In fields like healthcare and management, phrases such as “Improvement Science” or “Science of Improvement” have become increasingly common. However, their widespread use has led to significant confusion, with different people using the same terms to refer to vastly different ideas, concepts, and guiding principles. This ambiguity hinders progress and makes it difficult for professionals to build a shared understanding of how to drive meaningful and sustainable change.
The article “Seven Propositions of the Science of Improvement: Exploring Foundations” by Rocco J. Perla, Lloyd P. Provost, and Gareth J. Parry seeks to resolve this confusion. The authors argue that to advance the field, professionals need a standard set of core principles and a common lexicon grounded in the historical and philosophical origins of improvement theory. They trace these origins to the work of W. Edwards Deming and his System of Profound Knowledge, presenting it as the bedrock upon which a true science of improvement is built.
The authors propose that this system provides a scientific foundation for seven core propositions that, together, define the essence of improvement science. This article is for anyone—leaders, researchers, or frontline staff—who wishes to move beyond surface-level tools and understand the “raw materials” that make improvement a true scientific discipline.
The Foundation: Deming’s System of Profound Knowledge
First used in the 1996 book The Improvement Guide, the term “science of improvement” was built upon Deming’s framework. This system is composed of four interrelated parts that must be understood in concert:
- Appreciation for a System: This involves seeing an organization not as a collection of separate parts, but as a network of interdependent processes and people working together toward a common aim. Management’s role is to understand and optimize these interdependencies.
- Understanding Variation: This is the ability to distinguish between common cause variation (inherent, predictable noise within a process) and special cause variation (resulting from specific, identifiable events). Failing to understand this distinction leads to flawed decision-making, such as tampering with a stable process or failing to investigate a real problem.
- Theory of Knowledge (Epistemology): This component examines how we know what we know and how our beliefs about knowledge impact learning and decision-making. It emphasizes that all improvement comes from developing, testing, and implementing changes in an iterative cycle.
- Psychology: This involves understanding human behavior and the interpersonal and social dynamics that impact system performance.
Building on this foundation, the authors articulate seven propositions that provide a robust framework for the science of improvement.
The Seven Propositions of the Science of Improvement
Proposition 1: Improvement is Grounded in Testing and Learning Cycles
At its core, science is a process of testing claims and predicting outcomes. A key principle is falsifiability, introduced by philosopher Karl Popper, which states that a claim is only scientific if it can be tested and potentially proven false. This iterative cycle of hypothesis, experiment, and observation has been the norm of scientific proof for centuries, dating back to the physicist Ibn al-Haytham in the 11th century.
The Plan-Do-Study-Act (PDSA) cycle is the modern embodiment of this scientific method. It is not a “lesser” way of knowing but a direct application of scientific principles. Each cycle requires a prediction (hypothesis), data collection (test), and analysis (study) to generate learning that informs the next action. This reinforces the idea that applied science is about continuous testing and learning, not waiting for a single “perfectly designed experiment”.
Proposition 2: Its Philosophical Foundation is Conceptualistic Pragmatism
The science of improvement is deeply influenced by the philosophy of conceptualistic pragmatism, primarily developed by C.I. Lewis. Pragmatism shifts the focus from what is universally “true” to what is “useful” for making predictions and taking action. Lewis argued that all knowledge is filtered through our past experiences and mental models (“conceptualistic”) and that its primary function is to help us transition from the present to a desired future (“pragmatic”).
This philosophy directly supports Shewhart’s control charts, which use data from past experience (the “actual present”) to predict the likely range of future performance. It also justifies an approach where change is not about implementing a rigid “magic bullet” but about local reinvention, where teams use general theories and tailor them to their specific context.
Proposition 3: It Embraces a Combination of Psychology and Logic (“Weak Psychologism”)
Historically, philosophy and logic (which prescribe how people should think) were kept separate from psychology (which describes how people do think), a position known as antipsychologism. The science of improvement, however, rejects this rigid separation. It embraces a “weak form of psychologism,” which acknowledges that both logical frameworks and an understanding of human psychology are essential for effective decision-making.
We need objective standards and evidence-based practices, but we must also recognize that people do not always act logically and that innovation often requires creativity beyond formal rules. Deming’s System of Profound Knowledge inherently embraces this view, integrating disciplines to create a more complete picture of how to drive change.
Proposition 4: It Considers Both the Contexts of Justification and Discovery
Building on the previous point, science involves two distinct but complementary activities: the context of discovery and the context of justification. Discovery is the creative, intuitive, and often messy process of generating new ideas and hypotheses. Justification is the rigorous, logical process of testing those ideas with data to determine if they work.
A common mistake is to view only the justification phase as “real science”. The science of improvement provides a lens to bridge both contexts. It honors the knowledge and hunches of subject matter experts (discovery) while insisting that those ideas are systematically tested using methods like PDSA cycles (justification). This iterative loop between creativity and validation is what maximizes learning.
Proposition 5: It Requires the Use of Operational Definitions
Words like “safer,” “better,” or “efficient” have no communicable meaning until they are defined in operational terms. An operational definition, a concept Deming borrowed from physics, puts meaning into a concept by specifying the test, sampling method, and criteria for measuring it.
Without operational definitions, team members may think they are working toward the same goal when, in reality, they have different interpretations. Establishing a shared understanding of terms, measures, and goals is a bedrock principle for effective communication and collective action in any improvement effort. It is the only way to ensure everyone is on the same page.
Proposition 6: It Employs Shewhart’s Theory of Cause Systems
Walter Shewhart’s control chart is far more than a statistical tool; it is a theory of variation and cause systems. It provides a scientific method for distinguishing between common cause variation (the natural, random fluctuation within a stable system) and special cause variation (unexpected outcomes from a specific, “assignable” cause).
Understanding this distinction is critical for effective management. When a process shows only common cause variation, it is stable and predictable. The appropriate action is to change the underlying system if a different level of performance is desired. Attributing a single outcome to a special cause in this situation is “tampering” and often makes performance worse. Conversely, when special cause variation appears, it is economical to investigate and address its root cause. The control chart provides a rational basis for knowing when to act and when to leave a process alone.
Proposition 7: Systems Theory Directly Informs the Work
Finally, the science of improvement requires systems thinking. Instead of dissecting problems into isolated parts, it focuses on how interdependent components—people, processes, equipment, policies—work together as a whole to accomplish an aim.
An organization’s outcomes are a property of its systems; as Deming would say, “systems are perfectly designed to achieve the results they get”. Therefore, to improve outcomes, one must improve the system. This perspective helps reveal that most problems are not caused by individual workers but by the structure of the systems they work within. It also highlights the complexity of change, where interventions can have delayed impacts and unintended consequences, making a holistic view essential for sustainable improvement.
Conclusion: A Call for Deeper Understanding
The seven propositions provide a cohesive, philosophically grounded framework that elevates improvement from a collection of tools to a scientific discipline. For practitioners, especially in high-stakes fields like healthcare, understanding these foundations is crucial. It allows them to move beyond simply applying methods to integrating complex ideas and adapting them to solve real-world problems. Ultimately, the authors suggest, this scientific knowledge is most powerful when combined with the ethical and moral imperative to improve the lives of others—linking the “character” of improvement science with its practice.
Perla, R. J., Provost, L. P., & Parry, G. J. (2013). Seven propositions of the science of improvement: Exploring foundations. Quality Management in Health Care, 22(3), 170–186. https://doi.org/10.1097/QMH.0b013e31829a6a15