What would happen if Americans were not required by law to respond to census surveys? This is of interest because Canadian government officials have decided to drop the mandatory long form in next year’s census and replace it with a voluntary household survey. All Canadian households will still be required to answer basic questions on the census short form.

The Census Bureau tested this idea a few years ago when it researched how response rates to the American Community Survey would change if survey response was not required. The American Community Survey, a monthly survey that has replaced the census long form, includes questions about topics that include educational attainment, commuting, citizenship, income and housing costs. The bottom line: There is a “significant drop in mail cooperation” when respondents are not required to respond, Census Bureau researchers reported.

Experiment Design and Results

Researchers came to this conclusion after testing four different types of messages as part of American Community Survey mailings in March and April 2003 to about 138,000 addresses. Three-quarters of the sample was evenly split between two “voluntary mail treatments,” and the remaining 25% was split evenly between two “mandatory mail treatments.” Results were calculated not only for the national sample, but also for areas that had had high and low response rates to the 2000 Census.

Among households that were told survey response was mandatory, the “mail cooperation rates” (mail-back rates from contacted, occupied households) were 59.2% and 57.3%, depending on how the message was framed. Among those sent a voluntary message, response rates were 38.8% and 34.8%, depending on the wording of the letter accompanying the survey form; the lower cooperation rate came from households that received a more direct message stating in the first paragraph that participation was voluntary. Mail cooperation rates dropped the most in areas that had high cooperation rates during the 2000 Census, compared with low-cooperation areas.

If households do not return their forms after being reminded in a follow-up postcard, the Census Bureau tries to reach them for a telephone interview and then contacts a sample of non-responding households for face-to-face interviews, which cost more than processing a mailed-back form. Researchers said they found more resistance to those interviews from households mailed the voluntary form than from those mailed the mandatory form.

For the mandatory survey, the combined weighted response rate from mailed-back forms and the sample of follow-up interviews was about 98%, but was only about 93% for the voluntary survey. Researchers say the voluntary survey also could result in lower data quality — unless sample size were increased — because its greater reliance on personal interviews contributes to an increase in sampling error as the result of weighting issues.

Higher Costs of Voluntary Survey

If the American Community Survey were made voluntary, the Census Bureau concluded that it would have to mail survey forms to a larger number of households and conduct a larger number of follow-up phone or personal interviews in order to assure the same quality of results that it promises from the mandatory survey. Thus, a voluntary survey would cost at least $59.2 million more, in 2005 fiscal-year dollars, than a mandatory survey, the bureau concluded.

Interestingly, survey respondents who mailed back their forms were about equally likely to fill out all items on the form whether response was mandatory or voluntary. Among those interviewed by telephone or in person, researchers reported an increase in missing data from households that received the voluntary survey, compared with those that received the mandatory survey, but the difference was small.

The experiment, conducted at the request of Congress, may have reduced any pressure for the bureau to consider a voluntary American Community Survey. But it did result in one operational change. In comparing the two mandatory-survey messages it tested, the Census Bureau concluded that the one it had been using produced a slightly lower response rate than a more user-friendly version, in which some explanatory information was moved from a letter to a brochure. After reporting the results, the bureau implemented the more user-friendly mandatory mailing package.

 Find more about the ongoing decennial count and other census-related matters at All Things Census.

For more than a decade, the Census Bureau has been developing an alternative measure of poverty that is intended to better reflect the costs of basic living expenses as well as the resources people have to pay them. The bureau has just released results for 2010 from the alternative metric-called the Supplemental Poverty Measure (SPM) — that uses a wider range of factors than the official federal measure to determine poverty status.

The additional data used in the alternative measure include medical expenses, tax credits, non-cash government benefits (such as food stamps, housing subsidies and school lunch programs), and cost-of-living adjustments for different areas. The alternative measure is not intended to replace the official poverty measure, at least for now. For the foreseeable future, the Census Bureau will report two sets of numbers.

Compared with the official measure, SPM figures released by the Census Bureau show a higher national poverty rate for 2010, 16.0%, compared with the official poverty rate of 15.2%

When the alternative measure is used, a greater share of Hispanics in 2010 lived in poverty than any other group. By contrast, when using the official poverty rate, a greater share of blacks in 2010 lived in poverty than Hispanics or any other group. Even so, no matter which measure is used, Hispanics make up nearly three-in-ten of the nation’s poor-28.6% under the official poverty measure and 28.7% under the SPM.

Read the full report for more detailed information on the Census figures.

Related Story: Childhood Poverty Among Hispanics Sets Record, Leads Nation

The 2010 Census certainly did miss some people; in that respect, it is like every other census. The Census Bureau conducted an independent follow-up survey in order to estimate how many people were not included. Results will be published next year. Some other quality indicators already are out — for example, the bureau said it obtained at least some useable information from 99.62% of the nation’s housing units, an increase from 99.45% in 2000.

If the Census Bureau does not receive a completed form from a particular address, and census-takers cannot reach anyone there after multiple visits or calls, the agency tries to obtain information about the household from neighbors or a building manager. If that fails, the Census Bureau resorts to a last-ditch statistical technique called “imputation” to fill in the missing data. As this article on our “All Things Census” page explains, agency analysts use imputation when they do not even know whether someone lives at a particular address, when they know someone lives there but not how many people, and when they know how many people live somewhere but do not know their race or other characteristics. Imputation is based on what bureau analysts know about the size and type of neighboring households. In the 2010 Census, 1.16 million people were added to the count through imputation, or .39% (less than half a percent) of the total.

D’Vera Cohn, Senior Writer, Pew Research Center

The counts and characteristics of same-sex couples are among the most written-about data from the 2010 Census and American Community Survey. Yet, two decades after the Census Bureau began offering people the option to describe themselves as a same-sex “unmarried partner,” producing accurate numbers remains a challenge.

The quality of information about same-sex couples depends both on the questionnaire responses the Census Bureau receives, and the procedures the agency implements to collect and edit those responses. There have been changes to both during the years the bureau has been releasing same-sex couple data, and the agency has cautioned against comparing numbers from different years to produce trends.

The Census Bureau has just completed the release of state-by-state total counts of same-sex couples from the 2010 Census, which adds up to a national total of about 902,000. Later this year, the bureau will release a highly anticipated count of same-sex married couples from the 2010 Census, the first time it has done so from a decennial census. In conjunction with that release, the bureau will publish its own evaluation of data quality about same-sex couples. This posting describes what is known so far about data quality for both married and unmarried same-sex couples from past censuses, the American Community Survey and Census Bureau research.

The challenges of counting same-sex couples illustrate the difficulties of data collection and group identification in an era of rapid social change. The Census Bureau has a delicate balancing act as it tries to capture a demographic snapshot even as national norms are more akin to a motion picture.

Read the full article by D’Vera Cohn detailing the challenges of counting same-sex couples at pewsocialtrends.org. Keep up to date on population and demographic developments at All Things Census

When census-takers can’t reach anyone at a particular address or obtain information about occupants in other ways, they sometimes use a last-resort statistical technique called “imputation” to fill in missing data. One marker of the quality of a census is how much it relies on imputation to add people to the count.

In the most extreme cases, census-takers have only an address taken from a master list drawn up in cooperation with local officials. They may not even know that a housing unit exists at that address, much less who lives there. If the address is indeed found to be an apparent dwelling place, the census-taker may not be able to get anyone to come to the door, and neither neighbors nor building managers may be willing and able to supply information. Yet the Census Bureau’s orders are to count everyone living in the U.S. on April 1, Census Day.

So to meet the goal of having a complete and accurate census, the Census Bureau imputes the existence and number of people living at the address in question, a procedure known as “count imputation.” (The other kind of imputation, called “characteristic imputation,” is when the Census Bureau has a head count for an address but is missing race, age or other personal information.) The number of imputed people tends to be higher among hard-to-count groups such as ethnic and racial minorities.

In 2010, according to figures supplied by the Census Bureau, 1,163,462 people were added to the household population (the total excluding group quarters) via count imputation, or .39% (less than one half of one percent) of the total. This served to boost the household population to slightly more than 300 million; pre-imputation, it stood at 299,594,753. By comparison, in the 2000 Census, 1,172,144 people were added to the household population via count imputation, or .43% of the total.

Using another metric — addresses, not people — a slightly higher share in 2010 had some usable information available, so a slightly lower share required count imputation (.38% compared with .55% in 2000). Count imputation was performed on a slightly lower number of addresses in 2010 than in 2000 as well — 521,947 compared with 666,848 in 2000, according to Census Bureau figures.

Types of Count Imputation

There are three kinds of count imputation. “Status imputation,” the most extreme form, is when census-takers do not even know whether a particular address is a real livable residence (maybe it’s a business or in such disrepair that no one could live there), and, if so, whether the unit is occupied or vacant. “Occupancy imputation” is when the Census Bureau knows that an address is a real housing unit, but not whether anyone actually lives there. “Household-size imputation” is when an address is known to be a real, occupied home, but census-takers don’t know how many people live there.

The Census Bureau has released data comparing rates for the three different types of count imputation in 2010 and 2000, using home addresses as the denominator. Household-size imputation, the largest of the three categories in 2010, accounted for .24% of addresses, followed by status imputation (.12%) and occupancy imputation (.03%). In 2000, status imputation was the largest category (.23% of addresses), with occupancy and household-size imputation accounting for .16% each.

In 2010, of the 521,947 addresses without usable information, about 325,000 required household-size imputation in 2010, 38,000 required occupancy imputation and 159,000 required status imputation, according to figures supplied by the Census Bureau.

State Patterns

There was less variation among states in 2010 than in 2000 in the rate of count imputation, according to the Census Bureau. “We have lower variability in the data,” Census Bureau Director Robert M. Groves told a recent press briefing. “We like that result.”

In all but seven states and the District of Columbia, count imputation rates declined from 2000 to 2010, according to figures supplied by the Census Bureau; the states where rates did not decline include Arkansas, Colorado, Georgia, Louisiana, Mississippi, New Jersey and North Carolina.

The highest count imputation rate in 2010 (District of Columbia, at .93% of addresses) was notably lower than the highest rate in 2000 (Arizona, 1.37 %).

The state that had the highest number — as opposed to the highest proportion — of people added to its total via count imputation in 2010, according to Census Bureau figures, was Texas (143,813), followed by Florida (100,575), New York (92,600), Georgia (82,026) and then California (68,204). As a share of all count imputations, the most extreme kind — status imputation — made up the majority in a dozen states, generally smaller ones.
Imputation Technique and History

In carrying out imputation, the bureau applies what it knows about the size and type of neighboring households to fill in the number of people, or their characteristics, at the addresses with missing data. Imputation procedures have grown more sophisticated over the decades.

In the 1940s, imputation was based on a random-ordered set of values on a set of punched cards — a “cold-deck imputation” as it came to be known. Now, the bureau uses a “hot-deck imputation” technique that employs continually updated census data from similar people or households generally within the same census tract as the basis for assigning a value to a missing record.

The Census Bureau’s use of imputation to add people to the total count dipped in the 1990 Census, in part because the agency put on a large and expensive operation to reach non-responding households. Only 54,000 people were added to the census count in 1990, compared with 1.2 million in 2000. In 2000, the agency relied on imputation in part as a way to restrain the rising cost of contacting non-responding households.

A generally positive evaluation of the 2000 Census by the National Research Council described these 1.2 million imputations as “a problematic group with regard to accuracy,” but acknowledged that if they had not been included, the census “would undoubtedly have underestimated the true number of household residents (particularly when a unit was known to be occupied.).”

The National Research Council evaluation noted that the number of imputed people was low compared with the total size of the 2000 Census count, but the share of imputed people is higher among some hard-to-count groups such as renters and ethnic and racial minorities. The National Research Council panel noted that imputations at addresses that were not known even to be housing units tended to cluster in rural areas, such as the Adirondacks region in New York and parts of Arizona and New Mexico. Some of these addresses may have been fishing camps or other temporary recreational lodging, it noted.

Imputation and Statistical Sampling

Although imputation is a statistical technique, it differs from the politically controversial statistical sampling procedure that the bureau has considered using in the past. For 2000, the Census Bureau proposed to use statistical sampling as part of its non-response follow-up operation to estimate the size and characteristics of the population that didn’t return census forms. That plan was abandoned after the Supreme Court ruled that statistical sampling cannot be employed to produce census data used to apportion congressional seats among the states.

The court ruling did not bar the use of statistical sampling to adjust census numbers for the purpose of producing annual population estimates and distributing federal funds. But the Secretary of Commerce, who oversees the bureau, chose not to do so, on the recommendation of a study conducted by a committee of bureau officials.

The Census Bureau’s use of imputation was challenged in court after the 2000 Census by the state of Utah, after the state failed to gain a congressional seat, which instead went to North Carolina. Utah argued that it, not North Carolina, would have gained the seat had it not been for imputations-in particular, imputations of people in housing units for which household size was not known. Those imputations, Utah argued, amounted to illegal use of statistical sampling. In 2002, the Supreme Court rejected Utah’s lawsuit (Utah et al v. Evans).

The Supreme Court ruled that sampling and imputation differed in three key respects-the “nature of the enterprise,” “methodology,” and “immediate objective.” It described imputation as “inference,” not statistical sampling. The court also stated that imputation does not violate the Constitution’s requirement for an “actual enumeration,” which some have contended prohibits any methods other than a door-to-door count. Without imputation, the court stated, the result would be “a far less accurate assessment of the population.”

 Find more about the 2010 count at All Things Census

The average size of U.S. households has been declining for decades, but may have grown in recent years, at least in part because of an increase in multi-generational households. Data from the 2010 Census will supply additional information about whether, where and why this is happening.

Meanwhile, here is an early clue: Average household size in Maryland, which had declined in the 1970s, 1980s and 1990s, grew by a tiny amount from 2000 to 2010, according to numbers analyzed by the Maryland State Data Center. In 1970, average household size was 3.25 in Maryland; in 1980, it was 2.82; in 1990, 2.67; in 2000, 2.61. In 2010, it was again 2.61 (the 2000 and 2010 numbers look the same, but the data center analysis shows there actually was an increase of .01 before rounding). Household size grew in several large suburban counties, according to the analysis.

The Census Bureau plans to release statistics on household size in May, but they can be calculated now based on other data released by the bureau this week. To accommodate states such as Maryland that want to take account of prison populations in legislative redistricting, the bureau released detailed counts of populations for group quarters (including prisons). If group quarters counts are subtracted from the total population counts released earlier, the number of people living in households can be derived, then divided by the number of occupied housing units included in the bureau’s recently released redistricting files to produce average household size.

 Find more about the 2010 count at All Things Census

In addition to publishing detailed numbers from the 2010 Census, the Census Bureau has been releasing performance indicators from the count. They offer clues to help answer the question of how well the bureau did in counting the entire U.S. population, only once, and in the right place.

The most crucial evidence, an independent post-census quality survey, will not be published until next year. That survey will produce measurements of undercounts and overcounts of specific race and Hispanic groups. However, other information is available now that speaks to the completeness of the count as well as to its data quality. This posting explains some of these key metrics and places them in a chronological narrative of the census-taking process.

Mail Participation Rate: The Census Bureau mailed millions of forms in March 2010. An early important indicator was the mail participation rate: 72% of occupied households sent back their forms, the same as for the short form in 2000. This was a good result, both for cost and quality. Non-response follow-up is expensive, and mailed-in census forms tend to have more accurate information than can be obtained from interviews weeks or months after Census Day, April 1
.
Follow-up Counts: The Census Bureau began its non-response follow-up operation May 1. Census-takers made up to six attempts to contact each household from which forms had not been received-about 47 million of them. If they could not reach the occupants, they would try to get information about them from “proxies, ” secondhand sources such as neighbors or building managers.

Census-takers resorted to secondhand sources to obtain information about more than a fifth of those follow-up households — 22%. That was higher than in 2000, when data on 17% of follow-up households came from proxies. This has implications for data quality, because secondhand information is less reliable.

Overall Count Rate: By the time census-takers were done with the follow-up phase of the count, they had obtained at least some usable information from 99.62% of the nation’s housing units, either from census forms or interviews, according to Census Bureau statistics. That was slightly better than the 99.45% total in 2000.

If census-takers could not obtain usable information about a household address-in some cases, they did not even know whether the address was a housing unit-they used statistical methods to fill in the missing data. That procedure is known as imputation, which will be discussed in a later posting.

 Find more about the 2010 count at All Things Census.