Establishment Survey

Differences between Establishment and Household Surveys

Establishment surveys differ from household surveys in a number of notable ways, creating some unique challenges and considerations for survey researchers who are conducting establishment surveys. The most fundamental difference between establishment and household surveys is the unit of analysis. Whereas in a household survey, the unit of analysis is the household, family unit, or an individual, the unit of analysis for an establishment survey may be either an establishment or an enterprise. An establishment could be a business operating in a particular location, a business entity reporting unemployment insurance claims, a hospital or clinic location, and so on. An example of an enterprise would be a corporation that includes one or more locations or establishments. The term establishment survey is used as a generic or umbrella term to refer to surveys that collect data at either the establishment or enterprise level.

A second important distinction between establishment and household surveys is the use of informants. Whereas informants to a household survey are generally selected as part of the sampling design, informants to establishment surveys are often the person or people within an organization who are the most knowledgeable about the subject matter of the survey or those who have access to the required information. Also, although household surveys generally rely on one person to answer questions about the household overall or others living in the household, there may be multiple informants to an establishment survey, depending on the survey topics and the complexity of the establishment.

The complexity of selecting proper informants to respond to establishment surveys is usually correlated with the complexity of the sampled organizations. An example of an establishment with a simple structure for identifying a proper informant would be a nursing home, where the administrator of the nursing home would generally be the desired informant. Establishments with more complex structures, such as universities or larger businesses, may have a more decentralized structure that requires additional effort to determine who is the most appropriate informant and may include multiple informants. Surveys of larger or more complex establishments conducted via telephone may require additional time to identify appropriate informants; this is especially true of establishments for which there is no contact name provided from the sample frame. For this reason, some survey firms have invested in the additional effort of pre-contacting the establishment to obtain the name of a contact person (and potential informant) prior to the full survey being fielded. Given that the informants completing establishment surveys are answering questions about the organization as a whole, or about others within the establishment, survey questions for the establishment that ask for hard, factual data about the operations of the establishment tend to be more appropriate than questions that ask opinions or require subjective evaluation.

...