Term
|
Definition
| Links organization with its marketing environment; involves specification, gathering, analysis, and interpretation of info to help management understand the environment, identify problems and opportunities, and develop and evaluate courses of action |
|
|
Term
| 3 C's (Framework for Marketing Strategy |
|
Definition
| Analyze situation for customer, company, and competitor |
|
|
Term
|
Definition
| Segmentation -> Targeting -> Positioning (Set Strategy) |
|
|
Term
|
Definition
| Product, Price, Promotion, Place (Formulate tactical plan for each) |
|
|
Term
| Uncontrollable Environmental Factors |
|
Definition
Economy
Technology
Laws & Regulations
Social & Cultural Factors
Political Factors |
|
|
Term
| Controllable Marketing Variables |
|
Definition
Product
Pricing
Promotion
Distribution |
|
|
Term
| Heinz Ketchup in Brazil (example) |
|
Definition
| Heinz wanted to use learning from Mexico where it was successfully launched (focused on making the product available in neighborhood stores in Mexico). However, in Brazil, most grocery shopping is done in supermarkets |
|
|
Term
| Coca-Cola in India (example) |
|
Definition
| In the U.S., people consume soft-drinks with meals, but in India people mostly drank water with meals. Their brand proposition was to "enhance the moments we already enjoy everyday" through coke. |
|
|
Term
| First step in the Market Research Process |
|
Definition
| Define research problems (decisions/assumptions, and the inputs needed) |
|
|
Term
| Second step in the Market Research Process |
|
Definition
| Identify research design (Needs should drive design--subject to logistics) |
|
|
Term
| Third step in the Market Research Process |
|
Definition
| Measurement (useful data comes from good communication) |
|
|
Term
| Fourth step in the Market Research Process |
|
Definition
| Sampling (who can give us the needed answers?) |
|
|
Term
| Fifth step in the Market Research Process |
|
Definition
| Analysis (what can we learn) |
|
|
Term
|
Definition
-Determine the basis of segmentation
-Establish market potential and responsiveness for various segments
-Select target markets
-Create lifestyle profiles: demography, media, and product image characteristics |
|
|
Term
|
Definition
-Test concept
-Determine optimal product design
-Package tests
-Brand positioning and repositioning
-Test marketing |
|
|
Term
|
Definition
-Pricing policies
-Importance of price in brand selection
-Product line pricing
-Price elasticity of demand |
|
|
Term
|
Definition
-Copy decisions
-Media decisions
-Creative advertising testing
-Optimal promotional budget
-Optimal promotional mix
-Evaluation of advertising effectivess |
|
|
Term
| Way marketing research can go wrong: |
|
Definition
-asking the wrong research question
-drawing conclusions from unrepresentative (potentially small) samples ( issue with generalization_)
-Uncontrollable Environment Factors |
|
|
Term
| Ethical issues in field work |
|
Definition
Increasing discomfort level of respondents (misleading respondents, disrespecting their privacy)
-Following unacceptable field work procedures (inappropriate sampling procedures and sample size) |
|
|
Term
| Ethical issues in data preparation and analysis |
|
Definition
Identifying and discarding unsatisfactory respondents
Using questionable statistical techniques |
|
|
Term
| Ethical issues in interpreting results |
|
Definition
making incorrect conclusions and recommendations
incomplete or biased reporting |
|
|
Term
| Research is valuable if it is: |
|
Definition
-Relevant
-Timely
-Accurate
-Cost effective
-Designed with decision goals in mind |
|
|
Term
|
Definition
-Can reduce, but not eliminate uncertainty
-Does not replace decision making
-Pay careful attention to the assumptions |
|
|
Term
| What makes the quality control scenario "ideal"? |
|
Definition
-Random selection of items to inspect (how can selection of items to inspect be "biased"?)
-Population clearly known (We know in which "world" our statistical inferences hold)
-Errors (or deviations from the standard) follow a well-defined pattern |
|
|
Term
|
Definition
| Can be tabulated in a Frequency Table and put into a data distribution bar graph |
|
|
Term
| What do we use for continuous data? |
|
Definition
|
|
Term
| Normal Distribution (Gaussian) |
|
Definition
| A mathematical ideal that extends from negative infinity to positive infinity. The Y-axis is a probability (Probability density curve). Center line is mu, each segment is a standard deviation from mu (the mean). |
|
|
Term
|
Definition
| used to describe the basic features of data; ex: central tendency (typical value, e.g. mean, median mode), dispersion (how much "spread" around typical, e.g. a standard deviation), skewness |
|
|
Term
|
Definition
|
|
Term
|
Definition
|
|
Term
[image]
What is the circled number? |
|
Definition
|
|
Term
|
Definition
| Current state of the world (status quo) |
|
|
Term
| Alternative/Experimental Hypothesis (H1) |
|
Definition
|
|
Term
| What is a two-tailed test? |
|
Definition
| A test in which the alternative hypothesis (your hypothesis) is that the actual state of the world is higher or lower than the null hypothesis (status quo). [image] |
|
|
Term
| What is a one-tailed test? |
|
Definition
| The alternative hypothesis (your hypothesis) is that the actual state of affairs is EITHER higher or lower than the status quo (null hypothesis). >, < |
|
|
Term
| What do we use to "feel good" about our decision and feel confident that the new value is either part of the old null distribution or actually part of a new distribution of IQ scores? |
|
Definition
| Significance level (alpha) |
|
|
Term
| Significance Level (alpha) |
|
Definition
| How far in the NULL distribution are you willing to go to be confident that an observed value DOES NOT belong to the NULL distribution---> typically alpha = 0.05 or 5% |
|
|
Term
|
Definition
| Value at green bar belongs to the alternative distribution. This means we must UPDATE our knowledge of the current state of the world |
|
|
Term
|
Definition
| Value at green bar belongs to the NULL hypothesis; thus we didn't learn anything new |
|
|
Term
|
Definition
1. Right decision (Accept null, when null is true)
2. Type 2 Error (accept null when alternative is true)
3. Type 1 error (reject null when null is true)
4. Right decision (reject null when alternative is true) |
|
|
Term
| The Marketing Research Process for PRIMARY Research |
|
Definition
1. Identify a business problem
2. Translate problem into a research methodology
3. Design the research instrument e.g., surveys, questionnaires, or guidelines
4. Sampling plan
5. Data collection and coding
6. Analysis, interpretation and reporting |
|
|
Term
| In the first step of the Market Research Process we first identify the __________ problem, and then formulate the ___________problem. |
|
Definition
|
|
Term
| Potential Marketing Problems of Declining Market Share |
|
Definition
-Outdated product
-New competition
-Shifting demographics
-Inappropriate pricing |
|
|
Term
| Potential Marketing Problems of Declining Profits |
|
Definition
-Ineffective promotions
-Escalating distribution costs
-Improper channel structure |
|
|
Term
| Potential Marketing Problems of Inability to Gain Channel Participation |
|
Definition
-Lack of product differentiation
-Misdirected promotions
-Inferior product image |
|
|
Term
| Potential Marketing Problems of Heavy Turnover in Sales Force |
|
Definition
-Lack of proper sales incentives
-Improper allocation of territories
-Unrealistic sales quotes |
|
|
Term
| Potential Marketing Problems of Declining Company Sales |
|
Definition
-Decline in industry sales
-Increased competition |
|
|
Term
| Goal of Marketing Research |
|
Definition
| to provide relevant, accurate, reliable, valid, and current information to facilitate managerial decisions |
|
|
Term
| What are methods to performing exploratory research? |
|
Definition
| In-depth interviews, focus group discussions, observational studies |
|
|
Term
| What are some characteristics of exploratory research? |
|
Definition
| Flexible, versatile, but not conclusive |
|
|
Term
| What is exploratory research useful for? |
|
Definition
| Discovery of ideas and insights |
|
|
Term
| What are methods to performing descriptive research? |
|
Definition
|
|
Term
| What are characteristics of descriptive research? |
|
Definition
|
|
Term
| What is descriptive research useful for? |
|
Definition
| Describing opinions, attitude and market characteristics |
|
|
Term
| What are methods to performing causal research? |
|
Definition
|
|
Term
| What are some characteristics of causal research? |
|
Definition
| Treatment and Control (A/B) |
|
|
Term
| What is causal research useful for? |
|
Definition
|
|
Term
|
Definition
1. Exploratory Research Design
2. Conclusive Research Design |
|
|
Term
| What are the two types of research that fall under the conclusive research design |
|
Definition
| Descriptive Research, Causal Research |
|
|
Term
| What are the two types of research design methods that fall under Descriptive Research? |
|
Definition
| Cross-Sectional, and Longitudinal |
|
|
Term
| What are the two types of Cross-Sectional Research designs? |
|
Definition
| Single cross-sectional design, multiple cross-sectional design |
|
|
Term
| Uses of exploratory research |
|
Definition
| Formulate a problem or define a problem more precisely, identify alternative courses of action, develop hypotheses, isolate key variables and relationships for further examination, and establish priorities for further research (often done before CONCLUSIVE/formal/quantitative research is undertaken) |
|
|
Term
| What is the objective of Qualitative Research? |
|
Definition
| To gain qualitative understanding of the underlying reasons and motivations |
|
|
Term
| What does a typical sample for Qualitative Research look like? |
|
Definition
| Small number of cases; often non-representative |
|
|
Term
| What kind of data collection occurs in qualitative research? |
|
Definition
|
|
Term
| What kind of data analysis do we perform when observing qualitative research? |
|
Definition
|
|
Term
| What is the outcome of qualitative research? |
|
Definition
| Develop an initial understanding (often) |
|
|
Term
| What is the objective of Quantitative Research? |
|
Definition
| To quantify the data and generalize the results from the sample to the population of interest |
|
|
Term
| What does a typical sample for Quantitative Research look like? |
|
Definition
| Large number of cases; often representative |
|
|
Term
| What kind of data collection occurs in quantitative research? |
|
Definition
|
|
Term
| What kind of data analysis do we perform when observing quantitative research? |
|
Definition
|
|
Term
| What is the outcome of quantitative research? |
|
Definition
| Recommend a final course of action (often) |
|
|
Term
| Methods of exploratory research |
|
Definition
| Focus groups, depth interviews, observation studies, lead user analysis ("beta testing") |
|
|
Term
| When setting up a focus group to perform exploratory research, what should we do in terms of recruitment for the study? |
|
Definition
| Screen participants on qualifications, demographics, and segments |
|
|
Term
| What should be set up before staging a focus group? |
|
Definition
| Hosting multiple sessions, and having a moderator provide a loose outline of the session and prioritize questions |
|
|
Term
| How should a focus group typically flow? |
|
Definition
1. Introduction (set the tone, no right/wrong answers, videotaping, two-way mirror, clients interested in research, controlled but informal)
2. Controlling the session (everyone participates within reason, safe for all points of view)
3. Problem respondents are handled discreetly ("Phone call for you, please take all of your things") |
|
|
Term
| Focus Groups produce ______(quantitative/qualitative)___ data. |
|
Definition
| Qualitative, provides a consensus around major "trends", "Verbatims" often are used to substantiate future business recommendations, but NEVER quantify the data gathered. (i.e. NEVER try to set up an "x% of participants" conclusion) |
|
|
Term
|
Definition
| an unstructured, indirect form of questioning that encourages respondents to project their underlying motivations, beliefs, attitudes or feelings regarding the issues of concern. |
|
|
Term
| Common techniques used in qualitative research? |
|
Definition
-Projection
-Word association
-Sentence Completion (and paragraph/story completion)
-Picture interpretation (ZMET, Zaltman Metaphor Elicitation Technique) |
|
|
Term
|
Definition
| respondents are presented with a list of words, one at a time, and asked to respond to each with the first word that comes to mind. The words of interest, called test words, are interspersed throughout the list which also contains some neutral, or filler words to disguise the purpose of the study. Responses are analyzed by calculating the frequency with which any word is given as a response, the amount of time that elapses before a response is given. |
|
|
Term
|
Definition
| Sample word association results |
|
|
Term
|
Definition
| respondents are given incomplete sentences and asked to complete them. Generally, they are asked to use the first word of phrase that comes to mind. (Ex. "A person who shops at Sears is __________________.") |
|
|
Term
| ZMET (Zaltman Metaphor Elicitation Technique) |
|
Definition
| Participants are asked to bring pictures that they think represents a particular topic (e.g. a brand) to the interview (gives participants control of the research stimuli and a greater sense of involvement with the interview topic). The pictures are metaphors that serve as entry points into their thinking process. Exploring the meaning of these metaphors (by a trained interviewer) allows us to elicit many important ideas. |
|
|
Term
|
Definition
| Interviewing professional respondents, with scheduled interviews, at a lower cost than Focus Group discussions. However, less information is found through these discussions. |
|
|
Term
| What is the technique used in depth interviews in which the line of questioning proceeds from product characteristics to user characteristics, allowing the researcher to tap into the consumer's network of meanings? Example? |
|
Definition
Laddering
1. "Finished homework" - product characteristics
2. "So I can do well in school"
3. "So I can get a good job"
4. "So I won't be homeless"
Advertising theme for getting homework done: not being homeless --> "Get shit done. Don't have to shit outside." |
|
|
Term
|
Definition
| Observe customers in their native surroundings to understand unarticulated needs. Works best when developers are proposing solutions for an identified potential user population, whose needs are poorly understood. |
|
|
Term
| What are some issues researchers are faced with when performing observation studies? |
|
Definition
What is the "data"?
Transparency to "participants" |
|
|
Term
| Disadvantages of qualitative techniques. |
|
Definition
-Respondent fatigue (except observation studies)
-Serious risk of interpretation bias
-Requires trained interviewer
-Could be expensive |
|
|
Term
| Advantages of qualitative techniques |
|
Definition
-Starting point of more structured quantitative research
-Useful for intimate product categories like cosmetics, luxury undergarments. (uncover feelings, beliefs, values, and attitude) |
|
|
Term
| When do we use descriptive research? |
|
Definition
1. To describe the characteristics of relevant groups such as consumers, salespeople, organizations, or market areas
2. To estimate the percentage of units in a specified population exhibiting a certain behavior
3. To determine the perceptions of product characteristics
4. To determine the degree to which marketing variables are associated (relational statistics, e.g., correlation) |
|
|
Term
| What are the four main types of survey methods? |
|
Definition
| Telephone, personal, mail, electronic |
|
|
Term
| Factors for evaluating usefulness of survey methods (e.g. personal vs. email) |
|
Definition
1. Response rate(% of total attempted interviews completed)
2. Sample control: ability to reach the sample members effectively and efficiently
3. Control of data collection environment: the environment in which the respondent answers questions (includes control over field staff and supervisors)
4. Respondent factors: perceived anonymity and social desirability
5. Diversity of question types / stimuli that can be used |
|
|
Term
| Method of getting a random sample of phone interview participants |
|
Definition
| Select an area code (target market), replace the last three digits with a random number generated between 000 and 999, and repeat the process until the desired number of telephone numbers is obtained |
|
|
Term
| Single Cross-Sectional Design |
|
Definition
| only one sample of respondents and info is obtained from this sample only once |
|
|
Term
| Multiple Cross-Sectional Design |
|
Definition
| two or more samples of respondents and info from each sample is obtained only once |
|
|
Term
|
Definition
| A fixed sample of population elements is measured repeatedly on the same variables |
|
|
Term
| What is the difference between Longitudinal and Cross-Sectional designs? |
|
Definition
| Longitudinal design sample or samples remain the same over time |
|
|
Term
| Which type of descriptive research is better at detective change? |
|
Definition
|
|
Term
| What type of descriptive research is more accurate? |
|
Definition
|
|
Term
| What type of descriptive research is better at preventing response bias? |
|
Definition
|
|
Term
|
Definition
| companies that collect and sell common pools of data of known commercial value designed to serve a number of clients (Ex. Nielsen Media Panel, Scanner Data); Includes sources from both consumers and institutions such as retailers, wholesalers, and/or industrial firms. |
|
|
Term
| What is the Nielsen Media Panel? |
|
Definition
| 11,000 U.S. household are randomly selected and a "Nielsen People Meter" is placed on each TV in the household. Their tuning data is stored in the PM and automatically uploaded to Nielsen servers each night. This allows Nielsen to estimate TV viewing by demographic and socioeconomic characteristics like household income, education of head of house, occupation of head of house, household size, age of children, age of women, and geographic location, etc.. |
|
|
Term
| What are the two major types of errors that can go wrong in descriptive research? |
|
Definition
| Random/Sampling Error, Non-Sampling Errors |
|
|
Term
| What are the two major types of non-sampling errors that can occur during descriptive research? |
|
Definition
| Response Error and Non-Response Error |
|
|
Term
| What are the three major reasons for a response error in descriptive research? |
|
Definition
Researcher error: Measurement error, population definition error, sampling frame error, data analysis error
Interviewer Error: Respondent selection error, questioning error, recording error, cheating error
Respondent Error: Inability Error, unwillingness error |
|
|
Term
|
Definition
Descriptive research error due to the particular sample selected being an imperfect representation of the population of interest.
Error is defined as the variation between the true mean value for the sample and the true mean value of the population. |
|
|
Term
|
Definition
| Due to various reasons such as errors in the problem definition, approach, scales, questionnaire design, interviewing methods, and data preparation and analysis. Results in non-response/bad response errors. |
|
|
Term
| An increase in sample size (decreases/increases) random error, but also (increases/decreases) the non-sampling error. |
|
Definition
| Increase in sample size decreases random error (if considerable size from relevant population is taken), but also increases the non-sampling error. |
|
|
Term
| Describe the Concept of Causality |
|
Definition
|
|
Term
|
Definition
Concomitant Variation
Time Order of Occurrence
Absence of Other Possible Causal Factors |
|
|
Term
| Concomitant Variation (Conditions for Causality) |
|
Definition
| extent to which a cause, X, and an effect, Y, occur together or vary together in the way predicted by the hypothesis under consideration |
|
|
Term
| Time Order of Occurrence (Conditions for Causality) |
|
Definition
| states that the causing event must occur either before or simultaneously with the effect; it CANNOT occur afterwards. |
|
|
Term
| Absence of Other Possible Causal Factors (Conditions for Causality) |
|
Definition
| the factor or variable being investigated should be the only possible causal explanation. |
|
|
Term
|
Definition
| Getting the same result if measured again |
|
|
Term
| Internal validity (Causality) |
|
Definition
Measuring exactly what you think you're measuring.
Refers to whether the manipulation of the independent variables or treatments actually caused the observed effects on the dependent variables; Control of extraneous variables is a necessary condition for establishing internal validity |
|
|
Term
| External Validity (Causality) |
|
Definition
Can I generalize the result?
Refers to whether the cause-and-effect relationships found in the experiment can be generalized. (To what populations, settings, and times can the results be projected?) |
|
|
Term
| Limitations of Experimentation |
|
Definition
1. Experiments can be time consuming to design
2. Experiments are often expensive (businesses often look at the control group as a "waste")
3. Experiments can be difficult to administer. (Can be impossible to control for the effects of the extraneous variables, particularly in a field environment) |
|
|
Term
|
Definition
1. Laboratory/Survey
2. Field |
|
|
Term
| Research Instrument (3rd Step of Marketing Research Process for Primary Research) |
|
Definition
| a testing device for measuring a given phenomenon, such as a paper and pencil test, a questionnaire, an interview, a research tool, or a set of guidelines for observation |
|
|
Term
| Measurement (Part of 3rd step of Marketing Research Process for Primary Research) |
|
Definition
| assigning numbers to characteristics. Must follow some rules which would be standardized and applied uniformly, and not change over objects or time. Measurement is fundamental in all branches of science, both social (e.g. Economics) and natural (e.g. Physics) |
|
|
Term
| Examples of difficult things to measure |
|
Definition
| Attitude, perception, and satisfaction |
|
|
Term
| Examples of things than can be easily measured |
|
Definition
| GDP of a country, unemployment rate of a country in a quarter, height of a student, time taken to complete a 200m race |
|
|
Term
|
Definition
branch of measurement that involves the construction of an instrument that associates qualitative constructs with quantitative metric units.
Types of Scaling: Nominal, Ordinal, Interval, Ratio |
|
|
Term
|
Definition
| Nominal, Ordinal, Interval, Ratio |
|
|
Term
|
Definition
Numbers serve only as labels or tags for identifying and classifying objects
Only permissible operation is counting
Example: Meal Preferences can be 1-Breakfast 2-Lunch 3-Dinner |
|
|
Term
|
Definition
"Ranking"
**OFTEN INTERPRETED AS INTERVAL DATA FOR THE SAKE OF RESEARCH BY ASSUMING EQUAL INTERVALS
Can determine whether an object has more or less of a characteristic than some other object, BUT NOT HOW MUCH MORE OR LESS.
Permissible operations are counting and statistics based on centiles, like percentile, quartile, median.
Example:
Satisfaction of customers
Low/Medium/High
Customer 1: Low
Customer 2: Low
Customer 3: Low
Customer 4: Medium
Customer 5: High
Median is Customer 3's opinion: Low (must be in order to get median)
Placement in a percentile: grouping them by what percentage of the participants stated greater than or equal satisfaction than a particular participant. |
|
|
Term
|
Definition
Numerically equal distances on the scale represent equal values in the characteristic being measured.
Permits comparison of the differences between objects.
The location of the zero point is not fixed. Both the zero point and the units of measurement are arbitrary. (0 Celsius does not mean presence of no heat, it is arbitrary)
It is not meaningful to take ratios of scale values.
Permissible operations: counting, statistics based on centiles, arithmetic mean, standard deviation, and other statistics commonly used in marketing research |
|
|
Term
|
Definition
Possesses all the properties of nominal, ordinal, and interval scales
ABSOLUTE ZERO POINT
Meaningful to compute RATIOS of scale values (can say "Twice as high, or one-half as much")
All statistical techniques can be applied to ratio data. |
|
|
Term
|
Definition
1. Nominal
2. Ordinal
3. Interval
4. Ratio |
|
|
Term
What statistics methods should we use for the following example:
"1. When at home, which TV brand do you watch most often? (circle one)"
__Sony
__Toshiba
__Panasonic
__Samsung
__Phillips
__Other |
|
Definition
| Frequency tables, mode, percentages belonging to each group |
|
|
Term
What statistics methods should we use for the following example, and what type of scale are we using?:
"1. Please rank each television brand in terms of your preference. Place a 1 by your top choice, 2 by second choice, etc.
__Sony
__Toshiba
__Panasonic
__Samsung
__Phillips
__Other
" |
|
Definition
Frequency tables, mode, median
Ranking-----> ORDINAL (X is better than Y but can't say by how much) |
|
|
Term
What statistics methods should we use for the following example, and what type of scaling are we using here?:
"1. For each brand below, indicate how much you like each brand by circling the appropriate number:
1-Do Not Like at All
5- Like very much
Sony 1 2 3 4 5
Toshiba 1 2 3 4 5
Panasonic 1 2 3 4 5
" |
|
Definition
Frequency tables, mode, median, mean, standard deviation
RATINGS ---> Interval scale, X is n units different from Y, distance between 2 and 3 is the same as distance between 3 and 4 (ideally) |
|
|
Term
What statistics methods should we use for the following example and what type of scaling are we using?:
"1. How much would you be willing to pay for a 24" flat screen TV made by each company?
Sony $__________
Toshiba $_______
Panasonic $________" |
|
Definition
Frequency tables, mode, median, mean, standard deviation, geometric mean
Absolute magnitude are meaningful
RATIO SCALE---> $0 IS ABSOLUTE ZERO, $40 IS TWICE AS MUCH AS $20, $20 IS HALF AS MUCH AS $40) |
|
|
Term
|
Definition
Numbers serve only as labels or tags for identifying and classifying objects
Only permissible operation is counting
Example: Meal Preferences can be 1-Breakfast 2-Lunch 3-Dinner |
|
|
Term
|
Definition
"Ranking"
**OFTEN INTERPRETED AS INTERVAL DATA FOR THE SAKE OF RESEARCH BY ASSUMING EQUAL INTERVALS
Can determine whether an object has more or less of a characteristic than some other object, BUT NOT HOW MUCH MORE OR LESS.
Permissible operations are counting and statistics based on centiles, like percentile, quartile, median.
Example:
Satisfaction of customers
Low/Medium/High
Customer 1: Low
Customer 2: Low
Customer 3: Low
Customer 4: Medium
Customer 5: High
Median is Customer 3's opinion: Low (must be in order to get median)
Placement in a percentile: grouping them by what percentage of the participants stated greater than or equal satisfaction than a particular participant. |
|
|
Term
|
Definition
Numerically equal distances on the scale represent equal values in the characteristic being measured.
Permits comparison of the differences between objects.
The location of the zero point is not fixed. Both the zero point and the units of measurement are arbitrary. (0 Celsius does not mean presence of no heat, it is arbitrary)
It is not meaningful to take ratios of scale values.
Permissible operations: counting, statistics based on centiles, arithmetic mean, standard deviation, and other statistics commonly used in marketing research |
|
|
Term
|
Definition
Possesses all the properties of nominal, ordinal, and interval scales
ABSOLUTE ZERO POINT
Meaningful to compute RATIOS of scale values (can say "Twice as high, or one-half as much")
All statistical techniques can be applied to ratio data. |
|
|
Term
|
Definition
1. Nominal
2. Ordinal
3. Interval
4. Ratio |
|
|
Term
What statistics methods should we use for the following example:
"1. When at home, which TV brand do you watch most often? (circle one)"
__Sony
__Toshiba
__Panasonic
__Samsung
__Phillips
__Other |
|
Definition
| Frequency tables, mode, percentages belonging to each group |
|
|
Term
What statistics methods should we use for the following example, and what type of scale are we using?:
"1. Please rank each television brand in terms of your preference. Place a 1 by your top choice, 2 by second choice, etc.
__Sony
__Toshiba
__Panasonic
__Samsung
__Phillips
__Other
" |
|
Definition
Frequency tables, mode, median
Ranking-----> ORDINAL (X is better than Y but can't say by how much) |
|
|
Term
What statistics methods should we use for the following example, and what type of scaling are we using here?:
"1. For each brand below, indicate how much you like each brand by circling the appropriate number:
1-Do Not Like at All
5- Like very much
Sony 1 2 3 4 5
Toshiba 1 2 3 4 5
Panasonic 1 2 3 4 5
" |
|
Definition
Frequency tables, mode, median, mean, standard deviation
RATINGS ---> Interval scale, X is n units different from Y, distance between 2 and 3 is the same as distance between 3 and 4 (ideally) |
|
|
Term
What statistics methods should we use for the following example and what type of scaling are we using?:
"1. How much would you be willing to pay for a 24" flat screen TV made by each company?
Sony $__________
Toshiba $_______
Panasonic $________" |
|
Definition
Frequency tables, mode, median, mean, standard deviation, geometric mean
Absolute magnitude are meaningful
RATIO SCALE---> $0 IS ABSOLUTE ZERO, $40 IS TWICE AS MUCH AS $20, $20 IS HALF AS MUCH AS $40) |
|
|
Term
| Scaling Techniques can be ________ and ______________. |
|
Definition
| Comparative, non-comparative |
|
|
Term
|
Definition
involve the direct comparison of stimulus objects (ex. two brands compared along quality dimension)
Comparative data must be interpreted in relative terms and have only ordinal or rank order properties. |
|
|
Term
|
Definition
| each object is evaluated independently of the others in the stimulus set. (Ex. one brand is rated on a scale independent of other brands). Resulting data are generally assumed to be interval or ratio scaled |
|
|
Term
|
Definition
| Paired Comparison Question |
|
|
Term
| Advantages of Paired Comparison Questions |
|
Definition
Same known reference points for all respondents
Easily understood and can be programmed in a survey (including validating *wrong* responses) |
|
|
Term
| Disadvantages of Paired Comparison Questions |
|
Definition
| Ordinal nature of data (cannot say how much they prefer one over the other) Inability to generalize beyond the stimulus objects (They might have a preference for a product not listed in the columns/rows) |
|
|
Term
| What is this? [image] [image] |
|
Definition
|
|
Term
| What is this? [image][image] |
|
Definition
|
|
Term
| What types of comparative scaled questions are there? |
|
Definition
Paired Comparison
Rank Order
Constant Sum |
|
|
Term
| What are some types of non-comparative scaled questions? |
|
Definition
Continuous Rating Questions/Sliders
Itemized Rating Scale (Likert Scale, Semantic Differential Scale) |
|
|
Term
| Continuous Rating Questions/Sliders |
|
Definition
Respondents rate the objects by placing a mark at the appropriate position on a line that runs from one extreme of the criterion variable to the other. Used in non-comparative scaling.
Ex. "How would you rate Sears as a department store?"
Probably The Worst 0 --------> Probably the Best 100 |
|
|
Term
|
Definition
| The respondents are provided with a scale that has a number or brief description associated with each category. Categories are ordered in terms of scale position, and the respondents are required to select the specified category that best describes the object being rated. Commonly used itemized rating scales are the likert and semantic differential scales |
|
|
Term
|
Definition
requires the respondents to indicate a degree of agreement or disagreement with each of a series of statements about the stimulus objects. Analysis can be conducted on an item-by-item basis or total (summated) score can be calculated. Negative items with similar semantic meaning should be reverse coded when computing overall score.
[image] |
|
|
Term
| What can having odd scale-points on a Likert Scale do? |
|
Definition
| Allows participants to express indifference (but respondents might be tempted to not apply their mind and choose this) |
|
|
Term
| Advantage of having fewer items on a Likert test |
|
Definition
| easier to process/interpret by respondents |
|
|
Term
| Advantage of having more items on a Likert test |
|
Definition
| Respondents might be able to express minute differences in their answers |
|
|
Term
| What happens when we label all items in a Likert scale? |
|
Definition
| Consistent interpretation, but reads more ordinal-like than if we only labeled end points. |
|
|
Term
|
Definition
A type of itemized rating scale with end points associated with bipolar labels that have semantic meaning (usually the scale has 7 points)
Example:
Describe your mood-
Sad --:--:--:--:--:X:--: Happy |
|
|
Term
| What strategy do we use to control the tendency of semantic differential respondents to mark right- or left-hand sides without reading labels? |
|
Definition
| Positioning negative adjective sometimes on the left, and sometimes on the right |
|
|
Term
|
Definition
Occurs when participants gauge probability based off of how often they personally are put in specific situations.
Ex. what is a more likely cause of death in the U.S.: being killed by falling airplane parts, or being killed by a shark? --> Chance of dying form falling airplane parts is 30x more likely than getting killed by a shark |
|
|
Term
| Another word for "immediate exposures" that affect our attitude and perceptions. |
|
Definition
|
|
Term
|
Definition
| Enabling a person to discover or learn something for themselves. |
|
|
Term
| Heuristics can lead to these types of biases: |
|
Definition
| Confirmation, False Consensus, and Fundamental Attribution Error |
|
|
Term
|
Definition
| Remembering info that supports our beliefs |
|
|
Term
|
Definition
| Thinking our own preferences represent the majority |
|
|
Term
| Fundamental Attribution Error |
|
Definition
| Thinking other people's actions reveal their preferences/attitudes |
|
|
Term
Consider two different persons, Mr.A and Ms. B
Who is happier under the Prospect Theory?
A: $100 Tax return, $60 traffic ticket
B: $40 in March Madness office pool
Loss Aversion Coefficient (Lambda)= 2 |
|
Definition
Ms. B is presumably happier because the events are seen as separate. Mr. A sees the gain from event 1 ($100 tax return) and the loss from event 2 ($60 ticket) as separate events, where loss is twice as bad as an equivalent gain.
Mr A: v(100) + 2v(-60) = 100 - 20 = -20 (UNHAPPY, OVERALL LOSS)
Ms. B: v(40) = 40 (HAPPY, OVERALL GAIN)
However if Mr. A mentally combines the events, then...
Mr.A: v(100-60) = v(40)=40 (EQUALLY AS HAPPY AS MS. B) |
|
|
Term
|
Definition
described the common human tendency to rely too heavily on the first piece of info offered ("anchor") when making decisions
Ex. Credit card --> class tended to provide answers based off of the minimum balance due rather than the second piece of info stating they couldn't afford to pay greater than or equal to the minimum payment due. |
|
|
Term
|
Definition
| Collects measurement firm every member of the population. A complete enumeration of the almonds and population. No sampling error. Very costly - both money and time. |
|
|
Term
|
Definition
| A subgroup of the elements of the population selected for participation in the study. Cost effective - both in terms of cost and money. But this Indians it also comes at a cost (error that comes from inferring about population from a sample.) |
|
|
Term
We have:
small budget
short time available
large population size
small variance in the characteristic
low cost of sampling errors
high cost of nonsampling errors
destructive nature of measurement
Which should we use? A sample or a census? |
|
Definition
|
|
Term
We have:
large budget
long time available
small population size
large variance in the characteristic
high cost of sampling errors
low cost of nonsampling errors
nondestructive nature of measurement
Which should we use? A sample or a census? |
|
Definition
|
|
Term
|
Definition
1. Define the population
2. Determine the Sampling Frame
3. Select sampling technique
4. Determine the sample size
5. Execute the sampling process |
|
|
Term
|
Definition
| Specific list of all the people in your target population and their contact information |
|
|
Term
| What are the two main types of sampling techniques? |
|
Definition
| Nonprobability Sampling Techniques (Convenience, judgmental, quota, snowball), Probability Sampling Techniques (simple random sampling, systematic sampling, stratified sampling, cluster sampling |
|
|
Term
|
Definition
| do not allow for objective evaluation of the precision of the sample results, i.e. how confident we are about generalizability |
|
|
Term
|
Definition
| Attempts to obtain a sample of convenient elements; right place, right rime, (difficult to generalize inferences to a specific population) |
|
|
Term
|
Definition
| form of convenience sampling; population elements are selected based on the judgment of the researcher; difficult to generalize inferences to a specific population |
|
|
Term
|
Definition
ensures that the composition of the sample is the same as the composition of the population with respect to the characteristic of interest
2-Stage sampling procedure:
1. Determinet attributes that are important (income, gender, residence) and determine their quota/upper limits (often quotas represent incidence in the population)
2. Sample elements selected based on convenience or judgment until the quote is met |
|
|
Term
|
Definition
| an initial group of respondents is selected, usually at random through probability sampling; after being interviewed, these respondents are asked to identify others who belong to the target population of interest; subsequent respondents are selected based on the referrals; rare/low-usage products |
|
|
Term
| Simple Random Sampling (SRS) |
|
Definition
each element in the population has a known and equal probability of selection. This implies that every element is selected independently of every other element. (all non-probability samples had "dependent" elements)
Often difficult to construct a sampling frame that will permit a simple random sample to be drawn |
|
|
Term
|
Definition
the sample is chosen by selecting a random starting point and then picking every ith element in succession from the sampling frame. The sampling interval, i, is determined by dividing the population size N by the sample size n and rounding to the nearest integer.
Ex. need a 20 person sample from a population of 100 people, i = 100/20 = 5, choose a random number between 1 and 5, lets pick 3. Select every 5th person in the list starting from 3, i.e. 3rd, 8th, 13th |
|
|
Term
|
Definition
Two stage sampling
Population is partitioned into non-overlapping groups, called strata and a sample is selected from each statum
1. Strata should be mutually exclusive and collectively exhaustive
2. The elements within a stratum should be as homogenous as possible, but the elements in different strata should be as heterogenous as possible
Elements are selected from each stratum by a random procedure, usually Simple Random Sample. |
|
|
Term
| Proportionate stratified sampling |
|
Definition
| size of the sample drawn from each stratum is proportional to the relative size of that stratum in the total population |
|
|
Term
| Disproportionate stratified sampling |
|
Definition
| size also depends on other factors |
|
|
Term
|
Definition
1. target population is first divided into mutually exclusive and collectively exhaustive subpopulations, or clusters.
Then a random sample of clusters is selected, based on a probability sampling technique such as SRS. For each selected cluster, either all the elements are included in the sample, or a sample of elements is drawn, preferably probabilistically |
|
|
Term
| Characteristics of sampling clusters |
|
Definition
Elements within a cluster should be as heterogeneous as possible, but clusters themselves should be as homogeneous as possible.
Ideally each cluster should be a small-scale representation of the population. |
|
|
Term
|
Definition
|
|
Term
|
Definition
|
|
Term
| Which type of sampling offers a quantifiable measure of sampling error? |
|
Definition
|
|
Term
| Surveys can be used in both _______ and ___________ research. |
|
Definition
|
|
Term
|
Definition
| temporal displacement of an event whereby people perceive recent events as being more remote than they are and distant events as being more recent than they are; leads to over reporting of the frequency of events |
|
|
Term
Prevent telescoping bias:
"How many gallons of soft drinks did you consume during the last four weeks?" |
|
Definition
"How often do you consume soft drinks in a typical week? Please select from below:
1. ___ Less than once a week
2. ___ 1 to 3 times per week
3. ___ 4 to 6 times per week
4. ___ 7 or more times per week" |
|
|
Term
|
Definition
| open-ended questions that respondents answer in their own words |
|
|
Term
|
Definition
| specify the set of response alternatives and the response format; may be multiple-choice, dichotomous, or a scale |
|
|
Term
|
Definition
1. General, uninformative
2. New Information
3. Specific
4. Past behavior, demographics |
|
|
Term
|
Definition
refers to the testing of the questionnaire on a small sample of respondents to identify and eliminate potential problems
respondents should be drawn from the same population
best done via a personal interview, no matter the format of the actual survey; not always possible |
|
|
Term
|
Definition
Left: Bar chart
Right: Histogram |
|
|
Term
| Relative Frequency Histogram |
|
Definition
| the height of each bar represents the proportion of the data (in percentages) in that bar; the height of the "bars" sums to 100% |
|
|
Term
| probability density histogram |
|
Definition
the area of each bar represents the proportion of the data in that bar
total area inside the histogram is 1 |
|
|
Term
| Formula for mean of POPULATION distributions |
|
Definition
|
|
Term
|
Definition
| Standard Deviation of POPULATION Distributions |
|
|
Term
| Formula for mean of SAMPLE distributions |
|
Definition
|
|
Term
|
Definition
| often used when population SD is not available; "estimate" of population SD; the sample is likely to be less variable than population, therefore a correction is applied to "raise" the SD estimate for the population; the larger the value of sample size, N, the less effect will this correction have |
|
|
Term
|
Definition
| Standard Deviation of SAMPLE distribution |
|
|
Term
|
Definition
a frequency distribution that is:
-Symmetrical
-(+/-) SD from mean contains 68% of scores
-All of the area under the curve = 100% |
|
|
Term
|
Definition
numbers expressed in standard deviation units
a.k.a. standard normal or unit normal |
|
|
Term
|
Definition
|
|
Term
Scores on the Math portion of the SAT are normally distributed with a mean of 514 (mu = 514) and a std dev of 117.
What is the z-score for George, who received a score of 527? |
|
Definition
X= George's score
mu = mean, 514
stddev = 117
(527-514)/117 = .11 |
|
|
Term
We find that a student, George's z-score for his SAT math score of 527 is .11
If the mean is 514 and the std dev is 117, which percents of test takers did George do better than? |
|
Definition
Z-score table states the probability of scoring equal to or greater than George is .4562.
So the percent of test takers George did better than is 1 - .4562 = .5438
So George did better than 54.38% of his peers! |
|
|
Term
| Distribution of sample means |
|
Definition
imagine drawing multiple random samples from a population. For a particular survey question, you are able to collect data that we can measure the mean of, such as the average height of the participants.
If we keep repeating this process, finding more and more sample means, we can organize these sample means into their own distribution |
|
|
Term
|
Definition
EVERY distribution of sample means have the following characteristics:
1. mean will be equal to the population mean (mu)
2. standard deviation will always be equal to stddev/sqrt(N)
^^^^STANDARD ERROR OF THE MEAN = STANDARD DEVIATION OF THE SAMPLE MEAN DISTRIBUTION
The shape of the histogram will tend to be normal as N increases |
|
|
Term
| Standard Error of the Mean (SEM) |
|
Definition
The typical distance between all possible sample means and the population mean
stdev / sqrt(N)
stdev = population stdev
N = sample size |
|
|
Term
Consider a small population of 4 values: 10 , 12, 14, and 16
A. Find population mean
B. Find population standard deviation |
|
Definition
A. 13
B. Standard Deviation = 2.24
10 - 13 = -3
12 - 13 = -1
14 - 13 = 1
16 -13 = 3
-3^2 = 9
-1^2 = 1
1^2 = 1
3^2 = 9
9 + 1 + 1 + 9 = 20
20/n = 20/4 = 5
sqrt(5) = 2.236
2.24 |
|
|
Term
| Why is Central Limit Theorem (CLT) important? |
|
Definition
| We can just do one single sample draw from the population. Do not need to repeat the sampling process multiple times, thus allowing for hypothesis testing. |
|
|
Term
|
Definition
"Is the sample mean different from the population mean?"
Simplest case is the one-sample test |
|
|
Term
| FBI is investigating loan officers of a large US bank, for illegal mortgage practice. • The mean income of all bank loan officers is µ=$50,000. Also, the standard σ=$12,000 • FBI checks the paper of N=16 bank loan officers from this bank. The mean income of these 16 officers = $55,000 • Can we conclude that the bank indulged in illegal practice? |
|
Definition
N = 16
Sample Mean = $55,000
Standard Error of the Mean (Standard Deviation of sample mean) = stdev/sqrt(N) = 12,000/sqrt(16) = $3,000
Find where $55,000 lands in a distribution where the mean is $50,000, and stdev (sterror) is $3,000
Find the probability that one will find an officer with a mean salary greater than $55,000 merely by chance.
z-score = (X-mu)/sterror = (55000 - 50000)/3000 - 1.67
^^^ *** test statistic *** ^^
Probability in Z-Score table: 0.0475 = 4.75% chance of finding an officer who makes more than $55,000 a year "by chance"
In other words, the difference between the population mean and the sample mean merely due to chance is very unlikely
The bank is potentially guilty of fraudulent practice. |
|
|
Term
|
Definition
apriori = alpha
our level of significance (we decide it, but is typically 5% or 0.05) |
|
|
Term
|
Definition
| set by the alpha value. If alpha is .05, the critical value is 1.65 because 1.65 in the z-table shows the probability closest to 0.05. |
|
|
Term
|
Definition
The value obtained from the "observed" mean of the sample "drawn" from the population.
Used to compute the test statistic (z-score, it is the X value) |
|
|
Term
| How can we calculate the standard deviation of the population when only given the sample size, and a portion of the population? |
|
Definition
The Population Proportion
sqrt(pi(1-pi)) where pi = portion
SD of the sample mean or standard error then is...
sqrt((sqrt(pi(1-pi))/N) |
|
|
Term
|
Definition
|
|
Term
|
Definition
Standard Deviation of the sample mean
or
Standard Error |
|
|
Term
| How do we calculate the value of the sample proportion? |
|
Definition
divide the number of observations by the sample size
ex. sample size = 30, 17 have red hair, p = 17/30 =0.567 |
|
|
Term
|
Definition
| Z-Score of a value of a sample proportion |
|
|
Term
|
Definition
Represents probability that the null (status quo) is true and that this is not new information
ex. FBI investigation example, p = 4% so the probability that we AREN'T gaining new insights into the practices of the firm is very slim |
|
|
Term
|
Definition
| Represents the level of significance; exhibits the probability of making a type 1 error (rejecting null, when null is true) |
|
|
Term
| What describes IF a treatment worked (or if the result is likely due to a sampling error?) |
|
Definition
|
|
Term
| What describes HOW WELL a treatment worked? |
|
Definition
Effect Size
"size of the effect" |
|
|
Term
| Imagine you're testing to see if a new "treatment" worked. The difference between the means (mean difference) should be.... |
|
Definition
|
|
Term
| If Mbefore - Mafter NOT zero, what should we ask? |
|
Definition
| Is it due to chance or due to the treatment? |
|
|
Term
| When do we use a paired t test? |
|
Definition
| when the groups are related |
|
|
Term
| One-Tailed Hypotheses (Treatment example) |
|
Definition
Null: the population's mean is NOT higher after treatment
Research (Alternative): The population's mean memory score is higher after treatment than before taking the drug |
|
|
Term
| In what table do we use degrees of freedom (df) in determining the area to the right of the critical value? |
|
Definition
|
|
Term
| How do we calculate the degrees of freedom? |
|
Definition
|
|
Term
| How do we find the t value? (Test Statistic) |
|
Definition
t = (Mafter - Mbefore)/(stddev(sampling dist. of the difference)/sqrt(N))
N = sample size |
|
|
Term
|
Definition
| Difference (D) is used for paired t-test |
|
|
Term
|
Definition
| Provides a range of plausible values for a population parameter (e.g. population mean), about which we can be confident to a certain extent |
|
|
Term
|
Definition
| UB = sample mean + (t critical)(SEM) |
|
|
Term
|
Definition
| LB = sample mean - (t critical) (SEM) |
|
|
Term
| T Critical vs. T Test Statistic |
|
Definition
| T Critical comes from a table! T Test statistics are computed |
|
|
Term
|
Definition
|
|
Term
|
Definition
|
|
Term
The mean score on a test of scientific knowledge for a sample of 50 college graduates is M = 68.3 (SD = 12.4). Compute a 95% confidence interval around the mean.
What is the point estimate? |
|
Definition
68.3
The sample mean (i.e. your best guess about the population mean) |
|
|
Term
The mean score on a test of scientific knowledge for a sample of 50 college graduates is M = 68.3 (SD = 12.4). Compute a 95% confidence interval around the mean.
What are we looking for in terms of a range of values? |
|
Definition
A range between LB and UB likely to contain the population mean 95%
Basically we are saying the population mean will be somewhere close to the sample mean. By finding the upper and lower bounds we are stating that sample mean is a good indicator of the population mean because from it we are able to derive a region in the distribution where the actual unknown population mean will fall 95% of the time. |
|
|
Term
| What do we use when we do not know the population SD? |
|
Definition
|
|
Term
| What is the most plausible value of an unknown population mean? |
|
Definition
| Point estimate (i.e. the known sample mean) |
|
|
Term
| The further values are from the point estimate, the ___________ plausible they are. |
|
Definition
|
|
Term
| What is the most plausible value of an unknown population mean? |
|
Definition
| Point estimate (i.e. the known sample mean) |
|
|
Term
| The further values are from the point estimate, the ___________ plausible they are. |
|
Definition
|
|
Term
| Confidence Interval for a Mean Difference |
|
Definition
| difference between a sample mean and a population mean (as in, a single sample t test) |
|
|
Term
| What is the point estimate for a confidence interval for a mean difference? |
|
Definition
| The difference between the sample mean and the population mean (imagine a distribution with this difference at its center, where the x axis signifies the various differences between sample and population means). |
|
|
Term
| What does it mean when the 95% critical interval of the difference between the sample and population means contains 0? (as in a negative lower bound, positive upper bound) |
|
Definition
| It is quite likely that the difference between the sample mean and the population mean is 0. Thus, the observed mean was not significantly different. |
|
|
Term
| What do we use when we do not know the STANDARD DEVIATION of the POPULATION. |
|
Definition
|
|
Term
| When samples sizes are large, the ____________ is a better approximation of the _____________. |
|
Definition
| When sample sizes are large, the sample SD is a better approximation of the population SD. |
|
|
Term
| Why is the critical value for a t distribution higher than z for smaller samples? |
|
Definition
| There is more "noise" in a t test-statistic (particularly for lower N), and therefore we want to have a higher threshold before we reject the NULL. |
|
|
Term
| If we have a two tailed test, what happens to the level of significance (alpha)? |
|
Definition
It is divided equally into 2 equal parts
alpha = 0.05
a1 = 0.025 a2 = 0.025 |
|
|
Term
|
Definition
|
|
Term
|
Definition
|
|
Term
| What kind of test (one/two-tailed) would we use when we don't have any specific expectation, only a change? |
|
Definition
| Two-tailed hypothesis test |
|
|
Term
| When do we use chi-square tests? |
|
Definition
| When the data is nominal, and we are considering category membership; chi-square tests are used to analyze FREQUENCIES |
|
|
Term
| Two types of chi-square tests |
|
Definition
Goodness of fit -> when there is 1 nominal variable
Independence ->when there are 2 nominal variables |
|
|
Term
| Goodness-of-fit Chi Square |
|
Definition
| used to compare observed sample distribution with an expected probability distribution; was the observed value significantly different from the expected value? |
|
|
Term
| degrees of freedom for chi square tests |
|
Definition
| df = number of categories - 1 |
|
|
Term
|
Definition
|
|
Term
| How do we get critical value when performing a chi square test? |
|
Definition
| finding degrees of freedom (df=number of categories -1) and the alpha (0.05) and check a chi square table. |
|
|
Term
|
Definition
| Statistical examination of two or more variables which are described simultaneously |
|
|
Term
|
Definition
|
|
Term
| How do we find the degrees of freedom of a crosstab for a chi square test? |
|
Definition
| df = (number of categories - 1) * (number of categories - 1) |
|
|
Term
| How do we determine the expected frequencies (EF) if the NULL hypothesis is EXACTLY true? |
|
Definition
EF = (Row Total*Column Total)/Total number of observations
***DO THIS FOR EACH CELL |
|
|
Term
| What do we ask when we perform an independent samples t |
|
Definition
After treatments, if (measurement 1 - measurement 2 ) does NOT equal zero, is the difference likely due to chance or the treatment?
Does the treatment itself matter? |
|
|
Term
| Degrees of Freedom for an independent samples t test |
|
Definition
df(1) + df(2)
=
(N1 - 1) + (N2-1) |
|
|
Term
| which table do we consult for an independent samples t test's critical values? |
|
Definition
|
|
Term
| t test statistic for independent samples t test |
|
Definition
| t = (M1 - M2)/stddev(the sampling distribution of the difference) |
|
|
Term
| What does ANOVA stand for? |
|
Definition
|
|
Term
|
Definition
| measures how far a set of numbers are spread out. A variance of zero indicates that all the values are identical. |
|
|
Term
| Why do we need One-Way F tests (ANOVA) if we already have independent sample t tests? |
|
Definition
| Because independent sample t tests are for two samples. Any greater amount of samples means we cannot use independent t tests anymore! |
|
|
Term
| ANOVA (Analysis of Variance) |
|
Definition
A hypothesis testing procedure for evaluating mean differences between two or more treatments
**If used on only 2 groups, will bare same results as an independent samples t test |
|
|
Term
|
Definition
| a function of sum of squared deviation from some mean value |
|
|
Term
|
Definition
| Are the measurements very similar inside a particular group, and very different across groups? |
|
|
Term
| Sources of variability within treatments |
|
Definition
| individual differences, error |
|
|
Term
| Sources of variability across treatments |
|
Definition
| treatment effect, individual differences, error |
|
|
Term
|
Definition
used in ANOVA to compare the relative amount of variability created by treatment vs. individual differences and error
F = 1 if treatment has no effect; between and within group variability will be equal
F >1
If treatment has an effect, between group variability will be more than within group variability |
|
|
Term
| Degrees of freedom for ANOVA |
|
Definition
df(between) = # groups - 1
df(within) = N - # groups |
|
|
Term
|
Definition
| F= variability across treatments / variability within treatments |
|
|
Term
|
Definition
|
|
