Language investigation

Turn the vague aim 'I want to investigate language and power' into a focused, investigable research question.

1. 01Narrow the field

   Choose a specific context where power is enacted in language - say, a broadcast interview between a presenter and a politician.

2. 02Specify the feature

   Pick concrete features to analyse: control of the agenda, interruptions, and interrogative versus declarative moves - observable markers of power in discourse.

3. 03Tie it to theory

   Connect to Fairclough's power in discourse and Wareing's types of power, so the findings can be interpreted, not just described.

4. 04State the question

   'How is power in discourse enacted through turn-taking and questioning in a broadcast political interview?' - focused, answerable with a single transcript, and theory-linked.

Result: The vague aim becomes 'How is power in discourse enacted through turn-taking and questioning in a broadcast political interview?' - a focused question that specifies the data (one interview transcript), the features (turn-taking, questioning) and the theory (Fairclough, Wareing), making the whole investigation achievable.

You want to investigate a two-year-old's use of language during play. Plan the data collection and transcription.

1. 01Obtain consent

   As the participant is a child, obtain informed consent from the parent/guardian, explaining the purpose and use of the recording, and follow safeguarding - a compulsory ethical step.

2. 02Collect appropriate data

   Record a modest, focused sample (for example 15-20 minutes of naturalistic play with a familiar caregiver) - enough to yield features, small enough to transcribe and analyse in depth.

3. 03Choose transcription detail

   Use a transcription key marking pauses, the child's actual pronunciations (for phonological analysis), overlaps and caregiver turns - the level of detail the CLA focus requires - and anonymise the child's name.

4. 04Address the observer's paradox

   Record during ordinary play with a familiar adult in a familiar setting so the child behaves naturally, and acknowledge in the write-up that recording may still affect behaviour.

Result: The plan obtains parental consent and follows safeguarding, collects a focused naturalistic sample, transcribes at a phonologically detailed level with anonymisation and a stated key, and mitigates and acknowledges the observer's paradox - ethically sound, methodologically matched data collection.

In an interruption study you count that Speaker A interrupts 12 times and Speaker B twice. Show how to analyse this properly.

1. 01Present the quantitative pattern

   State the counts clearly (A: 12 interruptions; B: 2) - this establishes a pattern but does not yet explain anything.

2. 02Interpret qualitatively

   Examine the interruptions in context: are A's interruptions dominant (seizing the floor, changing topic) or supportive (collaborative overlaps, agreement)? The meaning depends on function, not just frequency.

3. 03Frame with theory

   Relate to the dominance model (Zimmerman & West) if the interruptions are floor-seizing, but note the critique (Beattie) that not all interruptions are dominant - testing the data against the theory rather than assuming it.

4. 04Answer the question

   Conclude what the combined evidence shows about power in the interaction, acknowledging that the raw count alone would have been misleading without the qualitative reading.

Result: The count (12 vs 2) identifies a pattern, but the analysis is completed by qualitatively examining whether A's interruptions are dominant or supportive and testing this against the dominance model and its critique - integrated mixed-method analysis that interprets rather than merely counts.

Your small study of one mixed-sex conversation found more male interruptions. Write a properly calibrated conclusion.

1. 01State what the data showed

   Report precisely: 'In this single conversation, the male speaker interrupted more frequently, and most of these interruptions were floor-seizing rather than supportive.'

2. 02Relate to theory

   'This is consistent with the dominance model (Zimmerman & West), in which interruption enacts conversational dominance.'

3. 03Acknowledge limitations

   'However, the data set is a single conversation between two individuals, so it cannot support any general claim about men and women; the pattern may reflect these individuals, their roles, or the observer's paradox rather than gender.'

4. 04Draw the modest conclusion

   'The findings therefore suggest a pattern consistent with the dominance model in this instance, and indicate that a larger, controlled study would be needed to test whether it generalises.'

Result: The conclusion states the specific finding, relates it to the dominance model, candidly notes that one conversation cannot support a general claim (confounds, the observer's paradox), and calibrates the claim accordingly - the evidence-proportioned, self-aware evaluation the NEA rewards over overclaiming.

A draft report's introduction promises to apply the dominance model, but its analysis only counts interruptions and its conclusion claims to have 'proved men dominate conversation'. Identify the problems.

1. 01Check introduction against analysis

   The introduction promises the dominance MODEL, but the analysis only COUNTS interruptions without qualitatively interpreting them or testing the model - a mismatch between promise and delivery.

2. 02Check the conclusion against the data

   The conclusion 'proved men dominate conversation' overclaims: a single small data set cannot prove a universal generalisation - the claim is not proportioned to the evidence.

3. 03Check for the evaluation

   There is no evaluation of the method's limitations (sample size, confounds, the observer's paradox), a required and rewarded section that is missing.

4. 04Prescribe the fixes

   Add qualitative interpretation testing the dominance model, recalibrate the conclusion to what the specific data showed, and add an honest evaluation of limitations - restoring consistency from question to evaluation.

Result: The draft is internally inconsistent (an introduction promising theory the analysis never applies, a conclusion overclaiming from limited data, and a missing evaluation); the fixes are qualitative theory-testing analysis, a calibrated conclusion and a candid evaluation - making the report a coherent whole from question to limitations.