MKT 444 - Exam 3
How do you clean data?
-Examine data for values out of the coded range -Examine data for values that are extreme -Compare responses that show inconsistency
What is included in the background?
-Overall goal of research -Decisions that need to be made -Company's strength and weaknesses regarding the issue in question
How is population standard deviation calculated?
-Use results from a previous study -Conduct a pilot study -Use secondary data -Use judgement
What are the key issues that should be addressed during a presentation?
-What are the objectives for the research? -What have we learned from the data? -What impact do they have? What do we need to do, given the information we now have? -How can future studies of this nature be enhanced?
What is included in an executive summary?
-Why the research was done -What was found -What those findings mean -What action, if any, management should undertake.
What are the 7 steps of a sampling plan?
1. Defining a population of interest 2. Choose a data collection method 3. Identify the sampling frame 4. Sampling methods 5. Determine sample size 6. Develop operational procedures 7. Execute the plan
How are open-ended questions coded?
1. List all responses 2. Consolidate responses 3. Set codes 4. Enter codes (match individual responses with consolidated list of response categories) Or use a word-cloud!
What are the 3 things to keep in mind about statistical significance?
1. Random samples are assumed (If not, significance tests are not appropriate) 2. Big data does not mean "good" data (No matter how much data, it must come from random samples) - ON EXAM 3. Do not over-rely on significance testing
What are the 5 steps of hypothesis testing?
1. State the hypothesis 2. Choose appropriate test statistics 3. Developing a decision rule 4. Calculating the value of the test statistics 5. Stating the conclusion
What steps are taken to calculate the t-stat?
1. Uses the appropriate formula to calculate the value of the statistic for the test chosen. 2. Compares the value just calculated to the critical value of the statistic (from the appropriate table), based on the decision rule chosen 3. Based on the comparison, determines to either reject or Fail to Reject (FTR) the null hypothesis H0.
What is convenience sampling?
A sample based on using people who are easily accessible. -True in an exploratory situation, where there is a pressing need to get an inexpensive approximation of true value -Researchers should strive to make it representative of the target population -Careful planning and careful execution of the sampling process can aid in producing a good convenience sample
What is judgment sampling?
A sample in which the selection criteria are based on the researcher's personal judgement about the representativeness of the population under study -Sample is not representative of the population
How are "don't know," "no opinion," and "not applicable" responses coded?
As 9 - then usually are eliminated from calculations
What is a hypothesis?
Assumption or theory that a researcher or manager makes about some characteristic of the population under study In hypothesis testing, the researcher determines whether a hypothesis concerning some characteristic of the population is likely to be true, and not due to sampling error or chance
What are 4 ways to determine the sample size?
Budget Available -A major factor -The larger the sample, the smaller the sampling error -The cost increases as the sample increases on a linear basis Rule of Thumb -Potential clients may specify they want a sample of a specific size -Sample size based on past experience and its sample sized used for a similar study -Client often requests samples of 300, 400, or 500 in a RFP Number of Subgroups Analyzed -Other things being equal, the larger the number of subgroups to be analyze, the larger the required sample size Traditional Statistical Methods -probability samples only -Use statistical formulas to determine the sample size under given variance, level of confidence, and desired error levels
What is the central limit theorem?
Central Limit Theorem: For any population, regardless of its distribution, the distribution of sample means or sample proportions approaches a normal distribution as sample size increases The properties of the normal distribution are crucial to classical statistical inference The larger the sample size, the distribution of averages becomes more normal and more average Take the average of several subgroups (100 rabbits in each)
What is a chi-square test?
Chi-Square Test: To evaluate how well the actual frequencies "fit" the expected frequencies Marketing researchers often want to examine if there are significant differences in the frequency counts of nominal and ordinal data. Both are non-metric scales It may involve a one-way frequency count or cross-tab. Researchers often call the single variable chi-square test a goodness-of-fit test because the procedure evaluates how well the actual frequencies "fit" the expected frequencies In using the chi-square test, it is important that each cell has a minimum expected count of 5.
How are scaled-response questions coded?
Closed-ended questions where the response choices are designed to capture the intensity of the respondent's feeling EX: How likely are you to use a gaming console? Very unlikely (1) - Very likely (5)
What is cluster sampling?
Cluster Sampling: Population of interest is divided into mutually exclusive and exhaustive subsets such as geographic areas. Each cluster is assumed to be representative of the population as a whole rather than homogenous groups. (HyVee has 2 clusters in Ames) Researcher selects a sample of subsets and then collects data either from all the elements in the subsets or from a sample of elements. Random sample of the subsets is selected Once the clusters are formed, then elements can be chosen through some type of random approach or through surveying all the elements within a cluster, which then is a census. EX: Doesn't have to have an even representation of each department, you can choose just 2 departments. It's still representative of Ivy.
What is a census?
Collection of data obtained from or about EVERY member of the population of interest. -Not often employed in marketing research -Cost and time required are too great. Availability is another issue -May be appropriate and feasible in settings where a firm has only a small number of customers -Census does not necessarily provide more accurate results than a sample
What are the 4 types of non-probability sampling methods?
Convenience Judgement Quota Snowball Probability of being selected for the sample is unknown.
What is correlation analysis? What is Pearson's product-moment correlation?
Correlation Analysis: Measures the degree to which changes in 1 variable are associated with changes in another variable -Measure the degree which changes in one variable are associated with changes in another variable -If you can find out the relationship of the variables, you can make predictions about future behavior -It is not for a linear relationship -Correlation is NOT causation Pearson's product-moment correlation: A correlation analysis technique for use with metric data
What are 5 statistical procedures talked about in class?
Correlation analysis Regression analysis Cluster analysis Factor analysis Conjoint analysis
What are cross tabulation tables?
Cross tabulation table: Examination of the responses to one question relative to the responses to one or more other questions Likely to be the next step in analysis. They represent a simple-to-understand, yet powerful analytic tool Cross tabulation includes frequencies and percentages, with the percentages based on column totals Need one question to have the nominal scale to run Cross Tabulation (Nominal scales partition data into categories that are mutually exclusive and collectively exhaustive) Graphic Representations of Data (Line, pie, and bar charts)
What are descriptive statistics? What are the types?
Descriptive Statistics: The most efficient methods of summarizing the characteristics of large sets of data. In a statistical analysis, the analyst calculates one number or a few numbers that reveal something about the characteristics of large sets of data Types: -Measurements of frequency -Measurements of central tendency (mean, median, and mode) -Measurements of dispersion or variance (range, variance and standard deviation)
What is a population of interest? How is it specified?
Entire group of people whom information is needed. Specified in terms of -Geographic characteristics -Demographic characteristics -Product/service use -Awareness Defining the population of interest is usually the first step in the sampling process and often involves defining the target market for the product or service in question: whose opinions are needed in order to satisfy the objective of the research
A correlation analysis between exam score and absence from classes results has R of -0.99. What does that mean?
Exam score and absence from classes has strong negative correlation
True or False: In marketing research, it is always preferable to take a census rather than a sample of the population under study
False
True or False: To test statistical difference of mean, we use T test, Z test, or ANOVA and Chi-Square.
False
True or False: In the research report, recommendation is a statement of action derived from the objective.
False
What is the null and alternative hypothesis for the following: Does the loyalty toward local stores vary by a person's age?
H0: Loyalty toward local stores does not vary by a person's age Ha: Loyalty toward local stores does not vary by a person's age
What is the null and alternative hypothesis for the following: Does the importance of service in selecting a clothing store vary by gender?
H0: The importance of service in selecting a clothing store does NOT vary by respondent's gender Ha: The importance of service inspecting a clothing store does vary by respondent's gender
What is included in methodology?
How research was done and why it was done that way -Research design -Secondary data used -Primary data collection method -Sampling procedure -Data collection process -Analytical techniques
What are mathematical differences?
If the numbers are not exactly the same, they are different - this does not imply statistical significance
ŷ =50+5.4X1 ŷ=estimated exam score in points 50=estimated constant X1= hours spent studying What is the independent variable? What is the dependent variable? When X1 is increased by 1 unit, how does this affect the estimate exam score?
Independent variable (X1): Hours spent studying Dependent variable (y): Exam score in points When X is increased by 1 unit, y increases by 5.4
What is a z-score?
Indicates the number of standard deviations a score is away from the mean. The areas under a standard normal distribution (reflects the percent of all observations) are for various Z-scores
What are line, pie, and bar charts used for?
Line charts: Simplest form of graphs. Useful for presenting a given measurement taken at several points over time. Pie charts: Appropriate for displaying percentages or proportions of a whole (100%) Bar charts: Good for side by side relationships/comparisons. Most flexible of the graphs.
Which graph would be simplest way to report the net income of Walmart over the 10-year period
Line graph
What is a sampling frame?
List of population elements from which units to be sampled can be selected or a specified procedure for generating such a list _____________ EX: American National Bank has 1000 current customers. The manager wants to launch a survey asking customers about its service quality. She wants to survey 100 customers Sampling Frame: 1000 current customers - phone numbers or email address
What are 4 ways of data collection?
Mail surveys - biases associated with low response rates Telephone surveys - growing problem with nonresponse, reduced use of landlines Internet surveys/web respondents - may not always be representative of the target population (issues when used on social media) Blending Sample - Trend that is based on interviews collected by different means such as mail-telephone-Internet panel, Internet panel-SMS (text), Internet panel-social media, etc. Delete duplicates.
If income gaps are wide for average income, what central measurement do you use?
Median
What is normal distribution? What are some of its characteristics?
Normal Distribution: Bell curve, a distribution that occurs naturally in many situations. Symmetrical, half of the data will fall to the left of the mean, and half will fall to the right Characteristics -Symmetric about its mean (it is not skewed and implies that the 3 measures of central tendency are all equal) -Uniquely defined by its mean and standard deviation -Total area under a normal curve is equal to one, meaning that it takes in all observations. -Standardized mean is 0. -68% in 1SD, 95% in 2SD, 99.7% in 3SD
What are the 2 types of hypotheses?
Null Hypothesis (H0): States no difference Alternative Hypothesis (Ha): Not the null EX: Does an average consumer buy a new computer every 3 years? H0: An average consumer purchases a new computer every 3 years. Ha: An average consumer does not purchase a new computer every 3 years.
For the clothing boutique stores, the manager would like to see if the e-mail campaign conducted this year yielded significantly better results than the direct mail campaign that was used last year. Both campaigns were conducted using the firm's database of customers that was obtained from individuals filling out loyalty cards. The mean of purchases for the direct mail campaign last year was 4.23. Use confidence level of 95%.
One-sample t-test example Result: The t-test shows there was a significant increase in the number of purchases using the e-mail campaign versus the direct mail campaign.
What are the 3 types of t-tests?
One-sample t-test: Used to test if the mean of a sample distribution is different from a specified or test mean. Independent-sample t-test: Test if there is a significant difference between two groups. Paired-sample t-tests: Used when a researcher wants to compare two different responses from an individual respondent
What are the 3 types of tabulation?
One-way frequency table Cross tabulation table Basic statistical measures
What are one-way frequency tables?
One-way frequency table: Most basic tabulation. A table showing the number of respondents choosing each answer to a survey question This table is the first summary of survey results. These tables typically indicate the percentage of those responding who gave each possible response to a question Most commonly, marketing researchers compute percentages for multiple-response questions on the basis of the number of respondents, reasoning that the client is primarily interested in the proportion of people who gave a particular answer
When do you reject and fail to reject the null hypothesis?
P>a - Fail to reject H0 P<=a - Reject H0
What is statistical significance?
Particular difference is large enough to be unlikely to have occurred because of chance or sampling error, the difference is statistically significant
What is factor analysis?
Permits the analyst to reduce a set of variables to a smaller set of factors or composite variables by identifying underlying dimensions in the data
What are the 2 types of sampling methods?
Probability Samples: Samples in which every element of the population has a known, nonzero likelihood of selection Non-probability Samples: Samples in which specific elements from the population have been selected in a nonrandom manner. -Not all members of the population have an equal or known chance of being selected -Selection is based on convenience because it is easy or inexpensive to reach
How do you determine sample size for each of the sampling methods?
Probability samples: Rely on factors like available budget, rules of thumb, and number of subgroups to be analyzed in their determination of sample size Non-probability samples: Use formulas to calculate the sample size required -Ability to make statistical inferences about population values based on sample results is a major advantage of probability samples. -Probability is more expensive to implement than nonprobability samples of the same size
What is data entry?
Process of converting information to an electronic format
What is coding? What should you always pre-code?
Process of grouping and assigning numeric codes to the various responses to a question Pre-code close ended questions.
What is conjoint analysis?
Provides a basis for estimating the utility that consumers associate with different product features or attributes Procedure used to quantify the value consumers associate with different levels of product/service attributes or features It is a popular procedure used by marketers to help determine what features a new product or service should have and how it should be priced. It can be argued that conjoint analysis has become popular because it is a more powerful, more flexible, and often less expensive way to address these important issues than is the traditional concept testing approach.
What is the correlation coefficient (R)? What is the coefficient of determination (R^2)?
R describes the degree of the association between independent variable and dependent variable(s).• R value is from strong positive correlation (1) to strong negative correlation (-1). R^2 measures how well the regression line fits the data. A measure of the percentage variation in the dependent variable explained by the variations in the independent variables.
What is random error? What is systematic error?
Random Error or Random Sampling Error: Error that results from chance variation, which is the difference between the sample value and the true value of the population mean. Systematic Error or Bias: Error that results from problems or flaws in research design or the execution of the research design.
What is regression analysis? What are 3 of its uses?
Regression Analysis: A procedure for predicting the level or magnitude of a (metric) dependent variable based on the levels of one or multiple independent variables (metric). Enables the researcher to project the level of magnitude of a dependent variable based on the levels of more than one independent variable. 3 Uses: Predictive analysis - Forecast future opportunities and risks is the most prominent application of regression analysis in business. Demand analysis - For instance, to predict the number of items which a consumer will probably purchase. New insights - find a relationship between different variables by uncovering patterns that were previously unnoticed.
How are multiple-choice questions coded?
Respondents can choose one or multiple answers in the list -0,1 for each of the choices -Not all possible alternatives are included to overcome include the question 'other' -Researcher must spend time generating the list of possible responses -Amount of information provided is more limited EX: Of the following gaming platforms, which ones have you played? (select all that apply) a. Online (0 or 1) b. Social networking (0 or 1) c. Gaming console (0 or 1)
A small clothing boutique store chain collected data from a sample of 221 respondents. Before analyzing the data, the researcher in charge of data collection wanted to see if the sample represented the firm's current target base. Using the retailers' database, the researcher was able to identify that 70% of its customers were female and 30% were male. Does the sample drawn representative of the customer base using a chi-square test, given 95% confidence level?
Result: At the 95% confidence level, there is no significant difference between the actual gender count and the expected values, which means the sample is representative of the customer base for this retail chain.
What is a decision rule?
Rule or standard used to determine whether to reject or fail to reject the null hypothesis. The significance level (α) is critical in the process of choosing between the null and alternative hypotheses. The level of significance—0.10, 0.05, or 0.01, to indicate the researcher will accept a 10%, 5%, or 1% chance of being wrong.
What is snowball sampling?
Sample in which additional respondents are selected based on REFERRALS from initial respondents -Dramatic reduction in search costs, especially for low incident rate study. However, the total sample is likely to be biased.
What is quota sampling?
Sample in which quotas, based on demographic or classification factors selected by the researcher, are established for population subgroups. EX: 10 shoppers who own X, 15 who one Y, and 12 who own Z -Use of quotas ensures that the appropriate subgroups are identified and included in the survey
What is a sample? What is sampling?
Sample: A subset of all the members of a population. Information obtained can be used to make estimates about various characteristics of the total population Sampling: Process of obtaining information from a subset (a sample) of a larger group (the universe or population) -The motivation for sampling is to be able to make these estimates more quickly and at a much lower cost. Has shown that a small percentage of a population can produce accurate estimates about the population -Critical that the sample is selected in a scientific manner to ensure that the sample is representative - that it is a true miniature of the population. However, achieving this goal in sampling form a human population is not easy
What is cluster analysis?
Segmentation - The general term for statistical procedures that classify objects, or people, into some number of mutually exclusive and exhaustive groups on the basis of two or more classification variables.
What are the 4 types of probability sampling methods?
Simple Random Systematic Stratified Cluster Every element of the population is selected in a random manner.
What is simple random sampling? What is it used for? What is the probability of someone being sampled?
Simple Random Sampling: Probability sample selected by assigning a number to every element of the population and then using a random number table to generate a list of random numbers. The elements of the population that correspond to those random numbers are then selected to be part of the sample Used for marketing research because every element has a known and equal chance of being selected. However, for many studies it is not possible to use the simple random approach because not every element can be identified. Probability of Selection = sample size / population size
In statistical testing, if a difference between two groups is large enough to be unlikely because of chance, we say that this is...
Statistically significant
What is stratified sampling? How is it calculated?
Stratified Sampling: A sample is forced to be more representative through a simple random sampling of mutually exclusive and collectively exhaustive subsets either proportionally or disproportionally. They are HOMOGENEOUS GROUPS. Researcher first selects a sample of elements from each subset Subsets - should be related to the characteristics of the population we are really interested in measuring (EX: 3 people from each department of Ivy - 18 total)
What are some strategies for minimizing measurement survey errors? -Surrogate information -Interviewer -Measurement instrument bias -Input -Non-response bias -Response bias
Surrogate information error - This error results from seeking and basing decisions on the wrong information. It results from poor design and can be minimized only by paying more careful attention to specification of the types of information required to fulfill the objectives of the research Interviewer error - This error occurs because of interactions between the interviewer and the respondent that affect the responses given. It is minimized by careful interviewer selection and training. In addition, quality control checks should involve unobtrusive monitoring of interviewers to ascertain whether they are following prescribed guidelines Measurement instrument bias - Also referred to as questionnaire bias, this error is minimized only by careful questionnaire design and pretesting Input error - This error can occur in the process of transferring data from questionnaires to the computer. It is the result of incorrect keystrokes from a respondent. Use software checks to find illogical response patterns or improperly scanned machine-scored questionnaires Nonresponse bias - This error results from the fact that those people chosen for the sample who actually respond are systematically different from those who are chosen and do not respond. It is particularly serious in connection with mail surveys. It is minimized by doing everything possible (e.g., shortening the questionnaire, making the questionnaire more respondent friendly, doing callbacks, providing incentives, contacting people when they are most likely to be at home) to encourage those chosen for the sample to respond Response bias - This error occurs when something about a question leads people to answer it in a particular way. It can be minimized by paying special attention to questionnaire design. In particular, questions that are hard to answer, might make the respondent look uninformed, or deal with sensitive issues should be modified.
What is systematic sampling? What is it used for? How is it calculated?
Systematic Sampling: Sample in which the entire population is numbered and elements are selected using a skip interval Used for direct marketing campaigns for testing direct offers. Most all software has capabilities to sort based on a systematic random sample Skip interval = population size/sample size (the "nth") -Use random first element as a starting point
How should a hypothesis conclusion be written?
The conclusion summarized the results of the test. It should be stated from the perspective of the original research question EX: Alpha=5% & p=17.8% There is insufficient evidence to reject the null hypothesis, as the two-tailed significance level of .178 exceeds the decision rule, or cutoff value, of .05. So, the researcher can conclude the study found no meaningful differences in gender in terms of their perceptions of the importance of service when selecting a retail clothing store
What is managerially important differences?
The differences are large enough to be meaningful to the manager Important differences for managers - One must be able to distinguish between mathematical differences and statistically significant differences in using the data analysis in managerial decision making -Not using significant difference as the only measurement. The obsolete value of the difference must be considered -The cost-to-benefit ratio must be evaluated
What are the components of a research report?
Title page Table of Content Executive Summary Background Methodology Findings, Conclusions and Recommendations Appendices
True or False If a particular difference is large enough to be unlikely to occurred because of chance or sampling error, then the difference is statically significant
True
How are dichotomous questions coded?
Two fixed alternatives - yes or no, agree or disagree (1,0) -Easy to administer -Usually evoke a rapid response -Fail to communicate intensity of feeling EX: Have you played video games on your smart phone? a. Yes (1) b. No (2)
What are the 2 types of hypothesis errors?
Type 1 Error (Alpha error): Reject the null hypothesis when, in fact, it is true Type 2 Error (Beta error): Failure to reject the null hypothesis when, in fact, it is false
What are operational procedures?
Used in selecting sample elements in the data collection phase of a product (develop and specify whether a probability or nonprobability sample is used and how). It is detailed, clear, and unambiguous
What is validation? What is editing?
Validation: process of ascertaining that interviews actually were conducted as specified. Select 10-20% of respondents to validation purpose -Was the data collection process properly followed? Did the interviewer use proper screening? Were there any fraud issues? Editing: A process of ascertaining that questionnaires were filled out properly and completely -Did questions get complete? Skip patterns followed? Any paraphrase of respondents' answers in open end questions? -Computer-assisted telephone interviewing (CATI), Internet, and other software-driven surveys have built-in logical checking. The editing process for paper surveys involves manual checking -The person doing the editing makes the judgement calls for the open-ended questions, as to what point particular answers are to be eliminated, or whether respondents should be re-contacted
What is variance? What is standard deviation?
Variance: sum of squared deviations from the mean divided by the number of observations for population variance. Sample Standard Deviation: Subtracting the mean of the series from each value in a series, squaring each result, summing the results, dividing the sum by the number of items minus 1, and taking the square root of this value
Assume that previous fast food research has shown that 80 percent of the consumers like curly French fries. The researcher wishes to have an error of 6 percent or less and be 95.44 percent confident of an estimate to be made about curly French fries consumption from a survey. What sample size should be used for a simple random sample?
Z = 95.44 --> 2 E = .06 P = .8 [2^2 * .8(1-.8)] / .06^2 n=177.8 --> A sample of 178 s needed for this project
You are making sure there's plenty of samples of chili for the patrons of the cook-off. The following standards have been set: -Confidence level: 99% -Error: 4 oz or less -SD: 3 oz What is the necessary sample size?
Z = 99% --> 2.6 E = 4 o = 3 (2.6^2 * 3^2)/4^2 n=3.8 --> A sample of 4 is needed, but a sample of 20-30 will allow for more powerful analysis
How is sample size calculated?
Z = level of confidence expressed in standard errors o = population standard deviation E = acceptance amount of sampling error P = proportion (use .5 if unknown)
What is a t-test? What is a z-test?
Z-test: used for larger samples of 30+ per cell T-test: used for small samples of <30 per cell or population standard deviation is unknown
All of the following should be included in the methodology section of a research report EXCEPT: a. detailed information on the statistical procedures. b. how the sample was selected. c. the number of people interviewed. d. how people were surveyed. e. why these people were interviewed.
a. detailed information on the statistical procedures.
All other things equal, larger samples will have less _______ than smaller samples. a. measurement error b. random sampling error c. time and cost d. non-sampling error e. interviewer error
b. random sampling error
The _________ is a list of the members or elements from which units to be sampled are selected. a. population b. sampling frame c. Census d. universe
b. sampling frame
What is the most important reason for the growing popularity of nonprobability samples? a. sampling error can be computed b. representative of population c. lower cost than probability samples d. less nonresponse bias than other types of sampling
c. lower cost than probability samples
If the researcher is expecting different responses between males and females in a study, which type of sample might be most useful? a. judgment b. snowball c. stratified d. systematic
c. stratified
What is a p-value?
p-value: Exact probability of getting a computed test statistic that is due to chance. P-values are from 0% to 100%. A p-value of .01 means there is only a 1% probability that the results from an experiment happened by chance. The smaller the P value, the smaller the probability that the observed result occurred by chance (sampling error).
