Mandy Sha

#ISWGHS @UNStats handbook #survey on households individuals foundations, what're emerging approaches? ✔️ machine translation + post-edit ✔️ #GenAI draft/eval items, coding open-ends ✔️ Tools for workflow (DataCTRL), data capture, documentation ✔️ CAVI 🗣️, web probing, crowdsource

#ISWGHS @UNStats handbook #survey on households individuals foundations, what're 3⃣ applied frameworks for quex design #translation instrumentation? 1⃣ cognitive response process model 2⃣ QDET ques design, dev, eval, testing 3⃣ repsondent-centered design: ⤵️ burden, UX-informed

Check out Chapter2⃣ by @NORCNews researchers in 🆕📖 𝘊𝘳𝘰𝘴𝘴-𝘊𝘶𝘭𝘵𝘶𝘳𝘢𝘭 𝘙𝘦𝘴𝘦𝘢𝘳𝘤𝘩 𝘪𝘯 𝘗𝘳𝘢𝘤𝘵𝘪𝘤𝘦 👇🏻 amazon.com/Cross-Cultural… The volume presents useful timely topics, methods, solutions to 📈 research quality & design across cultures, languages, countries.

@NORCNews .@NORCNews researchers contributed to several chapters in 📖𝘊𝘳𝘰𝘴𝘴-𝘊𝘶𝘭𝘵𝘶𝘳𝘢𝘭 𝘙𝘦𝘴𝘦𝘢𝘳𝘤𝘩 𝘪𝘯 𝘗𝘳𝘢𝘤𝘵𝘪𝘤𝘦. Chapter2️⃣ "Development of a Practical Translation Approach for Survey Research Projects" addresses survey translation issue with real-life solutions.

Poor survey translations don't just reduce response rates. They introduce measurement error and exclude key populations from data that shape policy decisions. NORC researchers offer a practical solution for projects with limited budgets and tight timelines in the new volume Cross-Cultural Research in Practice. Available now: go.norc.org/3Npcm6x

I want to grow up to be beautiful as Ms. Debra (and with those blue 👢)!

How does Category Entry Points (CEP) move brands beyond traditional brand funnels [linear model that maps customer journey from awareness to buy]? @quantilope M.Brandhorst @InsightsMRX "become the basis of brand health KPI"

Brand📈= physical + mental availability, what is Category Entry Points (CEP)? @quantilope MBrandhorst @InsightsMRX 🔎CEP, ask: what makes you buy this product? Generate from consumer qualitative interviews, expert interviews, past research, #GenAI

Human-#LLM partnership in large sample #QualitativeResearch, can we code at scale? @andrewmparker1 @RANDCorporation #CIPHER @USCDornsife Codebook draws on sources 1⃣literature 2⃣closed-ended items 3⃣human scan raw responses 4⃣MuseQDA for coding ➕ check inter-rater reliability

Survey #data contamination... can we predict #LLM generated answers to open narrative Qs? Jan Karem Hohne 1⃣ Very similar to human & meaningful 2⃣ No systematic differences regarding answer length 3⃣ BERT reliably predicts (almost 💯 correct) for @GeminiApp and Gemini+ persons

#AI to code other/specify responses, workflow? Ting Yan @AmeriSpeak @NORCNews #CIPHER @USCDornsife AI code other/specify verbatim ➡️automated rule-based edit ➡️ edited values ready for use - easy 🔎 where problems happen & update procedures - codes/edits at var not case level

#GenAI data donation, who how why from Australia? @ANU_PELab @ourANU #CIPHER @USCDornsife 🥇use @ChatGPTapp For 18-24 yrs old, incentive has significant impact Less likely to consent: females, perceived threat of #AI to democracy Generalized trust ≠ significant association

Does #AI matter in probability panel design? @trentbuskirk @ODU #CIPHER @USCDornsife ✔️ early problem detect, intelligent adapt, reduce burden ✔️ part of the loop to inform, not make decisions 🤔 pre-data collection assist, as interviewer, analyst

silos, low discoverabile/reusabile/inaccessible/irretrievable 🚫standardization, solution for research #data infrastructure #interoperability? @gesis_org @UniCologne A.Katsanidou @LAPOP_Lab 🔎Findable 🔘Accessible 🛞Interoperable: shared🗣️ ♻️Reusable #AI ready: machine workflow

What does the future hold for #DigitalTwins in survey research? @collect_intel E.Hadfield @WAPOR 1⃣ delegate in artificial superintelligence 2⃣ training: intentions hard but preferences easy 3⃣ trust = perfect mimicry 4⃣ volitional turing test 📏autonomous adaptive 🤔goalsetting

Not🆕 #DigitalTwins virtual model sim real-world chars/behaviors counterpart, so what's big deal? @collect_intel E.Hadfield @WAPOR #BigData models ➡️ gen-purpose #AI agents as compute cheapens ⚠️ compounding errors in scaled workflow 🗝️ robust system to eval fidelity/user trust

Human 🆚 #AI #SyntheticData, similar survey responses? @cnchapman Webinar @SawtoothSoft 1⃣ collect/reuse human sample 🫢convenience or leftover sample 2⃣ create #DigitalTwins 🤨prompt template valid? 3⃣ #LLM take survey 🤔 4⃣ compare results 🫠statistical assumptions not apply

How to prioritize + select claims? @Curion_Insights 1⃣ MaxDiff: choose most/least from long list ➡️ ranked order 2⃣ Sequential monadic eval: short list claims across 🗝️metrics ➡️ in-depth diagnostics - used together or just one - via survey Outcome🟰 select the highest-impact!

What's marketing/advertising CLAIM? @Curion_Insights statement re product chars, benefits, performance to 🆙buy, eg lower cholesterol, fat free 🔎 IDI, group, ethno + journey🗺️social👂 ≠ function, warning, puffery (legal term on promotional wording, 🚫fact or meant to mislead)

Check out Chapter 6⃣ in 🆕📖 𝘊𝘳𝘰𝘴𝘴-𝘊𝘶𝘭𝘵𝘶𝘳𝘢𝘭 𝘙𝘦𝘴𝘦𝘢𝘳𝘤𝘩 𝘪𝘯 𝘗𝘳𝘢𝘤𝘵𝘪𝘤𝘦 👇🏻 amazon.com/Cross-Cultural… The volume presents useful and timely topics, methods & solutions to enhance research quality and design across cultures, languages, and countries.