Business Research Flashcards
What is business research, and what ways are there to do business research?
A series of well though out activities that uses data analysis.
- Perform
- Steer
- Evaluate
Name all 6 hallmarks of business research, and their purpose.
- Purposiveness
- Rigor (nauwkeurigheid/precisie)
- Objectivity
- Pasimony (eenvoud)
- Replicibilaty
- Generalizability (mostly for fundamental research!)
What is the difference between fundamental and applied research?
Fundamental research is mostly done by universities. The focus lies on generating knowledge in general (for multiple organizations). Mostly open research.
Applied research focuses on a current problem faced by a manager. It applies to a specific company and is mostly classified.
What is inductive research?
What is deductive research?
Inductive: “From the bottom up”. Start with data and form a theory.
Deductive: “From the top down”. Start with a theory and seek confirmation.
Name the three types of research.
Causal, Correlational, Explorational.
What strategies belong to Causal Research? And to Correlation research and explorational research?
Also state whether the strategy is a deductive or inductive form of research.
Causal: Lab- and field experiments (deductive).
Correlational: Big data, archival- and survey research (deductive).
Explorational: In-depth interviews, focus groups and observation (inductive).
Name the relevant steps when defining a business problem.
- Map your actual and desired state.
- Explore the feasibility. (is it doable, is the focus not too big/small, is data available, are you using variables?)
- Explore the relevance. (is it worthwile/important for managers or academic?)
What is necessary to formulate a good business problem (problem statement)?
- Formulate in terms of variables and relations (between variables).
- It is an open-ended question.
- The question is stated clearly and unambiguously (dubbelzinnig).
- There is managerial and academically relevance.
What makes a good research question? Name three factors.
It should adress a problem of the problem statement.
First theoretical, then practical.
Stated clearly and unambiguously.
Name all three types of theoretical questions.
Context questions (what is…) - only if elaboration is needed.
Conceptualization questions - only key variables that NEED explanation.
Relationship questions (how does..) - explain the relations between variables.
Name two types of practical research questions.
Relationship questions (to what extend does X affect Y, and how much?)
Implication questions (how to implement your results?) - Open!
What is the correct way to give a variable a definition? Give three factors.
- Give an informative name (simple but clear/unambiguously).
- Use a definition without jargon, unless very obvious (based on literature).
- One or two supporting variable references per definition.
EXTRA (4): ALWAYS use the exact same names throughout your report.
Explain dependent, independent and control variables.
The independent variable influences the dependent variable in a positive or negative way (when X changes, so does Y).
A control variable is not the focus of the study, but is included to better understand the effect of a dependent variable on the independent variable.
What is a mediator, and what kinds of mediator are there?
A mediator explains the mechanism between the independent variable (IV) and dependent variable (DV).
Full mediator: X ONLY has an effect on Y through the mediator.
Partial mediator: X has an indirect effect on Y through the mediator, but also has a direct effect on Y.
What is a moderator and what kind of moderators are there?
A moderator ALTERS the effect of X on Y. It can alter the strenght, and sometimes even the direction (positive to negative).
Quasi moderator: Moderates the effect of X on Y, but also effects Y itself.
Pure moderator: Affects the effect of X on Y, but doesn’t effect Y itself.
What is a conditional proces?
An IV and DV with either a moderator or mediator in between.
What is a conditional process model?
A model that contains both mediator(s) and moderator(s).
What makes a good hypothesis?
It is:
- Testable
- Derived from theory (not “gut” feeling).
- Unambiguously phrased.
What is a directional hypothesis? And an unidirectional?
Directional: An indicated direction of the effect.
Unidirectional: There is an effect, but we do not indicate the direction.
How can you justify an hypothesis?
Base it on literature (academic relevance! Don’t use 1 author!)
What steps do you need to take when designing a research?
- Choosing between (deductive) research strategies (type and way of collecting data).
- Choosing between statistical techniques.
- Choosing between sampling designs.
How can you define causality? Name four things.
- X and Y: IV and DV should occur. A change in Y should be associated with change in X. There is a significant correlation.
- There is a logic explanation why X affects Y.
- X precedes Y in time.
- There is NO other cause that explains the co-occurence of X and Y.
What is the difference of desciptive and inferential statistical techniques?
Descriptive: Methods of summarizing data in an informative way.
- Measures of central tendancy: mean, median, mode.
- Measures of dispersion (range, st. d, variance, interquartile range).
Inferential: Methods to draw conclusions (testing the variables)
- Mean difference test (t-test);
- Chi Square test;
- Analysis of variance (ANOVA);
- Regression analysis;
- Logit analysis
Name four types of measurement scales.
- Nominal (using non-numerical groups with no order).
- Ordinal (using non-numerical values that are ranked/ordered).
- Interval (using numerical values with no natural zero point (time, temperature).
- Ratio (using numerical values with a natural zero (body weight, distance, age).
When do you use a t-test?
When comparing MEANS of TWO groups.
IV: Nominal/ordinal.
DV: Interval/ratio.
When do you use an ANOVA?
When you compare MEANS between MORE than two groups.
IV: nominal/ordinal
DV: interval/ratio
When do you use a Chi-square test?
When comparing variances.
IV: Nominal/ordinal
DV: Nominal/ordinal
When do you use a Logit Analysis?
IV: Interval/ratio
DV: Nominal/Ordinal.
When do you use a regression analysis?
IV: Interval/ratio
DV: Interval/ratio
What is the Likert scale?
Completely disagree (1) Partly disagree (2) Neutral (3) Partly agree (4) Completely agree (5)
What is a Semantic differential scale?
Organized _ _ _ X _ Unorganized
What is sampling?
The process of selecting a number of elements from the population. This results in a smaller group which we can draw conclusions for, generalizable for the entire population.
How does the sampling process look? Name the four steps.
- Define the population.
- Determine the sampling frame (physical representation of the population (e.g. database)).
- Determine the sampling design
- Non probability sampling
- Propability sampling - Determine sampling size.
What is probability sampling? What are it’s pros and cons?
Each element (person) has the same chance of being selected as a subject.
Pro: Results are generalizable.
Con: Costly
What is non-probability sampling? What are it’s pros and cons?
The elements of the population do not have the same chance of being selected as a subject.
Pro: Not time and resource intensive.
Con: Less generalizable to population.
Name four forms of probability sampling and explain.
Simple random sampling: each element has the same chance.
Systematic sampling: Every n-th element get’s selected.
Stratified samling: Divide the group into meaningful homogenous groups –> apply simple random sampling.
Cluster sampling: Divide the population in heterogenous groups and randomly select groups and its members.
Name four forms of non-probability sampling. Explain them.
Convenience sampling: Select subjects who are conveniently available.
Quota sampling: Fixed quote per subgroup.
Judgment sampling: select people based on certain elements, form of convenience sampling.
Snowball sampling: Keep asking if participants know other (new) participants.
When developing questions for a survey, what should you avoid? Name four things.
Double/ambiguous questions.
Leading questions.
Loaded questions
Double negatives
What does the item & response model look like?
See summary.
How should you decide where to hold a survey?
See summary.
What factors play a role when choosing the strategy where to hold a survey?
Measurement (interactivity, multi-media, interviewer presence or self-administratation)
Representation (coverage quality, sampling control, response rate)
Economics (sample size, questionnaire size, survey speed, survey costs)
What should the appearance of a questionnaire look like?
- A good introduction (and ending)
- Organizing questions and give guidence where necessary.
- Ask personal data at the end (age, gender, education, etc).
- Conclude the survey!
How do you calculate the response rate?
participants / sample.
How can you increase your respond rate?
Maximize reward of participation (appreciation, tangible).
Minimize cost of participation (time, money, feeling threatened).
Maximize trust of participation (anonimity, well known association, open communication).
What is validity?
Do you measure what you intend to measure?
What is reliability?
Is the data accurate and consistent? (does not have to be the right kind of data, that is validity!)
What is external validity?
Generalizability. Can you generalize the results back to the entire population? (Depends mostly on the sample method)
How can you test reliability in survey questions?
Cronbach’s Alpha (interrelatedness, > 0.7 is good).
How can you manipulate an IV?
- Presence vs Abscence (e.g. bonus vs no bonus)
- Frequency (e.g. high bonus vs low bonus)
- Type (e.g. punishment vs reward).
What is an extraneous variable?
Every other variable that can influence the DV, except for the IV.
What is a cofound?
A variable (Z) that threatens internal validity.
You can control it by including:
- Extraneous variables;
- Control variables.
What is the difference between a lab and field experiment?
A lab experiment is created in an artificial environment. The researchers can control the environment. A field research is a natural environment where manipulation is possible.
Lab experiment: high internal validity.
Field experiment: high external validity.
What is the history effect?
Effects outside of the experiment that influence the DV (e.g. the management team has changed).
What is the maturation effect?
Biological or psycological changes over time (respondent gets hungry, angry, sleepy, etc.)
What is the testing effect?
Prior testing effects on the DV. Pre-measurement influences the results.
What is the instrumentation effect?
The observed effect differs because of a change in the measurement.
What is the selection bias effect?
Incorrect selection of the respondents.
What is the mortality effect?
Drop out of participants during the experiment (e.g. due to frustration).
Name 6 effects that threaten internal validity for lab and field experiments?
History effect Maturity effect Selection bias Mortality effect Testing effect Instrumentation effect
Name three ways to increase internal validity.
Randomization
Design control (control group and extra group)
Statistical control
How does randomization affect the internal validity?
It controls for the history-, instrumentation-, selection bias- and mortality effect.
How do control groups and additional groups affect the internal validity?
Control: controls for history-, maturation-, instrumental- and statistical regression effect.
Additional: controls for testing- instrumental- and the statistical regression effect.
How does statistical control affect the internal validity?
It controls for the history- and selection bias effect.
Name three pre-experimental design groups.
One-shot case study
One-group pretest-posttest
Static group
Name three true experimental designs.
Pretest-posttest control group.
Posttest-only control group.
Solomon four-group.
Name two quasi experimental designs.
Time series.
Multiple time series.
Name three statistical experimental designs.
Randomized blocks.
Latin square.
Factorial.
What are strenghts of archival research? Name four.
Tapping into industry wisdom.
Examining effects across time & countries.
Examine socially-sensitive phenomena.
What is the golden rule of archival research?
Your IV should measure at the SAME or HIGHER unit of analysis level than the DV!
Name an example of the five unit of analysis levels.
Consumer - Brand - Firm - Industry - Country.
Name three reliability problems of archival research.
Missing data.
Inaccurately recorded data.
Fake data.
What is a questionable proxy?
Assuming the wrong measurement for a variable (small plane flyer as a measurement for thrill seeking).
How can you validate whether an archival measurement is a good one?
Provide precedence (has this been done before?).
Provide sound logic.
Provide evidence of a substantial correlation.
Name three advantages of field experiments.
Real world behavior –> high external validity.
It’s authentic (authenticaty) (context, treatment, participants, outcome measures).
Novel insights. Field experiments enable:
- Long-term effects;
- Answer questions that cannot be answered in a lab;
- Checking lab effects in the real world.
Name five disadvantages of field research.
Time consuming.
Challenging to implement.
Focus on observer behavior.
High degree of “noise”.
Ethical consideration.
What is the problem of field research (regarding validity)?
Internal validity is hard to control. A bad internal validity leads to no external validity.
Name 7 internal validity threats of field research.
Unexpected factors.
Poor timing.
Failure to randomize.
Non-compliance/failure to treat.
Insufficient sample size.
Spillovers
Side effects.
Name four ways to increase validity in field research.
Randomization checks (prior and post-hoc).
Unit of randomization.
Power calculations.
Outcome measures.
What is big data?
Big data is a field that uses enormous piles of data and systematically extracts information from that data, which is normally to large or complex for data software.
Name 3 advantages of Big Data.
Enormous amounts of data lead to very specific results.
Always on: collecting data is a constant.
Nonreactive: People normally alter their behavior when they know they are being observed, which isn’t a problem with big data.
Name 7 disadvantages of big data.
Incomplete: Data shows what customers do, but not why.
Inaccessible (in- and outside) due to legal reasons, other firms, ethical reasons or database that are not integrated.
Non-representative: You don’t collect data of a specific group.
Drifting: Big data often changes it’s platform of it’s users, which makes data change with it.
Algorithmically confounded: The design of a platform can influence behavior, introducing bias/noise.
Dirty: Big data is loaded with junk or spam information.
Sensitive: Some information is sensitive, making it inethical to use.
Name three exploratory research strategies.
- In depth interview.
- Focus groups.
- Observation.
What is an interjudge reliability?
Different “judges” or interviewers have a different view on matters. The interjudge reliability is the degree of agreement between judges.
Calculated by Cohen’s Kappa.
Name two validity problems with exploratory research.
Interviewer- and interviewee biases.
Name the nine differences of international research.
Factors:
- Cultural
- Ethnic
- Climate
- Economic
- Religious
- Historical
- Geographical
- Consumption patters
- Research conditions.
