Languages & Linguistics Research Questions

It names the language, the learners and the level, and its verb says how often you will measure. Asking whether high school ESL students acquire and retain vocabulary differently when taught through music and language integration activities, compared with traditional instruction, names both conditions, the population and the outcome. Acquire and retain commits you to measuring twice, so a single post-test wouldn't answer it. Treat the language pair as a variable rather than background: learners of Korean from Myanmar is a population, language learners is not.

Yes, at least once. Your examiner knows the acronym. The person who finds your abstract in a repository five years from now does not, and a question that can't be read cold is a question that gets skipped. Content and Language Integrated Learning, then CLIL. The Common European Framework of Reference for Languages, then CEFR. Words first, short form in brackets after, short form everywhere else. This field carries more acronyms than most, so it's worth a pass through your question set purely for them.

It depends what the question asks about. How teachers conceptualize an approach and how they implement it are two questions with different data behind them: belief you can interview for, practice you generally have to watch. Keeping the two apart is what lets one study claim both. If observation is out of reach, ask the belief question and say plainly that self-reported practice is what you have, rather than writing it up as though you saw it.

Yes, and a structural question often has to. "How is the passive formed in the source and target language respectively?" reads like background and isn't: any counting question that follows it is meaningless until the form has been described in each language. A question like that gets answered in its own right, with its own section in the findings. What doesn't work is a setup question you never actually answer, which a reader notices the moment a later question rests on it.

It is, when the question names both the material and the standard it will be checked against. "To what extent does the set of textbooks New Senior English for China cover the vocabulary in the wordlists required by the 2017 English Curriculum Standard?" is an audit: there is a right answer, and a reader knows what would count as one. Finite is what you need when there is no recruitment to fall back on. The same move works on campus signage, a translation, or a curriculum document, anything you can work through completely.

Through a gatekeeper, and earlier than feels necessary. A head teacher or program director says yes first, then the teacher whose lessons you want to sit in, then the learners themselves. Where any of the learners are minors, add parental consent and a slower ethics review to your timeline, and check what age counts as the threshold where the school is, because it isn't the same everywhere. One approved point of access usually buys you one named school or program, which is why most classroom studies stay inside one.

Often yes, with safeguards, and a lot of classroom research is done exactly this way. The usual conditions are that someone other than you collects consent, that taking part cannot affect anyone's grade, and that your dual role is stated plainly in the methods chapter. Your institution may add its own conditions, or rule it out entirely, in which case the nearest workable version is a colleague's class with the same learner profile.

Finite matters more than large. A corpus question works when the material is named and closed: one textbook series checked against one curriculum wordlist, the signage on two named campuses, the pronouns of address in one telenovela. Choose material you can work through completely, then report exactly what was in it. A partial pass over a huge corpus is weaker than a complete pass over a small one, and much harder to defend.

No. What gets asked is whether you can read the data accurately, which you answer by stating your own proficiency and bringing in a second rater wherever judgment is involved: coding, glossing, translation. If the language is one you're still learning, build that check into the method rather than leaving it to be raised in your oral defense.

Yes, if anyone outside your language pair is going to follow the argument. The Leipzig Glossing Rules are the usual convention: the original, a morpheme-by-morpheme gloss aligned to it, and an idiomatic translation underneath. Departments vary in how strictly they hold to it, so check your own style guidance before you set the format. Where the translation is itself what you're studying, say who produced it, because a rendering you made yourself is data you created.