How Do You Plan Around Ethics Approval, Recruitment, Data Access, and Other Dependencies?

What do you do when one part of the research cannot begin until something else happens?

You may have a finished questionnaire but no ethics approval. Ethics approval may be progressing, but the university where you intend to recruit participants has not yet granted access. Site permission may eventually arrive, but recruitment could still take several months. Or perhaps your entire analysis depends on a restricted dataset that you have requested but do not yet possess.

These are not simply tasks on a to-do list. They are dependencies: conditions that determine whether other parts of the project can begin, continue, or finish.

Research timelines become fragile when dependencies are hidden. A plan might allocate eight weeks for data collection while overlooking the fact that those eight weeks cannot begin until several approvals, access arrangements, instruments, systems, and recruitment preparations are complete. Good planning therefore asks not only, “How long will each task take?” but also, “What must happen before this task is possible?”

Map what depends on what before assigning dates

A research timeline is really a network of dependencies

A simple timeline displays activities against dates. A useful research plan also shows the relationships among those activities.

Consider a participant-based study:

research question → study design → protocol and instruments → required ethics review → site authorization → recruitment → data collection → data preparation → analysis → reporting

This sequence immediately reveals something that a flat task list may hide. If recruitment requires both ethics approval and site authorization, completing the recruitment materials early does not mean recruitment can begin. If analysis requires a completed and cleaned dataset, having the statistical software ready does not remove that dependency.

Actual projects are usually more complicated because some activities can occur simultaneously. The objective is therefore not to force the research into one straight line. It is to identify the constraints that genuinely determine when other work becomes possible.

What exactly is a research dependency?

A dependency exists when one activity, milestone, or decision relies on another condition being satisfied.

Some dependencies arise from the scientific logic of the study. You cannot analyze measurements that have not yet been collected. Others arise from ethics, regulation, institutional processes, contracts, resources, or logistics.

Once you start looking at the project this way, many apparent scheduling problems become dependency problems.

Identify the gate before estimating the date

Suppose you hope to begin participant recruitment on September 1.

Instead of placing that date directly on the calendar, ask:

What must be true before recruitment can begin?

The answer might include:

• the study design and recruitment procedures are sufficiently finalized;
• required ethics review has been completed;
• approved participant information and consent materials are ready;
• the participating institution has granted access;
• the recruitment channel is authorized;
• research staff have completed required training;
• the data collection system is operational; and
• the study has any other required authorization to activate.

September 1 is therefore not primarily a date. It is the intended result of several preceding conditions.

This way of thinking is particularly useful when working backward from a fixed submission or graduation deadline. Each milestone can be traced backward to the conditions required to make it achievable.

Do not treat ethics approval as the only external dependency

Researchers sometimes organize an entire pre-data-collection timeline around ethics approval and assume that everything becomes possible once approval arrives.

That may not be true.

Ethics review determines whether the proposed human-participant research satisfies the applicable ethical and regulatory requirements of the reviewing body. It does not necessarily grant access to a school, hospital, organization, proprietary database, records system, community, laboratory, or other resource.

A hospital may require administrative or data-governance approval. A school may require permission from its administration. A company may need legal review. A controlled dataset may require a formal data-access request and institutional certification. A commercial instrument may require a license. A community-based study may require appropriate engagement or local authorization beyond formal ethics review.

The two may interact, and one organization may require evidence of the other. They should nevertheless be mapped separately unless the responsible institution explicitly combines them.

For human-participant research, determine what approval is required before the relevant activity

Under U.S. Department of Health and Human Services regulations, investigators conducting nonexempt human-subjects research are responsible for obtaining Institutional Review Board approval before involving human subjects. Investigators must also conduct approved research consistently with the IRB-approved plan and obtain prior approval for modifications except where changes are necessary to eliminate apparent immediate hazards to participants.

Those rules apply to research governed by that regulatory framework and should not be generalized as the exact procedure everywhere. Institutions and jurisdictions may use different ethics systems, terminology, submission pathways, and requirements.

The planning principle is more general: determine exactly which activity requires which authorization. Do not merely write “ethics” somewhere before “data collection.”

For example, ask whether the applicable rules permit recruitment before final approval, whether pilot activities require review, whether site authorization must precede ethics submission, and whether changes requested by a gatekeeper would require an ethics amendment. The responsible ethics office or institutional authority should answer these questions for the specific study.

Submission is not the same milestone as approval

A common timeline error looks like this:

Week 4: submit ethics application
Week 5: begin recruitment

The plan has silently assumed the outcome and duration of the review.

Submission and approval are different milestones:

HHS guidance illustrates why the distinction matters. An IRB may require modifications or clarification before research can proceed, and approval conditions may need to be verified before particular activities begin. The exact process depends on the reviewing body.

Your timeline should therefore model the process, not merely the act of submitting paperwork.

Data access can have its own chain of dependencies

Existing data are sometimes described as though they eliminate data collection and therefore simplify the schedule automatically.

They may. They may also replace participant recruitment with another dependency chain.

Restricted data can require an application, institutional approval, a research-use statement, a data-use agreement, information-security arrangements, authorized users, ethics documentation where applicable, training, and approval by a data custodian or access committee.

NIH controlled-access genomic data provide a concrete example. Its current Data Use Certification Agreement ties use to an approved Data Access Request and approved research project and includes institutional certification responsibilities. Other NIH resources, such as the National COVID Cohort Collaborative, use institutional Data Use Agreements before researchers apply for data access.

These are specific systems, not universal requirements for secondary data. They demonstrate why “download dataset” can be a poor representation of the actual task.

Before designing a project around restricted data, investigate:

• who controls access;
• what application is required;
• whether institutional sign-off is necessary;
• whether ethics review or another determination applies;
• what data-use restrictions exist;
• what security environment is required;
• who may become an authorized user;
• whether fees, contracts, or training apply; and
• how long the process might reasonably take.

Most importantly, verify that the dataset actually contains the variables, observations, documentation, and coverage your research question requires before allowing the entire project to depend on obtaining it.

Recruitment depends on more than permission to recruit

Once authorization is obtained, recruitment does not necessarily begin producing participants at the required rate.

Recruitment itself may depend on:

• access to the eligible population;
• cooperation from gatekeepers;
• approved recruitment channels;
• participant interest;
• eligibility rates;
• participant availability;
• study burden;
• seasonal or academic calendars;
• competition with other studies; and
• retention where follow-up is required.

A sample target is therefore not a schedule. If you need 120 participants and can recruit approximately 12 eligible participants per week, the nominal recruitment period is ten weeks. If the actual rate becomes six per week, it doubles.

This is why recruitment should be represented as a process with intermediate checkpoints rather than one box labeled “recruit participants.”

Map dependencies that can occur in parallel

Dependencies do not mean that the entire project must proceed one task at a time.

Suppose ethics review and institutional access approval are separate processes and neither requires the other to be completed first. If the relevant authorities permit it, they may be pursued in parallel.

Likewise, while waiting for external approval, you may be able to prepare analysis code using simulated data, organize references, refine documentation, complete permitted training, develop data dictionaries, or draft sections of the methods.

A dependency map should distinguish:

This distinction can shorten the elapsed project time without shortening the scientific work itself.

But do not invent parallelism where authorization is required

Pressure from a deadline can make premature activity look like efficient scheduling.

If recruitment requires prior ethics approval, beginning recruitment while approval is pending is not a clever way to parallelize the project. If a data-use agreement restricts access to approved users, obtaining the files through an informal route does not remove the dependency. If institutional permission is required before entering a site, starting fieldwork while the request is being processed does not make the project more efficient.

Some dependencies are decisions rather than approvals

Not every project is waiting for another organization.

You may be unable to finalize the questionnaire because the primary outcome has not been defined. The sample-size calculation may depend on choosing the main comparison. The ethics application may depend on a recruitment procedure the team has not agreed upon. The data-management plan may depend on whether identifiable information will be retained.

These are decision dependencies.

They can create circular planning:

We cannot finish A until we decide B, but we are postponing B until A is clearer.

When this happens, explicitly identify the unresolved decision and determine what information is actually needed to make it. The next planning problem is no longer “finish the questionnaire.” It is “resolve the outcome definition that is blocking the questionnaire.”

If several parts of the study are caught in this situation, map which parts depend on decisions that have not yet been made rather than allowing the project to remain vaguely “in planning.”

Some dependencies are resource constraints

A task may be scientifically and ethically ready but still unable to begin because a required resource is unavailable.

Examples include:

• a laboratory instrument is booked by another project;
• a licensed software package has not been purchased;
• funds have not been released;
• a trained interviewer is unavailable;
• the statistician cannot begin until another project finishes;
• required supplies have not arrived; or
• the secure computing environment has not been configured.

These dependencies should appear in the project plan if they can determine the schedule.

A useful question is not simply, “Do we need this resource?” but “By what date must it be available, and what happens if it is not?”

Dependencies can continue after data collection

Researchers often map pre-collection dependencies carefully and then assume that everything afterward is under their control.

Analysis may depend on transcription, laboratory processing, database locking, external coding, data reconciliation, or receipt of a final dataset. Manuscript submission may depend on coauthor approval. Thesis completion may depend on supervisor review, examination scheduling, revisions, signatures, or institutional clearance.

The entire research project therefore benefits from dependency mapping, not merely the approval stage.

This becomes particularly important when identifying research tasks that may take longer than expected. Activities controlled partly by other people or systems can be short in workload but long in elapsed time.

Find the dependencies that can move the final deadline

Not every dependency deserves equal attention.

Suppose two activities are delayed by one week.

The first is formatting a literature-search log that is not needed until the final report. Other work can continue.

The second is approval required before recruitment. No participant can be recruited until it arrives.

Both tasks are late by seven days, but their schedule consequences are very different.

Focus especially on dependencies that:

• block several later activities;
• have uncertain completion times;
• are controlled externally;
• cannot easily be accelerated;
• occur near a hard deadline; or
• would force major redesign if they fail.

These are the dependencies most likely to require early action and schedule protection.

Create fallback plans for dependencies that might fail

Some dependencies are not merely uncertain in timing. They may never be resolved as hoped.

A school may decline participation. A proprietary dataset may be unavailable. Recruitment may remain far below target. A laboratory instrument may become unusable. A collaborating site may withdraw.

Where the risk is plausible and consequential, identify alternatives before failure occurs.

A contingency plan is not permission to switch methods casually. Alternatives may affect ethics approval, comparability, contracts, budget, methodology, or the interpretation of findings. The value lies in knowing what options could be investigated rather than inventing them under deadline pressure.

Use milestone dates for dependency resolution

A dependency should have more than a vague note saying “waiting.”

Identify when it needs to be resolved for the rest of the project to remain viable.

For example:

Required ethics approval obtained by May 15

may be a meaningful milestone if recruitment must begin by June 1. The two-week gap may provide time for final activation work and some schedule contingency.

Likewise:

Data-access authorization obtained by August 1

can function as a decision point. If access has not been obtained by then, the project may need to reassess whether the final deadline remains achievable.

This is where major research project milestones become useful management tools rather than decorative dates.

Build buffer around dependency uncertainty

Externally controlled dependencies are rarely good candidates for schedules that assume the earliest possible completion date.

If the responsible office provides a typical processing range, use it. If revisions are common, consider the possibility of iteration. If recruitment rates are uncertain, monitor them early and preserve schedule capacity before the latest safe completion date.

Do not automatically add the same percentage to every dependency. The amount of contingency should reflect uncertainty, recoverability, and consequences.

This is the logic behind building buffer into a research timeline: protect the project where variation can actually propagate through later stages.

Maintain a dependency register for complex projects

A small study may need nothing more than a page of notes or a simple table. A larger project may benefit from explicitly tracking dependencies.

For each important dependency, record:

• what is required;
• which task or milestone depends on it;
• who controls or owns it;
• what documents or decisions are needed;
• when the process can begin;
• the expected or estimated resolution date;
• the latest safe resolution date;
• current status;
• what happens if it is delayed; and
• any realistic contingency.

This does not need to become bureaucratic. The purpose is to make invisible constraints visible enough to manage.

Review dependencies when the research plan changes

A change in one part of the project can create new dependencies elsewhere.

Adding another research site may require new authorization. Changing recruitment procedures may require ethics review. Adding identifiable data may change privacy or security requirements. Changing an instrument may require licensing, translation, technical configuration, or additional testing. Switching data sources may create a new data-use agreement.

Under HHS-regulated nonexempt human-subjects research, investigators must obtain prior IRB approval for modifications to approved research except when necessary to eliminate apparent immediate hazards to subjects. Other systems have their own requirements.

Whenever the project changes materially, therefore, ask not only what task has changed but also which dependencies have changed with it.

See how several independent-looking tasks become one dependency network

The useful insight is that “ethics,” “school permission,” “survey,” “interviews,” and “analysis” are not five boxes that can simply be placed in consecutive months. Some processes can overlap. Others cannot. Several are controlled partly by external actors, and one delayed dependency can move everything downstream.

Once the relationships are visible, the researcher can start external processes earlier, use waiting periods productively, establish meaningful checkpoints, and recognize sooner when the thesis timeline is becoming infeasible.

Dependencies cause problems when they are mistaken for ordinary tasks

Plan from prerequisites to milestones instead of from dates to tasks

Take each major stage of your research project and ask what conditions must already be satisfied before it can begin. Then identify who controls those conditions and how uncertain their timing is.

For complex projects, this may reveal that your nominal starting date is not the date on which you begin collecting data. The real project begins much earlier, when you start resolving the conditions that make data collection possible.

This also helps answer whether the project is realistically ready to move forward. A study may have an excellent protocol and still be operationally unready because access, resources, approvals, or recruitment pathways remain uncertain. Those unresolved dependencies should inform whether the project is ready to move from planning into the next stage.

Before trusting your research schedule, map the dependencies

Common questions about research project dependencies

Make the prerequisites visible before they become bottlenecks

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.