Designing a Study You Can Defend
An apparent increase in diagnoses, student engagement, service utilization, or institutional performance may reflect changes in a digital system rather than changes in the underlying phenomenon. Learn how software upgrades, revised recording practices, and data migrations can create artificial trends, and how to distinguish genuine change from measurement artifacts.
Read Guide →
Designing a Study You Can Defend
Automated data capture can reduce manual transcription, improve consistency, and support real-time quality checks. Yet automation can also introduce systematic errors that affect thousands of records before anyone notices. Learn how to distinguish errors that automation prevents from those it creates, and how to design validation and monitoring procedures that protect research validity.
Read Guide →
Designing a Study You Can Defend
Digital trace data can reveal detailed patterns of recorded activity, but they rarely capture the full context of human behavior. Missing events, ambiguous clicks, offline activity, and changing platform rules can undermine research conclusions. Learn how to determine whether digital traces provide sufficient evidence for the intended construct and when additional data sources or a revised question are necessary.
Read Guide →
Designing a Study You Can Defend
Continuous sensor measurement can reveal short-term variation and events missed by occasional assessments. However, more observations do not automatically produce more valid or informative evidence. Learn how temporal resolution, measurement error, missing data, autocorrelation, and analytical requirements determine whether continuous monitoring improves a research study.
Read Guide →
Designing a Study You Can Defend
Choosing how frequently a sensor records data is a methodological decision, not simply a technical setting. Sampling too slowly may miss important events, while unnecessarily high sampling rates can increase noise, storage requirements, and battery consumption. Learn how to select and justify sampling frequency according to the signal, research question, device characteristics, and intended analysis.
Read Guide →
Designing a Study You Can Defend
Passive wearable monitoring is often assumed to observe behavior without influencing it. Yet wearing a device, knowing that activity is recorded, or receiving feedback may change what participants do. Learn how monitoring reactivity can affect research validity, how to distinguish observation effects from intervention effects, and when study design adaptations may reduce or clarify these influences.
Read Guide →
Designing a Study You Can Defend
High-frequency sensors can generate millions of observations, but a larger dataset does not necessarily provide stronger evidence. Excessive data volume may overwhelm storage, quality assurance, processing, and statistical analysis. Learn how to estimate sensor data requirements, distinguish useful temporal information from redundancy, and design a manageable analytical workflow without discarding essential evidence.
Read Guide →
Designing a Study You Can Defend
A research study can sometimes be conducted entirely remotely, from participant recruitment and consent to data collection and follow-up. However, feasibility depends on more than access to online tools. Learn how to determine whether remote procedures preserve measurement validity, participant inclusion, ethical protections, and the evidence required to answer the research question.
Read Guide →
Designing a Study You Can Defend
Remote recruitment can expand a study's geographical reach while changing who encounters the invitation, qualifies to participate, and completes data collection. Learn how digital access, platform membership, self-selection, and recruitment procedures influence sample composition, and how to assess whether the resulting participants support the intended research conclusions.
Read Guide →
Designing a Study You Can Defend
Poor internet access can prevent otherwise eligible participants from joining remote studies, interrupt assessments, and create systematic missing data. Learn how to design low-bandwidth and offline alternatives, support participants without compromising measurement validity, and distinguish accessibility accommodations from changes that alter the research evidence.
Read Guide →
Designing a Study You Can Defend
Remote research can reach participants across geographical boundaries, but it may systematically exclude people with limited connectivity, unsuitable devices, low digital literacy, or accessibility barriers. When these characteristics are related to the outcomes being studied, the resulting sample may produce biased estimates. Learn how digital exclusion operates through recruitment and data collection, why large samples do not eliminate the problem, and how to evaluate and reduce its consequences.
Read Guide →
Designing a Study You Can Defend
Combining remote and in-person participant data may increase sample size and improve access, but differences in measurement conditions and participant selection can complicate interpretation. Learn when pooling is defensible, why adding collection mode as a covariate may be insufficient, and how to choose between pooled, adjusted, stratified, and mode-specific analyses.
Read Guide →
Designing a Study You Can Defend
Remote data collection allows participants to contribute research evidence from different locations, but variations in devices, instructions, environments, timing, and technical support can affect measurement quality. Learn how to establish standardized remote procedures, distinguish essential controls from acceptable flexibility, document protocol deviations, and maintain comparability without unnecessarily excluding participants.
Read Guide →
Designing a Study You Can Defend
Online research can produce large samples quickly and inexpensively, but convenient recruitment may systematically favor participants who are accessible, digitally engaged, or particularly interested in the topic. Learn when convenience becomes a threat to representativeness, why sample size cannot eliminate selection bias, and how to determine whether online findings support population-level conclusions.
Read Guide →
Designing a Study You Can Defend
An adaptive study design allows researchers to make prospectively planned changes using accumulating study information. Understand which modifications qualify as adaptive, how they work, and why statistical and operational safeguards matter.
Read Guide →
Designing a Study You Can Defend
Adaptive designs are not automatically superior because they permit changes during research. Learn when planned adaptations provide meaningful scientific benefits and how to distinguish methodological value from operational convenience.
Read Guide →
Designing a Study You Can Defend
A credible adaptive design requires more than announcing that a study may change. Learn which objectives, adaptation rules, statistical methods, interim procedures, and governance safeguards should be established before relevant results become available.
Read Guide →
Designing a Study You Can Defend
Adaptive designs can preserve statistical validity, but poorly planned modifications may increase false-positive findings. Understand how repeated testing, outcome-driven sample size changes, and treatment selection affect Type I error, and how researchers can prevent misleading conclusions.
Read Guide →
Designing a Study You Can Defend
Adaptive designs can require statistical expertise beyond conventional sample size calculations and hypothesis testing. Learn which adaptive procedures demand specialist support, what statisticians contribute, and how researchers can assess whether their team is prepared.
Read Guide →
Designing a Study You Can Defend
Stopping a study for futility may prevent unnecessary recruitment when achieving its scientific objective has become unlikely. Learn how futility decisions work, why they differ from evidence of no effect, and which statistical and ethical safeguards are necessary.
Read Guide →
Designing a Study You Can Defend
An interim analysis examines accumulating study information before the planned final analysis, but its purpose and timing must be carefully defined. Learn when interim analyses are useful, how they affect statistical inference, and what researchers should establish before conducting them.
Read Guide →
Designing a Study You Can Defend
Not every study benefits from an interim analysis. Learn when preliminary testing adds unnecessary complexity, why delayed outcomes and short recruitment periods limit its usefulness, and how to distinguish essential safety monitoring from optional interim efficacy testing.
Read Guide →
Designing a Study You Can Defend
A research study can measure several outcomes, but the number it can meaningfully support depends on its objectives, sample size, measurement quality, and analytical plan. Learn how to prioritize outcomes without sacrificing scientific breadth or creating unsupported conclusions.
Read Guide →
Designing a Study You Can Defend
A study can conduct many statistical comparisons, but not every comparison can support an equally strong conclusion. Learn how sample size, multiplicity, statistical power, and the definition of confirmatory hypotheses determine which comparisons are defensible.
Read Guide →
Designing a Study You Can Defend
A clearly designated primary outcome or comparison helps establish what a study is principally intended to demonstrate. Learn when primary designation is necessary, how it shapes statistical planning, and why some research designs legitimately require more than one primary objective.
Read Guide →