Why do frequencies for IPUMS samples sometimes differ from official results?

INTERNATIONAL
1

Why do frequencies for IPUMS samples sometimes differ from official results?

2

There are a number of reasons why the IPUMS sample data may yield different counts from official census results:

  1. Sample error. IPUMS microdata are samples, not the full-count data.

  2. Sample bias. The IPUMS sample unit is usually the household. This introduces a slight bias in estimating statistics for individuals – greater bias for characteristics more commonly shared by households members (ethnicity, religion, birthplace) than by individuals (age, sex, marital status).

  3. Statistical disclosure controls. Measures to safe-guard the privacy and confidentiality of individuals, households and other entities, can introduce error in the samples.

  4. Omission of special populations, such as homeless, collective dwellings, or other non-private households (usually documented on sample description pages)

.

  1. Omission of areas of the country due to loss of microdata, lack of coverage, security concerns, etc. (usually documented by a flag on the sample selection page).

  2. Imputations and adjustments to the official figures by the National Census Office that are not encoded in the microdata files provided to IPUMS.

Users who require official census results should consult web pages of the National Census Offices.