How Plain Paycheck's numbers are verified

Every federal and state figure on this site comes from a published government document, not a summary of one. That covers brackets, standard deductions, exemption credits, and payroll-tax rates. This page walks through how a number gets checked before it ships, the mistakes that checking has already caught, and what the calculators do not try to know.

Every figure is cross-checked against at least two independent sources before it ships. The state's own Department of Revenue document wins over any secondary summary. For federal figures that means the IRS or the Social Security Administration. When a state publishes its formula, its rate table, and its standard-deduction worksheet as three separate documents, all three get checked against each other, not just against one PDF. Each state's page links the exact documents its numbers came from, so the citation sits one click from the figure it backs.

Checking a number once is not enough to keep it correct as the site grows, so every jurisdiction's math is pinned by a test fixture. Each fixture is a full worked example, with a salary, a filing status, and every deduction line, recomputed from the primary documents and saved next to the tax engine's own test suite. If a later edit to the engine or the data changes what that fixture produces, the test fails before the change ships, not after a reader notices their numbers look off. The second wave of states went a step further. One person built the figures, and a second person recomputed the same worked example from the source documents with no view of the first answer, to catch anything a single pass would wave through.

The two-source rule and the independent recompute have both changed numbers before publication. Three cases:

• Georgia cut its flat income-tax rate from 5.19% to 4.99% in May 2026, retroactive to the start of the year. Other calculators and summaries were still carrying the old 5.19% weeks later. Catching it meant reading the enacted bill instead of trusting an aggregator's cached rate.
• Oregon's 2026 withholding-formula document has a worked example that still prints the prior year's exemption credit, $256, even though the live 2026 rate tables in the same publication resolve to $263. The independent recompute caught the mismatch by working from the tables rather than copying the document's own stale example.
• The federal "no tax on overtime" and "no tax on tips" phase-out math comes straight from the statute text and the Schedule 1-A worksheet, not a secondary description of them. The statute specifies a floor-rounded divisor for the phase-out, which is easy to round the wrong way if you work from a plain-English summary instead of the worksheet itself.

A paycheck estimate is not the same thing as an actual pay stub, and the gap is on purpose. Every calculator here leaves out W-4 adjustments: extra withholding, dependents, and multiple-job elections all change what gets withheld without changing the tax actually owed. Benefit premiums for health, dental, and vision are not modeled, because they vary by employer and plan rather than by tax law. City and county income taxes are flagged where a state has them but not computed, because local rates and rules vary too much within a single state to model generically. And none of this adjusts for cost of living, so take-home reads the same whether it is spent in a cheap metro or an expensive one. Each state's page and the home page say all of this in plain sight instead of hiding it in fine print.

Estimated spending taxes are an optional, separate estimate of what state and local sales tax would take out of a year of everyday retail spending: apparel, entertainment, furnishings, personal-care products, and dining out. Groceries are left out because how states tax them varies too much to fold into one rate. The estimate is built from two different kinds of number, and the page keeps them distinct instead of blending them into one anonymous figure.

The first is each state's base sales-tax rate. That is a verified fact in the same sense as the income-tax figures here: cross-checked against a state Department of Revenue document plus an independent aggregator, with several of the largest states spot-checked directly against their own DOR pages. The second is the local-tax add-on and the all-states comparison figure. Both come from a single source, the Tax Foundation's population-weighted state and local sales tax rate table, dated as of January 1, 2026. No government agency publishes a population-weighted combined rate. A state's own revenue department gives you the base rate and, at most, a range of local rates, never one blended number. So that figure cannot clear the same two-source bar the rest of the site holds itself to. It ships as an editable field you can type over, labeled as an estimate, and as a separately labeled comparison mode, never presented with the same certainty as the verified state rate.

The default spending amount comes from the Bureau of Labor Statistics' 2024 Consumer Expenditure Survey, the most recent edition published as of this writing. It sums the average annual spending on food away from home, apparel and services, household furnishings and equipment, entertainment, and personal-care products. It leaves out groceries, housing, healthcare, insurance, and transportation. Vehicle purchases are a lumpy, infrequent expense rather than steady annual spending, and gasoline already carries its own excise tax in every state. Like the local-tax figure, this is a starting point built for editing, not a number to take as personal fact. Your own spending in any of these categories is very likely different from a national average.

A single state average cannot capture the variance inside a state. Alaska is the clearest case. It has no state sales tax at all, but its largest cities, Anchorage and Fairbanks, charge no local sales tax while small towns charge 7% or more, so a statewide average badly overstates what most Alaskans actually pay. The same problem shows up in every state with home-rule cities or local-option taxes. Louisiana's combined rate can run past 11% in some parishes, Washington's ranges from 6.5% to over 10.6%, and local rates change every quarter in most states, while this figure is a snapshot dated January 1, 2026. The estimate is right for a state taken as a whole and wrong for almost any specific address, which is why it is the one figure on the site built to be edited rather than trusted as printed.

Federal and state tables are rebuilt for each tax year as the IRS, the SSA, and state revenue departments publish their new figures, which for most jurisdictions lands between October and December for the following year. Some items are flagged as pending, like a rate tied to a legislative trigger or a threshold waiting on an agency's annual indexing. Those get rechecked on their specific statutory or administrative dates rather than left until the next annual refresh, so a mid-year change does not sit uncorrected for months.

All of this verification happens before a number ever ships as data. The calculators themselves run entirely in your browser. Your salary, filing status, and every other input are computed on your own device and never sent anywhere. The site counts pageviews for its own analytics and records nothing about what you entered.

• IRS Rev. Proc. 2025-32, the 2026 federal brackets and standard deduction
• SSA, the 2026 Social Security wage base
• California EDD, 2026 withholding schedules , checked byte-for-byte against the state's published tables
• Georgia HB 463 , the enacted 4.99% rate cut, retroactive to January 1, 2026
• Oregon DOR, 2026 Withholding Tax Formulas , cross-checked against the department's own rate tables
• Every other state's figures are verified per state, with sources linked on each state page