diataxis_type	how-to
diataxis_goal	Provide better writing samples during a voice elicitation interview

How to Provide Better Writing Samples

Which questions ask for writing samples

The following questions prompt you to produce a writing sample during the interview:

Question ID	Module	Prompt Summary
M01-Q10	Orientation	Baseline writing sample
M02-Q11	Voice Personality	Personality-revealing writing
M03-Q08	Emotional Tone	Emotionally characteristic writing
M04-Q04	Complexity	Technical or complex writing
M04-Q12	Complexity	Extended complexity sample
M05-Q08	Narrativity	Narrative writing sample
M12-WS01	Writing Samples	Unconstrained writing sample 1
M12-WS02	Writing Samples	Unconstrained writing sample 2
M12-WS03	Writing Samples	Unconstrained writing sample 3
M12-WS04	Writing Samples	Unconstrained writing sample 4

Not all questions appear in every session. Adaptive branching determines which modules are active based on your writer type.

The pipeline analyzes four categories of features from your writing samples:

Lexical diversity: Vocabulary richness, type-token ratio, hapax legomena ratio, word frequency distributions. These features capture the range and specificity of your word choices.
Syntactic complexity: Sentence length distribution, clause embedding depth, dependency parse patterns, subordination ratio. These features capture how you construct sentences.
Pragmatic markers: Hedging frequency, stance markers, evidentiality signals, reader-orientation cues. These features capture your relationship to your claims and your audience.
Discourse cohesion: Paragraph structure, transition patterns, information density distribution, rhetorical move sequences. These features capture how you organize ideas across spans of text.

Write as you normally would. The pipeline measures your natural habits, not your performance under observation. Attempting to write differently than usual degrades the accuracy of the observed profile.
Shorter samples are acceptable. A 100-word paragraph written naturally is more valuable than a 500-word essay written self-consciously. The pipeline extracts features from samples as short as 50 words.
Do not edit extensively. First-draft writing reveals more about unconscious style patterns than polished prose. Light corrections for typos are fine. Restructuring or rewriting defeats the purpose.
Choose familiar topics. Writing about something you know well produces more natural prose than writing about an unfamiliar subject. The topic itself does not affect scoring; only the style features matter.
Use your typical register. If you normally write formally, write formally. If you normally write casually, write casually. The system measures what you do, not what you think you should do.