PRC common mistakes
PRC · Understand the traps and how to avoid them.
Treating 'research' as a single activity and applying one set of conduct rules to every project, instead of first classifying the work (clinical trial, secondary data analysis, quality improvement, program evaluation, market research) and then identifying which professional-conduct standards attach to that category.
Candidates memorize a generic list of ethical principles and then force every scenario into it. Because the exam's scenarios are deliberately ambiguous about the category of research, this leads to selecting a plausible-but-wrong governing standard, and the candidate cannot explain why a given rule applies to a given project.
What to do instead
Before answering any scenario, write a one-line classification: what is being studied, who/what is the source of data, whether an intervention occurs, and whether results are intended for generalizable knowledge. Then map that classification to the applicable conduct requirements. Practice on 15 scenarios by classifying first and only then selecting an answer.
Confusing the roles of the researcher, the sponsor/funder, the institution, and the oversight body, and attributing an obligation to the wrong party (for example, assuming the researcher is responsible for a sponsor's monitoring duty, or that the institution approves a protocol change that only an oversight committee can approve).
Role boundaries are rarely memorized as a matrix; candidates learn obligations as a flat list. Under time pressure they pick whichever party 'sounds responsible' rather than the party that actually holds the duty, and distractors exploit exactly this ambiguity.
What to do instead
Build a four-column role matrix (researcher / sponsor / institution / oversight body) with rows for protocol changes, adverse-event reporting, consent, data custody, and conflict-of-interest disclosure. For each practice question, name the role before reading the options.
Answering conflict-of-interest questions by disclosing only financial interests, and overlooking non-financial interests such as career advancement, institutional pressure, editorial roles, or personal relationships with participants or co-authors.
Training materials emphasize money, so candidates develop a narrow mental model. The exam tests whether the candidate understands that the purpose of disclosure is to let others judge whether judgment could be biased, which is not limited to payments.
What to do instead
Adopt a two-question test for every COI item: (1) could this interest reasonably be seen to bias judgment or create an obligation? (2) has it been disclosed to the party who needs to know, in the form and timeframe required? Apply the test to non-financial examples until it is automatic.
Selecting 'obtain informed consent' as the answer whenever a scenario involves human participants, even when the scenario is about a different obligation (data security, authorship, record retention, or reporting to an oversight body) or when consent is not the operative requirement.
Consent is the most rehearsed concept in research conduct, so it becomes a default answer. Candidates stop reading the stem carefully and pattern-match on the word 'participant', missing the specific obligation the question is testing.
What to do instead
Read the final sentence of the stem first and identify the exact question being asked (what must the researcher do next / what went wrong / which requirement applies). Eliminate any option that addresses a real but different obligation. Drill 20 items where consent is a distractor.
Describing data-handling failures only in terms of 'keeping data secure' and ignoring the linked requirements: documented custody and access, de-identification adequacy, retention and destruction schedules, permitted secondary use, and transfer to third parties.
Candidates treat confidentiality as a single yes/no attribute. In reality the exam probes the chain of custody and the specific permissions attached to data, and a generic 'secure storage' answer is usually a distractor designed to feel safe.
What to do instead
For each data scenario, trace the data lifecycle: collection, storage, access, sharing, retention, destruction. Identify which link the scenario breaks. Memorize that de-identification strength, permitted uses, and retention periods are governed by the protocol and applicable requirements, not by convenience.
Handling authorship questions by ranking contribution size alone, without applying the criteria for authorship, the requirement for approval of the final version, accountability for the work, and proper acknowledgement of non-author contributors.
Candidates import a 'first/last author by seniority' heuristic from their own field and ignore that authorship criteria are conjunctive and that ghost, gift, and guest authorship are distinct failures. Distractors are written to reward the seniority heuristic.
What to do instead
Memorize the authorship criteria as a checklist and apply all items to each named person. Separately memorize the definitions of ghost, gift, and guest authorship, and the rule that contributors who do not meet criteria belong in acknowledgements with permission.
Treating fabrication, falsification, and plagiarism as the whole of research misconduct, and failing to distinguish them from other unacceptable practices (duplicate publication, salami slicing, image manipulation, selective reporting, citation manipulation) and from honest error.
The classic three-part definition is over-learned. Candidates then mislabel questionable practices as misconduct or, conversely, dismiss serious misrepresentation as 'just poor practice', and they miss the intent and significance elements that separate misconduct from error.
What to do instead
Build a three-tier table: misconduct (fabrication, falsification, plagiarism), other unacceptable practices, and honest error. For each practice, record the defining feature and the typical evidence. Practice classifying 20 short scenarios into the three tiers.
Answering record-keeping questions by choosing the longest retention period offered, or by assuming raw data may be discarded once analysis is complete and results are published.
Candidates lack a mental model of why records are kept (verifiability, audit, dispute resolution, reproducibility) and so default to 'more is safer'. This produces answers that conflict with specified retention requirements and with the principle that data must remain available for verification.
What to do instead
Learn the rationale first: records exist so that findings can be verified and challenged. Then learn that retention periods and custody are set by the protocol, funder, institution, and applicable requirements, and that destruction must be documented and authorized rather than assumed.
Reading a scenario and answering from personal or local workplace habit ('this is how our lab does it') rather than from the stated professional standard, especially on mentoring, supervision, delegation, and who may sign off on a protocol or report.
Experienced candidates are the most exposed: tacit local norms feel authoritative and are fast to retrieve, while the formal standard requires deliberate recall. The exam deliberately includes options that match common local practice but violate the standard.
What to do instead
For each domain, write down one local habit you hold and the formal requirement beside it. In practice, justify every answer by naming the requirement it rests on; if you cannot name it, flag the item and revisit it rather than trusting familiarity.
Managing time by answering long scenario items immediately and marking nothing for review, so that later, easier items are rushed and earlier uncertain items are never revisited with fresh reasoning.
Scenario-based conduct exams front-load reading load; candidates underestimate reading time per item and do not build a review buffer. Fatigue then degrades exactly the judgment-heavy items that carry the most marks.
What to do instead
Do two passes. On pass one, answer only items you can justify in one sentence and flag the rest. On pass two, return to flagged items with the role matrix and classification step. Rehearse with timed sets of 30 scenario items to calibrate reading speed before exam day.