Methodology

Primary Data Source: Census HTOPS

The primary data source for How Is America is the U.S. Census Bureau Household Trends and Outlook Pulse Survey (HTOPS). The Household Pulse Survey (HPS) was officially relaunched as part of the Household Trends and Outlook Pulse Survey (HTOPS) in January 2025. Our data comes from HTOPS Wave 2506 (March 2026).

HTOPS is a rapid-response survey designed to measure how emerging social and economic issues are impacting American households. It covers a broad range of topics including AI usage, food security, housing, employment, health insurance, household spending, and transportation.

How Is America uses data from March 2026 (Wave 2506).

Data Dictionary

Key HTOPS variables used throughout How Is America, with their coding and descriptions:

Census Sampling & Weighting

HTOPS is a probability-based survey, meaning every adult in the U.S. has a known, non-zero probability of being selected. This distinguishes it from opt-in online panels and makes it one of the most methodologically rigorous rapid-response surveys available.

• •Sample frame: The Census Bureau draws the sample from the Master Address File, the same frame used for the decennial census and American Community Survey.
• •Sample size: Approximately 7,500 respondents per wave, drawn from all 50 states and DC. Response rates vary by wave but typically fall between 5–10%.
• •Weighting: Each respondent receives a PWEIGHT (person weight) that adjusts for nonresponse and calibrates the sample to match known population totals by age, sex, race/ethnicity, education, and state. When applied, the ~7,500 respondents represent approximately 260 million American adults.
• •Collection mode: Internet self-response, with follow-up by phone for non-responders in some waves. The online-first approach may undercount populations with limited internet access.
• •Population: Civilian, non-institutionalized adults aged 18+ in the United States. Excludes those in prisons, nursing homes, military barracks, and other group quarters.

Margin of Error

As with any survey, HTOPS estimates carry sampling error. Key considerations:

• •National estimates: With ~7,500 respondents, national percentages carry a margin of error of roughly ±1–2 percentage points at the 90% confidence level.
• •Division-level estimates: Sample sizes per Census division range from roughly 330 to 1,700 respondents. Smaller divisions (e.g., East South Central with ~330) have margins of error of ±5–8 percentage points, while larger divisions are more precise.
• •Subgroup estimates:Cross-tabulations (e.g., AI usage by income within a specific age group) can produce small cells with high margins of error. We flag estimates based on small sample sizes (n<50) where they appear.
• •State-level data: The HTOPS public-use file does not include state identifiers. State-level data on How Is America comes from supplemental sources (ACS, BLS, USDA) rather than HTOPS directly.

Wellbeing Index

The American Wellbeing Index is a composite 0–100 score computed for each Census division. It combines six HTOPS metrics and one CDC PLACES health metric, each normalized relative to the worst-performing division:

The composite score is the weighted sum of all seven dimension scores. The worst division on each dimension scores 0 for that dimension; all others score proportionally higher. The CDC Population Health component is computed by averaging state-level CDC PLACES health scores within each division.

Squeeze Index

The Squeeze Index measures financial pressure on households in each Census division. It combines three HTOPS metrics:

• •Rent delinquency (rentBehind): % of renters behind on rent
• •Food insufficiency (foodInsufficient): % reporting not enough to eat
• •Expense difficulty (expenseDifficult): % reporting difficulty paying expenses

Each metric is normalized to 0–100 relative to the maximum across divisions, then the three normalized scores are averaged. The resulting squeeze score is converted to a letter grade:

Geographic Granularity

The HTOPS Public Use File (PUF) identifies respondent geography at the Census Division level only — it does not include state-level identifiers. This means all regional analysis on How Is America is at the division level (9 divisions), not the state level.

Comparison to Other Federal Surveys

Understanding how HTOPS compares to other major surveys helps contextualize the data:

HTOPS fills a unique niche: it's faster than the ACS, broader than the CPS, and more methodologically rigorous than most private polling. Its main limitations are smaller sample size (limiting geographic granularity) and periodic rather than continuous collection.

Supplemental Data Sources

How Is America supplements HTOPS data with state-level data from five additional federal sources:

• •USDA Economic Research Service: State-level food insecurity rates (3-year average 2021–2023). Source: USDA ERS Food Security
• •Bureau of Labor Statistics: State unemployment rates (March 2025, seasonally adjusted). Source: BLS Local Area Unemployment Statistics
• •U.S. Census Bureau ACS via FRED: Median gross rent and cost-burdened renter percentages by state (2023 American Community Survey). Source: Federal Reserve Economic Data (FRED)
• •CDC PLACES: State-level health outcomes and risk factors (2023 data, 2024 release) from the Behavioral Risk Factor Surveillance System (BRFSS). County-level estimates are aggregated to state-level population-weighted averages. Measures include obesity, diabetes, depression, physical inactivity, smoking, binge drinking, insufficient sleep, lack of insurance, and annual checkup rates. Source: CDC PLACES
• •U.S. Census Bureau ACS 1-Year Estimates:State-level demographic profiles (2023) including median household income, median gross rent, median home value, median age, poverty rate, educational attainment (bachelor's degree or higher), and unemployment rate. Source: Census Bureau Data

Limitations & Caveats

• •Geographic resolution: HTOPS public-use data is only available at the Census Division level. Division-level averages can mask significant variation between states within a division.
• •Sample size: With ~7,500 respondents split across 9 divisions, some divisions have small sample sizes (as low as ~330), which increases the margin of error for those estimates.
• •Self-reported data: HTOPS is a survey. Responses are self-reported and subject to recall bias, social desirability bias, and other survey limitations.
• •Composite indexes: The Wellbeing and Squeeze indexes are constructed by How Is America, not the Census Bureau. The choice of weights, normalization method, and included metrics reflects editorial judgment and is not an official government statistic.
• •Multi-select questions: For topics like AI tools, transportation modes, and insurance types, respondents could select multiple options. Percentages for these questions sum to more than 100%.
• •Point-in-time snapshot: HTOPS Wave 2506 captures a single point in time (March 2026). Conditions may have changed since data collection.
• •Supplemental data timing: State-level supplemental data comes from different time periods (USDA 2021–2023 average, BLS March 2025, ACS 2023) and may not align perfectly with the HTOPS wave.
• •Nonresponse bias: Despite PWEIGHT adjustments, some systematic differences between responders and non-responders may persist. Populations with limited internet access, language barriers, or distrust of government surveys may be underrepresented.

Trust, Price Stress, Childcare & Transportation Detail

Wave 2506 introduced several new topic modules analyzed on How Is America:

• •Institutional Trust: TRUST1 measures overall government trust. TRUST2_1 through TRUST2_9 measure confidence in 9 specific institutions (Congress, Supreme Court, President, military, police, media, public schools, banks, medical scientists). Additional variables capture trust in federal statistics. All coded on Likert scales; -99 and -88 excluded as missing.
• •Price Stress: PRICECHNG captures perceived price direction (increased/same/decreased). PRICESTRESS measures stress level among those perceiving increases. PRICECONCRN measures overall concern about rising prices. All weighted by PWEIGHT.
• •Childcare: CHILDCARE asks whether childcare was disrupted in the past 4 weeks (among households with children). CHILDCARE_RSLT1 through CHILDCARE_RSLT9 capture consequences (cut hours, unpaid leave, left job, supervised while working, etc.). -99 and -88 excluded.
• •Transportation Detail: TRANSPORT1 through TRANSPORT12 capture modes used (personal vehicle, bus, rail, rideshare, walk, bike, etc.). ACCESS_TRANSP measures adequacy of available transportation. NEEDS_TRANSP1 through NEEDS_TRANSP7 capture reasons for unmet needs (cost, safety, availability, disability). All use PWEIGHT for weighted estimates.

Further Reading

• •Census Bureau HTOPS Program Page
• •HTOPS Technical Documentation
• •HTOPS Public Use Files (PUF)