Consultancy Research Marketing PDF

lOMoARcPSD|35253986 Consultancy Research Marketing Marketing Research Project (HEC Montréal) Scanne pour ouvrir sur Studocu Studocu n'est pas sponsorisé ou supporté par une université ou un lycée Téléchargé par Olivia Jalby ([email protected]) ...

lOMoARcPSD|35253986 Consultancy Research Marketing Marketing Research Project (HEC Montréal) Scanne pour ouvrir sur Studocu Studocu n'est pas sponsorisé ou supporté par une université ou un lycée Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 Consultancy Research Marketing Class 1 : Introduction to Marketing Research Proposal Marke琀椀ng research helps managers 1. Specify what informa琀椀on is needed 2. Collect and analyze that informa琀椀on 3. Interpret the informa琀椀on with respect to the objec琀椀ves of the study Help managers make be琀琀er decisions 1. Provides a compe琀椀琀椀ve edge in compe琀椀琀椀ve markets 2. Enhances pro昀椀tability 3. Translates data into useful informa琀椀on 4. Provides insights on consumers (exis琀椀ng and poten琀椀al) Marke琀椀ng Research can be used to assist all aspects of marke琀椀ng 1. Product (concept tes琀椀ng for new products, monitor brand equity, assess test markets, select package designs) 2. Pricing (price elas琀椀city, market demand, sales forecas琀椀ng) 3. Distribu琀椀on (web site tes琀椀ng, channel performance, coverage, convenience) 4. Promo琀椀on (media and ad copy e昀昀ec琀椀veness, sales force compensa琀椀on and territories) 5. Buyer behavior (segmenta琀椀on, awareness, sa琀椀sfac琀椀on, preferences, purchase behavior, online "liking" propensi琀椀es) 6. General corporate research (market trends, diversi昀椀ca琀椀on opportuni琀椀es, percep琀椀ons about compe琀椀tor posi琀椀oning). Ques琀椀ons Marke琀椀ng Research Can Address - What features of the product do consumers value most? - Are customers sa琀椀s昀椀ed with our products? - How does the public perceive our company? - What kinds of people buy our products? Two Complementary Approaches to Marke琀椀ng Research 1. Projects executed to address speci昀椀c problems (ad hoc) Emphasis of this class A snapshot in 琀椀me Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 2. Ongoing Marke琀椀ng Intelligence systems (MIS) Aka Decision Support System (DSS) Con琀椀nuous monitoring of consumer and market behavior Core of customer rela琀椀onships management (CRM) e昀昀orts This is BIG DATA Marke琀椀ng Research Process 1. Formulate the Problem 2. Determine the Research Design 3. Design Data Collec琀椀on 4. Design the Sample and Collect Data 5. Analyze and Interpret the Data 6. Prepare the Research Report Formulate the problem It's only when the problem or opportunity has been clearly de昀椀ned and the objec琀椀ves of the research stated precisely that research can be designed properly. 3 sources of marke琀椀ng research problems/opportuni琀椀es 1. Unan琀椀cipated 2. Planned 3. Serendipity Marke琀椀ng Problem (Opportunity) Example: St-Hubert wants to enter the vegetarian market Turn the marke琀椀ng problem into a research problem In prac琀椀ce, the research problem is formulated into objec琀椀ves Measure the interest of St-Hubert among vegetarians Iden琀椀fy the needs, expecta琀椀ons, and percep琀椀ons of vegetarian op琀椀ons for the menu Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 Test the products that resonate with the target market to de昀椀ne a preliminary o昀昀er Marke琀椀ng Problem (3 Sources) turned into research problem CIBC PLANNED: Rise in Bitcoin, introduc琀椀on of Bitcoin investment accounts Research Problem: Are Canadians interested in opening a Bitcoin investment account? Canada’s Wonderland SERENDIPITY: Development by the R&D department of a new augmented reality ride Research Problem: Are consumers interested in the new augmented reality ride? Montréal Canadiens UN-ANTICIPATED: Increase in the number of immigrants unaware/un-interested in Hockey Research Problem: How can we increase immigrants’ awareness of hockey? Convert the Marke琀椀ng Problem into a research problem Determine the research design - How much do we already know about the problem? - Determine the research design o Exploratory o Descrip琀椀ve o Causal Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 - Determine the appropriate methodology o Qualita琀椀ve o Survey / Quan琀椀ta琀椀ve o Observa琀椀on o Experiment Design Data Collection - Depending on the methodology chosen, which data collec琀椀on to use? o Open ended? Structured? Unstructured? o Survey? Ques琀椀onnaire? Focus group? o Online? in-person? In the mall? Design the Sample and collect data - Sampling frame and selec琀椀on process o List of popula琀椀on elements from which the sample will be drawn - Who will we ask? o What is the target popula琀椀on? o Think carefully of who the marketers are targe琀椀ng - Sample size o How large should the sample be? - Once the sample design is speci昀椀ed, data collec琀椀on can begin. Analyze and interpret the data - Need to analyze the data to capture 昀椀ndings to address the problem - How will the data the coded? - What analysis techniques will be used? - Did we answer the objec琀椀ves? - Do we need more data? - What went wrong? Prepare the research report - Submi琀琀ed to management that summarizes the research results, with conclusions and recommenda琀椀ons Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 - Who will read the report? Know your audience - The research report needs to be clear, accurate, and concise Class 2: Research Design and Data Types Two Types of data Primary Secondary Original data created Data collected that by the researcher already exists Primary Data The types of primary data to be collected are nearly endless. Good for awareness and a琀 tudinal research Demographics & Socioeconomic: age, educa琀椀on, occupa琀椀on, marital, status, gender, income, ethnicity, or social class. Psychographic & Lifestyle: personality traits, ac琀椀vi琀椀es, interests, and values A琀 tudes & Opinions: person's ideas, convic琀椀ons, or liking Awareness & Knowledge: ex brand awareness, their exper琀椀se Mo琀椀va琀椀on: involves determining why people behave as they do Inten琀椀ons: an琀椀cipated or planned future behavior Behavior: have done or are doing (ex: purchase and usage behavior) Collec琀椀on technique o Communica琀椀on: Ask o Observa琀椀on: Watch Secondary Data Informa琀椀on that was previously gathered for some other purpose than the current research project. Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 Faster less expensive to acquire than primary data Provides insights into: Clarify, guide or rede昀椀ne the research to be conducted Answers some or all the organiza琀椀ons research needs Assist in the research design Interpret primary data with more insights Compara琀椀ve data to check reliability of data gathered via primary research Provide informa琀椀on that cannot be collected Good for market size, market condi琀椀ons and buying behaviors, etc Limita琀椀ons: Not designed speci昀椀cally to meet the researcher’s needs Availability, the informa琀椀on you need is not available Applicability “data 昀椀t problem” where the data is not in units or measurements to 昀椀t the research objec琀椀ves or in the current 琀椀me frame – could conduct data conversion which provides es琀椀mates Accuracy, research could be biased – always seek out the original source and assess the reputa琀椀on of the organiza琀椀on that conducted the research (who paid for the research?) Comparability is a problem when integra琀椀ng and examining from di昀昀erent data sources Secondary data – sources Internal External Sourced from within the organiza琀椀on Sourced from outside the organiza琀椀on requiring the research requiring the research Always worth checking internal data 昀椀rst Newspapers, trade associa琀椀ons / but it might not be usable or reliable journals, published reports, directories, governments, country info consultants, research 昀椀rms etc Examine 昀椀rst before conduc琀椀ng primary Google or online – be very clear on the research keywords used Customer databases, CRM, Loyalty Cards, Could be free or with a fee Sales Data etc Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 Examples of Secondary Data -External Examples: Census Data Exploratory research It’s the preliminary research – when rela琀椀vely li琀琀le is known about the phenomenon to be inves琀椀gated or when more informa琀椀on is needed The most 昀氀exible for the 3 research designs When to use: Formulate problems more precisely Develop Hypotheses Establish priori琀椀es for research Eliminate imprac琀椀cal ideas Clarify concepts Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 Types – primarily qualita琀椀ve research Literature search Focus groups In-depth individual interviews Projec琀椀ve techniques Interview Focus Group More sensi琀椀ve or personal topics Exchange between par琀椀cipants When you want a complete view of Facilitates idea琀椀on and new ideas that a person's experience researchers didn’t think of Allows explora琀椀on of unintended We o昀琀en group people with similar characteris琀椀cs themes together to avoid confronta琀椀ons of ideas that prevent us from moving forward. Between 15 and 30 interviews, depending on the subject Between 4 and 8 groups of 8 par琀椀cipants Descriptive Research A good descrip琀椀ve study presupposes much prior knowledge about the phenomenon studied It is more rigid than exploratory, problem is clear When to use: WHO? Describe characteris琀椀cs of certain groups / segments WHAT? Es琀椀mate propor琀椀on of people who behave in a certain way WHEN? Make speci昀椀c predic琀椀ons HOW? To ensure that you’ll be ge琀 ng the right data, you should specify in advance the objec琀椀ve each ques琀椀on addresses, the reason it's included, and the analysis in which it will be used. Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 Types: - Longitudinal study: measured repeatedly used for tracking, trends. Trying to understand the meanings of the 昀氀uctua琀椀ons of the data over 琀椀me. The panels uses are: o True: same sample/ques琀椀ons o Omnibus: same sample / di昀昀erent ques琀椀ons - Cross-sec琀椀onal study: one-琀椀me, sample survey Causal Research We use this research method if we want to see the impact of one variable on another Provide evidence regarding causal rela琀椀onships, problem is clear When to use: Wish to answer the ques琀椀on, "Does X cause Y?“ Example: Price elas琀椀city: Will a 5% increase in price hurt customer demand Types: Laboratory experiment Field experiment Rela琀椀onship Among Research Design Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 Class 3: Qualitative Research (Planning and Collecting) Qualitative Research Qualita琀椀ve research is a type of research that aims to gather and analyze non-numerical data in order to gain an understanding of individuals' social reality, including understanding their a琀 tudes, beliefs, and mo琀椀va琀椀on. The informa琀椀on is obtained from small samples but is rich and detailed The ques琀椀oning is less structured Guides the research Con昀椀rms and enriches hypotheses 4 contexts: 1. unique research methodology 2. 昀椀rst step before quan琀椀ta琀椀ve research 3. con昀椀rm results of quan琀椀ta琀椀ve research 4. further explore the results of quan琀椀ta琀椀ve research Types: - Focus groups - In-depth individual interviews - Projec琀椀ve techniques - Observa琀椀on and ethnographic studies Bene昀椀ts : - 80% of communica琀椀on is nonverbal - Iden琀椀昀椀es emo琀椀ons - Iden琀椀昀椀es barriers - Extracts what consumers cannot verbalize - Extract what people might not want to reveal Limits : - The quality of the data depends on the ability of the par琀椀cipants to express their ideas and the ability of the moderator to extract informa琀椀on and manage the group - Cannot generalize results - Analysis is complex and involves interpreta琀椀on - Time and cost Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 Qualita琀椀ve research deals with words and meanings, While Quan琀椀ta琀椀ve research deals with numbers and sta琀椀s琀椀cs Process of exploratory research 1) Formulate the objec琀椀ves 2) Choose the op琀椀mum method 3) Create a sample structure 4) Prepare the research instrument 5) Choose the moderator and prepare the logis琀椀cs 6) Conduct the exploratory phase 7) Analyze the results 1) Formulate the objec琀椀ves a. What are the objec琀椀ves of the exploratory phase? i. What informa琀椀on should we collect from this research? ii. How can this help the company in solving its marke琀椀ng problem? Step 1: Objec琀椀ves of the exploratory phase Understand Canadians perspec琀椀ve on Sports Understand the needs and desires of Canadians for sneakers Understand Canadians assessment of current market o昀昀er (likes/dislikes/unmet needs) Gain insights about the Canadians assessment of the 4 Ps of Alyssa Felix's sneakers Guide the development of the survey for the descrip琀椀ve research Step 2: Choose the Op琀椀mum Exploratory research method - In-Depth interviews o One on one conversa琀椀on o Guided by an interviewer o In-depth understanding of a topic, poten琀椀ally a sensi琀椀ve topic 1) In-depth understanding of a topic 2) One on one interviews, 1 hour + (45 min for your project) 3) Interviewer ideally is the researcher 4) Familiarize with the topic prior Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 5) Prepare interview protocol 6) Record with permission 7) Interview is all about the informant 8) No right or wrong answers 9) Ac琀椀ve listening, taking notes 10) Transcribe as soon as possible o Guidelines 1) Funnel ques琀椀ons in a sequence from general to speci昀椀c a. Start by describing the topic, expecta琀椀ons of interview, dura琀椀on b. Start with grand tour ques琀椀ons/background / general ques琀椀ons 2) Ask “how” ques琀椀ons rather than “why” 3) Do not ask yes / no ques琀椀ons 4) Do not ask leading ques琀椀ons 5) Use probes strategically to get more elabora琀椀on 6) Try to circle back to earlier topics for greater depth is appropriate 7) Be willing to explore tangen琀椀al topics that the interviewer brings up but use good judgment to stay on topics o A few 琀椀ps 1) Listen and do not interrupt the par琀椀cipant 2) Do not be afraid of silence 3) Avoid making judgments about what is said 4) Conduct the interview in a neutral loca琀椀on 5) Be con昀椀dent and trustworthy 6) Make the par琀椀cipant feel comfortable 7) Be curious - Focus Group o Small group interac琀椀on o Guided discussion by moderator o Insights into the sources of complex behaviors and mo琀椀va琀椀on 1) Research technique that collects data through group interac琀椀on 2) Seeks to gain insights through interac琀椀ons 3) Ac琀椀ve role of the moderator 4) Popular qualita琀椀ve method, especially in business 5) Tend to be quick, easy and inexpensive 6) Be careful about group think – each member to express themselves 7) Quality of ideas – discussion in focus groups greater than the sum of separate individual interviews – interac琀椀on is a unique strength Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 8) Element of homogeneity of the group o Focus group par琀椀cipants and dynamics 1) Face to Face focus groups: be琀琀er to be homogeneous within the group (age, gender etc.). 2) Strangers are be琀琀er recruits than friends 3) Moderator needs to provide a safe environment for par琀椀cipants to express themselves 4) Moderator to be mindful that certain par琀椀cipants may dominate, need to encourage others to par琀椀cipate 5) Moderator to be mindful of group think – can speci昀椀cally ask for other opinions or point of view 6) Request to record (video and audio) and transcribe as soon as possible o How to prepare 1) Plan the mee琀椀ng me琀椀culously (guide) 2) Ensure to have a 昀氀ow between themes that makes sense 3) Each theme should be “broad” to encourage spontaneity 4) Probe, dig deeper when you no琀椀ce added value 5) Bring back to the subject if the conversa琀椀on diverges 6) Readjust ques琀椀ons as needed based on previous comments 7) Schedule a break 8) E昀昀ec琀椀vely recruit par琀椀cipants 9) Use all the means to record the maximum of data related to the interviews. o Tips to Succeed The moderator  Strong interpersonal skills  Respec琀昀ul  Dynamic  Curious  Open-minded  Ability to break the ice and build trust  Able to listen, prepare and adjust  Extract value added insights During the focus group  Ques琀椀on design: logical order and easy to follow Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986  Cover the topic in depth  E昀케cient alloca琀椀on of speaking 琀椀me between par琀椀cipants Recruitment of par琀椀cipants: Group homogeneity – relevant depending on research objec琀椀ve  Gender  Age  Ethnicity  Income  Educa琀椀on  Etc. Loca琀椀on  Ensure par琀椀cipants are comfortable  Ability to record ALL par琀椀cipant reac琀椀ons Test  Test the discussion guide prior to session Limita琀椀ons of In-Depth Interviews and Focus group 1) Process issues (par琀椀cipants/moderator) 2) Not generalizable to the larger popula琀椀on, small samples 3) Analysis is challenging and can be biased 4) Time consuming to organize / schedule 5) Limited interac琀椀ons/lack of diverse perspec琀椀ves (in-depth interviews) Step 3: Choose the Op琀椀mum Exploratory research method - Who are we interviewing? Homogeneity versus heterogeneity How many par琀椀cipants per group? Face-to-face? Online? - How many people are we interviewing? How many focus groups are we conduc琀椀ng? How many people in each group? How many in-depth/ individual interviews are we conduc琀椀ng? Focus Group Sample Protocol: Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 1) Select target: Purposive sample selected for suitability and ability to provide insights. Design a recrui琀椀ng protocol to screen poten琀椀al par琀椀cipants 2) Determine the number of groups: 4-12 depends on research problem 3) Determine the # of par琀椀cipants in each group: Physical (8-12) Online (5) 4) Decide on the length of a session: Generally, 1 ½ hours to 2 hours 5) Select the site: A loca琀椀on that is neutral, comfortable and quiet. Convenient loca琀椀on and easy to access 6) Decide on the 琀椀ming of session: Depends on target sample 7) Recruit the par琀椀cipants: Generally, a survey to match the selected target Step 4: Preparing the research instrument 1) Start with the sec琀椀ons/ headings (10-12 topics) 2) Start with the most general, to more speci昀椀c, end with a general sec琀椀on 3) Write the ques琀椀ons (general to speci昀椀c) within each sec琀椀on 4) Determine the dura琀椀on of each sec琀椀on/ques琀椀on 5) Don’t ask yes/no ques琀椀ons 6) Don’t ask why. Ask how come. 7) Be Flexible 8) Word ques琀椀ons carefully, open-ended, no leading ques琀椀ons 9) Pre-test the instrument Step 5: choosing the moderator and preparing the logis琀椀cs 1) Moderator is instrumental in the success of the focus group 2) Seen as a catalyst for discussion, not an interviewer 3) Choose the moderator carefully 4) Possess a number of skills, especially strong communica琀椀on, genuine interest in others, the ability to express their own feelings, to be animated and spontaneous, have a sense of humor, empathy, ability to express thoughts clearly and to be 昀氀exible 5) Ability to manage opinions and group dynamics, di昀昀erent personali琀椀es – extract the most insights Step 6: conduc琀椀ng the exploratory phase More 琀椀ps for a successful Focus Group! Choose your moderator WITH CARE! Keep the discussion guide simple Start by bonding with par琀椀cipants The client must be invisible and silent Remember, this is a quick snapshot to recommend ac琀椀ons; the results are not 昀椀nal. Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 Be curious and open mind Projective Techniques - Projec琀椀ve methods help informants say things indirectly that are di昀케cult to say directly - Techniques to get the par琀椀cipant to project their feelings, emo琀椀ons, beliefs, concep琀椀ons, etc. through various s琀椀muli - For sensi琀椀ve topics or those with socially desirable responses, projec琀椀ve methods are less threatening and less apparently self-revealing. - Concepts, impressions, images or emo琀椀ons that are di昀케cult to ar琀椀culate; social or psychological blockages. - The s琀椀muli act as a “medium” between the par琀椀cipant and the researcher Types of projec琀椀ve techniques 1) Word Associa琀椀on a. Give par琀椀cipants a series of unrelated words one at a 琀椀me and have them respond with the 昀椀rst thing that comes to mind 2) Sentence Comple琀椀on a. Respondents are given incomplete sentences and asked to complete the thought 3) Cartoon Tests a. Present a simple drawing of a situa琀椀on and ask par琀椀cipants to describe what they think is going on 4) Symbol Matching a. Get a quick visual metaphor for a person, brand, b. Example: If HEC were one of these cars, which one would it be?’ Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 5) Object Personi昀椀ca琀椀on a. Like symbol matching, except usually non-visual b. In object personi昀椀ca琀椀on, we might ask ‘If this brand were a television star or movie star, who would it be?’ Or we might use more anthropomorphic representa琀椀ons such as animals or cars c. Example (Belk and Tumbat 2005). i. Apple would be a Volkswagen ‘beetle’ or van ii. PC would be a non-descript Chevrolet that was falling apart 6) Shopping List a. One of the oldest marke琀椀ng studies using projec琀椀ve methods 7) Dream Exercises a. The researcher takes advantage of the fact that dreams are a fantasy realm where a wide variety of things are possible. 8) Collage Construc琀椀on a. Par琀椀cipants are tradi琀椀onally given scissors, glue s琀椀cks, magazines, and a poster board and are asked to do a collage that expresses a theme of interest Class 4: Qualitative Research Analyzing The Bigg 6 – Methods of Qualita琀椀ve Data analysis (QDA) 1) Content Analysis a. Evaluates pa琀琀erns within a piece of content or mul琀椀ple pieces of content 2) Narra琀椀ve analysis a. Analysis of people’s stories to derive meaning 3) Discourse analysis a. Analysis of the use of language in conversa琀椀on within its social context 4) Thema琀椀c analysis a. Iden琀椀昀椀es key themes, pa琀琀erns of meaning in a data set - example experiences, opinions and perspec琀椀ves 5) Grounded theory a. Create a new theory (theories) using the data at hand, through a series of “tests” and “revisions” 6) IPA Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 a. Understand the personal experiences concerning a major life event, an experience or a situa琀椀on Thema琀椀c Analysis - Pa琀琀erns of meaning in data sources - Focus on people's opinions and perspec琀椀ves - Focus on similari琀椀es and di昀昀erences  most common themes to make sense and drive context - Itera琀椀ve process, not done at once The steps of thema琀椀c Analysis Class 6: Quantitative Research (Planning & Collecting) Prepare the Formulate Create a sample research Objec琀椀ves structure instrument (surey) Analyze the results Collect the data (SPSS) Oreo cases 1) Formulate objec琀椀ves Marke琀椀ng problem: The company’s budget is limited. They can only launch one 昀氀avor in the Montreal market. Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 The research problem: Evaluate preference for 昀氀avors and recommend the one with the highest probability of success. STAGE 1: Formulate the objec琀椀ves - Evaluate the preference for each 昀氀avor - Es琀椀mate purchase intent for each 昀氀avor - Determine which one has the most poten琀椀al 2) Prepare the research instrument Start with Qualita琀椀ve research 昀椀ndings Exploratory primary research undertaken prior to the quan琀椀ta琀椀ve research can help: Clarify what should be asked Iden琀椀fy the best way of tackling each topic Determine the most relevant wording for ques琀椀ons Determine the statements to be used in ra琀椀ng scales. To consider the administra琀椀on of the ques琀椀onnaire: How will the informa琀椀on be gathered? Administered by mail, phone, email, or in person? - 4 Types of measurement scales Example of measurement scales Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 3) Create a Sample structure De昀椀ne the Iden琀椀dy the Select a Select the Determine the Target Sampling Sampling Sample Sample Popula琀椀on Frame Procedure Characteris琀椀cs Structure - De昀椀ne the Target Popula琀椀on o The Target popula琀椀on is who you want to survey o It needs to be speci昀椀c for accurate results Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 o Simpler the de昀椀ni琀椀on the higher the incidence rate - Iden琀椀fy the sampling frame o Sampling Frame: the list of elements from which the sample is drawn  From a list of customers  From voter’s registra琀椀on list  From a list of postal codes  From a list of a research 昀椀rm  From university directory  Etc. - Select a sampling procedure o Probability  Simple Random:  Every element has equal probability to be selected  Stra琀椀昀椀ed  Involves the division of a popula琀椀on into smaller subsets known as strata based on member’s shared a琀琀ributes such as gender, income or educa琀椀on  1st popula琀椀on is divided into mutually exclusive subsets  2nd simple random sample is chosen from each subset  Cluster  Involves the division of a popula琀椀on into clusters that represents the popula琀椀on as a whole  1st popula琀椀on is divided into mutually exclusive subsets  2nd simple random sample of the subsets is selected Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 o Non-Probabiity  Convenience:  Easy to access, no way of knowing if they represent the target popula琀椀on  Judgment  Handpicked since serves the purpose  Snowballing  Start and recruit via referral  Quota  Select to represent the popula琀椀on via category or quotas - Determine the sample size o Comparability o Budget availability o Sta琀椀s琀椀cal method Types Of Errors Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 - Iden琀椀fy the Error Type (Individual) 1) If you increase the sample, it will decrease 2) Even if you increase the sample, it might not be solved 3) It skews the popula琀椀on one way or the other 4) It varies from the popula琀椀on above or below, but the average will s琀椀ll equal the popula琀椀on value 5) Because of these errors, researchers believe the data re昀氀ects reality, although they recognize that there may be varia琀椀ons in what is reported. 6) Respondents failed to answer your survey 7) You ignored part of your target popula琀椀on 8) Can be be琀琀er solved if you had a sampling frame Class 7: Quantitative Research (Analyzing) Analyze the results (SPSS) SPSS: Sta琀椀s琀椀cal Package for the Social Sciences - Preparing Raw Data: Data Coding o Data coding is the process of conver琀椀ng data into a form that can be analyzed Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 - Sta琀椀s琀椀cs Intro – Applies To Con琀椀nuous|Numeric Data o Mode  The mode is the number that occurs most o昀琀en in a data set o Median  The median is the middle value when a data set is ordered from least to greatest. o Mean  The mean (average) of a data set is found by adding all numbers in the data set and then dividing by the number of values in the set - Measurements Scales, Analysis, Graphics Univariate analysis Objec琀椀ve: Describe and synthesize research results by analyzing one variable at a 琀椀me. Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 Done as the 昀椀rst step of the analysis to ensure no errors in the 昀椀le and to describe the pa琀琀ern of response to the variable. Organize your ques琀椀ons according to the research objec琀椀ves Includes: the range of values and the central tendency of the values. - Frequency | Analyze (Ordinal) - Descrip琀椀ve | Analyze (Scale) Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 Class 8: Quantitative Research (Analyzing Bivariate) Bivariate Analysis Objec琀椀ve: Test the rela琀椀onship between variables. Is there a link that exists or not between the variables? Dependent and independent variables Age => purchase inten琀椀on => actual purchase Interdependent rela琀椀onship A琀 tude towards cars ↔ frequency of bicycle use Types of di昀昀erences between Responses 1) Mathema琀椀cal: If the numbers are not the same, they are di昀昀erent 2) Sta琀椀s琀椀cal: If a par琀椀cular di昀昀erence is large enough to be unlikely to have occurred because of chance or sampling error, then the di昀昀erence is sta琀椀s琀椀cally signi昀椀cant 3) Managerial: the di昀昀erence is important if results or numbers are su昀케ciently di昀昀erent Bivariate Analysis Process Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 Choose the Iden琀椀fy the analysis ques琀椀ons you Iden琀椀fy the Decide to technique Conduct the Start from the want to analyze measurement conduct (Cross- sta琀椀s琀椀cal test research and iden琀椀fy the scales for each bivariate tabula琀椀ons, (Chi-aquare, T- problem variable related related to the analysis comparisons of test, F Test) to each ques琀椀on means, ques琀椀ons Correla琀椀ons) Bivariate Analysis – Purpose, Scales & Tests Hypothesis Testing Hypothesis tes琀椀ng examines the rela琀椀onship between variables The null hypothesis: H0 : There is no rela琀椀onship between the two variables in the popula琀椀on. The alterna琀椀ve hypothesis: A rela琀椀onship exists between the two variables in the popula琀椀on. - Reject: p is less than or equal to 0.05 (≤ ) or 0.025 (t-tests) o We can reject the null hypothesis and accept the alterna琀椀ve hypothesis that there is a sta琀椀s琀椀cal di昀昀erence or rela琀椀onship between the two variables - Cannot Reject: p is greater than 0.05 (>) or 0.025 (t-tests) o We accept the null hypothesis that there is no sta琀椀s琀椀cal di昀昀erence or rela琀椀onship between the two variables Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 How to perform Sta琀椀s琀椀cal Tests Calculate the Determine Iden琀椀fy Ho strenght and the (null Calculate the Decision: direc琀椀on of appropriate hypothesis) sta琀椀s琀椀c (and reject or Interpret the the analysis and and Ha the p value) don't reject result rela琀椀onship test to (alterna琀椀ve to test Ho Ho (when perform hypothesis) applicable) Cross Tabulations - Comparing frequencies between variables - 2 nonmetric – nominal and/or ordinal - Is there a signi昀椀cant rela琀椀onship between the variables? o Gender (Q27) Ever ordered from Avery website(Q3)  The null hypothesis: H0 : There is no di昀昀erence among genders when it comes to their previous purchase behavior purchasing online from Avery  The alterna琀椀ve hypothesis: There is a di昀昀erence among genders when it comes to their previous purchase behavior purchasing online from Avery  Test at 0.05 signi昀椀cance - The chi-square test makes it possible to check whether a rela琀椀onship between two non-metric variables exists in the popula琀椀on that the sample represents. Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 Comparison of Mean & ANOVA – 3 Types - Comparison two independent means: Independent T-Test - Comparison of two matched means: Paired samples T-test - One-way analysis of variance: the ANOVA test The null hypothesis H0: the means are the same for both variables ( 1=2) The alterna琀椀ve hypothesis H1: the means are not the same between the two variables ( 1 ≠ 2) Reminder: Normal distribu琀椀on! Normal Distribu琀椀on – Comparing means Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 Idependent Samples test vs. Paired vs. ANOVA Independent Sample T-Test - Two variables evaluated by di昀昀erent people - A metric (scale) variable, a non-metric variable with only two categories (dichotomous) Gender (Q27) Avery sells high-quality merchandise (Q7) The null hypothesis: H0: the means are the same for both genders ( 1=2) The alterna琀椀ve hypothesis: H1: the means are not the same between the genders ( 1 ≠ 2) Test at 0.025 levels of signi昀椀cance Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 Paired Samples T-Test - Two variables evaluated by the same people - One metric (scale) variable, one non-metric variable regardless of the number of categories, or two metric (scale) variables Purchase online saves 琀椀me (Q15) Delivered to my door (Q19) The null hypothesis: H0: the means are the same for both variables ( 1=2) The alterna琀椀ve hypothesis: H1: the means are not the same between the two variables ( 1 ≠ 2) Test at 0.025 levels of signi昀椀cance ANOVA - Two variables evaluated by di昀昀erent people - A metric (scale) variable, a non-metric variable with more than two categories Marital status (Q26) Con昀椀dence spor琀椀ng goods purchase online good quality (Q4) The null hypothesis: H0: the means are the same for both genders ( 1=2) The alterna琀椀ve hypothesis: H1: the means are not the same between the genders ( 1 ≠ 2) Test at 0.05 levels of signi昀椀cance Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 Correlation - Strength and direc琀椀on of the linear rela琀椀onship between variables - 2 metric (scale) variables – if there is a rela琀椀onship, the next test is causa琀椀on via regression Descrip琀椀ons on-site accurate (Q10) Most companies can be trusted to deliver the product as pictured on the site (Q14) The null hypothesis H0: There is no rela琀椀onship between the two variables The alterna琀椀ve hypothesis H1: There is a rela琀椀onship between the two variables Test at 0.05 levels of signi昀椀cance Linear Regression Téléchargé par Olivia Jalby ([email protected])  lOMoARcPSD|35253986 - The impact of one variable on another - Rela琀椀onship of dependence, one variable impacts another - 2 metric (scale) variables - Run a correla琀椀on 昀椀rst, if signi昀椀cant run regression Income (Q23) Willingness to purchase from Avery’s website (Q2) The null hypothesis H0: X has no impact on Y The alterna琀椀ve hypothesis H1: X has a (posi琀椀ve/nega琀椀ve) impact on Y Need to specify the dependent and independent variable Dependent = willingness to purchase Independent = income Correlation & regression: 2 metric Variables Téléchargé par Olivia Jalby ([email protected])