interviewer reads questions, either face to face or over phone, to the respondent and records their answers. (mall intercepts, phone surveys, in office/executive interviews, in home interviews)
ex-mall intercept. pros- able to create a rapport w/ respondent, able to show product/probe/clarify instructions/adapt. cons- in person, human error, interviewer bias.
mall intercept- sample may be skewed, only mall patrons are surveyed, respondent irritation and discomfort answering in a mall
interviewer verbalizes questions while relying on computer technology to facilitate interview work.
ex- telephone interview using CATI (computer-assisted telephone interviews). respondent is contacted bu phone; interviewer is the voice of the computer.
pros-computer assisted reduces data errors, customizable, tabulations may be run during study. telephone surveys are inexpensive for geographic reach, yields a high quality sample, and quick turn around. cons- unable to show product prototypes/ads/packaging. can't observe body language and make eye contact. limited to quantity of info that can be obtained. ability to reach respondents.
computer plays integral role in posing question and recording respondents' answers.
ex- fully online survey. respondent answers questions using an internet-based questionnaire.
pros- respondent not seeking to please interviewer. ease of creating and posting. flexible design, customizable (skip logic patterns, show video). fast turnaround and quick tabulation of results. relatively inexpensive. cons- respondent must have internet access. survey fatigue- low cooperation rates. lack of internal expertise. professional panel members
ex-mail survey. questionnaire is mailed to respondent who fills it out and returns it by mail.
pros- respondent is in control-completed in their own time and schedule. respondent completes survey on their own without agent. confidentiality- respondent ease with sensitive topics(no interviewer apprehension). reduced cost for geographic reach. cons- high questionnaire requirements (no interviewer to clarify). non-response rates are high. time to return and tabulate. self selection bias.
studies in which conditions are controlled so that one or more independent variables can be manipulated to test a hypothesis about a dependent variable. independent variables are directly controlled by researcher (price, packaging, distribution, product features). dependent variables aren't directly controlled by researcher (sales, customer satisfaction, market share)
provides insights into general nature of a problem, possible decision alternatives, and relevant variables.
provides snapshot of some specific aspect of marketing environment-who what where and why
examines whether change in one variable is likely caused by observed change in another
degree to which an experiment actually measures what researcher was trying to measure. internal validity is ability of the experiment to show casual relationships unambiguously. external validity is applicability of experimental results to situations external to the actual experimental context.
real world resting of a new product or some element of the marketing mix. single market, group of markets, or regions. goal is to evaluate proposed national program on a smaller and less costly sale. used for forecasting success of new product- estimates market share and volume, tests hypotheses about elements of marketing mix, identify weaknesses in product and packaging design, and analyzing consumer characteristics
traditional test markets
tests elements of marketing mix using regular distribution channels- best analog of true marketplace
controlled test markets
store panels managed by research suppliers
simulated test markets
consumer sample is selected and recruited to central testing location
rolling roll-out test market
product launched fully in one region, adjustments made
simulated test markets
millennium marketing research runs simulated test market studies (STMs) in malls in representative cities using simulates stores. Respondents view ads and then indicate which ads they recalled (unaided awareness). They are then given millennium marketing dollars with which they can purchase products from simulated store. result- a measure of the trial rate and relative contribution of marketing communications.
test market cost factors and dangers
competitor intelligence, advertising expenses, point of purchase materials, coupons and sampling, diversion of sales from existing products, bad press/public reaction, sample results may not be representative of general population
gathering facts, opinions and attitudes using a structured questionnaire. most popular method for gathering primary data. 94% companies surveyed use survey data. 88% of companies surveyed use online methods to conduct survey based market research
advantages of surveys
standardization, ease of administration, suitability to tabulation and statistical analysis, sensitivity to subgroup differences.
determining survey method
timing, budget, respondent interaction, questionnaire length, incidence rate, structure, and cultural and infrastructure considerations
process of assigning numbers or labels to objects (person, brand, company) in accordance with specific rules for representing quantities or qualities of properties. income, age, attitudes, brand loyalty, likelihood to recommend. rule: the guide, method, or command for assigning a number to observation to provide an accurate description (grades a-f, assign numbers 1-5 to people according to their inclination to do household chores)
process of measurement
develop a construct- brand loyalty, marital roles, social class
conceptual definition- put boundaries around it
operational definition- observable characteristics to be measured
develop a measurement scale
nominal level of measurement
used to represent identity:
gender. marital status. to categorize or separate respondents. mutually exclusive. collectively exhaustive categories. stats- frequencies, mode.
ex- which best describes where you live (geographic area): urban_ rural_ suburban_
ordinal level of measurement
used to represent order and relative standing. ranking preferences and ratings of foods. stats- frequencies, median, mode.
ex- indicate preferences for following restaurants 1-4- noodles and co, panera, chipotle, wild onion
interval level of measurment
intervals between numbers have meaning. equal intervals between points show relative amounts, ex- on a scale from 1-7. preferred measure of more complex concepts. common use- level of knowledge about brands. stats- entire range may be used.
ex- on a scale from 1-7, how would you rate importance of water bottle being sold on campus?
ratio level of measurement
interval scale with meaningful zero point (shares all characteristics). ability to say how many times greater or smaller one objects is from another. distances, weights, age, income (not range)
ex- how old are you? ______
measurement- reliability and validity
reliability is degree to which measures are free from random error and therefore provide consistent data. validity is degree to which what researcher was trying to measure was actually measured
attitude measurement scales
attitudes are learned predispositions to respond to an object or class of objects in consistently favorable or unfavorable way. construct that exists in mind of consumer. not directly observable. measured in attempt to predict behavior.
semantic differential scale
used to measure meaning of object to a person. ranked between dichotomous (opposite) pairs of words or phrases. quick means of examining strengths and weaknesses of product or company image. hale effect may dominate respondents overall impression of concepts being rated.
series of statements (attitudes) are examined. uses statements expressing a favorable or unfavorable attitude toward concept. only one single set of uniform replies is used. respondent is asked to indicate level of agreement with each statement by assigning a numerical score.
purchase intent scales
likelihood the customer will buy. often evaluated at each stage of product development.
ex- "next time you purchase a hand-crafted beverage, how likely are you to go to starbucks?" (unlikely, somewhat likely, likely, very likely)
ease of administration and ease of use, balanced vs non balanced scale, number of categories and neutral option, forced vs non forced choice, motivates action.
set of questions designed to generate data necessary to accomplish the objectives of research project. used in every form of survey research. a critical link to the customer.
function of questionnaire
translates research objectives into specific questions to solicit info from respondents. standardizes wording and sequencing of questions (participants respond to identical stimuli). imposes uniformity on data gathering process. wording, question flow, appearance can motivate respondent esp online.
dangers of bad questionnaire design
using language likely to be misunderstood or misinterpreted. poor sequencing of questions. double barreled questions (two questions in one). question bias- leading or loaded wording.
criteria for a good questionnaire
does it provide decision-making info? (refer back to research objective and management decision. supply data needed to make decision.)
doe it consider the respondent? (poor design, length, confusing surveys have turned off thousands of respondents. design explicitly for intended respondent, be brief, use simple grammar, be clear, avoid jargon and business terms)
considers analysis (skip patterns, editing to ensure skip patterns were followed and questionnaire is completed, coding)
advantages- rich info, in real world terms, based on respondent's frame of reference, effective for copywriting, effective for copywriting, a means of interpreting closed ended questions.
challenges- editing and coding, vague responses and interpretations, bias toward articulate respondents.
respondents required to choose from a list of answers. (dichotomous, multiple choice)
advantages- jog respondent's memory, interviewer bias eliminated, coding and data entry automatic.
challenges- generating a comprehensive list, measurement error- respondent answer not available
effective questionnaire design
uses intro to set stage (describe general purpose, incentives, how were they selected). keeps research objectives in mind and decision-making criteria. knows and speaks to audience. assists in building rapport with respondents. leaves time for testing. doesn't expect rocket science- simple is good, keep it short as possible.
process of obtaining info from subset (sample) of a larger group (population). ex- all prospective car buyers, all prospective mba school students, all households with children living in houston.
small samples can be every accurate, depending on how individuals were selected. seeking a true miniature of population, goal is to make estimates of characteristics of larger group.
entire group of people about whom info is needed- define population of interest, whose opinions are needed in order to satisfy objectives of research, often based on characteristics of current or target customers
data collection from or about every member of population of interest. also known as canvassing population by asking everyone a set of questions.
non-sampling or measurement- all error other than sampling error (data analysis, non response)
sampling- error that occurs because the sample selected isn't perfectly representative of the population.
everyone in population has a known, nonzero, likelihood of selection. advantages are representative cross section of population of interest, sampling error can be computed, results may be projected to total population. disadvantages is that more expensive than non probability samples of the same size.
samples in which members of population have been selected in non random manner. advantages include lower costs, esp when accuracy isn't critical, quickly gather data, can be reasonably representative if thoughtful and thorough about screening members. disadvantages include uncertain how representative of population and sampling error can't be computed.
simple random samples
purest form of probability sampling. each population member has equal probability of being selected. assign number to each element of population. select numbers using some method of random selection (lottery drawing, random digit dialing in phone surveys)
uses random starting point then sample is selected using a skip interval. (every 100th name is selected). advantage- for use with hard copy lists, simpler, less time consuming than simple random sampling. disadvantage- sample frame error, possibility that hidden patterns are inadvertently pulled into sample.
involves dividing population into subgroups. random sample of subgroups is selected and all members of subgroups are interviewed. often selected from a number of small geographic areas to reduce data collection costs. advantages include very cost effective and useful when subgroups can be identified that are representative of entire population. disadvantage includes cluster specification error- when clusters aren't homogenous.
simple random sampling of mutually exclusive and exhaustive subsets (heavy, medium, light users). seeking subgoups that are homogenous=strata. assures that population is accurately represented. advantage- more statistically efficient because one source of variation has been eliminated. disadvantage- little may be known about demographic characteristics of consumers.
sample based on using people who are easily accessible (mall intercept). increased use as research is becoming focused on hard to find categories. advantage is obtain info quickly and inexpensively. disadvantage is lower quality sample, some population elements are eliminated.
selection criteria are based on expert's judgement in identifying representative sample of population for study. ex- focus groups
sample in which additional respondents are selected based on referrals from initial respondents. appropriate when reaching small, specialized populations.
convenience sample with constraints. quotas are a minimum number from each specified subgroup in the population-is selected. often based on demographic or classification factors known about the population. match sample to population based on demographic characteristics.
tend to be convenience sampling. unless using client sample (email lists of entire customer population). increasing use of online panels- ability to draw simple random samples, stratified samples, quota samples.
a set of methods and techniques used to obtain info and insights from data. helps avoid erroneous judgements and false conclusions. can constructively influence research and objectives and research design.
data analysis process
validation and editting- inspect and correct each questionnaire
coding-transforming raw data into numbers for tabulation
data entry and file creation- enter data into computer system to facilitate statistical analysis
tabulation and statistical analysis- generate descriptive statistics to summarize and describe responses for each question
close ended- assign numbers to predetermined categories (including missing data). often precoded- numeric codes have already been assigned to responses.
open ended- unstructured, determine categories of responses and assign numbers to each individual's response based on categories generated. challenging, subjective, and time consuming to code. a lengthy list of possible responses is often generated.
open ended unstructured coding process
1. read through responses and highlight key phrases
2. generate list of actual responses
3. consolidate responses into categories
4. set code
5. assign codes
6. enter and tabulate
data entry and file creation process
raw data files- constraints raw data for all questionnaires. system file- includes variable names, labels. assign number to uniquely identify the questionnaire. enter number for response to each question.
frequency tables shows number of respondents choosing each answer to survey question. most often illustrated by a number or as a percentage. response categories may be combined for some questions- should result in categories with a worthwhile number of respondents.
statistical analysis technique to study relationships among and between variables. sample is divided into subgroups in order to learn how independent variable varies between subgroups. frequency distribution for each subgroup is compared to the frequency distribution for total sample. two variables that are analyzed must be nominally or ordinally scaled.
single number used to make inferences about entire population of interest. measures central tendency- mean: average number dividing sum of responses to a question by sample size. mode: value that occurs most often.
descriptive statistics- measures of dispersion
range: simplest measure, distance between smallest and largest values of frequency distribution. standard deviation: quantitative index of a distribution's spread, or variability.
graphical representations of data
presentation of tables can be cumbersome and confusing the audience. line charts: for presenting a given measurement take at several points over time. pic charts: for presenting segments of a whole. bar charts: useful for providing side by side comparisons.
what decision makers are seeking- generalizations that answer the questions raised by research objectives. statement or series of statements that simply communicate the results of the study. merge info and paraphrase it in a few descriptive statements that generalize the results. most likely doesn't include data. don't suggest a course of action
ex- target audiences are seeking CAMP promises- fun and friendship, faith and the outdoors.