3D-LAP: Three-Measure Protocol in STEM
- Three-Measure Protocol (3D-LAP) is an analytic tool that evaluates whether assessment tasks can potentially elicit integrated three-dimensional learning across scientific practices, crosscutting concepts, and core ideas.
- The protocol systematically analyzes exam items to determine if tasks encourage active reasoning and integration, distinguishing traditional assessments from research-based approaches.
- Developed by interdisciplinary experts, 3D-LAP offers a validated framework that guides curricular transformation and improves the design of college science assessments.
The expression “Three-Measure Protocol” most plausibly refers to the Three-Dimensional Learning Assessment Protocol (3D-LAP) introduced by Laverty and colleagues. In that paper, there is no protocol formally named “Three-Measure Protocol”; the instrument is consistently called 3D-LAP. It was designed to characterize and support the development of assessment tasks in biology, chemistry, and physics that align with transformation efforts in college science education, and it evaluates whether a task has the potential to elicit evidence of three-dimensional learning, not what students actually do in response to it (Laverty et al., 2016).
1. Terminology and conceptual setting
3D-LAP emerged from a specific problem in higher-education STEM reform. Many courses had adopted more ambitious pedagogies, including active-learning approaches, while assessments often remained centered on factual recall, algorithms, and isolated procedures. The protocol addresses that misalignment by treating assessment as a central component of curricular transformation rather than a downstream administrative instrument. In this setting, assessment is used both to align with the learning environment and to assess the extent of the transformations (Laverty et al., 2016).
The term three-dimensional learning is adopted from the National Research Council’s Framework for K-12 Science Education. In this formulation, science learning integrates scientific and engineering practices, crosscutting concepts, and disciplinary core ideas. The paper emphasizes that these dimensions are not supposed to be taught separately. Instead, the educational vision is that students should use scientific knowledge through practices and crosscutting forms of reasoning in ways that are integrated in both instruction and assessment (Laverty et al., 2016).
A recurrent misunderstanding is to treat 3D-LAP as a measure of student achievement or as a rubric for grading student responses. The instrument does not operate in that way. Its object of analysis is the assessment task itself: whether the task is structured so that students could demonstrate three-dimensional learning.
2. The three dimensions of learning
The first dimension, scientific and engineering practices, concerns what students should be able to do with knowledge. The examples explicitly listed in the paper’s framing include Asking questions, Developing and using models, Planning and carrying out investigations, Analyzing and interpreting data, Using mathematics and computational thinking, Constructing explanations, Engaging in argument from evidence, and Obtaining, evaluating, and communicating information. These are action-oriented and therefore require assessment tasks that solicit performance, reasoning, or representation rather than mere recall (Laverty et al., 2016).
The second dimension, crosscutting concepts, consists of ideas that apply across disciplines. The examples cited are Patterns, Cause and effect, Scale, proportion, and quantity, Systems and system models, Energy and matter, Structure and function, and Stability and change. In 3D-LAP, these concepts function as lenses through which phenomena or explanations are organized, rather than as isolated topical labels.
The third dimension, disciplinary core ideas, refers to the central concepts of a discipline that organize knowledge and explain a wide range of phenomena. For college introductory courses, the authors do not simply import the K–12 grade-band articulation. Instead, they construct discipline-specific core-idea lists for biology, chemistry, and physics. To qualify as a core idea in 3D-LAP, a concept must have disciplinary significance, the power to explain a wide range of phenomena, and potential for generating new ideas (Laverty et al., 2016).
This suggests that 3D-LAP is best understood as an operationalization of a curricular vision in which meaningful assessment must coordinate epistemic activity, transdisciplinary conceptual structure, and disciplinary explanatory content.
3. Coding logic of the protocol
3D-LAP is a coding protocol for assessment tasks. A task is analyzed to determine whether it can elicit evidence of each of the three dimensions. A task counts as three-dimensional only if it meets the appropriate criteria for a scientific practice, a crosscutting concept, and a core idea (Laverty et al., 2016).
| Dimension | How 3D-LAP codes it | Key feature |
|---|---|---|
| Scientific practices | Criteria-based coding | Depends on constructed response or selected response |
| Crosscutting concepts | Brief descriptive statements | Focuses on what the task must involve |
| Core ideas | Alignment to main idea or sub-bullets | Uses discipline-specific lists |
For scientific practices, the criteria depend on task format because practices are action-based. The paper distinguishes constructed response from selected response. In the practice Developing and Using Models, a constructed-response item must: present an event, observation, or phenomenon for explanation or prediction; give or ask for a representation; ask for an explanation or prediction; and ask for the reasoning linking the representation to the explanation or prediction. A selected-response version parallels these requirements, except that the student selects rather than constructs the representation, explanation, and reasoning. The authors explicitly note that selected-response items can be coded as potentially eliciting a practice, but they do not provide the strongest evidence of three-dimensional learning (Laverty et al., 2016).
For crosscutting concepts, the protocol does not use a checklist of formal criteria. Instead it uses brief descriptive statements describing what a task must involve. For Structure and Function, for example, a task should ask the student to predict or explain a function or property based on structure, or identify a structure that leads to a function or property.
For core ideas, coding is based on whether the task aligns with the main idea or one of its sub-bullets. Because introductory college courses do not map neatly onto the Framework’s K–12 organizational structure, the lists of core ideas were newly developed for each discipline.
The paper’s biology example illustrates the integrated coding target:
Create a diagram that shows the molecular structure of the lipid bilayer in a typical cell membrane. Use the diagram to explain why oxygen (O₂) can easily pass through the membrane but sodium ions (Na⁺) cannot.
This task was analyzed as Developing and Using Models, Structure and Function, and a biology core idea about structure and function. The example also shows that some overlap between crosscutting concepts and core ideas is unavoidable.
4. Development and disciplinary adaptation
The development process was explicitly interdisciplinary. The team included disciplinary experts and discipline-based education researchers in biology, chemistry, and physics. Their process involved reviewing the Framework, NGSS, and supporting literature, negotiating shared interpretations of scientific practices and crosscutting concepts, applying draft criteria to many assessment tasks, revising criteria when they did not fit actual exam items, and involving broader faculty groups in defining the disciplinary core ideas (Laverty et al., 2016).
Several notable adaptations were made in order to operationalize the Framework for college assessment. Constructing Explanations and Engaging in Argument from Evidence were merged into one practice because the distinction was difficult to operationalize reliably in assessment tasks. Obtaining, Evaluating, and Communicating Information was reduced to Evaluating Information. Planning and Carrying Out Investigations was reduced to Planning Investigations. Asking Questions was removed for selected-response tasks because the authors could not construct a meaningful example. For crosscutting concepts, Scale, Proportion, and Quantity was split into two concepts: Scale and Proportion and Quantity (Laverty et al., 2016).
The project also experimented with implicit versus explicit coding, but that distinction was abandoned. The authors concluded that, for their purposes, tasks that only weakly or indirectly relate to a dimension are not useful for guiding three-dimensional assessment design.
A plausible implication is that the protocol’s final form reflects not only a theoretical framework but also repeated encounters with the practical constraints of real exam items in large-enrollment introductory courses.
5. Validation, reliability, and comparative use
The paper does not claim universal validity across all settings. Rather, it argues for validity and reliability for their dataset and purpose. The validity argument includes face validity, because experts in biology, chemistry, and physics negotiated and refined the criteria; content validity, because the criteria were grounded in the Framework and in actual assessment tasks; and expert consensus from disciplinary faculty on the core ideas (Laverty et al., 2016).
For reliability, the authors evaluated inter-rater reliability by having researchers code exam tasks and then comparing agreement. Their agreement rule is methodologically important: two coders were considered in agreement if both said that a task did elicit a scientific practice, crosscutting concept, or core idea, even if they identified different specific practices or concepts, so long as both judged that some dimension was present. A disagreement was counted only when one coder said a dimension was present and another said it was absent. They report percent agreement and set a lower threshold of 75%. For biology and chemistry, where there were multiple coders, pairwise agreements were averaged (Laverty et al., 2016).
The protocol was also shown to distinguish between different styles of assessment. The paper contrasts “A” exams, described as typical textbook-style assessments with few three-dimensional opportunities, and “B” exams, drawn from courses using research-based curricula and containing many more tasks with one, two, or three dimensions. The distinction does not rest on correctness or difficulty. It rests on whether the task structure invites a student to use knowledge in a scientific way.
This comparative function makes 3D-LAP useful not only for coding single items but also for characterizing entire exams or homework sets, comparing assessments across time or courses, and supporting faculty development in writing better tasks.
6. Scope, limitations, and significance
The protocol is careful about its own limits. It does not measure what students actually think or write in response. It does not judge the quality of student evidence. It does not infer instructor intent. It does not account for all item features that affect student response, such as wording bias or prompt structure. It was developed for introductory college biology, chemistry, and physics lecture courses, so broader use requires caution (Laverty et al., 2016).
The paper also identifies assessment-type patterns. Constructed-response tasks are generally stronger for eliciting 3D learning. Selected-response items can sometimes be coded as dimension-rich, but are usually weaker evidence. Tasks that simply require isolated skills such as drawing a Lewis structure, making a Punnett square, or drawing a free-body diagram may be zero- or one-dimensional. The authors do not prescribe an ideal numeric mix of item types, but they argue that three-dimensional assessments should contain a meaningful array of multi-dimensional tasks (Laverty et al., 2016).
A common misconception is that three-dimensional assessment requires every task to be open-ended or elaborate. The protocol does not make that claim. Instead, it distinguishes between tasks that integrate knowledge with scientific practice and tasks that remain canonical, fragmented, procedural. The significance of 3D-LAP lies in converting that distinction into a usable analytic instrument.
In that sense, the protocol’s major contribution is methodological. It operationalizes the abstract idea of three-dimensional learning into a practical tool for analyzing college science assessments. The name “Three-Measure Protocol” is therefore best treated as a mislabel for 3D-LAP rather than as the title of a separate instrument.