A Taxonomy for Learning, Teaching, and Assessing: A Revision of Bloom's Taxonomy of Educational Objectives

The revision converts Bloom's single hierarchical list of six cognitive categories into a two-dimensional table. The horizontal dimension is the Cognitive Process Dimension—Remember, Understand, Apply, Analyze, Evaluate, Create—with 19 specific cognitive processes distributed across these six categories [citation:9]. The vertical dimension is the Knowledge Dimension: Factual, Conceptual, Procedural, and Metacognitive [citation:5]. An educational objective is analyzed as a verb (cognitive process) acting on a noun (knowledge type), producing 24 possible cells for classifying any objective [citation:5]. This matters because the original taxonomy conflated content and process; the revision separates them, allowing teachers to see precisely whether an objective asks students to remember factual knowledge or evaluate conceptual knowledge—a distinction invisible in the 1956 framework.

The revision makes three changes. First, categories shift from nouns to verbs: Knowledge becomes Remember, Comprehension becomes Understand, Application becomes Apply, and Analysis becomes Analyze [citation:19]. Second, Synthesis and Evaluation swap positions and rename: Evaluation becomes Evaluate, and Synthesis becomes Create, which now occupies the highest position [citation:5]. Third, the hierarchy is relaxed rather than strictly cumulative; Anderson and Krathwohl note that only three of Bloom's levels—Understand, Apply, and Analyze—have empirical support for sequential accumulation, while the strict requirement that lower levels must precede higher ones is softened [citation:2]. The revision thus retains a general developmental sequence while abandoning the rigid staircase model that critics argued constrained instructional design.

The taxonomy table is a grid with the six cognitive processes on one axis and the four knowledge types on the other. Teachers classify their stated objectives by locating the verb (process) and noun (knowledge type) in the appropriate cell. They then examine whether instructional activities and assessment tasks occupy the same cells as the objectives [citation:5]. If an objective sits in Apply Procedural but the assessment only tests Remember Factual, the misalignment becomes visible. Anderson reported that the table is generic: by replacing specific topics with knowledge types, it can be used across all subject matters [citation:5]. The purpose is not to prescribe one sequence but to expose discrepancies between what teachers intend, what they teach, and what they test.

Critics raise several issues. The hierarchical structure, though relaxed, still implies that some cognitive processes are more complex or valuable than others, which may discourage teachers from using higher-order tasks with students who have not yet demonstrated lower-order mastery [citation:15]. Category boundaries can be ambiguous: placing a specific task into one cell often involves subjective judgment [citation:15]. The taxonomy focuses heavily on the cognitive domain and does not adequately address affective or psychomotor learning [citation:15]. Empirical evidence that using the taxonomy improves educational outcomes is limited [citation:15]. Some argue the structured framework restricts teacher flexibility and creativity, and that it does not sufficiently address 21st-century skills like digital literacy and collaboration in complex real-world contexts [citation:15].

Evaluation to Improve Learning (1981) argued that classroom assessment should serve formative and diagnostic functions rather than sorting students. The 2001 revision supplies the conceptual apparatus for that argument. The two-dimensional table makes alignment concrete: teachers can diagnose whether a formative assessment actually targets the cognitive process and knowledge type specified in the objective. Where the 1981 book made the case for evaluation as a feedback mechanism, the 2001 taxonomy provides the classification system to implement that feedback with precision. Bloom died before the revision was completed; Anderson and Krathwohl explicitly framed the work as a continuation of Bloom's project of making educational objectives operational and assessable [citation:7][citation:12].

The revision was edited by Lorin W. Anderson and David R. Krathwohl, with contributions from Peter W. Airasian, Kathleen A. Cruikshank, Richard E. Mayer, Paul R. Pintrich, James D. Raths, and Merlin C. Wittrock [citation:7]. Anderson was a Carolina Distinguished Professor at the University of South Carolina and had authored Bloom's Taxonomy: A Forty-Year Retrospective (1994) [citation:6]. Krathwohl had co-authored the original Taxonomy's affective domain handbook with Bloom and Masia in 1964 [citation:1]. The revision was thus executed by scholars with direct intellectual lineage to Bloom's original project, which lends the revision a degree of continuity and fidelity that a completely independent reconstruction would lack. Bloom himself died in 1999, two years before publication.