BUSM Exam 2 Study Set

Class 8: Thinking About Data (1)  

Making inferences from data. Understand why there is uncertainty about making inferences from data. 

Basic univariate statistics. Understand what a mean, median, mode, and standard deviation are.  

Distributions. Understand what distributions are for, and the three broad categories of distributions we discussed in class (normal distribution, right-skewed, left-skewed).  

Will Rogers Phenomenon. Understand what it is, examples, as well as its applications to marketing.  

Correlation. Understand what a correlation coefficient is, what the graph of a certain correlation might look like, and how correlations could be sensitive to outliers.  

Anscombe’s Quartet. Understand what it is, and what it implies for marketers.  

Berkson’s Paradox. Understand what it is, some examples (both from class as well as from other situations not discussed in class), and its implications for drawing good inferences from data.  

Simpson’s Paradox. Understand what it is, some examples (both from class as well as from real-life situations), and its implications for drawing good inferences from data. 

Class 9: Think About Data (2) 

Correlation and causation. Understand why correlation does not imply causation, be able to apply this idea to real-world examples. Understand the notion of a confounder and be able to apply it to real-world examples.  

Different types of causes. Understand necessary, sufficient, and probabilistic causes and be able to think about examples of each.  

• Necessary cause 

  • A condition that must be present for an event to occur, but it may not be enough by itself ex: oxygen for fire 

• Sufficient cause 

  • A condition that, of present, guarantees that the effect will occur, but other factors could also lead to the effect 

    • No wifi – so can not join a zoom meeting 

• Probabalistic cause 

  • A condition that increases or decreases the probability of an effect occurring, without guaranteeing taht the event will occur (or being necessary for teh event to occur) ex: flu shot doesnt necessarily mean you will not get the flu 

    • Most causes in marketing are probabalistic 

Conditions to establish causality. Understand what is needed to establish causality empirically.  

• Correlation 

  • A and B must be correlated – when A changes, B also changes 

• Temporal Precedence 

  • A must happen before B, – the effect cannot happen before the cause 

• No alternative explanations (hardest condition to demonstrate empirically) 

  • There must be no third variable C explaining the association between A and B 

Misconceptions about causality. Understand the five misconceptions about causality we discussed in class.  

1. Sample Size: a large sample doesn’t eliminate confounds 

2. Size of the correlation: large correlation may emerge even when no direct causal relationship exists

3. Causal Story: a convincing story does not prove causation – empirical evidence is needed to rule out alternative explanations (third variables)

   1. Coffee and longevity: “caffeine keeps people mentally alert, reducing cognitive decline. Coffee alos encoyrages social interactions, which are linked to lower stress and longer life expectancy.”

4. Meaningfulness: not all causal effects are meaningful 

   1. Small effect: 

   2. Temporal effect: 

5. Small vs. Large Scale 

   1. Scaling up a policy or marketing intervention introduces new factors that might change their impact – ex: universal basic income  

Selection and randomization. Understand the problem of self-selection in causal inference and how random assignment solves it.

• Selection: people with different levels of a potential cause may be fundamentally different from each other 

  • People self-select into different “conditions”—they choose whether to pursue education, they choose whether to use social media, etc.

    • Those who pursue education may be fundamentally different from those who don’t (in ways other than education)

    • Those who use social media may be fundamentally different from those who don’t (in ways other than social media).

  • Hidden factors (i.e., “lurking variables” or “confounders”) create pre-existing differences between people with different levels of a causal factor.

  • Solution: random assignment

    • a method to assign participants to different groups in a way that ensures each person has an equal chance of being placed in any condition.

    • random assignment removes pre-existing differences between groups

Class 10: Marketing Research (1)  

Five-step approach to marketing research. Understand the steps involved in marketing research. Understand the distinction between primary and secondary data. Understand what the characteristics of a good recommendation are.  

• 5-step marketing research approach 

  • Define the problem 

    • Set research objectives

      • Exploratory research: provides ideas about vague problem or question 

      • Descriptive research: finding the frequency with which something occurs or the extent of a relationship between two factors 

      • Causal research: determine the extent to which the change in one factor changes another one 

    • Identify possible marketing actions

      • Measures of success: criteria or standards used in evaluation proposed solutions to the problem- different research outcomes leads to different marketing actions 

  • Develop research plan 

    • Specify constraints

    • Identify data needed for marketing actions

    • Determine how to collect data 

      • Concepts: ideas about products or services 

      • Methods: approaches that can be used to collect data to solve all or part of a problem 

      • Special methods vital to marketing: sampling and statistical inference 

  • Collect relevant information  

    • Obtain secondary data: facts and figures recorded prior to project; divided into two parts: internal and external referring to if data comes from inside or outside organization 

      • Marketing input data: effort expended to make sale (budget reports, expenditures, call reports, etc)

      • Marketing outcome data: result of marketing efforts (accounting records on shipments and sales/repeat sales, broken down by representative, industry, geographic region; emails, phone calls, social media posts from customers 

    • Obtain primary data: facts and figures newly collected for project; can be divided into observational, questionnaire or other sources of data

  • Develop Findings 

  • Make Recommedndations / Take Marketing Actions 

• Primary Data: collected first-hand (ex: surveys) 

• Secondary Data: data already out there/collected (ex: social media data) 

Causal research: experiments. Understand what an experiment is, differences between between-subjects and within-subjects experiments and advantages/disadvantages of each, understand the ecological fallacy, 

understand order effects in within-subjects experiments, differences between online, lab, and field experiments and advantages and disadvantages of each.  

• Setting Experiment: Online 

  • Websites, social media, apps 

  • EX: a company randomly assigns customers to see one of two different versions of their website to  measure which drives more clicks 

  • Advantges: easier to collect large, diverse samples, Easy to ranomdize conditions, Cost effective 

  • Disadvantages: law control, potentially biased samples, cannot observe real behavior 

• Setting Experiment: Lab

  • Universities, research labs 

  • EX: a researcher bring participants to a lab and randomly assigns them to different rooms with different colors to see their emotional repsonces

  • Advantages: high control over external factors, best to isolate precise causal factors, easy to replicate

  • Disadvantages: participants may behave differently due to the artificial setting 

• Setting Experiment: Field 

  • Stores, companies, street

  • EX: a grocery store testing how shelf placement affects product sales by rearranging items for different weeks 

  • Advantages: measures real-world behavior 

  • Disadvantages: hard to control for external influences 

• Between-Subject Experiments

  • Each participant is exposed only to one condition 

  • Comparison are made between different groups of participants 

  • EX: a researcher shows two different versions of an ad to two different groups of consumers 

  • Advantages: participants are less likely to guess the hypothesis, easy to implement in real-world settings 

  • Disadvantages: requires more participants, can only estimate average effects  

• Ecological Fallacy 

  • Making inferences about an individual based on an average effect 

  • Example

    • Experiment: each participant sees a different version of an ad 

    • Meausure: How likely would you be to buy this product?

  • Do not draw individual-level conclusions from group-level data 

• Within-Subjects Experiments 

  • Each participant is exposed only to more than one condition 

  • Comparisons are made within each participant rather than between groups

  • EX: a researcher shows different versions of an ad to all participants 

    • The participant sees both ad A and B and the researcher compares the participants intent to buy after seeing ad 1 and 2

  • Advantages: requires smaller samples 

  • Disadvantages: participants may guess the hypothesis 

  • Order effects 

    • Practice effects 

      • Improve at a task because they have done it before 

    • Fatigue effects 

    • Carryover effects 

      • Exposure to one condition influences responses in subsequent conditions 

Exploratory research: focus groups. Understand what a focus group is, when to use them and when not to use them, understand the best practices for running focus groups, knowing other types of exploratory research techniques available. 

• Focus Group 

Class 11: Marketing Research (2)

Descriptive research: surveys. Understand what a survey is, when to use them, and when not to use them.  

• When to conduct a survey?

  • To measure attitudes and opinions: How to people feel about a brand?

  • To track trends over time 

  • To compare different groups: How do prefeerneces vary by age and income?

  • To gernalize insights: whenever you want to generalize insights to a broder population 

• When NOT to conduct a survey? 

  • Causality: whenever you want to establish causality 

  • Memory: whenever participants may not be able to recall past behaviors, emotions, or decisions 

  • Social desirability

Social desirability bias. Understand what it is, why it is a problem, and understand the techniques we discussed to solve it.  

• Social desirability

  • The tendency for survey respondents to provide answers that portray them in a favorable light instead of answering honestly 

  • Leads to: over-reporting of “good behavior”, under reporting of “bad behavior”

Non-response bias. Understand what it is, why it is a problem, and understand the technique we discussed to try to solve it. Understand why it is hard, if not impossible, to solve.  

• Non-responce bias

  • People who respond to a survey may be fundamentally different from those who do nto respond 

  • EX: US elections: polls often underestimate support for certain candidates because certain groups are likely to respond 

• How to solve non-responce bias 

  • Potential solution: survey weighting – adjusting results to match known demographics. Often fails because you don’t know which variables predict whether someone responds to the survey or not 

Acquiescence bias. Understand what it is, why it is a problem, and understand the technique we discussed to solve it.  

• Acquiescence bias 

  • The tendency for respondents to agree with statements regardless of their opinion 

  • Happens because of politeness, social pressure, inattention 

  • EX: do you like this new product?

• To solve it .. 

  • Ask experienced based questions – instead of “do you support recycling?”, “in the  past month, how many times have you received plastic items?” 

Attention checks. Understand what attention checks are and what they are for. 

Question wording. Understand the ways in which question wording might affect people’s answers in a survey.  

Conjoint analysis. Understand what it is, what it is for, when to use it and when not to use it, and the best practices for running an effective conjoint analysis. 

Class 12: Advertising and Promotion  

The promotional mix. Understand the elements of the promotional mix we discussed in class.  

• Promotional Mix

  • The combination of marketing communication strategies that companies use to reach, engage, and persuade their target audience 

Effectiveness of the promotional mix. Understand the key findings of the paper we discussed looking at advertising and promotions.  

Communication process. Understand the elements of the communication process and the idea that the source of a message influences the recipient. 

Integrated marketing communications. Understand what it is and why it is important.  

Key marketing communications decisions. Understand the decisions that marketing managers must make regarding marketing communications.  

Key marketing communications decisions—mission. Understand the different goals marketers might want to accomplish with their marketing communications.  

Key marketing communications decisions—message. Understand the different types of advertising appeals we talked about. 

 Key marketing communications decisions—media. Understand what factors need to be considered when selecting a media type.  

Key marketing communications decisions—money. Understand the different methods available to set a marketing communications budget, and the advantages and disadvantages of each.  

Key marketing communications decisions—measurement. Understand the types of measures available to assess the effectiveness of marketing communications.

 Class 13: Principles of Persuasion  

Commitment and Consistency. Understand the principle of commitment and consistency in persuasion, the example we talked about in class, and applications to marketing contexts.  

Social Proof. Understand the principle of social proof in persuasion, the example we talked about in class, and applications to marketing contexts. 

 Scarcity. Understand the principle of scarcity in persuasion, the example we talked about in class, and applications to marketing topics.