The target population of the 2017 Economic Census of Island Areas consists of all establishments that are:

• In operation for at least some part of 2017
• Located in Puerto Rico, the U.S. Virgin Islands, Guam, Commonwealth of the Northern Mariana Islands, or American Samoa
• Have paid employees
• Area classified in one of eighteen sectors defined by the North American Industry Classification System (NAICS)

        o   21—Mining, Quarrying, and Oil and Gas Extraction

        o   22—Utilities

        o   23—Construction

        o   31-33—Manufacturing

        o   42—Wholesale Trade

        o   44-45—Retail Trade

        o   48-49—Transportation and Warehousing (not  including 482—Rail

             Transportation or 491—Postal Service)

        o   51—Information

        o   52—Finance and Insurance (not including 525—Funds, Trusts, and Other

            Financial Vehicles)

        o   53—Real Estate and Rental and Leasing

        o   54—Professional, Scientific and Technical Services

        o   55—Management of Companies and Enterprises

        o   56—Administrative and Support and Waste Management and Remediation

             Services

        o   61—Educational Services (not including 6111—Elementary and Secondary

             Schools, 6112—Junior Colleges, or 6113—Colleges, Universities, and

             Professional Schools)

        o   62—Health Care and Social Assistance

        o   71—Arts, Entertainment, and Recreation

        o   72—Accommodation and Food Services

        o   81—Other Services (Except Public Administration) (not including 8131—

            Religious Organizations, 81393—Labor Unions and Similar Labor

             Organizations, 81394—Political Organizations, or 814—Private Households)

The frame for the Economic Census of Island Areas comes from the U.S. Census Bureau’s Business Register.  The Business Register contains information on the physical location of establishments, as well as payroll, employment, receipts (value of shipments), and industry classification data obtained from prior censuses and surveys or obtained from the administrative records of the Internal Revenue Service (IRS) and Social Security Administration (SSA) under special arrangements which safeguard the confidentiality of both tax and census records.  Information from the Bureau of Labor Statistics on industry classifications is also used to supplement the classification information from the IRS and SSA.

To be included in the frame, an establishment must satisfy the following conditions:

• It must be classified into an in-scope NAICS industry
• It must be an active establishment of a multi-unit firm, or it must be a single establishment firm with at least one quarter of 2017 administrative payroll
• It must be located in Puerto Rico, the U.S. Virgin Islands, Guam, Commonwealth of the Northern Mariana Islands, or American Samoa

The Economic Census of Island Areas is a complete enumeration of establishments located in the Island Areas.  Therefore, the accuracy of tabulations is not affected by sampling error.

A new frame is created every five years for each iteration of the Economic Census of Island Areas.

Sample selection for the 2017 Economic Census of Island Areas was performed in two phases.  The initial frame was created in November 2017.  Any establishments that appeared on the Business Register with 2017 payroll not on the original frame were added in March 2018.

2017 Economic Census of Island Areas questionnaires can be found at: https://bhs.econ.census.gov/ombpdfs/.  The reference period is the year 2017.

The key data items are the sales, value of shipments, or revenue; annual payroll; first-quarter payroll; and number of employees.

The Economic Census of Island Areas is a mandatory survey and response is required by law.

The 2017 Economic Census of Island Areas utilized a collection strategy offering both electronic and paper reporting options, depending on the type of respondent. Single-unit businesses were initially mailed a letter with instructions on how to report electronically, along with a paper copy of the questionnaire. Follow-up mailings consisted only of letters with electronic reporting instructions, although paper forms were available on demand throughout the collection period. Multiunit businesses received only letters with electronic reporting instructions, in both the initial and follow-up mailings.

Approximately 48,000 establishments were contacted in the initial mailing. The collection strategy included three mail follow-ups and an outgoing telephone follow-up.

The data collection unit is the establishment; there are approximately 51,000 establishments across the Island Areas.

The Economic Census of Island Areas utilizes a procedure that detects erroneous and/or questionable survey data called edits.  In other words, editing is a procedure for detecting incorrect (keyed or respondent) data.  Editing does not include changing the data. If data fails an edit the data is either referred to an analyst for review and correction or automatically imputed.

There are different editing methods: data validation edits, consistency edits, and ratio edits.  Data validation edits check if the value of a data item comes from a given set of acceptable values. Examples of range validation edits are range edits and list directed edits.  Range edits verify that data fall within a pre-specified range for continuous data; for example 0 ≤ percent from business ≤ 100. List directed edits verify that existing data values for a specified data item match predefined values provided in the questionnaire.  For example, all of the yes or no questions on the forms can only have values of 1 or 2 for yes and no.

Consistency edits check the response item against other associated responses or items to determine if a predefined relationship is followed.  Examples of consistency edits are balance edits and skip pattern validation edits.  Balance edits verify if the sum of detail items is equal to a data item total.  An example of a balance test is when annual payroll details do not equal to the total annual payroll.  The skip pattern edits verify that values of skip items are consistent with skip instructions provided in the questionnaire.  An example of a skip pattern is the e-commerce question. If the establishment had any e-commerce transactions, the respondent would go to provide the percent, if not the respondent would move on to the next question.

Ratio edits verify if the ratios of data items are within a certain range.  Examples of ratio edits are current cell ratio edits and historic cell ratio cells. The current cell ratio edits verify that the ratio of two different items from the same stat period and same ID are within a certain range.  For example, ratio edits help determine ratio inconsistencies between sales and annual payroll, annual payroll and hours, and annual payroll and number of employees.  The historic cell ratio edits verify within an observation that the ratio of an item in the current statistical period compared to a prior statistical period is within a certain tolerance.

Other required edits for Island Areas verify that a data value exists (i.e. the value is not blank) for a specified item.  Edit tests are also needed to summon general imputation if the question was left unanswered.  If a variable needs to be filled by general imputation, it must have an edit test to summon it. Survey rule edits are free-form tests that validate complex inter-item relationships.  Ratio edits are often implemented as survey rule tests. Most common edit tests used for Island Areas are range tests, list directed tests, balance tests, and survey rule tests.

Nonresponse is defined as the inability to obtain requested data from an eligible survey unit.  Two types of nonresponse are often distinguished.  Unit nonresponse is the inability to obtain any of the substantive measurements about a unit.  In most cases of unit nonresponse, the Census Bureau was unable to obtain any information from the survey unit after several attempts to elicit a response.  Item nonresponse occurs either when a question is unanswered or the reported data is unusable.

Nonresponse is handled by estimating or imputing missing data. Imputation is defined as the replacement of a missing or incorrectly reported item with another value derived from logical edits or statistical procedures.  The primary methods for imputing missing basic data items (such as receipts, sales, payroll, and employment) are:

• Using administrative data
• Deriving the missing data item from the establishment’s other data (either reported or administrative)
• Deriving the missing data item by statistical methods using data reported by similar establishments.

For many items, missing data is filled in using an auxiliary ratio method, which assumes a correlation between data items. For example, the auxiliary ratio equation for deriving missing receipts is as follows:

Receipts_i = Payroll_i · Σ(Receipts_j) / Σ(Payroll_j)

where Receipts_i is the missing receipts value being derived for establishment i, Payroll_i is the annual payroll value for establishment i, and Σ(Receipts_j) and Σ(Payroll_j) are the sums of all respondent receipts and annual payrolls within the imputation base.  If there are enough respondents which have passed certain edits tests in the same 5-digit NAICS as establishment i, then these respondents make up the imputation base for the item being imputed. Otherwise, the same check of establishments is done within the same 4-digit NAICS as establishment i.  The procedure continues up to the sector level until there are enough establishments that pass edit tests to form an imputation base.  Nonresponse for some other items is adjusted at the estimate level using expansion factors.  See the “Estimation” section for more information on this method.

The tabulation unit is the establishment.

Most estimates are simple summations of reported data, administrative data, and imputed data from all in-scope establishments.  The most common source of imputed data is administrative data from the IRS.  Dollar values are published in current dollars.  In tables that compare the current Census to prior Census, no adjustment has been made to the estimates to account for inflation during the intervening period.

The NAPCS tables, e-commerce tables, employment by function tables (Puerto Rico only), and storage space tables use expansion factors to account for nonresponse.  The remainder of this section outlines how the expansion factors are calculated and used.

An expansion factor is used to account for NAPCS product nonresponse.  The nonresponse adjustment factor is calculated as follows:

FACTOR = [Total Receipts] / [Reported Broad Line NAPCS Total]

In any industry with complete response for NAPCS, [Reported Broad Line NAPCS Total] = [Total Receipts] and FACTOR = 1.  Otherwise FACTOR > 1, and multiplying the total reported NAPCS by

FACTOR will expand the NAPCS total, such that total adjusted NAPCS equals total receipts.

A similar expansion factor is used on the e-commerce tables and storage space tables.  For those tables, the nonresponse adjustment factor is calculated as follows:

FACTOR = [Total Receipts] / [Total Receipts from Respondents Only]

The use of this factor makes the assumption that the level of response for e-commerce receipts is similar to the level of response for total receipts, and likewise for the level of response for storage space.

A separate expansion factor for establishment counts is calculated for the NAPCS tables for the count of establishments with a given NAPCS line.  The NAPCS establishment count expansion factor is calculated as follows:

FACTOR = [Total Establishment Count] / [Count of Establishments with Reported NAPCS]

The counts of establishments with a reported value for a given NAPCS product are multiplied by these factors to account for nonresponse.

A similar factor is used for counts of establishments with storage space:

FACTOR = [Total Establishment Count] / [Establishment Count, Respondents only]

All expansion factors are calculated at the 4-digit NAICS industry level.

Nonsampling error encompasses all factors other than sampling error that contribute to the total error associated with an estimate. This error may also be present in censuses and other nonsurvey programs. Nonsampling error arises from many sources: inability to obtain information on all units in the sample; response errors; differences in the interpretation of the questions; mismatches between sampling units and reporting units, requested data and data available or accessible in respondents' records, or with regard to reference periods; mistakes in coding or keying the data obtained; and other errors of collection, response, coverage, and processing.

Although no direct measurement of nonsampling error was obtained, precautionary steps were taken in all phases of the collection, processing, and tabulation of the data in an effort to minimize its influence.  Precise estimation of the magnitude of nonsampling errors would require special experiments or access to independent data and, consequently, the magnitudes are often unavailable. The Census Bureau recommends that individuals using these estimates factor in this information when assessing their analyses of these data, as nonsampling error could affect the conclusions drawn from the estimates.

The Economic Census of Island Areas used the following means to maximize response:

• Public awareness campaign to assist businesses and the public in understanding the importance of the 2017 Economic Census of Island Areas by providing news stories through trade and professional associations, chambers of commerce, and business and general media;
• Assignment of account managers to large businesses to provide personalized assistance;
• Redesigned census questionnaires and information sheets with standard formats and terminology that try to simplify reporting and minimize response burden;
• Electronic reporting capabilities that allow businesses to report using the Internet;
• Mailing materials that emphasize the mandatory and confidential nature of economic census reports, as provided by Title 13 U.S.C.;
• Toll-free assistance and/or a web-based help desk for any business that had questions about completing the EC of IA; and
• Systematic mail follow-up for nonresponse, supplemented by telephone follow-up for selected firms
• A respondent portal to provide answers to frequently asked questions, send secure emails to Census Bureau staff, report electronically, and perform other self-service options without having to pick up the phone.

For the 2017 Economic Census of Island Areas, the Census Bureau produced new response metrics in accordance with Census Bureau standard response rate calculations, in order to monitor data collection and to provide additional indicators of data quality.  These are the Unit Response Rate (URR), the Total Quantity Response Rate (TQRR), the Quantity Response Rate (QRR), the Administrative Data Rate (ADR), and the Imputation Rate (IR). (For definitions, see the Census Bureau Statistical Quality Standards, Appendix D3-B: Requirements for Calculating and Reporting Response Rates: Economic Surveys and Censuses, at https://www.census.gov/about/policies/quality/standards/appendixd3b.html.)

To produce these rates, the Census Bureau implemented a more detailed method for documenting the sources of data used for correcting estimated or inconsistent data. These correction sources align with those used in the Census Bureau’s annual economic surveys, and are defined as follows:

• corrections obtained directly from respondents
• data from administrative sources
• estimates by industry analysts based on data reported by similar establishments

When calculating the standard response metrics, the first two types of corrections are treated in the same manner as “reported” data.  For the third type of correction, the data are treated as imputed.  Census Bureau analysts’ estimates were treated as reported data for prior Economic Census of Island Areas reporting periods.  Data users should be advised that imputation rates calculated for the 2017 Economic Census of Island Areas are larger than similar metrics calculated for prior economic censuses, due to these procedural changes and to unit and item nonresponse in the classical sense.  The 2017 Economic Census was hampered by funding issues and delayed as a result of the partial government shutdown in January 2019.  Imputation rates are indicated with tabulated 2017 Economic Census data using a coding scheme like that presented with 2012 Economic Census data, as follows:

• 0: Imputation rate is less than 10%
• 1: Imputation rate is greater than or equal to 10% but less than 20%
• 2: Imputation rate is greater than or equal to 20% but less than 30%
• 3: Imputation rate is greater than or equal to 30% but less than 40%
• 4: Imputation rate is greater than or equal to 40% but less than 50%
• 5: Imputation rate is greater than or equal to 50% but less than 60%
• 6: Imputation rate is greater than or equal to 60% but less than 70%
• 7: Imputation rate is greater than or equal to 70% but less than 80%
• 8: Imputation rate is greater than or equal to 80% but less than 90%
• 9: Imputation rate is greater than or equal to 90%

Disclosure is the release of data that reveals information or permits deduction of information about a particular survey unit through the release of either tables or microdata. Disclosure avoidance is the process used to protect each survey unit’s identity and data from disclosure. Using disclosure avoidance procedures, the Census Bureau modifies or removes the characteristics that put information at risk of disclosure. Although it may appear that a table shows information about a specific survey unit, the Census Bureau has taken steps to disguise or suppress a unit’s data that may be “at risk” of disclosure  while making sure the results are still useful.

The Economic Census of Island Areas uses noise infusion as its primary method of disclosure avoidance.

Noise infusion is a method of disclosure avoidance in which the quantitative data values for each tabulation unit are perturbed prior to tabulation by applying a random noise multiplier to the magnitude data, such as sales or payroll.  Each record is assigned a single noise factor, which is applied to each quantitative data item as part of the estimation process.  Disclosure protection is accomplished in a manner that causes the vast majority of published cell values to be perturbed by, at most, a few percentage points.

The Economic Census of Island Areas employs a noise infusion method that balances the noise on receipts at the 3-digit NAICS level.  This means that using the receipts item, the noise factors are assigned in such a way to attempt to minimize the final total noise in that cell.  This is accomplished by sorting the establishment records within a cell by receipts and sequentially assigning noise at each stage to have the opposite sign of the total effect of noise in the cell.  This does not completely eliminate noise, but it can significantly reduce the effect of noise at the 3-digit levels and above.

Each published cell value has an associated noise flag indicating the relative amount of distortion in the cell value resulting from the perturbation of the data for the contributors to the cell.  The flag for ‘low noise’ (G) indicates the cell value was changed by less than 2 percent by the application of noise, and the flag for ‘moderate noise’ (H) indicates the value was changed by at least 2 percent but less than 5 percent.  Cells that have been changed by 5 percent or more are suppressed from the published tables.  Though some of these suppressed cells may be derived by subtraction, the results are not official and may differ substantially from the true estimate.

The number of establishments in a particular tabulation cell is not considered a disclosure; therefore, this information may be released without the addition of protective noise.  For an introduction to the noise disclosure avoidance method, see Using Noise for Disclosure Limitation of Establishment Tabular Data [PDF] by Timothy Evans, Laura Zayatz, and John Slanta in the Journal of Official Statistics (1998).

For sample-based tabulations, the estimated relative standard error for a published cell includes both the estimated sampling error and the amount of perturbation in the estimated cell value due to noise.

The Census Bureau has reviewed the data product for unauthorized disclosure of confidential information and has approved the disclosure avoidance practices applied.  (Approval ID: CBDRB-FY20-106).

For more information on disclosure avoidance practices, see FCSM Statistical Policy Working Paper 22.

The Economic Census has been taken at 5-year intervals for the following areas, except where noted below:

• American Samoa — Since 2002
• Guam — Since 1958
• Commonwealth the Northern Mariana Islands — Since 1982
• U.S. Virgin Islands — Since 1958
• Puerto Rico — Economic Censuses were conducted beginning with a census of manufactures for 1909 and continuing at 10-year intervals through 1949, excepting 1929. Wholesale and retail trades and services industries were included as part of the economic censuses for 1939. Starting with 1949 through 2002, the censuses of wholesale and retail trades and service industries have been conducted concurrently with the census of manufactures. The census of construction industries has been included since 1967. Congress has authorized the economic censuses to be taken at 5-year intervals covering years ending in "2" and "7."

The range of industries covered in the economic census for American Samoa, the Commonwealth of the Northern Mariana Islands, Guam, the U.S. Virgin Islands, and Puerto Rico was broadened in 2002. Sectors added for the first time in 2002 include information, finance and insurance, real estate, health care, and other service industries.

Printed statistical reports from the 2002 and earlier censuses provide historical data for the Commonwealth of the Northern Mariana Islands, Guam, and the U.S. Virgin Islands, and are available in some libraries.  For American Samoa only the 2002 printed report provides historical data.  Printed reports from 1997 and earlier censuses provide historical data for Puerto Rico.  Reports for 1992, 1997, and 2002 are also available in portable document format (PDF) on the Internet.