03 · What You Need to Know
Designing Research Procedures for High Information Value and Low Unnecessary Burden
What Does Maximizing Research Information Actually Mean?
Maximizing research information means obtaining evidence that is sufficiently relevant, accurate, precise, and interpretable to answer the study's intended questions.
It does not mean collecting every variable that could potentially be useful.
For a randomized trial, useful information may concern a prespecified treatment effect at a meaningful endpoint. For a descriptive survey, it may involve sufficiently precise estimates for the intended population. For qualitative research, it may involve accounts that support a rich and credible understanding of the phenomenon.
These objectives require different designs. There is no universal information-efficiency measure that applies equally to all research approaches.
The first task is therefore to identify what the study must establish and what evidence is necessary to support that inference.
Why Is More Data Not Always More Informative?
Additional observations may provide little useful information when they duplicate existing measurements, are poorly aligned with the research question, or introduce substantial measurement error.
For example, administering several similar satisfaction questionnaires may increase the number of recorded responses without substantially improving understanding of satisfaction.
Likewise, measuring a relatively stable characteristic every week may add little information if the characteristic is not expected to change during the observation period.
However, apparently repetitive observations may be essential when the study investigates reliability, fluctuations, or temporal dynamics.
Researchers should evaluate the incremental contribution of each measurement rather than assume that repeated or overlapping data are necessarily redundant.
How Should Research Objectives Determine Measurement Priorities?
Every important measurement should have an identifiable purpose within the study.
Researchers can map each proposed instrument or procedure to a primary outcome, secondary objective, necessary covariate, safety requirement, implementation question, or justified exploratory analysis.
This mapping makes it easier to identify activities that provide limited information relative to their demands.
| Measurement priority |
Scientific contribution |
Burden-reduction consideration |
| Primary outcome |
Directly addresses the central research question |
Preserve valid measurement and remove unnecessary administration delays |
| Key secondary outcome |
Supports an important additional interpretation |
Assess timing and the availability of suitable shorter instruments |
| Safety assessment |
Supports participant protection |
Retain necessary monitoring while improving operational efficiency |
| Analytical covariate |
Supports justified estimation or interpretation |
Collect at the time and frequency required by the model |
| Implementation outcome |
Explains delivery, feasibility, acceptability, or adoption |
Select measures aligned with defined implementation objectives |
| Exploratory variable |
Supports additional hypothesis development |
Evaluate incremental value against burden and analytical complexity |
| Administrative information |
Enables appropriate study operations |
Avoid unnecessary collection and repeated requests |
Measures should not be retained merely because they are easy to add to an online form. Even small additions can become consequential when repeated across many assessments.
What Is the Difference Between Data Quantity and Information Quality?
Data quantity concerns the number of observations, variables, or measurements collected. Information quality concerns how well those data support the intended scientific inference.
A large dataset can contain substantial measurement error, incomplete population coverage, or variables poorly aligned with the research question.
Conversely, a smaller dataset collected through appropriate sampling and valid measurement may provide stronger evidence for a clearly defined objective.
Statistical precision is only one component of information quality. Researchers must also consider bias, construct validity, and the assumptions connecting observations to conclusions.
How Can Researchers Identify Measurements With Low Incremental Value?
One approach is to ask what would become unknown or less certain if a proposed measurement were removed.
If removing an instrument would prevent estimation of a primary outcome, its contribution is substantial. If it would eliminate only a weakly justified exploratory variable already approximated by another measure, its incremental contribution may be limited.
Researchers should review conceptual overlap, existing validation evidence, timing, and analytical requirements.
High correlations alone do not establish that two instruments are interchangeable. They may measure different aspects of a construct or serve different purposes.
The decision to remove nonessential measurements should therefore be grounded in their scientific function rather than their duration alone.
Can Better Measurement Instruments Reduce the Need for Additional Questions?
Appropriately selected instruments may provide the required information with fewer items or less administration time.
For example, a validated short form may measure a defined construct sufficiently well for a particular research purpose without requiring the full-length instrument.
However, shorter instruments may offer less precision, narrower content coverage, or different score interpretations.
Researchers should examine reliability, validity evidence, intended population, administration conditions, and licensing requirements before substituting instruments.
A shorter measure is efficient only when it retains the information necessary for the intended use.
How Does Assessment Timing Affect Information Efficiency?
The timing of measurements can be as important as their number.
For example, an intervention evaluation may require a baseline assessment, an immediate post-intervention outcome, and a later follow-up to examine maintenance. Additional intermediate observations may be unnecessary if the study does not investigate the trajectory of change.
By contrast, a study examining daily fluctuations in academic stress may require more frequent measurements because widely spaced observations would miss important variation.
Researchers should determine the number and timing of follow-up visits according to the temporal information required.
The most efficient schedule is not necessarily the one with the fewest assessments, but the one whose observations contribute meaningfully to the intended analysis.
Can Statistical Design Improve Information Without Increasing Burden?
Appropriate statistical design can improve the precision or relevance of estimates without necessarily requiring more measurements from each participant.
Depending on the research question, useful strategies may include justified covariate adjustment, stratified sampling, appropriate allocation, or efficient repeated-measures models.
For example, collecting a reliable baseline measure may improve the precision of an intervention comparison when baseline values are predictive of the outcome and the analytical model is appropriate.
However, statistical adjustment does not automatically correct measurement error or selection bias.
Design decisions should be based on explicit assumptions and evaluated through suitable sample size calculations or simulation where necessary.
What Is the Role of Optimal Experimental Design?
Optimal experimental design uses formal criteria to select design features that provide information efficiently under a specified statistical model.
Depending on the problem, these criteria may concern the precision of parameter estimates, uncertainty in a treatment contrast, or the ability to distinguish competing models.
For example, researchers studying a response over time may use model-based design methods to identify assessment times that are particularly informative about the parameters of interest.
However, a mathematically optimal schedule under one model may perform poorly if its assumptions are incorrect or if participants cannot complete the required procedures.
Researchers should therefore incorporate feasibility and robustness considerations rather than treat statistical optimality as sufficient justification for a design.
Can Adaptive Designs Reduce Unnecessary Data Collection?
Adaptive designs permit prespecified modifications based on accumulating information while maintaining appropriate control of statistical properties and research integrity.
In some settings, adaptive allocation, sample size reassessment, or early stopping may reduce the number of participants or observations required to answer a research question.
However, adaptive methods do not necessarily reduce burden for every participant and may introduce additional operational complexity.
They require careful prospective planning, appropriate statistical methods, and relevant oversight.
Researchers should not describe informal changes made after inspecting results as a valid adaptive design.
Can Existing Records Reduce Direct Participant Burden?
Existing administrative, educational, clinical, or other records may provide information that would otherwise require direct reporting.
For example, a study of course completion may use institutional records rather than repeatedly asking students whether they completed particular modules.
However, records may not capture the same constructs as self-report. Completion records indicate documented activity but do not necessarily explain motivation, learning experiences, or barriers.
Researchers should assess data accuracy, completeness, definitions, timing, privacy, and applicable access permissions.
Record-based measurement can reduce direct participant burden while introducing other forms of data quality and governance complexity.
Can Passive or Automated Data Collection Improve Efficiency?
Automated systems may collect certain observations without requiring repeated manual entry.
For example, a learning management system may record login activity or assignment submission times, reducing the need for participants to maintain separate activity diaries.
Yet digital traces do not necessarily measure engagement, understanding, or meaningful participation. A login event is not equivalent to learning.
Passive collection also raises privacy, consent, security, and data minimization considerations.
Researchers should ensure that the collected information is necessary, appropriately authorized, and suitable for the intended interpretation.
How Can Participant Burden Affect the Effective Information Obtained?
Burden can influence whether participants enroll, complete assessments, and remain in the study.
Adding measurements may increase the potential information collected from each participant while also increasing fatigue or noncompletion.
Under some circumstances, a shorter protocol may produce more usable observations and better population coverage, even though each participant completes fewer questions.
However, these effects cannot be assumed. The consequences depend on the measurements, participants, and missing-data mechanisms.
Researchers should examine whether participant burden contributes to attrition and threatens research validity when evaluating the overall efficiency of a protocol.
Can Information Efficiency Be Expressed Numerically?
Researchers may use quantitative criteria to compare alternative designs, but there is no universally valid measure of information per unit of participant burden.
One practical planning indicator is the amount of scheduled participant time required by a procedure. Another may be the expected precision of a primary estimate under alternative designs.
For a specified statistical model, Fisher information or other formal information measures may support comparisons. Such measures depend on the model and do not automatically capture ethical burden, accessibility, or external validity.
A simple ratio of collected variables to participant minutes would therefore be a poor general measure of scientific efficiency.
Researchers should select evaluation criteria appropriate to their actual research objectives.
How Can Researchers Compare Two Measurement Strategies?
Alternative strategies should be compared using their expected information contribution, measurement validity, participant demands, and operational feasibility.
| Evaluation criterion |
Question for the researcher |
| Scientific alignment |
Does the strategy provide the evidence required by the primary and important secondary objectives? |
| Validity |
Do the instruments and procedures support the intended interpretations? |
| Precision |
Is the expected uncertainty acceptable for the intended estimates? |
| Temporal information |
Are observations collected at the times needed to understand the phenomenon? |
| Participant burden |
What time, effort, expenses, fatigue, and practical disruption are required? |
| Accessibility |
Can relevant population groups participate through the proposed procedures? |
| Missing data |
How might the strategy affect completion and the assumptions required for analysis? |
| Feasibility |
Can the strategy be implemented reliably with available resources and safeguards? |
The preferred design should provide sufficient evidence without imposing unnecessary demands. A strategy that minimizes time but produces invalid measurements is not efficient in any scientifically meaningful sense.
How Can Pilot Testing Inform Protocol Optimization?
Pilot and feasibility work can identify which procedures consume more time than expected, create confusion, or contribute to missed assessments.
Researchers can examine actual completion times, usability, accessibility, and participant feedback.
Where appropriate, alternative questionnaire formats or assessment schedules can be compared before the main study.
However, small pilots may provide imprecise estimates of effect sizes or measurement properties. They should be designed around the specific uncertainties that need resolution.
Changes to instruments or schedules may require further evaluation and applicable protocol approvals.
Why Should Participants Help Evaluate Information-Burden Trade-Offs?
Researchers may know which measurements are scientifically important but underestimate their practical demands.
Prospective participants can identify inconvenient schedules, inaccessible materials, repeated interruptions, or costs that are not apparent from the protocol.
Consultation may therefore reveal opportunities to improve efficiency without removing essential measurements.
The role of participants and community partners in identifying overlooked research burdens is particularly relevant when the intended population has varied circumstances.
Participant preferences should inform decisions but do not independently establish scientific validity or ethical acceptability.
How Can Researchers Avoid Over-Optimizing for Convenience?
Burden reduction becomes counterproductive when it changes the research question, eliminates essential information, or produces measurements that cannot support the intended inference.
For example, replacing a validated learning assessment with a single satisfaction question may reduce completion time but measure a different construct.
Likewise, removing all intermediate observations from a study of learning trajectories would prevent the researcher from examining the shape of change.
Researchers should distinguish burden reduction that preserves measurement quality from simplification that compromises it.
The scientific purpose should determine the minimum information requirements, while ethical and practical considerations should guide how those requirements are met.
How Should Information Efficiency Be Evaluated During Implementation?
Researchers should monitor whether actual data collection performs as expected.
Relevant indicators may include completion time, missing responses, appointment attendance, technical difficulties, participant feedback, and the availability of essential outcome data.
Unexpected burdens may justify reviewing the protocol, but modifications can introduce differences between participants assessed under different procedures.
Material changes should be evaluated for their effects on measurement, analysis, safety, and applicable approval requirements.
Researchers should avoid changing outcomes or analytical priorities selectively in response to favorable or unfavorable interim results.
What Is the Most Important Principle for Efficient Research Design?
The most important principle is to align every substantial participation demand with a defensible scientific or protective purpose.
This does not require eliminating exploratory research or restricting data collection to the primary outcome alone.
It requires deciding which information is sufficiently valuable, how accurately it can be measured, and whether the burden of obtaining it is justified.
Researchers should evaluate acceptable participant burden relative to scientific value across the complete study rather than considering measurements in isolation.
Watch Out
Do not equate research efficiency with collecting the fewest variables, achieving the shortest questionnaire, or minimizing the number of appointments. A procedure is efficient only when it reduces unnecessary demands while preserving the information, validity, and participant protections needed for the study.
