AQA A-Level English Language: Methods of Language Analysis in the NEA (7702)

In the AQA A-level English Language (7702) non-exam assessment, Language in Action, methods of language analysis are the tools you use to study your own data in the Language Investigation and to reflect on your own writing in the commentary. The specification says each task will have “a distinctive approach to analysis” shaped by the research pathway or theme you choose, but that every analysis will cover relevant aspects of six language levels: phonetics, phonology and prosodics; graphology; lexis and semantics; grammar, including morphology; pragmatics; and discourse.

This guide teaches section 3.3.3 Methods of language analysis (4.3.3 in the specification PDF) of the A-level subject content in the AQA AS and A-level English Language specification (7701/7702), Version 1.2 (14 October 2021). The NEA is A-level only: it has a word count of 3,500 (a 2,000-word investigation excluding data, and a piece of original writing and a commentary of 750 words each), is worth 100 marks and 20% of the A-level, and is assessed by teachers and moderated by AQA. For the tasks themselves, see the Language in Action NEA study guide, the NEA revision notes and the NEA practice tasks.

What does the specification say?

Specification point (3.3.3)	What it means for you
Both the investigation and original writing have a distinctive approach to analysis, shaped by the research pathway or theme	Choose the levels that fit your research questions or writing purpose; you do not need to apply all six equally
Underlying any analysis will be coverage of some relevant aspects of the language levels	Your analysis must still be systematic and use the levels’ terminology
The six levels: phonetics, phonology and prosodics; graphology; lexis and semantics; grammar, including morphology; pragmatics; discourse	These are the frameworks the moderator expects to see applied accurately

The preparation list in the specification also asks you to “transcribe spoken data where appropriate”, “use language concepts and ideas”, “work in greater depth and with greater range” and “reflect on the writing process using methods of language analysis”. Those are the skills this guide builds.

How do you choose methods for your investigation?

Start from your research question, not from the list of levels. Ask which levels will give the most revealing evidence for that question.

Investigation focus (examples from the specification)	Most useful levels	What you might count or examine
Features of a regional dialect, based on a survey	Phonology; lexis and semantics; grammar	Pronunciation variants; dialect words; non-standard verb forms
Gendered talk	Pragmatics; discourse; grammar	Interruptions and overlaps; hedges; tag questions; topic control
The language of sports commentary	Grammar; lexis and semantics; prosodics	Ellipsis and present-tense verbs; specialist lexis; pitch and pace
How travel guides represent a community	Lexis and semantics; grammar; discourse	Evaluative adjectives; agency in clauses; how sections are ordered
The writing of two children aged 8	Grammar, including morphology; orthography (graphology); discourse	Clause types and connectives; spelling patterns; narrative structure
How turn-taking works in real-time writing online	Discourse; pragmatics; graphology	Adjacency pairs; repair; emoji and punctuation used for tone

The specification’s four kinds of question help too: a genre-based investigation (what are the distinctive features?), a function-based one (what is the language used to do?), an attitudes-based one (how do people feel about it?) and a user-based one (who uses it?). Each suggests different methods. An attitudes-based study may analyse questionnaire responses; a genre-based study may compare features across several texts.

How do you apply each level well?

Level	Precise terminology to use	Pitfall to avoid
Phonetics, phonology and prosodics	Phonemes, elision, assimilation, glottal stop, intonation, stress, pause	Describing sounds with spelling alone; use a consistent transcription
Graphology	Typeface, layout, colour, image, positioning, white space	Describing visuals without linking them to meaning
Lexis and semantics	Semantic field, connotation, register, jargon, neologism, collocation	Listing words with no pattern or count
Grammar, including morphology	Clause types, sentence functions, modal verbs, tense and aspect, voice, affixes, compounds	Naming only word classes; look at phrase and clause structure
Pragmatics	Implicature, politeness and face, presupposition, speech acts	Guessing at intentions without evidence from context
Discourse	Structure, cohesion, turn-taking, adjacency pairs, openings and closings	Ignoring how the whole text or conversation is organised

How do you make analysis systematic?

Moderators reward analysis that is methodical and linked to context. Three habits help.

1. Code your data. Decide categories in advance (for example “hedges”, “tag questions”, “interruptions”) and mark every instance in your transcript or texts. Keep clean copies of data in the appendices, as the specification requires.
2. Combine quantitative and qualitative evidence. Counts and percentages show patterns; close analysis of examples explains them. The appendices should hold evidence that supports any quantitative approach.
3. Relate features to context. Explain how audience, purpose, genre, mode and the relationship between participants shape the features you find.

Worked example (quantitative to qualitative). In 20 minutes of transcribed group talk, Speaker A produces 14 hedges (“sort of”, “I think”) in 52 turns and Speaker B 4 hedges in 48 turns. As percentages of turns: A = 14 / 52 x 100 = 26.9%, B = 4 / 48 x 100 = 8.3%. The numbers show a difference; the analysis must then explain it, for example by examining whether A’s hedges soften disagreement (a politeness strategy) or express uncertainty, and whether the roles of the speakers (for example, who is chairing) explain the pattern. The conclusion should be cautious, as a small sample cannot support broad claims about all speakers.

How do methods apply to the original writing commentary?

In the commentary you use the same levels to reflect on your own writing and your style model. The specification says the commentary lets you “consider and evaluate the style model, the writing process and the effectiveness of the final piece of writing”.

• Identify features of the style model at several levels (for example the short declaratives and second-person address of a persuasive speech).
• Explain the choices you made in your own piece, using terminology: why you chose a particular sentence type, lexical field or structure, for your audience and purpose.
• Evaluate effectiveness honestly: what worked, what you redrafted and why.

Worked example. “In my travel article I followed the style model’s use of present-tense narration (‘The market wakes slowly’) to create immediacy, and I chose concrete sensory lexis from the semantic field of food to help a non-specialist reader picture the scene. In my second draft I replaced several adjectives with verbs, because the style model relied on dynamic verbs for pace.”

Frequently asked questions

Do I have to use all six language levels in my investigation? No. The specification says the approach is shaped by your research pathway, with coverage of relevant aspects of the levels. Choose those that best answer your research questions, and apply them in depth.

Should my investigation use numbers? Quantitative evidence is often useful, and the specification expects appendices to contain evidence supporting quantitative approaches. But numbers must be interpreted with qualitative analysis and linked to context.

Do I need to transcribe spoken data? Where your data is spoken, yes: the specification lists transcribing spoken data where appropriate. Use a consistent set of conventions and include a key.

Is the commentary just a description of my writing? No. It should analyse the style model and your choices using methods of language analysis, and evaluate how well your writing achieves its purpose.

Common mistakes

• Applying every level superficially instead of the most relevant levels in depth.
• Counting features without explaining what the pattern means.
• Making claims bigger than the data supports.
• Using terminology loosely (for example calling any short word a “monosyllable” without purpose).
• Writing a commentary that narrates the writing process with no linguistic analysis.
• Leaving out clean copies of data or evidence for quantitative claims.

Official syllabus

AQA AS and A-level English Language (7701/7702) specification, Version 1.2 (14 October 2021), A-level subject content 3.3 Language in action, section 3.3.3 Methods of language analysis, with the NEA requirements set out in 3.3 and the scheme of assessment.

Section numbers. This site follows the A-level (7702) subject content on aqa.org.uk, where Paper 2 content is 3.2 and the NEA is 3.3. The combined AS and A-level specification PDF (Version 1.2, 14 October 2021) puts the AS content in section 3, so the same A-level sections are numbered 4.1, 4.2 and 4.3 there.