The Carl DeMaio voter guide is a candidate scorecard published by Reform California, a taxpayer-advocacy organization DeMaio founded, that rates officeholders and candidates on fiscal policy votes and signed pledges rather than on party label or social issues. It functions as a filtering instrument for fiscally conservative voters, not as a neutral voter-information service, and its ratings should be read alongside official sample ballots and county voter guides. This explainer covers how the scoring works, what the numbers mean, where the method is vulnerable, and how to use the guide without being misled.
How the scorecard mechanism works
The guide is built on two inputs: recorded legislative behavior and signed pledge documents. For incumbent legislators, Reform California compiles floor votes and bill co-sponsorships on a designated set of bills it classifies as tax, spending, regulation, or transparency measures. Each selected vote is weighted, and the weighted tally produces a percentage rating, typically banded into letter grades.
For challengers with no voting record, the organization substitutes a candidate questionnaire and a signed "Taxpayer Protection Pledge" against new taxes. That substitution is the mechanism's central design choice. A challenger's score therefore reflects stated intentions, while an incumbent's score reflects revealed behavior. The two numbers appear side by side in the same guide but are not measuring the same thing.
The bill-selection step carries most of the analytical weight. Whoever chooses which votes count determines the ranking. Selection criteria are published in general terms but the full weighting formula is not typically disclosed at the individual-vote level, which means a reader cannot independently recompute a score from public roll-call data without the organization's own bill list.
In plain terms: the guide converts a large, messy legislative record into a single sortable number, and the sorting rules belong to the publisher.
Quantified context: what the ratings and scale look like
Reform California has graded California legislators across multiple two-year sessions, and the guide's output is a ranked list rather than a random sample. The practical effect is compression. In a legislature where most fiscal votes split close to party lines, a weighted score tends to produce a bimodal distribution: a cluster of high scores on one side and a cluster of low scores on the other, with few members in the middle.
That shape matters for interpretation. A 15-point gap between two candidates near the top of the scale may represent a single symbolic vote, while the same 15-point gap near the bottom may represent several substantive tax votes. Ordinal rankings hide this. The guide is most informative at the extremes and least informative in the middle of the pack.
Three structural numbers frame the whole exercise. California's legislature has 40 senators and 80 assembly members, so a full guide covers up to 120 incumbents in a general election year, plus challengers. State legislative terms run four years for the Senate and two for the Assembly, meaning Assembly scores refresh twice as often as Senate scores. And the guide's usefulness decays with distance from the publisher's bill list: the further a race is from tax-and-spending legislation, the less the score explains.
One counterintuitive detail: a candidate can score well while voting against the organization's position on a marquee bill, because the weighted average absorbs single votes. Voters who care about one specific measure should check that measure's vote directly rather than trusting the composite.
Where the method breaks down: edge cases
Several candidate categories sit outside the scoring logic entirely.
Appointed incumbents who served less than a full session may have too few recorded votes to generate a stable percentage. Candidates in nonpartisan local races, such as city council or school board, often have no comparable legislative record at all. Judges and judicial candidates are generally excluded because their decisions are not fiscal votes. Ballot-measure positions are sometimes folded in, but a Yes/No on a proposition is not the same kind of datum as a roll-call vote. Candidates who decline the questionnaire may receive a default low rating, which reflects nonparticipation rather than hostility to the organization's positions.
That last point is the professional caveat most thin explainers miss. A missing data point is not a negative data point, but in a ranked list the two look identical. Treat any score derived from nonresponse as unknown, not as evidence.
Common misconceptions about the guide
The most frequent error is treating the score as an ideological composite. It is not. The guide is deliberately narrow: taxes, spending, regulation, and governmental transparency. A candidate's positions on abortion, firearms, immigration, or education curriculum are generally not inputs. A voter who reads a high score as a general ideological endorsement is reading past what the instrument measures.
The second error is treating the guide as official. It is published by a private advocacy organization and is not a Secretary of State product, not a county elections office document, and not part of the official voter information guide mailed to registered voters. The official sample ballot remains the authoritative source for candidate names, party preferences, and ballot measure text.
The third error is assuming a single national methodology. Reform California's guide is a state-level instrument with state-level bill selections. Guides circulated under similar branding in other states or for federal races may use different criteria, and scores are not comparable across them. A 90 in one guide does not equal a 90 in another.
The fourth error is forgetting who is publishing. DeMaio is a political figure who has run for office and raises funds for the organization. The guide is an advocacy communication with a disclosed point of view. That does not make it inaccurate, but it does mean the selection of what counts is itself a persuasive act.
Reading a score against other voter information sources
| Source | What it measures | Who publishes it | Best use | Main limitation |
|---|---|---|---|---|
| Reform California guide | Selected fiscal votes and signed pledges | Private advocacy organization | Sorting candidates on tax and spending posture | Weighting not fully transparent; nonresponse penalized |
| Official state voter guide | Candidate statements, ballot measure text | State elections office | Authoritative ballot language and candidate self-description | No independent scoring |
| Newspaper editorial endorsements | Editorial board judgment | News organization | Narrative context and interviews | Opinion, not measurement |
| Party sample ballot | Party recommended slates | County political parties | Party-aligned shorthand | Reflects party, not fiscal record |
| Legislative roll-call databases | Every recorded vote | State legislature | Verifying a specific vote yourself | No composite; requires manual work |
| Campaign finance filings | Donations and expenditures | State disclosure agency | Identifying funding coalitions | Says nothing about policy positions |
The table's practical message: the DeMaio guide is one lens among six, and it is the only one in the list that produces a single comparable number. That convenience is exactly why it needs cross-checking.
Timeline: how a guide cycle unfolds
| Stage | Timing relative to election | What happens |
|---|---|---|
| Bill selection | Months before publication, after session ends | Organization designates scored votes and weights |
| Incumbent scoring | Weeks after final scored votes | Roll-call tallies converted to percentages |
| Candidate outreach | During primary and general campaign windows | Questionnaires and pledges sent to challengers |
| Draft compilation | Before public release | Scores assembled and ranked into grades |
| Publication | Typically in the weeks before ballots mail | Guide released online and through affiliated channels |
| Ballot mailing | State deadline for official materials | Official voter guides arrive separately |
| Vote-by-mail window | Weeks before election day | Most ballots cast during this period |
| Post-election review | After results certify | Scores revisited against next session's votes |
Two timing facts matter for use. First, the guide usually lands late enough that early mail voters may already have their ballots, so check for an updated version. Second, incumbents' scores reflect a completed session, while challengers' pledges are forward-looking. Comparing them is a comparison across time as well as across method.
Practical steps for using the guide
Use the score as a first filter, then verify. Concrete sequence:
Confirm your district and the exact race using your county elections office lookup, not the guide's own indexing. Locate the official sample ballot and read the candidate statement for anyone you are seriously considering. For incumbents, pull two or three of the specific scored votes from the legislature's roll-call database and read the bill titles yourself. For challengers, read whether the score came from a questionnaire, a signed pledge, or nonresponse. Nonresponse should be treated as unknown. Cross-check against at least one source from a different category in the table above before finalizing. Check the guide's publication date against your ballot's arrival date and look for revisions.
A threshold rule that helps: if two candidates are within roughly 10 points of each other on the guide, treat them as functionally tied on fiscal posture and decide on other grounds. The composite is not precise enough to justify a confident choice inside that band.
Red flags worth noticing:
A score presented without any indication of which votes were counted. A challenger rated low solely because no questionnaire was returned. A guide circulated with a date that predates the current session's key votes. Any version that presents itself as an official government document. Local races scored using state-legislative criteria.
Limitations and how far the evidence reaches
The guide measures what its publisher chooses to measure, on a scale whose internal weights are not fully public. It cannot tell you how a candidate would vote on an unanticipated bill, how they behave in committee or in negotiations, or how they handle constituent service. It says nothing about judicial temperament or local land-use decisions.
Composite scores also lose information by design. Converting dozens of votes into one percentage means the reader cannot see trade-offs, and two candidates with identical scores may have arrived there through opposite priorities. The honest reading is directional: high means generally aligned with the organization's fiscal agenda, low means generally opposed or unmeasured, and the middle is noise.
A final caution on freshness. Specific scores, rankings, and pledge counts change every cycle and should be taken from the current published guide rather than from any secondary description, including this one. The mechanisms described here are stable; the numbers are not.
This article is general educational information about how a political scorecard is constructed and read. It is not an endorsement of any candidate or organization, not legal advice on ballot or voting procedures, and not a substitute for the official voter information guide issued by your state or county elections office. For questions about your registration, ballot, or voting deadlines, contact your local elections official directly.