Information Quality

Leading organizations no longer treat information quality as an unknown by-product of the operation of information systems. With the explosive growth of interconnected information sources and the nearly complete reliance on information systems for critical operations, enterprises today face serious financial consequences as a result of bad data.

Information quality is highly context dependent. It is directly tied to the knowledge that the information is intended to impart to the user. The problem with understanding information quality comes when we realize that it is multi-dimensional. The various attributes to information quality that make defining a measurement of that quality a challenge.

For example, suppose I need the temperature of the outside air for two purposes. The first is to determine if the air temperature is within the tolerable range for my prized orchid. The second is as an input into one of the microprocessors that controls the mixture of air and fuel vapors in my car. In the first case is need to know the temperature within a degree or two (not very accurate), but I need to know it now. In the second, my car needs its input in real-time as well, but it also needs it to two decimal point accuracy, +/- .01 degree. So, if I get the temperature for yesterday, it has low quality for my decision, but if my car gets it today and it is not accurate as required, then it is also of low quality. Now, one could argue that we are talking data value and quality. However, the context dependency of the quality of the data is what determines its value.

There are many possible attributes of data quality. Two were discussed in the above example: timeliness and precision. But there are others. The following is a list of generally accepted information quality attributes:

Accurate	Information achieves its intended purpose. The methodology for data collection or creation including contextual metadata is available for users to interpret, evaluate and understand the level of accuracy. The information is unbiased and allows users to make fully informed decisions. If information represents opinion, this should be made clear. Information does not lose its own accuracy or that of its metadata as it moves through formats, hardware or software.
Complete	Comprehensive to fully address all intended uses and is not misleading due to omissions or undue complexity. Where appropriate, includes metadata that provides the provenance and context (e.g. author, intended use context, access controls, information for audit, etc.) for the information or service being described.
Reliable	From an authoritative or commonly accepted sources and has been created or compiled using commonly accepted methods. The results it demonstrates can be reproduced, especially where scientific data is concerned. Can withstand the scrutiny of processes such as peer review. Is protected from unauthorized change.  Ability to authenticate the source of information, including any approvals or amendments. Clearly attributable to an originating authority
Understandable	Clear in meaning and easy to read. The level of precision is such that purpose for the intended audience is achieved. In general, plain language is preferred to complex language Complex language and structure may be appropriate depending on the purpose and intended audience.  For example, a scientific research paper.
Relevant	Created with an intended purpose in mind and meet that purpose without burdening the recipient with undue detail or extraneous content.
Current	Up to date for the purpose intended. The timeframe in which information was created must be evident to users. Users need the most recent information or may need historical information created many years ago.
Accessible	Presented in a medium that its audience can access. Easy to navigate, whether in paper or electronic form. Easy to locate using the document structure, metadata, taxonomies, search engines, and other tools and techniques provided to help.
Timely	Available in an expedient manner that responds to client needs.