MR Glossary 1

Types of Surveys Opinion Polls: forecast, polling station survey, estimation of election HR: employee surveys (satisfaction, commitment…), management… Corporate: Awareness, E-Reputation… Marketing: U&A, A&U = Attitude & Usage A&A = Attitude & Awareness Concept test = Test of Product ideas CBC, ACA = Conjoint Based Choice, Adaptative Conjoint Analysis Shelf test = Packaging test with product photos in shop shelf Virtual shopping = Simulated shop in 3D Use test = Test of Product usage Sniff test = Test of smell (perfumes) HUT, IHUT = in Home Usage test Post-test = Test of a Advertising campaign (see BrandSpector) Copy test = pre-test of a Ad message Focus group = Qualitative research with 8 to 15 participants Bulletin Board = Quali research without live participation CSM = Customer satisfaction measurement CRM = Client relationship management Mystery Shopping = Measurement of the Quality of Service with fake clients

Main surveys for Market Research Marketing: Marketing Mix / 4 P esp. Audience Market share Media Market (single source TGI) Awareness & image, brand equity U&A, segmentation NPD : concept, conjoint, product, price, packaging Ad Tests : pre & post-test, tracking CRM : satisfaction, loyalty, quality

Qualitative vs. Quantitative Research Qualitative Research Quantitative Research Objective To gain a qualitative understanding of the underlying reasons and motivations To qualify the data and generalize the results from the sample to the population of interest Sample Small number of non-representative cases Large number of representative cases Data collection Unstructured Structured Data Analysis Statistical Non-statistical Outcome Recommend a final course of action Develop an initial understanding

Data collection & Samples CATI / CAPI / CAWi = Computer Assisted Tel Interview / Portable / Web Nat Rep /Census Rep/ Gen Pop = National Representative according to Demo profiles Internet Rep = Internet Users Rep Quali vs Quanti = Text with interaction vs Closed and OEQ questions Quotas vs Random = Rep pattern vs random probability sample Omnibus = around 1000 interviews per week (F2F, Tel, Web) Barometer = same questionnaire for several waves Tracking = same Q-re per week/month, phases with same resp. (continuous) Panel = same Q-re & same respondents Access panel = large collection of profiles for building samples Ad hoc = 1-shot sample & survey vs tracker, omnibus Screener = Survey or filter questions of a specific target Syndicated research = costs shared by several Clients Diary study = same respondents for daily notes of their behaviors Call & recall = same respondents with 2 periods Control & Test = Test vs Control groups with same profiles Longitudinal survey = mid term evolution of same respondents Census = usually exhaustive national survey for Demographics Establishment survey = big survey to measure IR and profiles of a target River panel, RTS = live recruitment of non-panellists

Panel vs. Panel Continuous panel: One recruitment for several Q’res at different dates Consumer panels : AG Nielsen, GFK, TNS… Distribution panels (retail & wholesale) : AG Nielsen, GFK, IRI Special Panel: audience (TV, radio…), IT, healthcare… Access panel: Sampling base for selecting new samples for different surveys. The whole access panel may be not representative but we can build representative samples from it (Nat Rep, Internet…)

Additional Sampling Vocabulary Incidence rate = Target / Whole population = (completes + quotas full) / (completes + quotasfull + screenout) Response rate = respondents / invites = (completes + quotasfull + screenout + incompletes) / invites Open rate = invites open / invites sent Drop rate = incompletes / start survey Quota full = enough completes in the sample cell Quota open = need more completes in the sample cell Hard quota = quota without tolerance (or less than 3 points) Soft quota = quota with higher tolerance or just control Interlocking quota / Nested quota = crossed quotas as gender*age Non-interlocking quota = RIM sample as gender, age, region Screen out = respondent who didn’t fit the Survey target definition Screener = questions or Q’re to filter the Survey target Category exclusion = exclude respondents of a previous similar survey recently Dedupe = exclude duplicate respondents in the sample Boost, booster = extra sample for a subgroup (over-sampling, Augment sample) Category usage = IR of a product among Consumers Feasibility = likelihood to achieve the number of completes asked by Client Incentive = reward for survey respondents (hard vs soft incentives) Census data = demographic profiles from National Census