Skip to content
Back to search results

Critical Incident Technique in Information Behavior

Introduced by
John C. Flanagan (technique); applied to information behavior in LIS
Year
1954
Type
Incident-based qualitative technique for studying effective and ineffective behavior
Family
Information behavior / qualitative method
Font size

What does it do?

The Critical Incident Technique (CIT), set out by John Flanagan in his 1954 Psychological Bulletin article, is a flexible set of procedures for collecting first-hand accounts of specific, observable incidents that were especially helpful or unhelpful in achieving some aim, and content-analyzing them to understand effective and ineffective behaviour.

In library and information science it has become a widely used method for studying information seeking and use: instead of asking people in the abstract how they look for information, researchers ask them to recount concrete recent episodes — a time they urgently needed information, succeeded or failed to find it, encountered a barrier, or used a source to good or ill effect. Analyzing many such incidents yields grounded categories of information needs, behaviours, barriers and source roles. A 2025 Libri review documents the technique's broad application across LIS information-behaviour research.

The intuition behind the method

If you ask someone 'how do you usually find information?' you get a vague, idealized answer that rarely matches what they actually do. CIT sidesteps this by asking for stories instead: 'Tell me about a specific time recently when you needed information and it really mattered — what happened, what did you do, how did it turn out?' Concrete, memorable episodes are easier to recall accurately and far richer than generalizations. By gathering many such critical incidents — some where things went well, some where they went badly — and then sorting them into patterns, a researcher builds an evidence-based picture of how information seeking really works for a group: which needs recur, where people get stuck, which sources save the day and which let them down. The technique's power is that it grounds every claim in something that actually happened.

History

John Flanagan developed the Critical Incident Technique out of studies of aircrew performance in the U.S. Army Air Forces Aviation Psychology Program during the Second World War, distilling the methods into his foundational 1954 Psychological Bulletin article. Originally aimed at identifying behaviours critical to job performance, CIT spread across psychology, management, nursing, education and service research. In library and information science it was taken up to study reference encounters, information needs, source use and service quality, becoming a staple qualitative method in information-behaviour research; a 2025 Libri literature review by Sadeghi, Nowkarizi and Tajafari surveys this body of LIS applications. Its enduring appeal lies in grounding the study of information seeking in concrete episodes of what actually helped or hindered people.

When to use it

Use CIT in information-behaviour research when you want grounded, concrete evidence of how people actually seek and use information, including where it succeeds and where it breaks down, rather than abstract self-reports. It is well suited to exploratory and needs-assessment studies, to evaluating services and systems by what really helped or hindered users, and to surfacing barriers and effective practices in a defined context. It is less appropriate for estimating prevalence in a population, for studying continuous routine behaviour that does not partition neatly into discrete incidents, or where participants cannot recall specific recent episodes. Because it relies on recalled first-hand accounts, applying it well means eliciting specific, recent incidents and analyzing enough of them to reach category saturation.

Limitations and considerations

  • Depends on participants' memory of specific incidents, so recall and selection bias can distort the corpus.
  • Salient, dramatic incidents are over-reported while routine or minor information behaviour is under-represented.
  • Inductive categorization is interpretive and analyst-dependent, requiring reliability checks to be credible.
  • Yields qualitative patterns rather than prevalence estimates, limiting quantitative generalization.

Source conditions do not replace independent scientific review or validation by an analysis engine.

Strengths

  • Grounds findings in concrete, recalled episodes rather than vague generalizations, improving accuracy and richness.
  • Deliberately contrasts effective and ineffective incidents, surfacing barriers and breakdowns as well as successes.
  • Inductive, data-driven categorization lets unexpected information needs and behaviours emerge from the evidence.
  • Flexible and economical, adaptable to interviews, questionnaires or diaries across many information-behaviour contexts.

Applications

  • Assessing information needs by collecting episodes where information was urgently sought or used.
  • Evaluating library and information services through incidents that markedly helped or frustrated users.
  • Identifying barriers and breakdowns in information seeking via ineffective critical incidents.
  • Studying the roles of human, documentary and digital sources in concrete seeking episodes.
  • Generating grounded requirements for systems, training and service redesign from recurring incident patterns.

How it works

How it works

1

CIT begins by specifying the general aim of the activity under study and defining what counts as a critical incident relative to it. In Flanagan's terms an incident is an observable, sufficiently complete human activity, and it is 'critical' when it makes a significant positive or negative difference to the aim. In an information-behaviour study the aim might be resolving an information need or completing a task, and a critical incident is an episode of seeking or using information that markedly helped or hindered that aim. Pinning down the aim and the incident criteria up front is what keeps the collected episodes focused and comparable.

Formula
\text{incident } I:\ \text{observable, complete act};\ I\ \text{critical}\iff |\text{effect on aim } A| \ \text{high}
2

The researcher then collects accounts of concrete incidents, typically through interviews, questionnaires or diaries, asking participants to describe specific, recent, first-hand episodes — the context in which it occurred, what they actually did, and how it turned out. Flanagan stressed recency and specificity to maximize accuracy of recall and to avoid vague generalizations. The result is a corpus of grounded episodes (context-action-outcome) rather than abstract opinions, which is the raw material the method depends on.

Formula
D=\{I_1,\dots,I_n\},\ I_k=(\text{context},\text{action},\text{outcome}),\ \text{first-hand, recent, specific}
3

The collected incidents are content-analyzed inductively to derive a classification of the behaviours, needs or factors they reveal. Categories emerge from the data rather than being imposed in advance, and the analyst iterates the frame until incidents can be reliably assigned, often checking inter-rater reliability. This bottom-up categorization is the analytic core of CIT: it turns a pile of stories into a structured account of the kinds of information behaviour that recur in the studied context.

Formula
C=\bigcup_k \text{categorize}(I_k),\quad \text{inductive frame of behavior types},\ \text{IRR}\to\kappa
4

A defining feature of CIT is that it deliberately gathers and contrasts both effective and ineffective incidents — episodes that helped achieve the aim and episodes that hindered it. Separating the categories into what worked and what failed sharpens the findings: it reveals not only how people succeed in finding and using information but the specific barriers, breakdowns and dysfunctional behaviours that derail them. This positive/negative contrast is more diagnostic than studying success alone.

Formula
C=C^{+}\cup C^{-},\quad C^{+}\!:\text{helped aim},\ C^{-}\!:\text{hindered aim}
5

Finally, the categorized effective and ineffective incidents are interpreted to draw conclusions about information behaviour and to inform practice. From the patterns the researcher infers recurring information needs, the roles different sources play, the barriers that block seeking, and requirements for systems, services or training. In LIS this interpretive step translates grounded incidents into actionable insight — for example, redesigning a service to remove a barrier that recurred across many negative incidents, or reinforcing a source or behaviour that drove many positive ones.

Formula
\text{requirements}=\phi(C^{+},C^{-}),\quad \text{infer needs, barriers, source roles in seeking}

Common pitfalls

  • Accepting vague or hypothetical accounts instead of insisting on specific, recent, first-hand incidents.
  • Collecting only positive (or only negative) incidents and losing the diagnostic effective/ineffective contrast.
  • Imposing a predefined category scheme rather than letting categories emerge inductively from the incidents.
  • Treating incident counts as population prevalence, when salience biases which episodes get reported.

Frequently asked questions

What makes an incident 'critical' in CIT?

An incident is a concrete, observable, reasonably complete episode of behaviour, and it is critical when it makes a significant positive or negative difference to the aim under study. In information-behaviour research the aim is usually resolving an information need or completing a task, so a critical incident is an episode of seeking or using information that clearly helped or hindered that aim — a search that succeeded against the odds, a barrier that blocked someone, a source that proved decisive. The significance to the aim, not drama for its own sake, is what makes the incident worth collecting.

Why collect both effective and ineffective incidents?

Because contrasting success and failure is far more diagnostic than studying either alone. Effective incidents show what behaviours, sources and conditions help people find and use information well; ineffective incidents expose the barriers, breakdowns and dysfunctional behaviours that derail them. Together they let researchers and designers reinforce what works and remove what does not. This deliberate positive/negative sampling is a defining feature of CIT and a key reason it is valued for service evaluation and needs assessment in information-behaviour research.

How is CIT applied to information behaviour rather than job performance?

The procedures are the same but the aim is reframed around information needs and use. Instead of asking which behaviours are critical to a job, the researcher asks participants to recount specific recent episodes in which they needed, sought or used information and it markedly helped or hindered them. These incidents are then content-analyzed into categories of information needs, behaviours, barriers and source roles, and the effective/ineffective contrast is interpreted to inform services, systems and training. The 2025 Libri review documents how extensively LIS researchers have adapted Flanagan's technique in exactly this way.

Related methods

Sources

  1. Flanagan, J. C. (1954). The critical incident technique. Psychological Bulletin, 51(4), 327-358.

    Open DOI reference
    Copied.
  2. Sadeghi, H., Nowkarizi, M., & Tajafari, M. (2025). Applications of the Critical Incident Technique in Library and Information Science Research: A Literature Review. Libri, 75(2), 145-157.

    Open DOI reference
    Copied.