explicitly identifies Cisneros among Democratic council members; no party-primary record located.
Larry V. Green	Democratic	reported	source	Reportedly sought the 2016 Democratic precinct-chair nomination for Harris County commissioner. This is party-process evidence, not a formal primary filing; do not confuse with Ronald C. Green.
Lee P. Brown	Democratic	reported	source	Contemporaneous Houston Chronicle identifies Mayor Brown as a Democrat and notes his 2000 Democratic National Convention speech.
Letitia Plummer	Democratic	party_record	source	2018 Democratic primary candidate for U.S. House; 2026 Democratic nominee for Harris County Judge.
M. J. Khan, P. E.	Republican	reported	source	Chronicle directly identifies former council member MJ Khan as a Republican contender in the 2023 nonpartisan mayoral race.
Mario Castillo	Democratic	reported	source	Listed as Democratic on the current council roster.
Mark Ellis	Republican	party_record	source	Chronicle candidate profile records Republican primary candidacy for justice of the peace; later Republican state Senate primary candidate.
Mark Goldberg	Republican	reported	source	Contemporaneous Chronicle item directly calls former District C council member Goldberg a Republican; not inferred from a later namesake primary candidate.
Martha Castex-Tatum	Democratic	reported	source	Listed as Democratic on the current council roster.
Mary Nan Huffman	Republican	party_record	source	Official results list her in the 2020 Republican primary for Harris County DA; also GOP-endorsed for council in 2023.
Melissa Noriega	Democratic	party_record	source	Texas Legislative Reference Library records her state House service as Democrat.
Michael Berry	Republican	reported	source	Contemporaneous Houston Press account identifies him as Republican but notes he avoided partisan labels in a council campaign.
Michael Kubosh	Republican	reported	source	Contemporaneous Texas Monthly profile identifies him as Republican and explains his strategic 2006 Democratic Senate candidacy. A 2019 Houston Chronicle profile reports eight of nine Republican primary ballots in 2004⁠-⁠2012 and GOP endorsement, while noting that he rejects party labels. No Republican candidacy or party office is documented; see Data/Houston/party_evidence/party_research_kubosh_2026⁠-⁠09⁠-⁠25.md.
Mike Knox	Republican	party_record	source	Republican precinct chair; 2024 Republican primary candidate for Harris County Sheriff.
Mike Laster	Democratic	reported	source	Contemporaneous Houston Chronicle report explicitly calls him a Democrat; UH 2009 election guide also lists D.
Mike Sullivan	Republican	party_record	source	Filed in Republican primary for Harris County tax assessor-collector.
Oliver Pennington	Republican	reported	source	University of Houston 2009 nonpartisan election guide lists him as R; later reporting calls him Republican.
Orlando Sanchez	Republican	party_record	source	Harris County official 2018 Republican primary cumulative report lists Orlando Sanchez for county treasurer.
Pam Holm	Republican	reported	source	University of Houston 2009 nonpartisan election guide lists her as R; contribution to Republican women is only corroborative.
Peter Brown	Democratic	reported	source	Chronicle calls him a progressive Democrat; UH 2009 election guide lists D.
Richard Nguyen	Democratic	party_organization	source	Contemporaneous report lists District F incumbent Nguyen on the Harris County Democratic Party's 2015 slate of Democratic candidates and on the Meyerland Democrats slate. Party release not independently located; no reliance on a possible 2023 namesake candidacy.
Rob Todd	Republican	reported	source	Contemporaneous local election commentary explicitly identifies former Houston council member Rob Todd as a Republican; no partisan candidacy found.
Robert Gallegos	Democratic	reported	source	Biography lists Democratic affiliation; no party-record source located.
Ronald C. Green	Democratic	self_identified	source	Identified himself as a loyal Democrat and received Harris County Democratic Party backing, despite a conservative-radio endorsement.
Sallie Alcorn	Democratic	reported	source	Listed as Democratic on the current council roster.
Shelley Sekula-Gibbs, M.D.	Republican	party_record	source	FEC identifies her congressional candidacy as Republican.
Stephen C. Costello	Republican	reported	source	Contemporaneous University of Houston-Downtown article describes him as Republican in 2015 mayoral field; UH 2009 guide lists R. A payment to Harris County Democrats is not counterevidence of identity.
Steve Le	Republican	reported	source	Contemporaneous Houston Chronicle report directly calls him a Republican council member.
Sue Lovell	Democratic	party_record	source	Elected by Texas Democrats to the Democratic National Committee; UH 2009 guide lists D.
Sylvester Turner	Democratic	party_record	source	Long-time Democratic state representative; later Democratic member of Congress.
Tarsha Jackson	Democratic	reported	source	Listed as Democratic on the current council roster.
Tiffany D. Thomas	Democratic	reported	source	Listed as Democratic on the current council roster.
Toni Lawrence	Republican	reported	source	Chronicle explicitly identifies her as Republican despite nonpartisan council ballot.
Twila Carter	Republican	party_organization	source	Harris County GOP endorsement, At-Large 3, December 2023 runoff.
Wanda Adams	Democratic	reported	source	University of Houston 2009 nonpartisan election guide lists her as D.
Willie Davis	Republican	party_organization	source	Harris County GOP endorsement, At-Large 2, November and December 2023.

The Houston assessment calls this candidate review-ready, not research-reviewed. Original-page spot checks, independent service-date verification, and two-build reproduction remain open upstream. The three local corrections need reconciliation with the published parser.

The separate Houston–Texas bridge diagnostic applies the same six-person linear-mapping step used for New York. All six bridges are Democrats and cover a restricted ideological range; the slope is consequently weakly determined and unstable under leave-one-out checks. This does not by itself establish ideological change across offices. Mechanically mapped positions are retained only as explicitly labeled diagnostics and are not an input to the present report.

Reproduction

From the Local project root, with the verified ZIP in the path expected by the preparation script:

Rscript Code/61_prepare_houston_2026_10_12.r
Rscript Code/62_fit_houston_party_start.r
Rscript Code/38_houston_2000_party_evidence.r
HOUSTON_SOURCE_VERSION=2000 Rscript Code/31_houston_endorsement_validity.r
Rscript Code/44_houston_mayoral_districts.r
Rscript Code/39_houston_state_bridge.r
quarto render houston-expanded-report.qmd
quarto render houston-expanded-technical-appendix.qmd
quarto render houston-dashboard.qmd

Code/61 overwrites the September 25 matrix in Results/Houston-2000/; that fit and its diagnostics are archived in Results/houston-2000-sept25-2026-10-12/ and are not reproduced by this sequence. Seeds, starts, and session information are saved alongside each fitted output.