ENADE Sistemas de Informação common mistakes
ENADE Sistemas de Informação · Understand the traps and how to avoid them.
Writing a discursive answer that only restates the question and lists keywords without building an argument, so the answer never reaches the 'excellent' band of the grading rubric.
Candidates trained on multiple-choice drills treat the discursive item as a recall test and dump definitions instead of answering the actual command verb (analisar, comparar, propor).
What to do instead
Before writing, underline the command verb and the object of the question, then outline a thesis sentence plus two or three argument steps plus a conclusion that explicitly answers the verb; check that each paragraph advances the thesis rather than defining terms.
Answering a discursive question with a generic essay that ignores the specific scenario (a company, dataset, or system described in the prompt) and its stated constraints.
Time pressure and anxiety push candidates toward memorized templates that feel safe but do not engage the particulars the rubric rewards.
What to do instead
Extract the concrete givens from the prompt into a short list (actors, constraints, goal) and make at least one paragraph explicitly reason about those givens, naming them as they appear in the text.
Confusing algorithm complexity classes, claiming an O(n log n) sort is linear, or asserting that a hash table guarantees O(1) worst case.
Candidates memorize the average-case headline for each structure but never internalize the difference between average, worst, and amortized analysis.
What to do instead
For each core structure and algorithm, write a three-row table (melhor, médio, pior caso) from memory and justify each row; specifically rehearse that hash tables degrade to O(n) under collisions and that comparison sorts are Ω(n log n).
Treating normalization as a mechanical checklist ("this table is in 3NF because it has no transitive dependency") without verifying functional dependencies or identifying the actual candidate keys first.
Textbook procedures are memorized as sequences of steps, and candidates skip the dependency analysis that gives the steps meaning.
What to do instead
Always start by listing the functional dependencies and deriving all candidate keys, then test each normal form against those keys; practice with a table where the intuitive answer is wrong because of a non-obvious candidate key.
Describing a software process or architecture using only buzzwords ("usar Scrum e microsserviços") without connecting the choice to the requirements, scale, or team constraints stated in the question.
Candidates learn frameworks as labels and are never asked to justify a choice against alternatives, so the answer becomes a name-drop.
What to do instead
Structure the answer as decision, justification, trade-off: state the choice, tie it to a named constraint in the prompt, and name one cost or risk of the alternative you rejected.
Mixing up the layers of a computer system or network model, e.g. placing routing at the transport layer or claiming TCP provides message boundaries.
The layers are studied as isolated definitions rather than as a stack whose responsibilities are distinguished by what each layer can and cannot guarantee.
What to do instead
Build a one-page reference mapping each layer to its protocol data unit, addressing scheme, and one guarantee it does not provide; test yourself by classifying ten concrete phenomena (retransmission, port multiplexing, framing) to the correct layer.
Writing code or pseudocode in the programming question that is syntactically plausible but fails on boundary conditions (empty input, single element, maximum index) or has an off-by-one error.
Candidates rehearse the happy path and never trace their own code by hand against edge cases under exam conditions.
What to do instead
After writing, dry-run the code on at least three inputs: empty, one element, and the boundary case named in the prompt; explicitly state the loop invariant or termination condition.
Analyzing an ethical, legal, or social impact question (data privacy, LGPD, algorithmic bias, professional responsibility) purely in technical terms, omitting the normative dimension the rubric requires.
The technical curriculum dominates candidates' preparation, so they default to engineering solutions and never invoke legal or ethical frameworks.
What to do instead
Prepare a compact reference of relevant norms (LGPD principles, the code of ethics of the profession, responsibility for produced systems) and practice writing one paragraph that names the applicable principle and applies it to the scenario.
Misreading the scope of an interdisciplinary item, e.g. answering a question about information systems in organizations with pure programming content, or vice versa.
Candidates classify the item by a keyword rather than by the actual object of the question, then answer the question they expected instead of the one asked.
What to do instead
Read the stem twice and write a one-line paraphrase of what is being asked before answering; verify that the paraphrase names the same object and level of analysis (organizational, system, or code) as the stem.
Treating the multiple-choice items as pure recall and choosing the longest or most technical-sounding option when uncertain, instead of eliminating options by logical contradiction.
Under time pressure candidates fall back on test-taking heuristics rather than reasoning about each option's truth value.
What to do instead
For each option, write a one-word verdict (verdadeiro/falso) with the specific reason before selecting; if two options seem true, re-read the stem for the qualifier (sempre, apenas, exclusivamente) that makes one false.