How the Standards Test Works

The Core Model

The calculator estimates the proportion of the US adult population (18–64) that simultaneously satisfies all of your selected criteria. Each criterion is applied as a conditional probability, and the results are multiplied together:

P(match) = P(Age) × P(Race) × P(Politics | Race) × P(BMI | Race)
         × P(Education | Race) × P(Children | Race, Age)
         × P(NeverMarried | Race, Age) × P(IntactFamily | Race)
         × P(ZeroPartners | Age)

The model treats each factor as conditionally independent once you account for race and age — a simplification that slightly overestimates joint probabilities for correlated variables (e.g., education and marital status), but keeps the calculation transparent and replicable. The interactive calculator shows the per-factor contribution in the breakdown panel so you can see exactly where your pool shrinks.

Data Sources

1. US Census Bureau — American Community Survey (ACS) 2022

Source for age distribution, racial and ethnic composition, educational attainment, and marital status of the US adult population aged 18–64, broken down by sex. The ACS is a continuous survey covering approximately 3.5 million addresses per year; the 2022 5-year estimates were used for stability. Total population figures:

• US women aged 18–64: approximately 104.8 million
• US men aged 18–64: approximately 101.2 million

2. CDC — National Health and Nutrition Examination Survey (NHANES) 2017–2020

Source for body mass index (BMI) distributions by race and sex. NHANES combines interviews and physical examinations; the 2017–2020 pre-pandemic cycle covers roughly 15,560 participants. BMI categories used:

• Healthy weight: BMI < 25
• Not obese: BMI < 30

BMI prevalence differs substantially across racial groups. For example, approximately 54% of Asian women have a BMI below 25, compared to roughly 36% of White women and 20% of Black women — the calculator accounts for this by weighting BMI probability by the racial composition of the user's selected pool.

3. Pew Research Center — Political Survey 2024

Source for self-identified conservative political affiliation by race and sex. Pew's annual political typology surveys cover approximately 5,000–10,000 US adults. Conservative identification rates vary widely by racial group and are applied as conditional probabilities given the selected race composition.

4. CDC — National Survey of Family Growth (NSFG) 2017–2019

Source for lifetime sexual partner counts and family structure data. The NSFG covers approximately 9,000 adults aged 15–49. The "zero prior partners" proportion is stratified by age group and sex; it is the single most restrictive filter in the calculator for most age groups.

How Race Conditioning Works

When a user selects multiple racial groups, the calculator first computes the combined racial proportion (sum of selected groups' population shares), then weights each conditional factor by the relative size of each selected group within that pool. This means, for example, that selecting "White" and "Asian" together produces a BMI probability that is a weighted average of the White BMI rate and the Asian BMI rate — not an arithmetic mean.

Formally, for any factor F applied to a selected race set R:

P(F | R) = Σᵣ [ P(F | race=r) × P(race=r | r ∈ R) ]

where P(race=r | r ∈ R) is the normalized proportion of race rwithin the selected subset.

Age Band Logic

The model maps the user's chosen age range to one of four age bands for applying age-conditional factors (children, marital status, partner count):

• 18–24 (midpoint < 25)
• 25–34 (midpoint 25–34)
• 35–44 (midpoint 35–44)
• 45–64 (midpoint ≥ 45)

The age proportion itself is calculated by summing the overlap between the user's range and each of the five 10-year census bands, prorated by the fraction of each band covered.

Worked Example

A user looking for a woman aged 22–32, White, conservative, healthy BMI (<25), bachelor's degree, no children, never married, raised by both parents, zero prior partners:

At 0.004%, approximately 4,100 women in the US would satisfy all nine criteria simultaneously. Each criterion individually is reasonable; stacked together, the math compounds rapidly — which is the calculator's core insight.

How to Interpret Your Result

The calculator returns two numbers: a percentage and anabsolute count. Both matter, and they tell different stories.

The percentage shows how rare or common your criteria are across the adult US population. Anything above 10% is genuinely common. Between 1% and 10% is selective but realistic — there are still millions of people in that range. Below 1% is where the pool starts to feel thin, and below 0.1% means you have stacked enough rare criteria that each person who qualifies is statistically unusual.

The absolute count puts the percentage in perspective. Even 0.1% of 104 million adult women is 104,000 people. A dating market of 104,000 is enormous — far more than any person could realistically meet in a lifetime of active dating. The insight the calculator delivers is not "this is impossible" but "here is the actual math behind your intuition."

The breakdown panel below the main number shows each factor's individual contribution. The factors that drop below 20% are the ones compressing your pool the most — those are the criteria worth examining most carefully.

Why Probabilities Multiply So Dramatically

The most common reaction to the calculator is surprise at how fast the number falls once a few filters are applied. The reason is straightforward set theory: each filter selects a subset of the previous subset.

Say you want someone in a specific age range (21% of adults), of a specific racial background (60%), with a bachelor's degree or higher (41%), and no prior children (42%). Those four criteria look modest individually. But 0.21 × 0.60 × 0.41 × 0.42 = 2.17%. You haven't added any particularly strict filters yet and you're already down to roughly one person in 46.

Add a political leaning requirement (~43% for conservative, given the racial composition above) and healthy BMI (~36%): the number becomes 2.17% × 0.43 × 0.36 ≈ 0.34%. That's roughly 350,000 adult women in the US — still a huge real-world pool, but visually alarming as a percentage.

The single factor that causes the most dramatic drop is the zero-prior-partners filter. Only about 4–6% of adults aged 25–34 report zero lifetime sexual partners in CDC surveys (and roughly 1–2% by their late 30s). Requiring this criterion alone eliminates 94–99% of the remaining pool instantly. This is the "nuclear option" of dating filters, and its effect on the result is the most commonly shared outcome from the calculator.

How the Model Handles No Race Selection

If a user selects all racial groups (or deselects everything), the model treats the full US adult population of the selected gender as the base pool — raceProp equals 1.0. The conditional factors are then weighted by the actual US population share of each group, which is equivalent to applying a national-average probability for each factor. This is the correct behavior: "no preference" means "everyone in the country is eligible," and the probability of each downstream factor reflects the population as it actually exists.

Factor Definitions

Each filter in the calculator maps to a specific, measurable survey construct — not a vague impression:

• Age range — age in years, applied against Census age bands (prorated by overlap).
• Race / ethnicity — Census racial and ethnic categories: White (non-Hispanic), Hispanic, Black, Asian, Other / Multiracial.
• Political leaning — self-identified political affiliation from Pew Research surveys; the model applies the "conservative" identification rate.
• Body weight — measured BMI from NHANES physical examinations (not self-report), bucketed as under 25 (healthy weight) or under 30 (not obese).
• Education — bachelor's degree or higher, from ACS educational attainment data.
• Children — whether the person has children, from ACS fertility and household data, stratified by age band.
• Marital history — the ACS "never married" category, stratified by age band.
• Raised by both parents — grew up in an intact two-parent household, from NSFG family-structure data.
• Prior sexual partners — NSFG lifetime partner count; the zero-partner proportion is stratified by age band and sex.

Where a factor is stratified by both race and age, the model applies the race-conditional rate for the age band selected by the midpoint of your chosen range.

Limitations

• Correlation ignored between factors. Education and marital status are positively correlated; income and education are correlated; political affiliation correlates with religion and lifestyle. The model treats them as conditionally independent once race and age are controlled for, which slightly overestimates the joint probability for combinations of correlated variables. A fully specified multivariate model would require microdata not available in public-use summary tables.
• Data lag. ACS 2022, NHANES 2017–2020, and Pew 2024 reflect conditions at time of publication; demographic shifts in education, marital patterns, and political affiliation are ongoing, particularly among younger cohorts.
• Binary filters only. Real dating preferences involve continuous dimensions — height, income, physical attractiveness, personality traits — that cannot be captured in a survey-based probability model. The calculator deliberately covers only factors for which reliable population data exists.
• Self-reported data. Survey-based sources (Pew, NSFG) rely on self-reporting. Partner-count data is particularly subject to social desirability bias: men historically over-report and women under-report. The NSFG attempts to correct for this with anonymous audio computer-assisted self-interview (ACASI) methodology, but some bias likely persists.
• US-centric. All population figures and conditional probabilities are based on the US adult population. Results are not applicable to other countries or to cross-national dating contexts.
• Not a forecast. The calculator estimates what share of the population meets your stated criteria, not the probability that you will find and date such a person. Geographic concentration, social network access, mutual attraction, and compatibility factors not in the model all affect real-world outcomes significantly.

The model is designed for educational insight and entertainment — not as a quantitative forecast of dating outcomes. The goal is to make abstract preferences concrete through real population math. If you want to examine the underlying data or propose a correction, the full probability table is available insrc/data/demographics.json in the project repository.