21.6 User Adjustment Behavior

User adjustment behavior is the set of actions through which a human user modifies their approach to a machine system in response to feedback received during interaction. It encompasses the full range of adaptive responses a user makes when interface outputs signal that the current approach is not producing the desired results — from minor parameter tweaks and rephrased queries to wholesale strategy revision and mode shifts. User adjustment behavior is the corrective action component of the interface control loop: it is how the human side of the feedback cycle responds to error signals, and its quality determines whether the interaction converges on the user's goal or stalls in repeated unsuccessful attempts.

The Adjustment Cycle

User adjustment behavior unfolds in a recognizable cycle tied to the interface control loop. The user acts on the system with an expectation about the result. The interface feedback signals indicate whether the expectation was met. If it was, no adjustment is needed and the user continues. If it was not, the user must generate an adjusted action — one that departs from the previous approach in some way that is expected to produce a better result. The adjusted action generates new feedback, which either confirms that the adjustment was effective or indicates that further adjustment is needed.

The quality of this adjustment cycle depends critically on the quality of the feedback that signals the need for adjustment and guides its direction. Feedback that clearly identifies what went wrong and why enables targeted, accurate adjustment. Feedback that only signals failure without specifying its nature leaves the user guessing about what to change. Feedback that actively misleads the user about the cause of failure produces adjustments that move in the wrong direction. The design of interface feedback signals is therefore inseparable from the design of conditions for effective user adjustment behavior.

Types of User Adjustment Behavior

User adjustment behaviors range in scope and in the depth of change they represent:

Surface-level adjustments modify the form or parameters of an action without changing the underlying strategy — rephrasing a search query using different words, adjusting a slider by a smaller increment, resizing a window. Surface adjustments are the first response to failure and require the least cognitive effort; they assume that the strategy is correct and that only the specific formulation needs modification.

Procedural adjustments modify the sequence or selection of actions without changing the overall approach — trying a different path through a menu hierarchy, using a keyboard shortcut instead of a menu item, performing steps in a different order. Procedural adjustments are appropriate when the goal is correct but the specific path to it needs revision.

Strategic adjustments represent a more fundamental revision of the approach to the task — recognizing that the current strategy will not achieve the goal and selecting a different strategy. Strategic adjustments require the user to step back from the immediate action level and reconsider their overall approach: using a different tool, decomposing the task differently, changing the representation of the problem.

Model adjustments are revisions to the user's mental model of how the system works — updates to the user's understanding of the system's capabilities, constraints, or behavior that correct a misunderstanding that has been producing repeated failure. Model adjustments are the most cognitively demanding form of adjustment and are often resisted because they require the user to revise established beliefs. They are, however, the form of adjustment that produces the most durable improvement, because they address the root cause of repeated failures rather than their surface manifestations.

Factors That Shape Adjustment Behavior

Several factors determine the quality and speed of user adjustment behavior:

Feedback quality is the most directly controllable factor from the interface design perspective. Users who receive specific, timely, informative feedback about the nature of their errors can make targeted adjustments; users receiving only generic failure signals must experiment broadly to discover what is wrong. The specificity of feedback has the largest single effect on adjustment efficiency.

Mental model accuracy determines how well users can interpret feedback and generate effective adjustments. Users with accurate mental models of system behavior can draw on that understanding to diagnose failures and identify appropriate adjustments. Users with inaccurate mental models may interpret failures incorrectly, generating adjustments that address their incorrect model of the problem rather than the actual problem.

Error tolerance and persistence affect how long users continue attempting adjustments before abandoning a task or seeking help. Users with high error tolerance and persistence will continue exploring adjustment strategies through extended failure sequences; users with low tolerance will abandon tasks or escalate to help-seeking after few failed attempts. Both extremes have costs: excessive persistence in an unproductive adjustment strategy wastes time, while premature abandonment prevents goal achievement.

Expertise and experience with the system and with similar systems determine the library of adjustment strategies available to the user. Expert users have encountered a wide range of failure modes and have learned effective responses to them; novice users have few adjustment strategies and must discover new ones through trial and error.

Adjustment Behavior and System Design

User adjustment behavior is substantially shaped by the design of the machine system. Systems that are designed with legible state, clear error signals, and transparent cause-and-effect relationships enable effective adjustment; systems with opaque state, ambiguous errors, and unpredictable responses impede it. The design principle of discoverability — making system capabilities and constraints legible through the interface — directly supports adjustment behavior by giving users the information they need to diagnose failures and identify effective modifications.

Systems that record and analyze user adjustment behavior provide designers with valuable insight into where users consistently encounter difficulty, where failures cluster, and where the adjustment strategies users attempt diverge from those that the system was designed to support. This behavioral data, when fed back into the design process, enables iterative improvement of the interface based on observed adjustment patterns rather than assumptions about what users will do.

Communication and Adjustment in Natural Language Systems

In natural language interaction systems, user adjustment behavior is particularly prominent and revealing. When users' queries or commands fail to produce useful results, they exhibit characteristic adjustment patterns: adding specificity, removing specificity, reformulating in a different vocabulary, providing context, breaking complex requests into simpler components, or explicitly correcting the system's interpretation. These adjustment patterns reflect users' theories of how the natural language system processes inputs and what modifications are likely to improve results. Studying these patterns reveals both how users model the system and where their models diverge from the system's actual capabilities and limitations.