Professional Responsibility Flashcards: What to Put on Them

The MPRE tests the ABA Model Rules of Professional Conduct and the ABA Model Code of Judicial Conduct as written, plus a small amount of law governing lawyers (agency, malpractice, evidentiary privilege). It is 60 scored questions plus 10 unscored, two hours, scaled 50–150, with passing scores set by each jurisdiction in roughly the 75–86 range. Because the exam is rule-text driven, your cards should track rule text and its structure, not a professor's gloss.

Your law school PR exam is a different animal, and this is where practice genuinely varies. Some professors test the Model Rules verbatim; some test their own state's adopted rules (California's are notably different in numbering history and in the substantive treatment of confidentiality); some test the Restatement (Third) of the Law Governing Lawyers alongside the Rules; and some run a policy-and-cases seminar where the graded work is an essay about the adversary system. Ask, in office hours, which source controls on the exam. That answer determines whether your cards say "Model Rule 1.6(b)(1)" or "our state's rule."

Practical consequence: build one deck to Model Rules text (portable to the MPRE), then add a small overlay deck of "how our state differs" cards if your professor tests local rules. Do not merge them. Mixing jurisdictions on the same card is the fastest way to get both wrong.

Rank by yield. These rule families generate the overwhelming majority of tested questions, and each deserves a small cluster of atomic cards rather than one bloated card:

Every card front should be a question with one answer. "Rule 1.7" is not a card. "A concurrent conflict is consentable only if four conditions are met — name them" is a card, and the back reads: the lawyer reasonably believes she can provide competent and diligent representation to each client; the representation is not prohibited by law; it does not involve one client asserting a claim against another client in the same litigation; and each affected client gives informed consent, confirmed in writing.

Build the modal verb into the prompt. PR exams are won on "must" versus "may." Disclosure under 1.6(b) is permissive. Withdrawal under 1.16(a) is mandatory. Reporting another lawyer's misconduct under 8.3 is mandatory when the conduct raises a substantial question as to that lawyer's honesty, trustworthiness, or fitness — unless the information is protected by 1.6 or learned through an approved lawyers' assistance program. Write cards as "must, may, or must not?" and answer with the verb first.

Add a consequence tag to every card: discipline, disqualification, civil liability, fee forfeiture, or evidentiary suppression. Students lose points by answering a disqualification question with a discipline rule. A card that ends "— tested as: disqualification" trains the right output.

Make explicit comparison cards for the pairs students conflate: confidentiality under Rule 1.6 (all information relating to the representation, from any source, ethical duty, remedy is discipline) versus the attorney-client privilege (communications for legal advice, evidentiary rule, remedy is exclusion); ineffective assistance under Strickland v. Washington, 466 U.S. 668 (1984), versus a malpractice claim versus a disciplinary charge; and informed consent (oral is enough) versus consent "confirmed in writing" versus a "writing signed by the client" under Rule 1.8(a).

Conflicts cards should be built as decision trees split across several cards, not one card per rule. Card one: is the conflict direct adversity or a material limitation (1.7(a)(1) vs. (a)(2))? Card two: is it consentable (the four conditions above)? Card three: is it imputed to the firm under 1.10, and if the conflict is personal to one lawyer and presents no significant risk of materially limiting others, it is not imputed. Card four: can screening cure it — yes for a lateral hire under 1.10(a)(2) if the screen is timely, the screened lawyer gets no part of the fee, and written notice goes to the affected former client; yes for government lawyers under 1.11(b); yes for prospective clients under 1.18(d).

For former clients, card the trigger separately: 1.9(a) applies to the same or a substantially related matter where the new client's interests are materially adverse. "Substantially related" is a defined term in the comments — card the definition, not a paraphrase.

Card each 1.8 subsection alone. Business transactions with a client, use of client information, gifts from clients, literary or media rights, financial assistance in litigation, third-party payors, aggregate settlements, prospective malpractice limitation, and sexual relationships each carry different requirements, and only some are imputed under 1.8(k).

Yes, but a short list. For the MPRE, you need the constitutional overlay on advertising and solicitation — Bates v. State Bar of Arizona, 433 U.S. 350 (1977) (truthful advertising is protected commercial speech), Ohralik v. Ohio State Bar Ass'n, 436 U.S. 447 (1978) (in-person solicitation for pecuniary gain may be banned), and In re Primus, 436 U.S. 412 (1978) (political-expression solicitation by a nonprofit is protected) — plus Nix v. Whiteside, 475 U.S. 157 (1986) (refusing to assist perjury is not ineffective assistance) and Upjohn Co. v. United States, 449 U.S. 383 (1981) (corporate privilege beyond the control group). Card the holding in one sentence and the rule it interacts with.

For a doctrinal seminar, add whatever cases your professor assigned with a fact hook on the front, since essay credit comes from analogizing facts.

Judicial conduct is roughly ten to twelve questions on the MPRE and deserves its own small deck: disqualification standards under CJC Rule 2.11, the ex parte communication ban and its narrow exceptions under 2.9, public comment on pending matters, gifts and reporting, and the limits on political and extrajudicial activity. These are frequently skipped and cheaply learned.

Generative tools are good at reformatting text you supply and bad at recalling rule numbers from memory. The safe workflow is extractive, not generative: paste the actual text of Rule 1.6 or Rule 1.15 into the prompt and ask for cloze-deletion cards drawn only from that text, with no outside additions. Then spot-check every card against the rule.

The characteristic failure modes are worth naming: invented subsection numbers, comment language presented as black-letter rule, pre-2018 advertising rules (Rules 7.2 and 7.3 were substantially restructured), Rule 8.4(g) treated as universally adopted when many states have not adopted it, and California-specific answers offered as Model Rule answers. Any card citing a rule you cannot find in the current rule text should be deleted, not fixed.

Use AI for the parts that are genuinely generative and low-risk: turning a rule you already understand into a two-line hypothetical, writing distractor answer choices for self-quizzing, or producing a mnemonic. Do not use it to decide what the rule is.

A complete Model Rules deck runs roughly 350–500 atomic cards, plus 60–80 for judicial conduct. That is reviewable in 20–25 minutes a day in a spaced-repetition system. If you are past 800 cards, your cards are too granular or you have carded comment language that will never be tested.

Start six to eight weeks before the MPRE, or from week three of the semester for a course exam, and pair the deck with released MPRE practice questions or your professor's old exams. Flashcards encode the rule; questions teach you which rule a fact pattern is calling for. Neither works alone, and the second is where most students under-invest.

If you cite the Rules in a paper or exam, the Bluebook form is: Model Rules of Pro. Conduct r. 1.6(b)(1) (Am. Bar Ass'n 2023). The 21st edition abbreviates "Professional" as "Pro." and uses a lowercase rule designator; older editions used "Prof'l" and a capital "R." Check which edition your school follows. State rules follow the same pattern with the jurisdiction name. On an exam, most professors accept "Rule 1.6(b)(1)" without more — but confirm, because that too varies.