Lecture_1 Flashcards
(32 cards)
What is evidence-based management?
Evidence-based management requires valid empirical research, enabling systematic problem-solving and data-driven decision-making.
What is the key difference between empirical methods and marketing analytics?
Empirical methods:
- Focus on designing and executing studies using primary data
- Emphasize basics of data analysis
Marketing analytics:
- Analyzes secondary (big) data
- Focuses on advanced data analysis techniques
Why is it important to define the research problem clearly?
A clear research problem ensures:
- Relevance to specific issues
- Avoidance of missing critical data
- Focused and actionable results
- no time wasting
What are the three main types of research design?
-
Exploratory: Discovers insights and defines terms -> helps learn more about
the problem, terms and definitions, or
identify research opportunities - Descriptive: Describes market characteristics or functions -> describes the phenomena of interest
-
Experimental: Determines cause-and-effect relationships -> uncovers underlying
causes of a problem.
Each type of research design serves different objectives, with exploratory research helping to define problems, descriptive research providing detailed insights, and experimental research testing hypotheses through controlled conditions.
What is the difference between correlation and causation?
Correlation: A statistical relationship between two variables.
Causation: One variable directly influences another.
- Example: Increased cheese consumption correlates with deaths by bedsheets but does not cause it.
What ethical principles guide empirical research?
- Voluntary participation: No coercion.
- Informed consent: Participants understand risks.
- Anonymity/Confidentiality: Protect identities and data.
- Anonymity: Assure participants that no one, including yourself, will be able to link the
data to a specific individual (Not always possible. Then assure at least confidentiality). - Confidentiality: Assure participants that identifying information about them acquired
through your study will not be released to anyone outside the study.
No data fabrication or manipulation: - Be transparent about the data you collected
and how you collected it - Always report your data, your analysis and
your results objectively and honestly - Present your results in such a way that they
do not cause embarrassment, disadvantage
or harm to any of the participants - Do not distort data to fit your purpose or
favor a particular result or target group
What is the null hypothesis?
The null hypothesis (H₀) assumes no effect or difference. Conservative (devil’s advocate) position (“no difference”).
- Example: An ad campaign has no impact on sales.
What is the alternative hypothesis?
The alternative hypothesis (H₁) predicts an effect or difference. Expected to deliver new insights.
- Example: An ad campaign increases sales.
What are the deductive and inductive approaches in research?
-
Deductive: Starts with theory, narrows to specific hypotheses, confirms through data.
-> Theory first: In research, we often follow a deductive approach: we start with a broad
literature review and theory, narrow it down to specific hypotheses and test these
through the collection of data.
→ more confirmatory in nature! -
Inductive: Begins with observations, identifies patterns, forms hypotheses.
Data first: The inductive approach works the other way around: we move from very specific
(“interesting but unexplained”) observations in our data to a detection of patterns,
up to a formulation of tentative hypotheses and a theoretical framework.
→ more exploratory in nature!
Most projects involve both inductive and deductive reasoning at some point during
the process.
Empirical Research and Big Data
- Through advancements in technology, firms are increasingly collecting enormous amounts
of data → BIG data - But: Full of “noise”, you often cannot narrow your results down to the one factor causing
your problem of interest. - In the vast majority of cases, central constructs of interest (like firm performance, customer satisfaction, or
repeat purchase) can be explained by a small number of characteristics; big data may even mask the central
explanatory variable(s) - Overspecification may lead to seemingly “exact“ models with relatively poor predictive accuracy
- Often, constructs of central importance are NOT captured in big databases which are collected for other purposes (or for none at all)
constructs of central importance:- Marketing: Brand loyalty, consumer perception, purchase intention.
- Business & Management: Leadership effectiveness, job satisfaction, employee engagement.
central constructs of interest:
* Marketing: Brand loyalty, customer satisfaction, purchase intention.
* Business & Management: Organizational culture, employee engagement, leadership style.
Constructs of central importance are more general and fundamental to a field of study.
Central constructs of interest are more specific and directly tied to a particular research project.
- Then why still worry about collecting additional (primary) data yourself?
- Almost no “big” data provide causal evidence – why?
- Correlation (association) is not causality
- Central and often occurring problems of secondary observational data are:
- Endogeneity (in the form of sample selection, unclear causal relationships, important missing
variables, etc.) - Identification issues (problem of divergence (=Verschiedenheit/Abweichung) between measurement and conceptualization of a
variable (e.g., using repeat purchase as indicator of customer loyalty) - P-value problem:
- P-values in larger samples are more likely to be zero
- Statistical significance ≠ practical significance
What is Endogeneity
Endogenität bezieht sich auf eine Situation in der empirischen Forschung, bei der eine Variable sowohl Ursache als auch Wirkung innerhalb eines Modells ist, was zu verzerrten und inkonsistenten Ergebnissen führt. Dies tritt auf, wenn:
Ausgelassene Variablen: Eine fehlende Variable beeinflusst sowohl die unabhängige als auch die abhängige Variable, was zu einer scheinbaren Beziehung führt.
Beispiel: Die Kundenbindung wird ignoriert, wenn man Wiederholungskäufe untersucht.
Gleichzeitige Kausalität: Die unabhängige Variable beeinflusst die abhängige Variable, aber gleichzeitig beeinflusst die abhängige Variable auch die unabhängige.
Beispiel: Der Umsatz beeinflusst die Werbeausgaben, und Werbung beeinflusst wiederum den Umsatz.
Messfehler: Fehler bei der Messung von Variablen führen dazu, dass ungenaue Beziehungen geschätzt werden.
How to Conduct Empirical Research
- Define the Problem (if we start here: research trigger = classical approach (problem-driven, deductive, explore or confirm, phenomena/theories))
- Determine Research Design
- Data Collection Method and Form
- Design Sample and Collect Data
- Analyze and Interpret Data (if we start here: research trigger = big data (data-driven, inductive, explore, data sets))
- Prepare the Research Report & Presentation
Problem vs Opportunity
- Problems may be found in practice and in prior literature of a certain field.
- A problem exists when a gap exists between what was supposed to happen and what did happen,
i.e., failure to meet an objective or expected result.
➢ This is what we normally think of when we think of “a problem” - An opportunity occurs when there is a gap between what did happen and what could have happened
➢ This is called an opportunity because the situation represents a favorable circumstance or chance for progress or advancement
➢ For example, our sales were $X but could have been $Y had we introduced a new, more
competitive product
Problem vs. Symptom
- Be aware that the symptoms are not the problem but the signals that alert us of
the problem. - Symptoms: Sales are falling; employee satisfaction is down; product returns are
up; grades are down…. - Knowing prior literature and / or exploratory research can help identify possible
causes of the problem
Literature Review
It is critical to have an explicit inclusion and exclusion criteria (e.g., key variables, research
methods, research respondents).
* Include or exclude low quality studies? Consider findings with caution.
* Do the review early
* The literature review can help you
* Find your gap: Has the problem already been addressed?
* Theorize: How are key constructs related and which other
factors influence them?
* Operationalize: Which instruments and measurements?
Research question -> literature review -> hypothesis
Hypothesis
Stating the relationship between two
variables, the causing factor (independent
variable) and the affected factor
(dependent variable)
Practical vs. theoretical research
Practical empirical research should be hypotheses-driven just as much as theoretical research!
Main difference – practical research is often broader.
Theoretical research focuses on developing, exploring, and understanding theories, concepts, and ideas. It is not directly applied to solve real-world problems but aims to expand knowledge and understanding of a subject.
Practical research (also called applied research) aims to solve specific, real-world problems by applying theories, concepts, or empirical methods.
Types of Hypotheses
difference hypotheses:
“Ad campaign 1 results in lower sales than ad campaign 2.”
Types of Hypotheses
coherence hypotheses:
“The higher the ad spending, the higher the sales”
-> in the same direction
Types of Hypotheses
one-tailed (the direction of the mean difference or the correlation is
part of the hypothesis):
“Ad campaign 1 results in lower sales than ad
campaign 2.”
Types of Hypotheses
two-tailed (no assumption about the direction of the mean difference or
the correlation):
“The factor ad campaign has an influence on the sales volume”
Two options of research designs
Cross-Sectional Design
Longitudinal Design
