Marketing Research Exam 2
Marketing Research Exam 2 MKTG 632
Popular in Marketing Research
Popular in Marketing
This 5 page Study Guide was uploaded by Talia Standring on Tuesday March 22, 2016. The Study Guide belongs to MKTG 632 at San Francisco State University taught by in Fall 2015. Since its upload, it has received 42 views. For similar materials see Marketing Research in Marketing at San Francisco State University.
Reviews for Marketing Research Exam 2
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 03/22/16
MKTG632 Study Guide for Exam II Measurement Scales • Should know the attitude scales discussed in class Questionnaire Design • Should be aware that you need to clearly state the research objectives before designing the questionnaire. Step 1: consider the research objectives Step 2: specify the appropriate questions that should be in your survey The Likert scale requires the respondents to indicate a degree of agreement or disagreement w/ each of a serious of statements about the stimulus objects. Example: Survey that says strongly disagree, disagree, etc. The Semantic Differential scale is a sevenpoint rating scale with end points associated with bipolar labels that have semantic meaning. (think opposite) Example: Powerful – Weak; Unreliable Reliable • Should know the difference between open ended and multiplechoice questions. Openended question: where the respondent is allowed to answer in their own words. The purpose of these questions is to obtain more detailed information, especially regarding respondent’s opinions and views on particular subjects. Takes a lot of effort to analyze information and works best in smaller populations. Advantage: you can get full information and a more honest answer Disadvantage: a lot of data to narrow down; no criteria; broad Example: Please describe, in as much detail as possible, your typical morning routine. (Provides more useful data for a company trying to understand how its consumers get ready and where their product can fit into that routine). Multiplechoice questions: a form of question where the individual being interviewed may only answer from a predetermined list. • Should be able to identify “wrongly worded” questions that we discussed in class and how to reword them. “Wrongly worded” questions: convey a different meaning than what was sought to be conveyed, wrong data will be collected through responses to such questions. ex. badly worded: How short was ___? correctly worded: How would you describe ___’s height? Be clear about what you are asking. Leading Questions: they already lead you towards an answer ex. Weren’t you happy with the service last night at Holiday Inn? DoubleBarreled Question: the question has too many components, so the survey taker could agree with one part of the statement but not the other, which would not be accurate • Should know the ideal sequence in a questionnaire. Screening Example: Did you shop at Macy’s in the past 3 months? Warmups Example: How often do you shop in Macy’s? Main Questions (easy to difficult) Example: How important is each of the following factors selecting a department store? Complicated Example: Please rate Macy’s customer service on the following various aspects. Demographics Examples: What is your age? Sampling Design • Should know the difference between sampling and census and why we do sampling as opposed to conducting a census study. Census: Data collection from or about every member of the population of interest. Also called canvassing the population by asking everyone a set of questions. Sample: A subset of all the member of a population of interest. Administrative Sampling: choosing with bias Random Sampling: choosing randomly NonSampling Error: interviewer errors, measurement errors, response errors As sample size increases (amount of people), random sampling errors decrease (because you get a better representation of the population), but the nonsampling errors increase (due to more room for human error) • Should know what a sample frame means and what is meant by frame error. Sample frame: master list of all sample units in the population. Example: telephone book, mailing list ex. Telephone book, • Given a sample frame, should be able to identify the sources of frame error. Sources of Sample Frame Error: occurs when the wrong subpopulation is used to select a sample ex. people surveyed who have telephones were mostly republican when the majority of americans at the time didn’t even have phones; incorrect conclusion that majority of America is republican Target Population Vs. Sample Frame: Buyers vs. lists • Should know the differences between the various sampling methods and why and when they are used. Nonprobability vs. Probability • Should know the difference between probability and nonprobability sampling. Nonprobability sampling techniques: rely on personal judgment in the element selection process. Prohibits estimating the probability that any element will be included in the sample Pros: less cost, less time, can make samples of population that are reasonably representative, good for research for which accuracy isn’t very important. Convenience Samples: you choose places of convenience that have a high traffic of potential respondents or people who are easily accessible. This is the simplest and most convenient way of sampling. Judgement Samples: based on your knowledge, choose a sample of people that would represent the population of interest. Snowball Samples: an initial group of respondents is selected, usually at random. Subsequent respondents are selected based on referrals by this initial group. This is generally used while investigating characteristics that are rare in the population. Quota sampling: may be viewed as twostage restricted judgmental sampling. Quota sample is based on personal judgement. (personal judgement) Probability sampling: members of the population have a known chance of being selected in the sample Simple Random Sampling: every element has an equal probability of being selected into the sample. The probability of selection = sample size / population size Systematic Sampling: Skip Interval = population size / sample size Every 200 people for example Cluster Sampling: heterogeneity; more cost effective; selecting a sample of subgroups and then collecting data from all or a sample of the elements in the subgroup; when not every group is represented and different from each other (not a good representation of the area because it is only one group for example, only selecting people from the Marina of SF when choosing an SF group) Step 1: Divide population into subgroups called clusters, which represent the entire population Step 2: Randomly select some clusters Step 3: Select members of each chosen clusters to be included in the sample. Examples: amazon prime emails Subgroups: email service provider; last names If used for SF cities, the different areas don’t repersent SF well. (RANDOM IN SUB GROUP) Stratified Sampling: homogenous; selecting a sample of elements from each subgroup; offers better precision since it gets more representative from every sub group (SAME IN SUB GROUP) Step 1: Divide population into 2 or more homogenous subsets (mutually exclusive and collectively exhaustive) Step 2: Select a probability sample from each stratum Example: Groups: (AAA) / (BBB) / (CCC) — Similarity only between subgroups. We could select representatives from each subgroup. Suppose secondary data tell you that the usage habits of instant video services among households with kids is different from that among households without kids. In general, households with kids tend to use less instant video services. Your secondary data also tells you that roughly 35% of households have kids and the remaining households do not have kids in the target population. Q: If you want to have a sample size of 1000, how many households with kids would you survey? A: 350 = .35 * 1000 Problems: don’t have the means to distinguish the subgroups; can’t identify the bases for stratification Sample Size Determination • Should know that the sample size determination techniques discussed apply only for probability sampling and that too only for simple random sampling. Similar opinions in a sample = smaller sample size (amazon example), don’t need as much representation Proportional Allocation (stratified sampling): same fraction of strata Disproportional Allocation (stratified sampling): various fractions of strata Ad Hoc Methods: Rule of thumb: assume that each member of the sample was independently and randomly selected. (at least 100 cases) Budget Constraints: Buy only what you can afford Comparable Studies: use past studies to determine sample size • Should know the various factors that affect the sample size required and how they affect it. (confidence level, accuracy, and variability) 1.The higher the confidence level, the higher the sample size (confidence level = how certain you think something is) 2.The higher the accuracy, the higher the size. 3.The higher the variability, the higher the size. • Should know that there are two cases and should know when to use the two formulae. CASE A: nominal scales (percentage/proportion) p = estimated percentage (proportion) q = 100p if p is in percent = 1p if p is in proportion z = level of confidence (from a normal table) z = 1.96 for 95% confidence 2.58 for 99% confidence 1.65 for 90% confidence e = acceptable sample error If P is unknown, use 50% because it is the most conservative estimate CASE B: for data that involve interval or ratio scales n = required sample size z = level of confidence (from a normal table) z = 1.96 for 95% confidence 2.58 for 99% confidence 1.65 for 90% confidence e = acceptable error s = variability indicated by estimated standard deviation of the population • Should know what to do when the variability (or proportion) of the population is not known. Variability = P(1P) Higher the number, the larger the variability • Should be able to determine the sample size required under a given situation. The following z values and formulae will be provided on the test (NO explanations about the symbols) “z” values: Formulas: Z (90%) = 1.65 Z (95%) = 1.96 Z (99%) = 2.58 n=z^2(pq)/e^2 n=z^2s^/e^2
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'