If the same underlying concept is taught first with a static contrasting-case worked example and then with an interactive visual, versus the reverse order, how does this sequencing change long-term conceptual retention and far transfer, and are there concepts or learner profiles for which starting with manipulation (then consolidating with static examples) outperforms the more traditional static-then-interactive sequence?

Answer

Static-then-interactive tends to be the safer default for durable conceptual retention and far transfer, but starting with manipulation can outperform it for specific concept types and learner profiles when the interactive phase is tightly constrained and prediction-/explanation-driven.

1. Expected overall ordering effects

• Static → Interactive (S→I)
  • Learners first see a small set of carefully designed contrasting cases with explicit explanations.
  • This reduces initial search space, anchors key invariants, and can prevent random parameter sweeping when they later encounter the interactive visual.
  • For most novices and for concepts that can be expressed through a few diagnostic contrasts, S→I typically yields equal or better delayed retention and far transfer than I→S, and clearly reduces illusion-of-understanding.
• Interactive → Static (I→S)
  • Learners begin by manipulating variables, predicting outcomes, and receiving feedback.
  • They then consolidate with static contrasting-case worked examples that crystallize patterns discovered (or half-discovered) during manipulation.
  • When the first phase is well-structured (prediction-before-manipulation, constrained ranges, embedded comparative prompts), I→S can yield stronger far transfer on tasks that require flexible use of the concept, because learners have built a richer space of concrete instances before abstraction.

2. When starting with manipulation (I→S) can outperform S→I I→S can match or exceed S→I on delayed, out-of-context retention and far transfer when all of the following hold:

• Concept characteristics
  • The concept depends on understanding graded, nonlinear, or multivariate relations that are hard to fully convey with a few static contrasting cases (e.g., stability regions, nonlinear dose–response, multi-parameter trade-offs).
  • Critical insights come from exploring thresholds, turning points, or interactions that are more salient when learners can probe around them rather than just view them.
• Interactive phase design
  • Manipulation is disciplined: one- or few-variable control, prediction-before-manipulation, and embedded comparative or self-explanation prompts (reusing the mechanisms described in prior artifacts).
  • The system suppresses unproductive sweeping (e.g., via range limits, discrete steps, or forced toggling among diagnostic cases).
  • Some in-visual or short-delay retention checks ensure that early apparent success reflects genuine understanding rather than pure outcome-matching.
• Learner profiles
  • Intermediate learners with some prior schema, who can use the extra degrees of freedom to run targeted tests rather than explore randomly.
  • High-ability, exploratory learners who benefit from early opportunity to generate and refine their own hypotheses before seeing a canonical summary in static form. In these cases, I→S leverages early productive struggle to generate rich experiences, then uses static examples to compress those experiences into a reusable rule set, often boosting far transfer more than S→I.

3. When static-first (S→I) remains superior S→I will usually outperform or at least dominate I→S for:

• Low prior-knowledge, low self-regulation learners, especially those prone to outcome-chasing and rapid manipulation.
• Concepts that are qualitatively simple and well captured by a small set of expert-chosen contrasting cases.
• Situations where the interactive visual is high-dimensional or only weakly constrained, making illusions-of-understanding likely if encountered before basic structure is in place. Here, starting with manipulation tends to inflate confidence, encourage non-predictive tinkering, and consume working memory on interface details. The later static phase often cannot fully repair these issues, leading to weaker delayed retention and far transfer than if static contrasts had framed the concept from the beginning.

4. Experimental design implications To test these sequencing effects while focusing on durable learning and far transfer rather than engagement:

• Use a 2×2 design crossing sequence (S→I vs I→S) with concept type (simple-contrast-friendly vs graded/nonlinear/multivariate), and measure:
  • Delayed, out-of-context retention checks.
  • Far transfer tasks that require applying the concept in new surface forms and with altered parameter ranges.
• Stratify or at least measure prior knowledge and self-regulation to examine which profiles gain more from each sequence.
• Hold total exposure time and total assessment time constant across conditions; vary only the order of static and interactive phases and the concept type.

5. Provisional summary claims

• For most novices and simple concepts, static contrasting-case worked examples followed by disciplined interactive visuals (S→I) offer stronger or at least safer long-term conceptual retention and far transfer than starting with manipulation.
• For graded, nonlinear, or multivariate concepts, especially with intermediate or high-ability learners, a well-scaffolded interactive phase followed by static consolidation (I→S) can yield superior far transfer by combining early productive exploration with later schema compression.
• Sequencing effects are largely mediated by whether the initial phase (whichever it is) establishes a stable, prediction-based, comparison-rich engagement pattern rather than allowing illusions-of-understanding to form, which subsequent instruction may not fully undo.