Living in the “information age”, we use terms like “data” and “information” all the time, without necessarily being precise in our usage. But it can be helpful to think more carefully about the differences between these terms, and how they impact the way we sift through and apply all the resources that are now available to us.
In this more specific sense, we can distinguish three terms:

Data:

observations, measurements, or other facts. Data may be incomplete, redundant or irrelevant.

Information:

data that is selected and enriched with structure, context and meaning. “Connecting the dots” describes the process of moving beyond mere data to some larger picture, with larger meaning and significance. Information in this sense implies purpose: what’s information to you may not be information to me, at least at this point in time, if i have no use for it.

Knowledge:

Knowledge is information in action, applied toward solving a problem, answering a question, or accomplishing some other objective. Knowledge extends beyond a particular set of data and information, providing predictive power, thereby also extending over time from today to tomorrow. It can be individual (my personal view) or shared across a community.

These elements are often visualized in a hierarchical or pyramid structure, showing how information is built upon data, and knowledge builds further upon information. The pyramid also suggests a quantitative relationship: we need relatively more data to derive a smaller amount information, and similarly more information to derive knowledge. Some would add wisdom, representing knowledge coupled with values or fundamental principles at a higher level, as another layer on top (consequently, this is sometimes called the DIKW model).

Data moving toward information is more objective: we can look together at the data and talk about what they mean. Information moving into knowledge is more subjective: i can try to communicate my knowledge to you, but that process is complicated by our different contexts and experience, and we may disagree about what conclusions to draw from the facts we agree on.

In the words of Harvard zoology professor Louis Agassiz: “Facts are stupid things, until brought into connection with some general law.” (recast in the terms above, data and knowledge.) (Samuel H. Scudder’s “The Student, the Fish, and Agassiz“, from which this quote is taken, is a classic account of the value of first-hand observation and inductive learning) The hierarchy, and the importance of moving toward the upper levels, is famously echoed in T.S Eliot’s poem “The Rock”:

Where is the Life we have lost in living?
Where is the wisdom we have lost in knowledge?
Where is the knowledge we have lost in information?

One of the wonders of human intelligence is that we can move easily between the various levels of this model. We leap (sometimes too quickly!) from data through information up to knowledge. Our ability to retain extensive knowledge in our heads can make some otherwise insignificant piece of data suddenly meaningful (that elusive experience factor that schools can’t streamline the teaching of, and automated systems still can’t touch). In this way, experience can create a communications barrier between the knowledgeable and the novice. Knowledge and information in context makes an easy progression transparent to the experienced. The novice, however, can’t move between levels so easily, and requires explanation, background information, and other help to move up the stack.

The entire scientific enterprise is a constant process of acquiring data, forging information from it, deriving knowledge, and then using that knowledge to focus the search for additional data to expand our knowledge even further. One consequence of modern “information” technology is a data and information surplus, which naturally leads to an attention deficit, as more and more information stretches our ability to focus and prioritize.

To make this less abstract, here are some examples from the domain of learning about the Bible.

Word Study:

The occurrence of a particular Greek word in a verse of Scripture constitutes data. It is not the subject of debate: once past the manuscript evidence, we all agree on what the words are. Looking at the distribution of words across text moves toward information: we might observe that the words translated ‘tax collector’ and ’sinner’ frequently occur together in the Gospels (of the 21 verses that include ‘tax collector’ (τελώνης), 9 also include ’sinner’ (ἁμαρτωλός)), but that John’s Gospel never uses τελώνης at all. We have not yet reached a ‘why’ (knowledge) behind this information: for that, we might look to other resources (the data per se do not contain a direct ‘why’). For example, the New Bible Dictionary suggests the tax collectors’ extortion, support of the ruling power of Rome, and habitual contact with Gentiles were several of the reasons that rabbis taught a good Jew should not eat with tax collectors. This attitude may help explain the linguistic association, and suggests other investigation (requiring new data) to either confirm or deny the conclusion (for example, several verses that don’t include ’sinner’ do include semantically related terms like ‘prostitute’ or ‘Gentile’).

Morphology:

Like word observations, the particular tense of a verb or case of a noun are observable facts (data). Our knowledge of the significance of e.g. the aorist tense in a given passage comes from our general grammatical information (based on other instances of its usage), and our conclusions of the deeper understanding of a passage.

Syntax:

We can observe particular syntactic constructions in the Scripture (data): what, if anything, are their deeper functions (a knowledge question which can only be answered based on information)?

Discourse Structure:

An observed phrase (data) may be repeated, or put in a syntactically unusual position like the beginning of a clause (’unusual’ here assumes we have information about what ‘usual’ is based on observing patterns of syntax). This is only information until we determine what the function of this linguistic device might be: perhaps making a particular detail of the narrative stand out, or introducing new material. This hypothesized knowledge can then be tested against other passages to see whether the data support the conclusion.

These are all linguistic examples, though the data-information-knowledge framework is not restricted to the domain of language (you could do similar kinds of analysis within the field of systematic theology, for example). Learning to think within this framework can help us more clearly identify what we “know” and where that knowledge comes from (that is, its supporting data and information), as well as helping us more effectively communicate our knowledge to others.
(There are a few more websites on this subject under my del.icio.us tag ‘dikw’.)