Information literacy: evaluating evidence and sources

Information literacy is the ability to work out what information a question requires, find it, judge its quality and relevance, and use it responsibly. It matters whenever a researcher has more material than they can reasonably read and no single source answers the question completely. A long list of search results is not evidence of a sound investigation. The important work is deciding which results help answer a particular question and what each can legitimately support.

Suppose a student asks whether a workplace training programme improves employee retention. They can find company testimonials, survey results, published studies and opinion pieces within minutes. Those sources do different jobs. A testimonial can suggest a mechanism; a survey may show what employees report; a study with a comparison group may support a stronger claim about outcomes. Information literacy helps the student avoid treating every item as interchangeable merely because it appears in a search result.

Information literacy starts with a defined need

Start by turning a broad subject into a question with boundaries. “Does training work?” leaves the population, training type, outcome and time frame undefined. “Among newly hired retail staff, is a structured onboarding programme associated with retention over the first year?” gives the search a clearer direction. It also shows what may still be missing: the quality of onboarding, the comparison with other staff and reasons employees leave.

Write down the concepts separately. In this example they include new hires, onboarding and first-year turnover. List sensible alternatives for each, such as induction for onboarding and employee retention for turnover. Search systems may use different vocabulary, and studies may measure a related but narrower outcome. A concept map prevents the search from being trapped by the wording of the first query.

The question may change as evidence appears. Perhaps the available studies focus on healthcare rather than retail, or measure intention to leave rather than actual departure. That does not justify quietly redefining the outcome. Record the change and explain why a new question fits the available evidence. An information need is a working decision, not a slogan fixed before reading begins.

Build a search strategy that can be revised

Use several routes appropriate to the subject: a library catalogue for books, disciplinary databases for research studies, official statistics for a population measure, and organizational documents for a programme’s own details. Search terms can be combined by concept, then adjusted when results are too broad or too narrow. A useful search log records the database, date, terms, filters and a brief reason for changing the query.

Do not assume that the first page of results contains the strongest evidence. Ranking may reflect popularity, recency, matching words or a platform’s rules. Search the reference list of a relevant paper to locate earlier work, and check later publications that test or challenge its claims. These steps can reveal whether an appealing finding is isolated or part of a larger conversation.

Keep inclusion rules proportionate. A review of current onboarding practice may need recent work, but an explanation of a foundational theory may require an older original text. Limiting everything to the past few years can hide the basis of a concept; accepting everything ever written can overwhelm the analysis. Record why a time limit, language limit or study-design filter was chosen.

Evaluate evidence and sources in context

Source evaluation asks what a source was made to do. A company case study may accurately describe its own programme but cannot by itself establish that the programme caused improved retention. A survey’s percentages need a sample, question wording and a denominator. Research evidence in an article needs scrutiny of its design, population, measurement and limitations, regardless of whether it appears in a respected venue.

Distinguish credibility from relevance. An official national survey may be carefully conducted yet too broad to answer a question about newly hired retail workers. A small local report may describe the right population but have missing data and no comparison. In a good synthesis, each source contributes only what its method and scope support. The writer can use both while being clear about the different weight given to each.

Check how key terms are defined. One study may call a new hire retained after six months; another may count employees still present after a year. One may include voluntary departures only; another counts all departures. Combining those results as though they measure the same outcome creates false precision. These differences belong in the analysis, not in a footnote added after a conclusion is written.

Trace claims back to their origin

A striking statistic often travels through news reports, summaries and presentations. Before relying on it, locate the underlying report or dataset. Check its date, place, population and calculation. A secondary summary may omit a qualification or repeat an error; the original may show that a percentage applied only to a small subgroup.

Tracing does not mean rejecting all secondary sources. A well-constructed review can help map an unfamiliar field and identify competing interpretations. But a researcher should know whether a claim comes from new observations, an analysis of existing data, a review or an opinion. If the original cannot be found, avoid presenting the number as settled fact.

Consider who is absent from the available evidence. Staff who left quickly may not be in a workplace survey of current employees. A programme described by managers may look different to its participants. Gaps like these do not make the evidence useless; they define the limits of a defensible conclusion and point to the next observation worth collecting.

Use information ethically and make reasoning visible

Record bibliographic details and notes when reading, not from memory at the end. Separate an exact quotation from a paraphrase and from your own inference. When paraphrasing, reconstruct the point in your own reasoning and acknowledge the source; merely replacing a few words is not independent analysis. Respect permissions and privacy when working with interviews, workplace records or unpublished material.

The final argument should show how evidence led to the answer. For the onboarding question, a careful conclusion might say that a structured programme is associated with stronger first-year retention in some settings, while noting that differences in staffing, pay or selection may also matter. Then state which evidence would distinguish these explanations. That is a more useful result than a confident general claim assembled from loosely related sources.

Information literacy continues after publication. New evidence, corrected data and changed contexts can alter an earlier judgment. Keep a record of the search and the reasons for including sources so that a later reader can understand the boundaries of the conclusion and revisit it responsibly.