Alabama Secretary of State ↗
Official legislative, statewide-office, and presidential returns; the authoritative election source.
Alabama Candidate Margin Overperformance
Historical legislative candidate margins relative to district-level expected margins, 1994–2022.
Historical CMO architectureWAR-style residual
The headline score is the observed legislative-versus-ticket gap minus the gap predicted from incumbency, downballot lag, demographics, and campaign effort.
8Historical cycles
509Contested D vs. R races
3Map views
This model applies Split Ticket's WAR structure to Alabama legislative elections: measure the raw legislative-versus-ticket gap, predict its structural portion, and score the residual.
WAR-style CMO is the headline measure. Positive values indicate performance ahead of the model's structural expectation. Scores are two-party margin percentage points, are zero-sum within a race, and are not causal estimates of individual candidate quality.
Explore the results
The default view maps CMO in margin points. The two raw comparison views show the legislative margin relative to the same district's governor result or previous presidential result. Color intensity is linear on a symmetric scale capped visually at ±30 points; tooltips show uncapped values.
CMO, absolute margin points
Candidate results
The table separates the observed ticket gap, modeled structural expectation, and remaining WAR-style residual.
| Cycle | District | Candidate | WAR-style CMO ↕ | Raw ticket gap | Lag adjustment | Incumbency adjustment | Career pooled | Structural gap | Baseline margin | Actual margin | Votes |
|---|
Diagnostics
The structural model is selected with election-cycle holdouts. These diagnostics test the construction; they do not turn a retrospective residual into a forecast.
Cycle-held-out tournament
| Model | Ridge alpha | MAE | RMSE |
|---|---|---|---|
| Barebones | 3.0 | 19.4 | 23.8 |
| Barebones | 10.0 | 19.3 | 23.7 |
| Barebones | 30.0 | 19.1 | 23.6 |
| Barebones | 100.0 | 18.8 | 23.4 |
| Full | 3.0 | 17.3 | 21.7 |
| Full | 10.0 | 17.2 | 21.6 |
| Full | 30.0 | 17.0 | 21.4 |
| Full | 100.0 | 16.8 | 21.3 |
Construct checks
| Check | N | Value |
|---|---|---|
| Repeat Candidate Next Cycle: candidate war cmo | 77 | 0.039 |
| Repeat Candidate Next Cycle: candidate raw ticket gap | 77 | 0.293 |
The score has little repeat-candidate persistence in the present sample. Read it as a candidate-cycle structural residual, not a durable or causal measure of candidate quality.
Data sources and attribution
Credits describe how each source is used in CMO. Derived scores, allocations, matches, and errors are this project’s calculations and should not be attributed to the source organizations.
OpenElections ↗
Standardized secondary election files used for comparison, normalization, and documented fallback—not a replacement for official returns.
U.S. Census Bureau ↗
1990/2000 decennial Census SF3, American Community Survey estimates, Census blocks and VTD geography.
U.S. Census Bureau TIGER/Line and archived Alabama enacted-plan shapefiles ↗
Legislative boundary geometry used to render maps and allocate geographic features; the page identifies the plan vintage.
Voting and Election Science Team (VEST) ↗
Election-specific precinct files used in the modern presidential/geographic pipeline where documented.
Database on Ideology, Money in Politics, and Elections (DIME), Adam Bonica ↗
Recipient-level contribution totals used for pre-electronic-era resource coverage; missing records remain unknown.
Alabama Secretary of State FCPA ↗
Principal-campaign-committee summaries provide the preferred 2014-2022 fundraising observations; identified committees with no cycle activity are observed zeros, while unmatched candidates remain unknown.
FollowTheMoney / National Institute on Money in Politics ↗
Candidate fundraising totals remain a secondary comparison source rather than the canonical modern finance input.
Shor–McCarty state legislative data ↗
Serving-legislator roster and party evidence used in historical incumbency review.
Wikipedia election pages ↗
Archived pages used only to cross-check candidate names and vote totals; discrepancies do not overwrite official returns.
Attribution boundary: Source organizations provide underlying records or geography; none endorses this model. Alabama Secretary of State returns remain authoritative. OpenElections and Wikipedia are secondary checks. Finance missingness is never interpreted as zero.
Data and provenance
Build updated August 21, 2026 from CMO methodology v4. Download the current rows, components, tournament, diagnostics, and provenance manifest.
How to read CMO
The preferred baseline is the same-cycle federal ticket inside the district, with a documented same-cycle state-ticket fallback when necessary.
The structural model uses symmetric incumbency, era-specific downballot lag, limited demographics, and capped campaign effort. WAR-style CMO is the observed ticket gap minus that prediction; ideology is excluded so it can be tested afterward.
Read the full CMO methodology, view the 2026 forecast, or read the forecast methodology.
Model output:
cmo_v4_candidates.csv. Scores cover contested Democratic-versus-Republican races.