You have six weeks and a $25,000 cap to study activation with 18 interviews and a survey of 1,200 users across free and paid plans, new and returning users, before deciding whether to add an onboarding checklist; the target metric is 30-day activation. How would you sequence the program?

I would run a sequential exploratory mixed-methods program, using qualitative work to define mechanisms and quantitative work to estimate their reach before evaluating the checklist. - stratify 18 interviews across plan and tenure, translate observed barriers into survey items for 1,200 sampled users, then test the highest-priority checklist flow with a small purposive usability sample. - audit event definitions, compare survey respondents with the sampling frame, require interpretable subgroup bases, and triangulate stated barriers against 30-day activation behavior. - deliver a six-week evidence map, questionnaire, and checklist decision brief with recruitment spend; interviews explain mechanisms and the survey estimates associations, but neither proves the checklist will cause an activation lift.

Product analytics show a 37% onboarding completion rate among 40,000 users; you have four weeks, $18,000, and 20 moderated sessions across mobile and desktop plus first-time and returning users to decide whether to make profile setup progressive. How would you design the research?

I would use a sequential explanatory design: locate the behavioral breakpoints first, then sample sessions to explain them and evaluate progressive setup. - segment the 40,000-user funnel by device and tenure, purposively recruit 20 participants from high-drop and successful paths, and run task-based comparison sessions on the current and progressive flows. - validate event coverage and denominator rules, balance cells without treating 20 sessions as representative, and compare observed task completion, assistance, and time with the 37% baseline. - provide a funnel diagnostic, recruitment matrix, session protocol, and four-week decision memo within the $18,000 cap; the sessions reveal usability mechanisms but cannot forecast the production completion lift.

A SaaS team has five weeks and $35,000 for 24 generative interviews and an 800-person survey across SMB and mid-market customers, split between buyers and daily users, to decide between seat-based and usage-based packaging; checkout conversion and ARPA are the decision metrics. How would you structure the pricing program?

I would separate value discovery from package measurement so the survey tests defensible constructs rather than phrases invented in a workshop. - run 24 role-stratified interviews to map jobs, value units, and budget authority, convert stable findings into mutually understandable package concepts, then field a randomized monadic survey to 800 eligible respondents. - verify buyer and user identity, pilot attribute wording, inspect item nonresponse and segment balance, and report conversion intent and ARPA scenarios separately rather than collapsing them into one score. - produce a value-unit map, tested questionnaire, segment estimates, and five-week package recommendation with the $35,000 cost ledger; stated purchase intent is hypothetical and must not be presented as realized revenue.

You must plan a 12-week activation study with a $30,000 budget and a panel of 120 customers across self-serve and sales-assisted acquisition in North America and Europe, to decide which lifecycle messages to automate; the primary metric is weekly retained activation. What would your longitudinal design include?

I would build a prospective cohort panel linked to behavioral events, with measurement scheduled around meaningful activation milestones rather than arbitrary weekly surveys. - quota 120 recruits across acquisition path and region, collect a baseline, use short event-contingent diaries plus milestone surveys, and invite a purposive subsample for interviews at weeks 2, 6, and 12. - define retained activation before launch, monitor response and behavioral attrition by segment, test whether prompting changes message engagement, and preserve a contact-light comparison subgroup where feasible. - deliver a cohort matrix, cadence calendar, attrition dashboard, and message decision rule within $30,000; the 12-week window supports change patterns but not seasonality or causal claims about automation.

An enterprise product gives you seven weeks and $28,000 for 30 discovery interviews followed by 16 prototype sessions across first-time admins, switchers, and delegated operators, to decide whether to fund a guided setup service; time-to-first-value is the metric. How do you connect generative and evaluative evidence?

I would use discovery to define the setup jobs and failure mechanisms, then evaluate whether the guided service addresses those mechanisms for each role. - allocate 30 contextual interviews by role and prior-system experience, turn the resulting job and dependency hypotheses into a service blueprint, and test critical setup tasks with 16 role-balanced prototype sessions. - keep discovery and evaluation recruitment criteria explicit, test realistic permissions and data dependencies, and record task success, assistance, confidence, and projected time-to-first-value by segment. - produce a hypothesis trace, service blueprint, test protocol, and seven-week funding brief within $28,000; a prototype can establish comprehension and feasibility signals, not production adoption or actual time savings.

You have four weeks, $20,000, 16 interviews, and a 600-person monadic survey across monthly and annual subscribers with and without budget authority to decide which pricing explanation enters checkout; completion rate is the key metric. Which evidence comes first?

I would first identify how customers interpret the price and value unit, then quantify comprehension of finalized explanations in a randomized monadic test. - use 16 interviews across billing cadence and authority to expose vocabulary and misconceptions, cognitively pretest revised copy, then randomize 600 qualified respondents to one explanation each. - confirm respondents recently encountered checkout, test randomization balance, distinguish comprehension from preference, and estimate segment differences with intervals rather than ranking noisy point estimates. - ship an interpretation model, pretest log, analysis plan, and four-week copy recommendation within $20,000; a survey choice is not a checkout behavior effect, so production completion still needs experimental validation.

Eighty thousand new users show a permission-step drop, and you have five weeks, $22,000, and 24 contextual sessions across iOS and Android, privacy-sensitive and convenience-oriented users, to decide whether to remove the gate; D7 activation is the metric. How would you triangulate the evidence?

I would treat telemetry as the prevalence signal and contextual sessions as the mechanism signal, then evaluate a gated alternative without pretending either source alone settles the decision. - verify the 80,000-user funnel by OS, recruit 24 users across privacy orientation and completion outcome, observe permission reasoning in context, and test a delayed-gate concept on the same critical tasks. - inspect permission-event loss by OS version, avoid deriving attitudes from behavior alone, compare successful and dropped users, and track comprehension, task success, and privacy concern separately. - deliver an event audit, mechanism matrix, concept test report, and five-week gate recommendation within $22,000; contextual evidence cannot estimate D7 lift, and historical telemetry cannot isolate the gate's causal effect.

A B2B platform allows eight weeks and $40,000 for 36 role interviews and a survey of 300 accounts across admins, approvers, and end users in small and large organizations, to decide whether admin setup or end-user education gets the next quarter's funding; organization time-to-first-value is the metric. How would you design the program?

I would study the onboarding system at the account level because role dependencies, not isolated screens, determine organization time-to-first-value. - recruit 12 role-linked account triads where possible plus unmatched cases to reach 36 interviews, map handoffs and delays, then survey 300 accounts with one defined respondent role and linked account metadata. - separate individual from account-level inference, verify who can observe each delay, account for organization size and missing roles, and compare reported bottlenecks with setup timestamps. - provide a role-dependency map, account survey, funding scorecard, and eight-week recommendation within $40,000; unmatched interviews and single-informant accounts limit claims about complete organizational workflows.

You own a 12-week activation program with $45,000, 40 interviews, 20 evaluative sessions, and a 1,000-user survey across trial users, new customers, and recently churned customers; the decision is whether to prioritize education or trial redesign, and the metric is 14-day retained use. How would you stage the work?

I would use stage gates so each method answers a distinct uncertainty before the program spends money on the next phase. - begin with behavior-linked interviews across the three lifecycle segments, field a survey only for hypotheses needing prevalence estimates, then evaluate education and trial concepts with 20 task-based sessions. - publish phase-specific questions and stop criteria, keep churned respondents out of tasks they cannot realistically perform, audit survey coverage, and compare concept outcomes against defined retained-use mechanisms. - maintain a 12-week program charter, evidence ledger, phase-gate readouts, and $45,000 burn plan; the program can prioritize a direction but cannot attribute a future 14-day lift before a production test.

In three weeks you must propose a six-month, $60,000 onboarding research program using a 200-person panel across novice and experienced customers in regulated and unregulated industries, to decide the order of three roadmap bets; monthly retained usage is the metric. What plan would you present?

I would propose a longitudinal mixed-methods program with fixed decision points, a refreshment sample, and evidence tied to each roadmap bet. - recruit a 200-person baseline cohort across experience and regulation, combine monthly behavioral linkage with brief milestone surveys, interview rotating subsamples, and add fresh recruits at months 3 and 6 to diagnose panel conditioning. - predefine retention and roadmap criteria, monitor attrition and consent scope by segment, compare continuing and refreshment cohorts, and model repeated observations at the participant level. - present a three-week protocol, six-month cadence, capacity plan, and $60,000 budget with decision briefs at months 2, 4, and 6; the window misses annual cycles and refreshment cohorts do not fully repair nonrandom attrition.

You have 15 working days and $12,000 to recruit 24 participants across high and low product tenure, successful and unsuccessful setup, and three company sizes, to decide which onboarding barrier enters the next study; setup completion is the metric. How would you use purposive sampling?

I would use a maximum-variation purposive sample that exposes contrasting mechanisms, not a miniature representative population. - build a 24-seat matrix across tenure, outcome, and company size, prioritize information-rich intersections, and reserve several seats for cases that contradict the dominant pattern. - verify segment membership from product data where consent permits, track empty cells and source effects, and stop adding similar cases only when new interviews no longer change the decision-relevant explanation. - deliver a sampling matrix, recruitment disposition log, and 15-day barrier brief within $12,000; purposive coverage supports analytic comparison but cannot estimate how common each barrier is.

A six-week discovery has a $18,000 cap and an initial sample of 18 marketplace sellers across new and established, high and low volume, and domestic and cross-border segments; you may add up to 12 participants to decide whether trust or workflow deserves a program, using 90-day retention as context. How would theoretical sampling work?

I would let emerging explanatory categories determine the additional sample while preserving a transparent ceiling and decision purpose. - start with the 18-cell purposive sample, compare incidents around trust and workflow, then spend the 12 reserve seats on cases that refine weak categories, boundaries, and negative cases. - maintain category definitions and memos, distinguish repeated wording from conceptual sufficiency, test rival explanations across segments, and document why each added participant was theoretically useful. - produce a theoretical sampling log, category map, and six-week program recommendation within $18,000; the approach develops an explanation of variation but does not yield retention effect sizes or population prevalence.

A team claims saturation after 12 interviews, but you have one week and $6,000 left to decide whether to recruit eight more across screen-reader users, keyboard-only users, and users without assistive technology; successful account recovery is the metric. How would you assess the claim?

I would assess conceptual sufficiency by segment and research question, not accept 12 as a universal saturation threshold. - review the sampling matrix and codebook for meaning saturation, use the eight available seats to fill thin accessibility contexts and seek disconfirming recovery strategies, then compare what the new cases add. - separate code repetition from depth, inspect whether any segment has only token representation, document new-code and changed-explanation rates, and avoid claiming saturation for populations not sampled. - issue a saturation memo, gap matrix, and one-week recruitment decision within $6,000; even a stable qualitative explanation does not estimate account-recovery success rates.

You need 20 participants with a 4% incidence rate in four weeks for $24,000, split across rural and urban users, new and experienced claimants, to decide whether to fund assisted claims support; claim completion within 30 days is the metric. How would you recruit this hard-to-reach sample?

I would use multiple bounded recruitment frames and budget from expected incidence and yield rather than promise a generic panel sample. - combine consented customer-list outreach, community partners, and a specialist panel, set quotas across geography and experience, and calculate contacts needed from source-specific incidence and response assumptions. - use a short neutral screener, verify claimant status without collecting unnecessary details, monitor source-by-segment yield and completion, and compare recruited cases with the eligible frame where data exist. - maintain a source plan, funnel dashboard, quota exceptions log, and four-week feasibility brief within $24,000; partner and panel recruits may differ from unaffiliated claimants, limiting coverage.

A diary study has eight weeks, $26,000, and 60 participants across caregivers and independent patients, frequent and occasional users, to decide when medication reminders should appear; adherence-event completion is the metric. How would you design the panel?

I would use an event-contingent diary with a low-burden scheduled fallback so entries capture reminder contexts without turning the study into constant surveillance. - quota 60 participants by care arrangement and usage frequency, trigger brief entries around eligible reminder events, collect a weekly reflection, and interview a rotating subsample at weeks 2 and 8. - pilot burden and notification timing, monitor missing events and completion by segment, distinguish no event from no response, and compare diary timestamps with consented product events. - deliver a diary protocol, burden budget, panel-health dashboard, and eight-week reminder brief within $26,000; self-report and prompting can alter adherence, and occasional users provide fewer comparable observations.

A 16-week panel starts with 180 users across free and paid plans and three tenure bands, has a $38,000 budget, and expects 30% attrition before deciding whether to retain a coaching program; weekly active use is the metric. How would you plan for attrition?

I would plan retention and attrition analysis before recruitment, oversampling only from a defensible segment-level yield model. - model expected completes by plan and tenure, recruit 180 with staged reminders and proportional incentives, record withdrawal reasons, and define the minimum effective sample needed for the coaching decision. - compare attriters and continuers on baseline and behavioral variables, report response trajectories by segment, use weighting or sensitivity analysis when assumptions are credible, and avoid silently replacing dropouts. - maintain a retention protocol, attrition dashboard, sensitivity appendix, and 16-week decision brief within $38,000; weighting cannot recover bias from unobserved reasons for leaving.

You have ten weeks and $20,000 for 90 trial users across creators, managers, and viewers, surveyed after every session, to decide whether to keep daily prompts; task completion and repeated-measure bias are concerns. What would you change?

I would reduce and randomize measurement exposure because asking after every session can condition behavior and exhaust the panel. - assign the 90 users within each role to high-frequency, sampled-event, or end-of-week measurement schedules, keep core task-completion telemetry passive, and compare trajectories for evidence of reactivity. - test baseline equivalence, monitor prompt opens and dropout by arm, model observations nested within users, and inspect whether self-reports or product behavior diverge as exposure accumulates. - deliver an exposure schedule, reactivity analysis plan, and ten-week prompting recommendation within $20,000; differing schedules reveal likely measurement effects but do not eliminate all awareness or history effects.

A six-month study has $50,000 and 240 subscribers across monthly and annual plans, low and high engagement, to decide when to offer retention support; monthly satisfaction and renewal are tracked, but repeated surveying may bias both. How would you design the measurement?

I would use a planned-measurement design that varies survey exposure while keeping renewal and behavioral outcomes consistent for all 240 subscribers. - randomly assign subscribers to monthly satisfaction, quarterly satisfaction, or renewal-only contact schedules, while collecting the same consented product events and renewal outcome for every group. - analyze repeated observations with mixed-effects models, preserve assignment in the comparison, track attrition and item missingness, and test whether survey frequency itself changes engagement or renewal. - deliver a six-month exposure calendar, analysis plan, burden dashboard, and $50,000 cost model; the team chooses the least burdensome cadence that still detects decision-relevant change.

Within three weeks and a $9,000 budget, you must sample 18 customers across successful and failed migrations, small and large datasets, to decide whether migration guidance needs a separate program; migration completion is 72%. How would you use negative cases?

I would deliberately sample cases that violate the emerging explanation so the recommendation is built on tested boundaries rather than the majority story. - create an 18-seat matrix by migration outcome and data size, begin with contrasting cases, then reserve seats for successful customers with predicted risks and failed customers without them. - verify outcomes from logs, compare alternative mechanisms consistently, document changes to the explanation after each negative case, and keep the 72% completion rate separate from qualitative inference. - deliver a deviant-case log, boundary conditions map, and three-week program decision within $9,000; negative-case analysis strengthens explanations but does not measure each mechanism's population effect.

You need 30 undocumented gig workers in six weeks for $32,000, split across two languages, platform-dependent and multi-platform workers, to decide whether to fund an anonymous support channel; 90-day support uptake is the metric. Would respondent-driven sampling be appropriate?

I would consider respondent-driven recruitment only as one access route and would not claim population estimates unless its strong network assumptions were defensible. - start with diverse community-partner seeds by language and platform dependence, limit referral chains and coupons, add direct partner outreach, and track recruitment network and source for all 30 participants. - protect identity, detect duplicate referrals without collecting immigration data, inspect seed and chain dependence, compare segments across sources, and test whether participation pressure or gatekeeping appears. - produce a safe referral protocol, network-yield dashboard, and six-week channel recommendation within $32,000; a small nonprobability network sample supports access and mechanism learning, not prevalence or 90-day uptake forecasts.

UX Researcher interview questions

100 real questions with model answers and explanations for Middle candidates.

See a UX Researcher resume example →

Practice with flashcards

Spaced repetition · Hunter Pass

Questions

surveysonboardingmonitoring

I would run a sequential exploratory mixed-methods program, using qualitative work to define mechanisms and quantitative work to estimate their reach before evaluating the checklist.

stratify 18 interviews across plan and tenure, translate observed barriers into survey items for 1,200 sampled users, then test the highest-priority checklist flow with a small purposive usability sample.
audit event definitions, compare survey respondents with the sampling frame, require interpretable subgroup bases, and triangulate stated barriers against 30-day activation behavior.
deliver a six-week evidence map, questionnaire, and checklist decision brief with recruitment spend; interviews explain mechanisms and the survey estimates associations, but neither proves the checklist will cause an activation lift.

Why interviewers ask this: The interviewer is evaluating whether the candidate can sequence generative, quantitative, and evaluative evidence without overstating causal inference.

moderated-researchdesignonboarding

I would use a sequential explanatory design: locate the behavioral breakpoints first, then sample sessions to explain them and evaluate progressive setup.

segment the 40,000-user funnel by device and tenure, purposively recruit 20 participants from high-drop and successful paths, and run task-based comparison sessions on the current and progressive flows.
validate event coverage and denominator rules, balance cells without treating 20 sessions as representative, and compare observed task completion, assistance, and time with the 37% baseline.
provide a funnel diagnostic, recruitment matrix, session protocol, and four-week decision memo within the $18,000 cap; the sessions reveal usability mechanisms but cannot forecast the production completion lift.

Why interviewers ask this: The interviewer is checking whether the candidate starts from reliable behavioral data and uses qualitative evidence for explanation rather than prevalence claims.

generative-interviewssurveysmonitoring

I would separate value discovery from package measurement so the survey tests defensible constructs rather than phrases invented in a workshop.

run 24 role-stratified interviews to map jobs, value units, and budget authority, convert stable findings into mutually understandable package concepts, then field a randomized monadic survey to 800 eligible respondents.
verify buyer and user identity, pilot attribute wording, inspect item nonresponse and segment balance, and report conversion intent and ARPA scenarios separately rather than collapsing them into one score.
produce a value-unit map, tested questionnaire, segment estimates, and five-week package recommendation with the $35,000 cost ledger; stated purchase intent is hypothetical and must not be presented as realized revenue.

Why interviewers ask this: The interviewer is assessing whether the candidate distinguishes generative value research from quantitative package evaluation and handles hypothetical pricing evidence honestly.

activationdesignmonitoring

I would build a prospective cohort panel linked to behavioral events, with measurement scheduled around meaningful activation milestones rather than arbitrary weekly surveys.

quota 120 recruits across acquisition path and region, collect a baseline, use short event-contingent diaries plus milestone surveys, and invite a purposive subsample for interviews at weeks 2, 6, and 12.
define retained activation before launch, monitor response and behavioral attrition by segment, test whether prompting changes message engagement, and preserve a contact-light comparison subgroup where feasible.
deliver a cohort matrix, cadence calendar, attrition dashboard, and message decision rule within $30,000; the 12-week window supports change patterns but not seasonality or causal claims about automation.

Why interviewers ask this: The interviewer is evaluating longitudinal program planning, attrition controls, and the distinction between temporal evidence and causality.

delegationprototypesmonitoring

I would use discovery to define the setup jobs and failure mechanisms, then evaluate whether the guided service addresses those mechanisms for each role.

allocate 30 contextual interviews by role and prior-system experience, turn the resulting job and dependency hypotheses into a service blueprint, and test critical setup tasks with 16 role-balanced prototype sessions.
keep discovery and evaluation recruitment criteria explicit, test realistic permissions and data dependencies, and record task success, assistance, confidence, and projected time-to-first-value by segment.
produce a hypothesis trace, service blueprint, test protocol, and seven-week funding brief within $28,000; a prototype can establish comprehension and feasibility signals, not production adoption or actual time savings.

Why interviewers ask this: The interviewer is checking whether the candidate preserves traceability from generative findings to evaluative criteria without acting as the service designer.

surveysmonitoringpricing

I would first identify how customers interpret the price and value unit, then quantify comprehension of finalized explanations in a randomized monadic test.

use 16 interviews across billing cadence and authority to expose vocabulary and misconceptions, cognitively pretest revised copy, then randomize 600 qualified respondents to one explanation each.
confirm respondents recently encountered checkout, test randomization balance, distinguish comprehension from preference, and estimate segment differences with intervals rather than ranking noisy point estimates.
ship an interpretation model, pretest log, analysis plan, and four-week copy recommendation within $20,000; a survey choice is not a checkout behavior effect, so production completion still needs experimental validation.

Why interviewers ask this: The interviewer is assessing sequencing discipline and whether the candidate separates comprehension evidence from behavioral impact.

triangulationmonitoringsessions

I would treat telemetry as the prevalence signal and contextual sessions as the mechanism signal, then evaluate a gated alternative without pretending either source alone settles the decision.

verify the 80,000-user funnel by OS, recruit 24 users across privacy orientation and completion outcome, observe permission reasoning in context, and test a delayed-gate concept on the same critical tasks.
inspect permission-event loss by OS version, avoid deriving attitudes from behavior alone, compare successful and dropped users, and track comprehension, task success, and privacy concern separately.
deliver an event audit, mechanism matrix, concept test report, and five-week gate recommendation within $22,000; contextual evidence cannot estimate D7 lift, and historical telemetry cannot isolate the gate's causal effect.

Why interviewers ask this: The interviewer is checking triangulation, segment logic, and restraint around causal and prevalence claims.

surveysdesignmonitoring

I would study the onboarding system at the account level because role dependencies, not isolated screens, determine organization time-to-first-value.

recruit 12 role-linked account triads where possible plus unmatched cases to reach 36 interviews, map handoffs and delays, then survey 300 accounts with one defined respondent role and linked account metadata.
separate individual from account-level inference, verify who can observe each delay, account for organization size and missing roles, and compare reported bottlenecks with setup timestamps.
provide a role-dependency map, account survey, funding scorecard, and eight-week recommendation within $40,000; unmatched interviews and single-informant accounts limit claims about complete organizational workflows.

Why interviewers ask this: The interviewer is evaluating whether the candidate chooses the right unit of analysis for multi-role enterprise onboarding.

churnactivationprioritization

I would use stage gates so each method answers a distinct uncertainty before the program spends money on the next phase.

begin with behavior-linked interviews across the three lifecycle segments, field a survey only for hypotheses needing prevalence estimates, then evaluate education and trial concepts with 20 task-based sessions.
publish phase-specific questions and stop criteria, keep churned respondents out of tasks they cannot realistically perform, audit survey coverage, and compare concept outcomes against defined retained-use mechanisms.
maintain a 12-week program charter, evidence ledger, phase-gate readouts, and $45,000 burn plan; the program can prioritize a direction but cannot attribute a future 14-day lift before a production test.

Why interviewers ask this: The interviewer is assessing program ownership, method sequencing, and disciplined use of budget and decision gates.

roadmaponboardingmonitoring

I would propose a longitudinal mixed-methods program with fixed decision points, a refreshment sample, and evidence tied to each roadmap bet.

recruit a 200-person baseline cohort across experience and regulation, combine monthly behavioral linkage with brief milestone surveys, interview rotating subsamples, and add fresh recruits at months 3 and 6 to diagnose panel conditioning.
predefine retention and roadmap criteria, monitor attrition and consent scope by segment, compare continuing and refreshment cohorts, and model repeated observations at the participant level.
present a three-week protocol, six-month cadence, capacity plan, and $60,000 budget with decision briefs at months 2, 4, and 6; the window misses annual cycles and refreshment cohorts do not fully repair nonrandom attrition.

Why interviewers ask this: The interviewer is testing whether the candidate can turn a broad longitudinal request into a governed program with explicit decision cadence and limitations.

research-participantspurposive-samplingonboarding

I would use a maximum-variation purposive sample that exposes contrasting mechanisms, not a miniature representative population.

build a 24-seat matrix across tenure, outcome, and company size, prioritize information-rich intersections, and reserve several seats for cases that contradict the dominant pattern.
verify segment membership from product data where consent permits, track empty cells and source effects, and stop adding similar cases only when new interviews no longer change the decision-relevant explanation.
deliver a sampling matrix, recruitment disposition log, and 15-day barrier brief within $12,000; purposive coverage supports analytic comparison but cannot estimate how common each barrier is.

Why interviewers ask this: The interviewer is evaluating whether the candidate understands maximum-variation sampling and avoids statistical generalization from purposive samples.

research-participantstheoretical-samplingsampling

I would let emerging explanatory categories determine the additional sample while preserving a transparent ceiling and decision purpose.

start with the 18-cell purposive sample, compare incidents around trust and workflow, then spend the 12 reserve seats on cases that refine weak categories, boundaries, and negative cases.
maintain category definitions and memos, distinguish repeated wording from conceptual sufficiency, test rival explanations across segments, and document why each added participant was theoretically useful.
produce a theoretical sampling log, category map, and six-week program recommendation within $18,000; the approach develops an explanation of variation but does not yield retention effect sizes or population prevalence.

Why interviewers ask this: The interviewer is checking whether theoretical sampling follows emerging concepts rather than arbitrary quotas or a fixed saturation number.

saturationmonitoring

I would assess conceptual sufficiency by segment and research question, not accept 12 as a universal saturation threshold.

review the sampling matrix and codebook for meaning saturation, use the eight available seats to fill thin accessibility contexts and seek disconfirming recovery strategies, then compare what the new cases add.
separate code repetition from depth, inspect whether any segment has only token representation, document new-code and changed-explanation rates, and avoid claiming saturation for populations not sampled.
issue a saturation memo, gap matrix, and one-week recruitment decision within $6,000; even a stable qualitative explanation does not estimate account-recovery success rates.

Why interviewers ask this: The interviewer is evaluating methodological precision around saturation and inclusion of underrepresented accessibility contexts.

research-participantsmonitoring

I would use multiple bounded recruitment frames and budget from expected incidence and yield rather than promise a generic panel sample.

combine consented customer-list outreach, community partners, and a specialist panel, set quotas across geography and experience, and calculate contacts needed from source-specific incidence and response assumptions.
use a short neutral screener, verify claimant status without collecting unnecessary details, monitor source-by-segment yield and completion, and compare recruited cases with the eligible frame where data exist.
maintain a source plan, funnel dashboard, quota exceptions log, and four-week feasibility brief within $24,000; partner and panel recruits may differ from unaffiliated claimants, limiting coverage.

Why interviewers ask this: The interviewer is testing practical low-incidence recruitment planning and awareness of frame and source bias.

research-participantsdiary-studiesdesign

I would use an event-contingent diary with a low-burden scheduled fallback so entries capture reminder contexts without turning the study into constant surveillance.

quota 60 participants by care arrangement and usage frequency, trigger brief entries around eligible reminder events, collect a weekly reflection, and interview a rotating subsample at weeks 2 and 8.
pilot burden and notification timing, monitor missing events and completion by segment, distinguish no event from no response, and compare diary timestamps with consented product events.
deliver a diary protocol, burden budget, panel-health dashboard, and eight-week reminder brief within $26,000; self-report and prompting can alter adherence, and occasional users provide fewer comparable observations.

Why interviewers ask this: The interviewer is evaluating diary-method fit, burden controls, and correct handling of event-level missingness.

monitoring

I would plan retention and attrition analysis before recruitment, oversampling only from a defensible segment-level yield model.

model expected completes by plan and tenure, recruit 180 with staged reminders and proportional incentives, record withdrawal reasons, and define the minimum effective sample needed for the coaching decision.
compare attriters and continuers on baseline and behavioral variables, report response trajectories by segment, use weighting or sensitivity analysis when assumptions are credible, and avoid silently replacing dropouts.
maintain a retention protocol, attrition dashboard, sensitivity appendix, and 16-week decision brief within $38,000; weighting cannot recover bias from unobserved reasons for leaving.

Why interviewers ask this: The interviewer is checking prospective attrition planning and whether the candidate understands the limits of statistical adjustment.

surveyssessions

I would reduce and randomize measurement exposure because asking after every session can condition behavior and exhaust the panel.

assign the 90 users within each role to high-frequency, sampled-event, or end-of-week measurement schedules, keep core task-completion telemetry passive, and compare trajectories for evidence of reactivity.
test baseline equivalence, monitor prompt opens and dropout by arm, model observations nested within users, and inspect whether self-reports or product behavior diverge as exposure accumulates.
deliver an exposure schedule, reactivity analysis plan, and ten-week prompting recommendation within $20,000; differing schedules reveal likely measurement effects but do not eliminate all awareness or history effects.

Why interviewers ask this: The interviewer is evaluating whether the candidate recognizes panel conditioning and can design checks for repeated-measure reactivity.

retentionengagementdesign

I would use a planned-measurement design that varies survey exposure while keeping renewal and behavioral outcomes consistent for all 240 subscribers.

randomly assign subscribers to monthly satisfaction, quarterly satisfaction, or renewal-only contact schedules, while collecting the same consented product events and renewal outcome for every group.
analyze repeated observations with mixed-effects models, preserve assignment in the comparison, track attrition and item missingness, and test whether survey frequency itself changes engagement or renewal.
deliver a six-month exposure calendar, analysis plan, burden dashboard, and $50,000 cost model; the team chooses the least burdensome cadence that still detects decision-relevant change.

Why interviewers ask this: The interviewer is checking whether repeated measurement is treated as an intervention with planned exposure, participant-level analysis, and burden controls.

migrations

I would deliberately sample cases that violate the emerging explanation so the recommendation is built on tested boundaries rather than the majority story.

create an 18-seat matrix by migration outcome and data size, begin with contrasting cases, then reserve seats for successful customers with predicted risks and failed customers without them.
verify outcomes from logs, compare alternative mechanisms consistently, document changes to the explanation after each negative case, and keep the 72% completion rate separate from qualitative inference.
deliver a deviant-case log, boundary conditions map, and three-week program decision within $9,000; negative-case analysis strengthens explanations but does not measure each mechanism's population effect.

Why interviewers ask this: The interviewer is evaluating purposeful use of disconfirming evidence and separation of qualitative explanation from prevalence.

respondent-driven-samplingmonitoringsampling

I would consider respondent-driven recruitment only as one access route and would not claim population estimates unless its strong network assumptions were defensible.

start with diverse community-partner seeds by language and platform dependence, limit referral chains and coupons, add direct partner outreach, and track recruitment network and source for all 30 participants.
protect identity, detect duplicate referrals without collecting immigration data, inspect seed and chain dependence, compare segments across sources, and test whether participation pressure or gatekeeping appears.
produce a safe referral protocol, network-yield dashboard, and six-week channel recommendation within $32,000; a small nonprobability network sample supports access and mechanism learning, not prevalence or 90-day uptake forecasts.

Why interviewers ask this: The interviewer is checking whether the candidate understands both the access value and inferential assumptions of respondent-driven methods.

Locked questions

21
You have three weeks and $14,000 to survey 1,000 account owners across SMB and enterprise, active and dormant accounts, to decide whether to retire email reporting; monthly report use is the metric, but the sampling frame contains only opted-in email addresses. How would you handle coverage?
surveyscoveragemonitoring
22
A four-week, $16,000 survey invites 8,000 users and gets 960 responses across free and paid plans, mobile and desktop users, to decide whether to simplify notifications; notification disablement is the metric. How would you assess nonresponse bias?
surveysmonitoringnonresponse-bias
23
In two weeks and for $8,000, you survey 1,067 eligible customers across new and established, basic and premium segments; 42% prefer option A, and the team must decide whether to launch it using a five-point preference margin. How would you report the confidence interval?
confidence-intervalscsssurveys
24
A five-week, $22,000 benchmark compares 300 trial users per arm across novice and experienced segments; onboarding completion rises from 64% to 67% with p = 0.18, and you must decide whether a three-point lift justifies implementation. How would effect size guide the decision?
effect-sizeonboarding
25
You have six weeks and $30,000 to plan a survey experiment across self-serve and assisted customers, with baseline activation at 40% and sample size still open, to decide whether a new education module is worth building if it lifts activation by at least four points. How would you determine the sample instead of choosing a universal number?
surveysexperimentsactivation
26
A three-week, $12,000 survey of 2,400 users across free, pro, and enterprise plans finds a 0.3-point satisfaction difference on a 10-point scale with p = 0.03; you must decide whether to change support policy, and 90-day retention is the business metric. What does the result mean?
surveysmonitoringretention
27
Within four weeks and $18,000, you compare 12 onboarding messages among 1,800 users across mobile and desktop plus two tenure bands, to decide which message advances to production; activation intent is the metric. How would you manage multiple comparisons?
multiple-comparisonsonboardingmonitoring
28
You have five weeks and $20,000 to benchmark SUS with 80 participants per release across novice and expert users on web and mobile, to decide whether a redesign meets the usability gate; the target is a six-point improvement. How would you interpret the scores?
usability-testingresearch-participantssystem-usability-scale
29
A six-week, $24,000 usability benchmark includes 60 participants per segment across first-time and experienced users on desktop and mobile, to decide whether checkout clears an 85% unassisted task-success gate; completion time is secondary. How would you analyze it?
usability-testingresearch-participants
30
You have four weeks and $15,000 to survey 1,500 employees sampled from 50 customer companies across admins and end users, small and large accounts, to decide whether to change permissions training; confidence in permissions is the metric. How would clustering affect the analysis?
surveysmonitoringclustering
31
A team has six weeks and $32,000 for a MaxDiff study with 1,000 respondents across SMB and enterprise buyers plus admins and end users, to choose five of 14 onboarding benefits for packaging; trial conversion is the metric. How would you design it?
maxdiffdesignonboarding
32
You have four weeks and $17,000 to ask 500 consumers across new and loyal, budget and premium segments to rank 25 unrelated app features with MaxDiff, before deciding the next release; monthly feature adoption is the metric. When would you reject this plan?
maxdiffdecision-makingmonitoring
33
A pricing team gives you eight weeks, $45,000, and 1,200 qualified buyers across startups and enterprises in North America and Europe to choose among subscription packages using conjoint; paid conversion and revenue per account are the metrics. How would you define attributes and levels?
conjoint-analysismonitoringpricing
34
You have seven weeks and $40,000 for a choice-based conjoint across three market segments, with 18 candidate attributes, to decide a launch package and estimate share preference; the recruitment vendor promises 300 completes. How would you plan the sample?
participant-recruitmentpromisesestimation
35
A consumer app has five weeks and $25,000 for a Gabor-Granger survey of 800 respondents across current free users and lapsed subscribers, light and heavy users, to choose one of six monthly prices; paid conversion and monthly revenue are the metrics. How would you run and interpret it?
surveysmonitoringgabor-granger
36
You have six weeks and $21,000 for an advanced card sort with 72 participants across new and experienced customers in two regions, to decide the category model for 48 help topics; successful self-service resolution is the metric. What design would you choose?
research-participantscard-sortingdesign
37
A team has four weeks and $16,000 for a tree test with 240 participants across prospects and customers, mobile and desktop users, to choose between two navigation taxonomies; the decision gate is 75% direct success on eight tasks. How would you benchmark them?
research-participantstree-testing
38
You have eight weeks and $36,000 to segment 2,000 users across two countries and free, pro, and team plans, to decide which onboarding program gets localized; 30-day retained activation is the metric. How would you interpret a four-class latent segmentation?
activationonboardingmonitoring
39
In five weeks with $19,000, a team wants conjoint from 400 respondents across individual and team buyers to choose among concepts where security, support, contract length, and price are bundled into three fixed packages; annual conversion is the metric. Is conjoint valid here?
conjoint-analysismonitoring
40
A pricing program has ten weeks, $50,000, 30 buyer interviews, a survey of 1,000 prospects, and purchase data from 15,000 customers across SMB and enterprise, monthly and annual plans, to decide a price increase; conversion and 12-month retention are the metrics. How would you triangulate methods?
triangulationsurveysmonitoring
41
You need 36 cybersecurity administrators in five weeks for $42,000, split across regulated and unregulated companies plus small and large teams, to decide whether to fund an emergency-access workflow; incident resolution time is the metric. How would you build the recruitment plan?
participant-recruitmentincidentsmonitoring
42
A four-week, $18,000 unmoderated study needs 600 completes across new and current customers in three countries to decide whether identity verification is understandable; pass rate is the metric, but duplicate and fraudulent participants could distort it. What controls would you use?
unmoderated-researchresearch-participantsmonitoring
43
You inherit a 3,000-member panel and have six weeks plus $20,000 to set governance across free and paid customers, consumers and business buyers, before deciding whether it can support a quarterly 150-person sample; response rate and 12-month panel health are the metrics. What rules would you establish?
ownershipmonitoring
44
A domain team has $30,000 per quarter and needs recruitment SLAs for eight studies totaling 96 participants across general consumers, enterprise buyers, and a 3% incidence specialist segment; you must decide which studies can commit to two-week or six-week deadlines, and fill rate is the metric. How would you set the SLA?
research-participantsparticipant-recruitmentmonitoring
45
You have three weeks and $15,000 to recruit 48 participants across hourly workers and salaried managers in two countries for 30- and 90-minute sessions, to decide whether scheduling support should differ by segment; completion rate is the metric. How would you set fair incentives?
research-participantsparticipant-incentivesjobs
46
A six-week, $28,000 longitudinal study follows 100 users across EU and US, adults and 16-to-17-year-olds, to decide whether to retain voice diaries; weekly diary completion is the metric, and clips may enter a repository for future research. How would you design consent?
informed-consentdesignmonitoring
47
You have five weeks and $35,000 for 30 interviews with recent fraud victims across high and low financial loss and customers with and without account recovery, to decide whether to fund a trauma-informed support study; recovery completion within 60 days is the metric. What safeguards are required?
monitoring
48
A health product allows eight weeks and $44,000 for a diary panel of 72 participants across patients with fluctuating symptoms and caregivers, rural and urban residents, to decide whether daily symptom prompts are acceptable; 30-day retention and diary burden are the metrics. How would you protect participants?
research-participantsmonitoringretention
49
A domain has two researchers and one shared recruiter, $120,000 for the next six months, and requests 14 studies totaling 220 participants across consumer, enterprise, and low-incidence expert segments; you must decide which portfolio can support quarterly roadmap decisions, and median time-to-insight is the operating metric. How would you plan ResearchOps capacity?
monitoringresearchopscapacity
50
You must decide in four weeks how to allocate a $70,000 annual recruitment budget across a 1,500-member customer panel and external vendors, serving free and paid users plus administrators in three regions; the target is 85% on-time fill and no participant contacted more than once per month. What governance model would you propose?
research-participantsparticipant-recruitmentprocurement
51
Two researchers are synthesizing 18 onboarding interviews, but their labels for trust and confusion keep overlapping. How would you build a usable codebook without flattening new evidence?
qualitative-codingonboarding
52
You have three days to compare 12 enterprise interviews across administrators, buyers, security reviewers, and daily users. How would you use framework analysis to avoid a generic summary?
framework-analysisgenerics
53
You previously designed the onboarding flow you are now studying, and six participants criticize decisions you advocated. How would you manage reflexivity during synthesis?
research-participantsdesignonboarding
54
Two coders disagree on 14 of 30 excerpts from a pricing study, and the PM asks whether the agreement percentage is acceptable. What would you do?
pricingconflict
55
Nine of 12 trial users say setup guidance is clear, but the other three all abandon at identity verification. How would you handle these negative cases in synthesis?
56
A concept test includes eight new small-business admins and eight experienced enterprise admins, and their reactions point in opposite directions. How would you compare the segments?
react
57
Ten interviews suggest users cannot find the export control, but the existing event funnel shows no drop at that step. How would you triangulate the conflict?
triangulationfunnel
58
A sentiment model has labeled 6,000 open-text cancellation responses, and stakeholders want the labels in tomorrow's churn readout. How would you validate the model for research use?
churncommunicationvalidation
59
A team wants synthetic participants to evaluate an AI writing assistant before spending two weeks recruiting 15 real users. What evaluation rubric would you require?
research-participantsdecision-making
60
A bilingual study has 20 English and 20 Russian interviews, but translated transcripts collapse two distinct trust concepts into one English label. How would you repair the synthesis?
distinct
61
In a 12-session prototype comparison, every mobile participant saw concept A and every desktop participant saw concept B. You discover the confound after session eight. What do you do?
research-participantsprototypessessions
62
A vendor recruits 20 participants for a retention study, but 17 are high-NPS customers sourced from the advocacy community. The study starts tomorrow. How would you respond?
research-participantsretentionprocurement
63
Halfway through a 240-person concept survey, you learn the survey tool assigned 80% of respondents to concept A instead of randomizing evenly. What would you do?
surveys
64
A two-week diary study begins with 24 participants, but only 11 complete it and almost all dropouts are first-time users. How would you handle the attrition?
research-participantsdiary-studies
65
A pricing survey has collected 600 responses when you discover that a marketing email sent that morning revealed the proposed price and described it as a bargain. What do you do?
surveyspricing
66
During six observed privacy-setting tasks, users behave confidently, but screen recordings show they reverse the settings within an hour after the moderator leaves. How would you address the observer effect?
67
After five checkout tests, participants praise the new flow, but the prototype skipped payment authorization and loaded every screen instantly. Can the study support a launch decision?
research-participantsauthprototypes
68
A researcher accidentally uploads ten unredacted interview videos containing account numbers to a repository visible to 60 employees. What actions do you take?
69
Your usability plan says to stop if more than two of the first five participants cannot enter the seeded account, and the third failure occurs in session four. The PM wants to continue. What do you do?
usability-testingresearch-participantssessions
70
A pilot for a 30-person MaxDiff study shows that seven of eight participants interpret two attributes as the same benefit. The launch deadline is in ten days. How would you redesign?
research-participantsestimationmaxdiff
71
After eight moderated onboarding sessions, a PM asks you to prove that concept B is statistically significant so it can win the design review. How do you respond?
moderated-researchdesignonboarding
72
A product leader rejects findings from 14 small-business interviews because one large customer told them the opposite yesterday. How would you handle the disagreement?
conflict
73
Twelve cancellation interviews blame missing collaboration controls, while the analyst reports that accounts using those controls churn more often. How would you reconcile the evidence?
churnresilience
74
The roadmap locks in five working days, but recruiting the full 18-person sample for a new admin workflow takes three weeks. What trade-off would you propose?
roadmap
75
A director selects three positive quotes from 16 interviews for a launch deck and omits five severe trust concerns. What do you do before the meeting?
76
A PM has already chosen a dashboard design and asks for five user sessions next week to validate it for an executive demo. How would you avoid research theater?
validationsessionsdesign
77
Recruiting 20 low-incidence compliance officers takes four weeks, but the team can reach eight verified officers in ten days. How would you decide between sample size and cycle time?
sampling
78
A PM wants a rank order of four customer segments from 12 exploratory interviews so the roadmap can target the biggest opportunity. What would you say?
roadmapexploratory
79
A sales leader says three strategic accounts dislike your proposed workflow, while 15 recruited operations users completed it successfully. How would you address the skepticism?
80
A roadmap committee asks you to state that poor reporting causes churn because seven exit interviewees mentioned it. How would you frame the evidence?
churnroadmap
81
A Dovetail repository has 430 tags, including near-duplicates such as trust, trusted, and credibility, so researchers cannot compare studies. How would you redesign the taxonomy?
dedup
82
Your 600-person research panel has a 38% no-show rate and the same 45 members appear in most studies. How would you restore panel health?
83
A participant withdraws consent and asks for deletion after their clips have been copied into three Dovetail projects and a quarterly deck. How would you fulfill the request?
research-participantsinformed-consent
84
Product teams expect ResearchOps to recruit any audience in five days, but specialist enterprise studies routinely take 18 days. How would you define a credible recruitment SLA?
participant-recruitmentresearchops
85
Six product squads submit research requests in different formats, and one-third reach kickoff without a decision or target audience. How would you standardize briefs?
initiation
86
A recruitment vendor delivers 12 participants, but four fail identity checks and three repeat the same scripted answers. How would you manage vendor quality?
research-participantsparticipant-recruitmentprocurement
87
Eight PMs want to run their own customer interviews every week because the two researchers cannot cover all squads. What guardrails would you implement?
guardrailsdiscovery
88
A new insights repository launches, but after two months only three of 11 squads search it before requesting studies. How would you improve adoption?
decision-making
89
Panel members report receiving four invitations in one week, while rare accessibility participants receive none for months. How would you govern contact and coverage?
research-participantscoveragea11y
90
The repository contains 900 findings, but teams keep reusing three-year-old evidence after the product and market changed. How would you prevent stale insight adoption?
decision-making
91
Six onboarding interviews reveal that first-time users misread the workspace invitation, and the PM changes the copy before launch. How would you evaluate influence on activation?
onboardingdecision-makingactivation
92
Twelve churn interviews identify a permissions workaround, and the team ships a fix alongside a new annual discount. How would you assess whether the insight affected churn decisions?
churn
93
A pricing study with 16 interviews and a 400-person survey leads the team to remove an unpopular usage cap but keep the price unchanged. How would you record the research impact?
surveyspricing
94
Three months after a study, stakeholders disagree about whether its findings changed the roadmap. What would a useful decision log have captured?
conflictcommunicationroadmap
95
A navigation redesign based on ten usability sessions launches next month. Which follow-up measures would you define before release?
usability-testingsessions
96
A quarterly review shows that only four of 22 research findings led to a documented product action. How would you investigate low insight adoption?
decision-making
97
A team requests a six-week study on a feature whose build is already contracted and cannot change for nine months. Would you run it?
98
A junior researcher is leading their first 12-participant diary study, and the pilot entries are vague and burdensome. How would you mentor them without taking over?
research-participantsdiary-studiesmentoring
99
An activation metric rises 6% after a release that includes four product changes, one of which came from your research. How would you discuss attribution with leadership?
activationmonitoring
100
A six-month activation research program produced five studies, but the roadmap adopted only two recommendations and deliberately rejected three. How would you evaluate the program's influence?
decision-makingactivationroadmap

Questions

1You have six weeks and a $25,000 cap to study activation with 18 interviews and a survey of 1,200 users across free and paid plans, new and returning users, before deciding whether to add an onboarding checklist; the target metric is 30-day activation. How would you sequence the program?

2Product analytics show a 37% onboarding completion rate among 40,000 users; you have four weeks, $18,000, and 20 moderated sessions across mobile and desktop plus first-time and returning users to decide whether to make profile setup progressive. How would you design the research?

6You have four weeks, $20,000, 16 interviews, and a 600-person monadic survey across monthly and annual subscribers with and without budget authority to decide which pricing explanation enters checkout; completion rate is the key metric. Which evidence comes first?

7Eighty thousand new users show a permission-step drop, and you have five weeks, $22,000, and 24 contextual sessions across iOS and Android, privacy-sensitive and convenience-oriented users, to decide whether to remove the gate; D7 activation is the metric. How would you triangulate the evidence?

10In three weeks you must propose a six-month, $60,000 onboarding research program using a 200-person panel across novice and experienced customers in regulated and unregulated industries, to decide the order of three roadmap bets; monthly retained usage is the metric. What plan would you present?

11You have 15 working days and $12,000 to recruit 24 participants across high and low product tenure, successful and unsuccessful setup, and three company sizes, to decide which onboarding barrier enters the next study; setup completion is the metric. How would you use purposive sampling?

13A team claims saturation after 12 interviews, but you have one week and $6,000 left to decide whether to recruit eight more across screen-reader users, keyboard-only users, and users without assistive technology; successful account recovery is the metric. How would you assess the claim?

14You need 20 participants with a 4% incidence rate in four weeks for $24,000, split across rural and urban users, new and experienced claimants, to decide whether to fund assisted claims support; claim completion within 30 days is the metric. How would you recruit this hard-to-reach sample?

15A diary study has eight weeks, $26,000, and 60 participants across caregivers and independent patients, frequent and occasional users, to decide when medication reminders should appear; adherence-event completion is the metric. How would you design the panel?

16A 16-week panel starts with 180 users across free and paid plans and three tenure bands, has a $38,000 budget, and expects 30% attrition before deciding whether to retain a coaching program; weekly active use is the metric. How would you plan for attrition?

17You have ten weeks and $20,000 for 90 trial users across creators, managers, and viewers, surveyed after every session, to decide whether to keep daily prompts; task completion and repeated-measure bias are concerns. What would you change?

18A six-month study has $50,000 and 240 subscribers across monthly and annual plans, low and high engagement, to decide when to offer retention support; monthly satisfaction and renewal are tracked, but repeated surveying may bias both. How would you design the measurement?

19Within three weeks and a $9,000 budget, you must sample 18 customers across successful and failed migrations, small and large datasets, to decide whether migration guidance needs a separate program; migration completion is 72%. How would you use negative cases?

20You need 30 undocumented gig workers in six weeks for $32,000, split across two languages, platform-dependent and multi-platform workers, to decide whether to fund an anonymous support channel; 90-day support uptake is the metric. Would respondent-driven sampling be appropriate?

Locked questions