3. Data Collection and Editing Procedures

The mode of data collection may offer additional options for collecting detailed race and ethnicity data.

In electronic modes of collection, for example, agencies may use multiple screens to collect detailed data. The minimum reporting categories may be collected on an initial screen and detailed data for each minimum reporting category the respondent selected may be collected on follow up screens, whether through checkboxes, drop down menus, write-in areas, or another method.

If detailed race and ethnicity data are collected in an interviewer-administered setting, the minimum categories should be asked first, treating each category as a yes/no question, followed by the detailed categories associated with the selected minimum categories.

The method of data collection has implications for the quality and fitness for use of the resulting data.

Wherever possible, race and/or ethnicity data should be collected through self-report, where the respondents directly provide their own race and/or ethnicity.

In cases where self-report is not possible, data may be collected:

• By proxy reporting, where a person knowledgeable of another's race and/or ethnicity responds on their behalf;
• By record matching, where existing records on an individual that contain their race and/or ethnicity are used to supply the information; or
• By observer identification, where an observer uses their best judgement of the most appropriate race and/or ethnicity categories in which to report an individual.

When data are collected through visual observation, agencies are not required to collect detailed categories and are encouraged to instead use the minimum categories. For statistical survey reporting, agencies must maintain records on the mode and method of data collection, and how nonresponse or other missing data were assigned or allocated, and must make that information available to data users to allow them to evaluate the utility, objectivity, and integrity of the data. Agencies should also maintain and provide this information for administrative, grant, and compliance-related data collections whenever feasible. Agencies should use the terminology in this section when describing the method of collection and should make it a practice to describe the method of data collection in any reports on data collection design or methods.

When coding write-in data, imputing missing data, or otherwise editing responses, agencies must adopt practices that maximize comparability between data collected on forms and surveys with and without write-in fields. Doing so will improve the comparability of race and ethnicity data across Federal datasets. For statistical survey reporting, agencies must maintain records on data processing procedures (such as coding, editing, and imputation practices), and must make that information available to data users to allow them to evaluate the utility, objectivity, and integrity of the data. Agencies should also maintain and provide this information for administrative, grant, and compliance related data collections whenever feasible.