Exam preparations Flashcards
What is field notes?
Field notes are a qualitative research tool used to record observations, thoughts, and reflections during or after fieldwork. Field notes are a qualitative research tool used to record observations, thoughts, and reflections during or after fieldwork.
Often used from ethnographic studies. other things just that the verbal communication that is of interest.
Be careful with observer bias.
What is archival data?
A structured source of information that exists independent
of the researcher (one type of secondary data)
Not an objective mirroring of reality
Fragments and lost pieces
Pay attention to time (when written) and purpose (why
written) of a certain documen
What is grounded theory?
Grounded theory is a qualitative research methodology focused on developing theories directly from data rather than testing pre-existing theories.
Important features:
Data-driven
Constant comparision
Open-ended and minded
Often inductive, can be deductive.
conceptualisation of underlying patterns
What is trustworthiness and how can we make the research trustworthy?
what is trustworhtiness and how can we make the research trustworthy?
key compomemts:
credibility
transferability
dependability
What is ontology?
How we view the nature of reality. Assumption about the nature of reality.
“Does gravity exist?” Yes! our view on the world; how do we look upon reality
What is epistemology?
How knowledge is acquired.
“How do we know gravity exists?” through evidence!
Abductive reasoning can best be described in the following way:
An iterative process between data and theory
Can regression analysis detect causation?
Yes, with a casual research design.
What is a population regression function?
describes the relationship between a dependent variable (outcome) and one or more independent variables (factors) for the entire population.
ex Imagine you want to know how study hours (independent variable) affect exam scores (dependent variable) for every student in the world.
How can you check for linearity and homoskedasticity?
Scatterplots
How can you check for multicollinearity?
Correlation matrices
How can you check for autocorrelation?
Durbin-Watson
How can you check for normality?
Shapiro-Wilk and Kolmogoro-Smirnov
What are the consequences when the Linearity assumption is violated and what is the solution?
When the linearity assumption is violates, the relationship between X and Y is not linear.
Consequence: biased estimates
Fix: Use a nonlinear regression model
What are the consequences when the Homoskedasticity assumption is violated and what is the solution?
When its violated we have heteroskedasticity.
Fix:
- Transform the covariates prior to regression (log transformation)
- Robust Standard Errors such as White Standard Errors for larger samples
What are the consequences when the No Perfect Multicollinearity assumption is violated and what is the solution?
In that case we have multicollinearity: independent variables are highly correlated. This does not violate the assumption of “no linear dependence” because multicollinearity is not perfect collinearity.
Fix: Remove or merge correlated variables. Increase sample size as it increases SST.
What are the consequences when the Normality of Errors assumption is violated and what is the solution?
The residuals (errors) are not normally distributed. Coefficient estimates remain unbiased, but hypothesis tests (e.g., t-tests, F-tests) may be invalid, especially in small samples.
Fix: Use non-parametric methods (which are not dependent on the normality assumption), or check if large samples mitigate this issue (Central Limit Theorem). Or use large samples.
What are the consequences when the “No Autocorrelation” assumption is violated and what is the solution?
No autocorrelation = No Independence.
autocorrelation is just the term for when the independence assumption in ols regression is violated.
So, when violated, errors are correlated across observations.
Fix:
- Transform the covariates prior to regression. Re-specify model to incorporate path-dependency.
- Use Robust Standard Errors (like Newey-West) to correct for the issues in residuals.
It is necessary for independent and dependent variables to be normally distributed?
No - the independent and dependent variables do not need to be normally distributed. Regression can handle variables of any distribution, like skewed.
For the errors and residuals, yes. That is an assumption in OLS regression.
How can we detect non-normal distributions of the residuals? (3)
- Histogram. Non-normality appears as skewed distributions or outliers.
- Q-Q Plot. Compares the distribution of residuals to a normal distribution. Points should align along a straight diagonal line if residuals are normal. Deviations from the line indicate non-normality. It creates a smiley because of heteroskedasticity.
- Shapiro-Wilk/Kolgomorov-Smirnow
What is a good fit of a regression and how can we measure it?
A good fit = The model explains a large proportion of the variability in the dependent variable. Residuals (differences between observed and predicted values) are small and randomly distributed.
The model meets assumptions (e.g., linearity, independence, homoscedasticity).
Predictions are accurate for the data.
Measured by R squared. It ranges from 0 to 1. 1 = Perfect fit.
What are outliers and how can we mitigate their impact on a regression?
Outliers are extreme values. We can detect them via scatterplots or boxplots ex.
Three ways to handle it:
Transforming (reduce impact)
Trimming (remove them)
Winsorizing (replace them)
What is the benefit from simple regression → multiple regression?
A simple regression only accounts for one independent variable to explain the dependent variable. Reduces omitted variable bias.
What is Zero mean of the residuals?
Refers to the overall average of the residuals across all observations.
What is Zero conditional mean?
Exogeneity. Ensures that the independent variables are uncorrelated with the error term u.
The covariance between independent variables and the residuals is zero.
What is constant variance of the residuals?
Homoskedasticity
What is omitted variables and its consequences?
Omitted variables are important factors that influence the dependent variable Y but are not included as independent variables X in the model. This creates omitted variable bias.
How can you address omitted variables?
- For panel or longitudinal data, fixed effects can control for omitted variables that are constant within individuals or groups.
- Randomized Control Trials
What is overspecification and its consequences?
Overspecification occurs in regression analysis when the model includes too many independent variables, some of which are irrelevant or redundant. These extra variables do not improve the model’s ability to explain the dependent variable Y and can even harm its performance.
Consequences: Multicollinearity
Do dependent and independent variables need to be normally distributed?
No.
Vi kan ju ha en bra modell trots tex negative skewness.
What is a Q-Q plot?
A Q-Q plot (Quantile-Quantile plot) is used to compare the distribution of a dataset to a theoretical distribution (e.g., normal distribution) by plotting their quantiles.
Assess if residuals from a regression model are normally distributed.
What is a scatterplot?
It’s used to show the relationship between two variables by plotting their values as coordinate points.
Ex: Explore if study hours X and test scores Y have a linear relationship.
What is sampling error and how can you solve it?
Sampling error is the difference between a sample statistic (e.g., sample mean, sample proportion) and the corresponding population parameter (e.g., population mean).
For example:
You survey 1,000 people from a city to estimate the average income. The sample mean might differ from the actual population mean due to sampling error.
SOLVE IT BY INCREASING THE SAMPLE SIZE
What is random sampling?
Ensure the sample is chosen randomly so every member of the population has an equal chance of being selected. This reduces the likelihood of systematic bias.
What is stratified sampling?
Divide the population into subgroups (strata) based on characteristics like age, income, or region, and take random samples from each.
Ensure that each subgroup is proportionally represented in the sample.
What is cluster sampling?
Divide the population into clusters (e.g., geographic areas or naturally occurring groups) and randomly select clusters for sampling. Makes data collection more efficient by sampling whole clusters instead of individuals.
What is a quasi-experiment and how is it different from RCT?
A quasi-experiment lacks the random assignment of participants to treatment and control groups, which RCT has.
What is difference-in-difference?
Compares outcomes between a treatment group and a control group before and after an intervention.
What is panel data?
Panel data (also known as longitudinal data) is a type of dataset that contains observations of multiple entities (such as individuals, firms, countries, etc.) over multiple time periods. It combines elements of cross-sectional data (data collected at one point in time) and time-series data (data collected over time for a single entity).
What is fixed effect?
Fixed effects control for factors unique to each entity that do not change over time (time invariant, culture)
Its cross-sectional but with fixed time.
Time-invariant factors are automatically controlled in FE for because these models focus only on within-entity variation over time.
Fixed Effects (FE) is a method used in panel data analysis to control for unobserved factors that don’t change over time but could influence the dependent variable.
Key idea: It focuses on changes within each individual (like a person, firm, or country) over time, removing the effect of time-invariant characteristics.
Simple Steps:
1. Remove what doesn’t change: Fixed effects filter out any constant characteristics of individuals (e.g., gender, long-term preferences).
2. Focus on within variation: The analysis only uses differences inside each individual unit over time to estimate relationships.
3. Control for bias: This avoids bias from factors that are constant over time but differ between individuals.
Example:
Imagine studying how work hours affect productivity across workers. Workers might differ in skills (constant over time). Fixed effects handle this by only looking at how changes in work hours for the same worker impact their productivity, ignoring cross-worker differences in skills.
When to use:
- When you suspect unobserved variables (e.g., personality, location) might bias your results, but these factors don’t vary over time.
——————————————-
1. Students and Test Scores
Question: How does study time affect test scores?
Problem: Some students are naturally smarter than others, which might bias the results.
Fixed Effects Solution: Compare each student’s test scores across different tests they take.
This removes the effect of natural intelligence (which doesn’t change over time).
Focuses only on how changes in study time for the same student affect their scores.
2. Cities and Pollution
Question: Does traffic volume increase air pollution in a city?
Problem: Some cities are naturally more polluted than others due to geography or industry.
Fixed Effects Solution: Look at the same city over time and how changes in traffic volume affect pollution.
This removes constant factors like geography.
Focuses only on how variations in traffic affect pollution within each city.
What is random effect?
Time-invariant variables can be estimated because random effects assume that these factors are uncorrelated with the independent variables.
Is one way panel regression model the same thing as fixed effect?
Not exactly. A one-way panel regression model can be a fixed effects model, but it can also be a random effects model depending on how the unobserved effects are treated.
What is One-Way Panel Regression Model?
A one-way panel regression model accounts for unobserved effects that vary only across entities or only across time, but not both.
What is a Two-Way Panel Regression Model? (TWFE)
A two-way panel regression model accounts for unobserved effects that vary both across entities and over time. It includes both entity fixed effects and time fixed effects, making it more comprehensive.
Is randomization used for Difference-in-Difference?
Usually no, because its not needed.
Why is an event study considered a quasi-experiment for casual inference?
An event study is considered a quasi-experimental design because it has no randomization. It relies on naturally occuring events rather than a random assignment of treatment.
What is casual inference?
Causal inference is the process of determining whether and how a change in one variable (the cause) directly influences another variable (the effect). It aims to go beyond observing correlations or associations by establishing a cause-and-effect relationship between variables.
Causal inference revolves around estimating the treatment effect by comparing observed outcomes to counterfactuals, often using control groups or statistical methods to approximate the unobservable counterfactual.
What is a contrafactual?
A counterfactual is the hypothetical scenario representing what would have happened in the absence of the treatment.
What is a staggered rollout design?
In a staggered rollout design, the treatment or intervention is introduced to different entities at different points in time.
Ex. Different regions of the country implement the smoking ban in public places at different times:
Region A: January 2023
Region B: June 2023
Region C: December 2023
What is a non-staggered rollout design?
In a non-staggered rollout design, the treatment or intervention is introduced to all treated entities at the same point in time.
Ex. A country implements a nationwide smoking ban in public places on January 1, 2023, applying to all regions simultaneously.
What is event time?
Event time refers to a time scale centered around a specific event or intervention (e.g., a policy implementation, a treatment, or a natural disaster).
t = 0
Day 0: The day of the merger (event).
Time is measured relative to this event (e.g., Day -5, Day +1) (typ som att vi räknar 2025 för att något hände för 2025)
What is calendar time?
Calendar time refers to the actual date or time period when an observation occurs, using a universal, chronological scale (e.g., years, months, days).
Time is measured in absolute terms (e.g., January 2022, Q3 2023). Alltså vanlig tid vi snackar om.
What is pre-event window?
time periods before the event occurs.
Pre-Event Window: Directly examines what happens before the event.
do not confuse it with estimation window:
Estimation Window: Provides a benchmark for “normal” returns, used to calculate abnormal returns.
What is the event window?
Same as event time. refers to the time period during which the event occurs, t = 0
What is post-event window?
The post-event window includes the time periods after the event occurs
What is typically the contrafactual in a CAPM study?
What would have been if we didn’t have the treatment?
In the context of CAPM, the counterfactual is indeed what the return on an asset would have been if CAPM perfectly explained returns.
Only accounts for market risk (systematic risk), not firmspecific risk.
What is TWFE? (Two-Way Panel Regression Model)
Two-Way Fixed Effects accounts for both entity-specific and time-specific unobserved factors. It isolates the effect of interest by controlling for unobserved heterogeneity
How does a calendar-time portfolio method differ to an event study method?
The calendar-time portfolio method analyzes portfolio returns over regular calendar periods (e.g., monthly or yearly) to evaluate long-term performance or strategies. For example, you might group stocks into value and growth portfolios based on their book-to-market ratios and track their average returns month by month to see if value consistently outperforms growth.
The key difference is that the calendar-time method looks at overall trends across time, while the event study method zeroes in on the effects of a particular event.
What do we mean by portfolio sorts?
Portfolio sorts refer to a method used in finance to group assets (e.g., stocks) into portfolios based on certain characteristics or metrics, such as size, value, momentum, or other financial variables.
Once asset are sorted, they are grouped into portfolios and divided into quantiles or other bins.
What is the benefit to using portfolio sorts over a panel linear regression method?
Portfolio sorts are simple and intuitive, grouping assets into portfolios based on a characteristic (e.g., size or value) and comparing returns. They are robust to outliers, non-parametric (no linearity assumptions).
Panel regressions allow for controlling multiple variables and provide more precise estimates but can be sensitive to outliers and require assumptions about linearity.
What is some of the draw downs of portfolio sorts, and how might we mitigate them?
Sorting bias: sorting stocks into portfolios can be biased.
Fix: Larger data set.
Also:
Portfolio sorts difficult to apply with more than two factors.
Fix: Allow TWFE?
What is conditional portfolio sorts?
Assets are sorted into portfolios based on the primary characteristic while controlling for another characteristic.
Sorting order is important.
Först dela upp i stor, mellan, liten.
What is unconditional portfolio sorts?
Each parameter is independently sorted.
Assets are sorted into portfolios based on a single characteristic without considering any other variables.
The positivist perspective is characterized by objectivity, that the researcher and what they are studying are (DEPENDENT OR INDEPENDENT) of each other, and data is primarily quantitative.
Independent
What is the role of the research question?
To motivate and focus the research.
In social sciences, research questions may…
Motivate theory development
When evaluating a research question, it is important that it:
Is relevant, researchable, and represents a gap in knowledge.
What is a cross-sectional research design and how does it relate to causality?
Data is collected at a specific point in time from a cross-section of respondents, so concluding causal inference is weak.
What is the research design?
It is the plan for how a research project will be conducted.
With respect to Professor Wedlin’s lecture on research design, which statement DOES NOT describe why a research design is needed? A research design:
Facilitates the formation of a research question that is relevant and researchable.
Considers strategies and choices for what data to collect, how to collect it, and how to assess it.
Locates the study in a particular knowledge domain.
Turns a research question and objectives into a project.
Facilitates the formation of a research question that is relevant and researchable.
What is a natural experiment?
Tänk natural experiment: CAUSALITY
An event occurs to a specific group of people outside the control of the researchers, but in such a way as to resemble random assignment.
Data is collected from before and after the event, and causality is established.
How could you describe a research question and related hypotheses?
Both research questions and hypotheses must be stated such that there is at least one testable alternative.
A research question may be answered through related hypotheses that are stated as questions.
You may answer a research question by posing hypotheses. The hypotheses must be empirically testable.
Quantitative research questions are meant to be very specific, with a set of general hypotheses, which in turn answer the research question.
You may answer a research question by posing hypotheses. The hypotheses must be empirically testable.
Constructs can have theoretical definitions and operational definitions. What is the purpose of operationalizing a construct?
Operationalization specifies the theoretical domain of a construct.
Operationalization specifies constructs within the accepted structure of hypotheses.
Operationalization specifies how a construct will be measured.
Operationalization specifies how a construct will be analyzed.
Operationalization specifies how a construct will be measured.
The interview method is appropriate when:
The researcher wants to get deep knowledge about a certain phenomenon.
Which answer would best fit the inductive approach?
The researcher discovers that not much empirical work has been done on a specific topic. Then, based on what theory she finds, decides to investigate.
The researcher is testing a theory by collecting data that is found to be most suitable given the existing theory.
The researcher is using an approach where he/she structures the findings according to theory.
The researcher, through a literature review, finds that there is a lack of theory explaining a certain phenomenon. Then, decides to investigate.
The researcher, through a literature review, finds that there is a lack of theory explaining a certain phenomenon. Then, decides to investigate.
The method chapter:
Is a way of showing, and giving reference to, the theory of methodology, such as what ontology is, and when to expect different ontological arguments.
Is not the most important chapter of a research paper, compared to, for instance, the theoretical chapter.
Should in detail describe how the research has been conducted, such as who did what, how, and when.
Should briefly cover how the research has been conducted without going into detail into exactly who did what and when.
Should in detail describe how the research has been conducted, such as who did what, how, and when.
When creating a questionnaire, what is a good way to make sure you properly cover the dimensions of each construct?
Start by looking at existing questionnaire on similar topics or theories.
Use existing questionnaires, but adapt them enough to avoid plagiarizing other researchers.
Avoid using existing questionnaires so that you do not plagiarize other researchers.
Read everything you can about the theoretical context so that you can then make the questionnaire.
Start by looking at existing questionnaire on similar topics or theories.
What is recall bias?
You only remember certain things better than other which may bias your answers.
What is self-selection bias?
conducting a survey on companies and sustainability, and only the companies interested in that area joins
Assume that there are 500 people in a population and they are all on a list. You want a random sample of 50. You add the first 10 names to a hat and have a friend randomly pick one of the names. Starting with that name, you take every tenth person thereafter. In this way, you get a sample of 50. What kind of sample is this?
Probability sampling
What is non-probability sample?
not all members of the population have a known or equal chance of being included in the sample. aka no random selection.
What is the sampling frame?
EVERYONE
What is the sampling list?
sampling list is a practical subset of that frame used for selecting the actual sample in a study
Validity and reliability are important in science. With respect to measurement, what are they?
Validity is how well a measure reflects what it intends to measure, and reliability is about the consistency of measurements.
Which statement best describes qualitative data coding?
It is the process of organizing and labeling data.
When you code your data according to the Gioia, Corely & Hamilton method, which coding approach is most appropriate?
You move from data-text (empirics) to higher analytical levels by aggregating and condensing.
(Condense = To reduce large amounts of data or information into a more concise and summarized form while retaining the most essential parts
Aggregating = To combine or group multiple pieces of information, data points, or observations into larger units or categories based on shared characteristics.)
Which type of reliability assessment is associated with qualitative analysis?
R-squared.
Cronbach’s alpha.
Inter-rater reliability.
The Levene test for homogeneity of variance.
Inter-rater reliability
In an OLS regression analysis, based on the Coefficients table, if you were going to write out a regression equation so that a person could calculate, “for a given value of X , Y would equal ___”, you would use the:
Unstandardized beta coefficients for the statistically significant independent variables.
In regression, what is the error term?
Residual variance that is not explained by the regression coefficients.
With respect to simple OLS regression, which of the following statements is correct?
You can use dummy variables as the dependent variable.
The constant indicates the slope of the regression line.
The beta coefficient for the X variable indicates the slope of the regression line.
A dummy variable changes the slope of a regression line.
The beta coefficient for the X variable indicates the slope of the regression line.
In regression, what is the residual variance?
It represents all the errors explained by the independent variables.
It is the cumulative variance of errors made when measuring the independent variables.
It is the cumulative variance of errors made when collecting the data.
It is the variance that is not explained by the regression coefficients
It is the variance that is not explained by the regression coefficients
Why is the Adjusted R Square different from the R Square?
It adjusts downward for each additional independent variable.
How would you interpret the “R Square” statistic?
It is the explained variance in the regression equation.
What information do you get from the standardized beta coefficients?
The relative effect size of each independent variable on the dependent variable.
In regression, specification error refers to including irrelevant independent variables, not including important independent variables, or choosing the wrong functional form. In layman terms we talked about too long and too short models. Which of the following statements is true?
A too long model reduces the precision of the beta coefficients, whereas a too short model causes a systematic bias in the parameter estimates.
With respect to OLS regression, which of the following statements is correct?
In simple regression (one X variable), the standardized beta is the slope of the regression line.
A dummy variable changes the slope of a regression line.
You can use dummy variables as the dependent variable.
The unstandardized beta constant indicates the Y-intercept of the regression line.
In simple regression (one X variable), the standardized beta is the slope of the regression line.
What is a normative research question?
What should happen?
What is a descriptive rq?
What is happening?
What is a exploratory rq?
What might be happening?
What is an explanatory rq?
Why or how is it happening?
What is action research?
Research question/purpose: To Design/Control
Researcher: Attached inside
What is collaborative basic research?
Research question/purpose: To describe/explain. Co-Produce Knowledge with Collaborators.
Researcher: Attached Inside
What is design and evaluation research?
Research question/purpose: To Design/Control. Normative questions.
Researcher: Outside
this form of research goes beyond describing or explaining a social problem, but also seeks to obtain evidence-based knowledge or relative success. Evaluation researchers typically take a distanced and outside perspective of the designs or policies being evaluated. Inquiry from the outside is necessary.
What is informed basic research?
Research question/purpose:
To describe and explain a social phenomenon. Basic Science with Stakeholder Advice.
Researcher: The researcher is a detached outsider of the social system being examined. The researcher directs and controls all research activities.
How do we gain knowledge according to a constructivist paradigm?
Through collecting survey data
Through interaction and reflection.
Through interpretation and objectivity.
Through fact and reflection.
Through interaction and reflection.
Jane did a qualitative case study of a company de-merger process, the one between the confectionary producers Fazer and Cloetta that took place in 2008. Data was collected during 2017 from multiple sources, including documents, newspapers and interviews. Once she started to analyze the data she realized that the information she received from her interviewees did not correspond with the data published in newspapers and documents. What potential source of bias does she have to reason about in her methods section?
Recall bias
Interviewer bias
Errors in recording
Measurement bias
Recall bias
A longitudinal study is a study that:
Is used to map change, and to understand the mechanisms whereby change happens.
Is used to capture an isolated incident at one given point in time.
Is used to map change, and assess the most desired outcome of that change.
Is used to capture causal-effect relationships at a specific point in time.
Is used to map change, and to understand the mechanisms whereby change happens.
What counts as evidence in an academic argument?
Findings from previous research and investigations.
Authoritative statements and established knowledge.
All three answers.
Own research, data and analyses.
All three answers.
How would you best characterize the following research question?
“The aim of this paper is to examine the communication process of organizations by investigating the potential relationship between corporate social responsibility (CSR) and the quality of their financial reporting.”
Correlational
Exploratory
Casual
Descriptive
Correlational
Variance and process approaches are two main types of research in business studies. What characterizes these two approaches?
A variance approach is characterized by analysis of variables seeking answers to what causes what, whereas a process approach is characterized by studying interrelational between events seeking answers to how things develop and change over time.
Case study research can be many things. Typically, we think that case studies contain the following main characteristics:
Delimited in space and time, quantitative, context independent.
Longitudinal, context independent, and in-depth.
Delimited in space and time, in-depth, and context dependent.
Boundaryless, longitudinal, quantitative.
Delimited in space and time, in-depth, and context dependent.
According to Van der Ven (2007), there are several critical steps in crafting a research question. Which are the most important steps?
Identifying, measuring and selecting a question.
Problematizing, operationalizing the measurment and defining a question.
Arguing, grounding, operationalizing and defining a question.
Situating, grounding, diagnosing the problem and selecting a question.
Situating, grounding, diagnosing the problem and selecting a question.
In her study of board meetings Pernilla realizes that some of the board members seem to be more active than others in the discussions, and seem to be given much more space (relative others) to talk and to present their views and opinions. Particularly, she notices, female board members are called on much more often than men to provide input on financial statements. What additional information would she have needed/could she have collected to be able to do a trustworthy analysis of power structures during board meetings?
Interviews with all board members.
Taking field notes on who said what and when.
Getting background information on all board members.
All three answers.
Taking field notes on who said what and when.
In conducting an instrumental case study, you are primarily interested in capturing:
A specific phenomenon present in the case.
A specific instrument used by individuals in the case.
All three answers.
A specific organization studied in the case.
A specific phenomenon present in the case.
(An instrumental study is used to explore or understand broader phenomena by focusing on a specific case, event, or example. The primary goal is not the case itself but the insights it provides about a larger issue or theory.)
n Rennstam & Wästerfors (2018), three problems encountered when dealing with qualitative data; problem of chaos, problem of representation, and problems of authority. The latter, problems of authority, relate to:
The difficulty of getting authoritative respondents to interview and to provide truthful statements.
The difficulty in asserting ourselves in relation to other researchers and experts, or determining what we can claim is new and relevant with our research.
The difficulty of generalizing from qualitative data, or finding what is representative of the population.
The difficulty of developing new theory, and finding cause and effect relationships.
The difficulty in asserting ourselves in relation to other researchers and experts, or determining what we can claim is new and relevant with our research.
The concept of praxeology is described in Tsoukas and Chia (2011), but what is it? Praxeology is concerned with:
How knowledge is created among practitioners.
How theory provide solutions for practitioners.
How praxis develops, and how solutions are found.
How knowledge is related to action, and how theory is related to practice.
How knowledge is related to action, and how theory is related to practice.
How would you best characterize the following research question?
“This paper aims to map the experiences of “extreme sporting managers” views on their leadership practices.”
Normative
Casual
All three
Descriptive
Descriptive
When formulating a research question, it is according to Van de Ven (2007) important that it:
Is analytical, researchable, and permits more than one answer.
Includes the central constructs that will be studied in the research project.
Is stated in such a way that the reader understands whether it will involve qualitative or quantitative research.
Is creative, interesting, and contributes to society.
Is analytical, researchable, and permits more than one answer.
Reliability in quantitative research is best described in the following way:
Replicability.
Clearly stated variable measurements.
All three answers.
Clearly stated sample.
All three answers
In evaluating qualitative research, we seek the following quality criteria:
Trustworthiness, credibility and transferability.
Transferability, variable measurements, reliability.
Credibility, replicability and transferability.
Trustworthiness, credibility and replicability.
Trustworthiness, credibility and transferability.
Engaged scholarship can be practiced in many different ways and for different purposes. Van der Ven (2007) describes alternative forms of engaged scholarship. Informed basic research is best described as:
Undertaken to describe, explain or predict a social phenomenon.
What is methodology?
All three answers.
Includes choices of epistemology and ontology.
Includes a way of conducting research.
Includes techniques for collecting and analyzing data.
All three answers.
Charting the territory means to learn about the topic area, which includes looking for previous conceptual papers and research studies. In social sciences, research questions:
May motivate theory development.
What is the purpose of a nomothetic study?
Establishing general laws and empirical generalization.
With Poppers idea of falsification in mind, when is a theory scientific?
When it is has been proven to be true using empirical data.
When it is testable using empirical data.
When no one has been able to show that it is false.
When there are no alternative hypotheses that can be true.
When it is testable using empirical data.
How would you characterize the following research question? “How do subsidiary managers approach non-routine problem solving processes?”
Correlational and normative.
Cognitive and exploratory.
Normative and descriptive.
Descriptive and exploratory.
Descriptive and exploratory.
According to Rennstam & Wästerfors (2018), what are the three most common types of problems that qualitative researchers encounter in their research?
Problem of chaos, representation and authority.
What kind of data can be collected through a survey?
Both qualitative and quantitative data
What are the most important features of a qualitative case study?
A collection of information from a sample of individuals in an organization.
A clear focus on an organization and an intense examination of its aims and goals.
A method of gathering information using relevant questions from a sample of people with the aim of understanding populations as a whole.
A clear focus of a bounded situation or system, and intense examination of the setting
A clear focus of a bounded situation or system, and intense examination of the setting
Replicability is important in some research. Why?
In order for others to be able to conduct a similar study.
What is the key principle for first order analysis?
An analysis focusing on the terms and labels that are found in the empirical data.
NOT quotes and examples
Is casual research design the same as explanatory research?
Yes
How do you reach a good research design?
Have it connected to the research question –> good qualitative research design.
Which are the assumptions needed to fit a linear regression model?
- Homoskedasticity
- Normality
- No Autocorrelation
- Linearity
What happens when we fit regression with non-linearity? (Q-Qplot)
It creates a smiley because of heteroskedasticity
What is the Central Limit Theorem?
states that when you take sufficiently large random samples from a population with any distribution (e.g., uniform, skewed, or normal), the sampling distribution of the sample mean will approximate a normal distribution
How can we mitigate non-normality of the residuals?
- Larger sample size
- Deflate the variables, normalize the x variable so its non dependent of the size
- Take the natural logarithm
How do you set up a a capital market-based event study?
- Define the event and establish the event window. Should be short. usually 3 day = -1, 0, +1 cause the info to the stock market can be early or late
- Define the estimation window.
120 days - his example - Define the post-estimation window
Typically not very interested in this
as opposed to regression - Establish the firm selection criteria.
Make sure that the shares are frequently traded during the event window. Frequent trading ensures accurate pricing: If shares are not frequently traded, the observed price may be outdated and not reflect the true market value during the event
Example: A stock that trades infrequently might show no price change during the event, not because the event had no impact, but because there were no trades to update the price. - Estimate the model parameters using data in the estimation window.
alpha and beta hat - Measure the abnormal returns for the shares in the sample.
CAR - Conduct tests.
Define null and alternative hypotheses. Measure the abnormal returns. Determine levels of significance for tests. - Present results and diagnostics
- Interpret results and draw inferences and conclusions
How do you set up a TWFE event study?
controlling for entity-specific and time-specific fixed effects.
Focused on both dimensions: Units and time. Event study needs both unit and time.
More in detail how to set it up:
To set up a two-way fixed effects (TWFE) study, you aim to account for both entity-specific (e.g., individuals, firms, or countries) and time-specific effects (e.g., year, month). Here’s a step-by-step guide:
- Define Your Research Question
Identify what you want to study, such as the effect of a policy or treatment on an outcome. - Collect panel data.
Your dataset should have repeated observations for each entity over time (e.g., regions across multiple years).
Dependent variable (Y): The outcome of interest (e.g., employment rates).
Independent variable (X): The main variable of interest (e.g., minimum wage level).
+ Other factors that might affect the outcome (e.g., GDP, population).
+ Entity ID (e.g., region) and time period (e.g., year). - Specify the model
- Include Fixed Effects
- Ensure Robustness
Use clustered standard errors to address heteroskedasticity and serial correlation, typically clustered at the entity level.
Test for collinearity issues (e.g., perfectly collinear variables).
What is the process in scientific progress?
A scientific revolution refers to a fundamental transformation in the way science is conducted, leading to new paradigms or frameworks for understanding the world. Thomas Kuhn, a key figure in the philosophy of science, described it as a shift from “normal science”—routine work within an established framework—to revolutionary science, where anomalies and crises lead to a radical change in the underlying paradigm.
The steps are:
Normal Science: Researchers work within a shared paradigm, solving problems using its methods and assumptions.
Crisis: Anomalies or unresolved problems challenge the paradigm, causing doubt among the scientific community.
Revolution: A new paradigm emerges, reshaping scientific theories, methods, and goals.
Return to Normal Science: Work resumes under the new paradigm.
What is realism?
Objective reality independent of human thought, and systematic inquiry can reveal universal truths about this reality (Quantitative, genereliazible).
BUT can oversimplify complex phenomena by ignoring subjective or contextual nuances.
What is logical positivism?
Logical Positivism emphasizes objective observation and empirical verification. Only measurements to be valid (quantitative).
Ignores subjective (qualitative) phenomena that can’t be measured or tested, such as social constructs.
Rejects metaphysics.
What is pragmatism?
Focuses on problem-solving and evaluates theories based on their practical utility rather than absolute truth (Quant & Qual).
BUT can undervalue theoretical rigor.
What does it mean to be scientific? (4)
Being scientific means adhering to the rules of science, rather than the topic investigated.
- The goal is inference:
- descriptive inference
- casual inference - The procedures are public
- The conclusions are uncertain
- The content is the method
In the context of crafting a research question, what is grounding?
grounding is the foundational step that ties the research question to a solid base of evidence, context, or theory, ensuring its relevance
Why would you use a qualitative research design?
Want to go deeper, explore complex vague/new theme/phenomena.
Dynamics of a process
“Thick descriptions”
Study meaning
Detailed insights into why and how things happen
What are some commonly used methods for analysis?
Grounded theory
Creating structure in coding, based on emergent inductive ways of thinking. Open minded. conceptualisation of underlying patterns
Content Analysis
Good for large amount of text. Umbrella term: Word counts, spaces, important themes. Quantitative presentation
Discourse analysis
Looking for meaning, deep, language as a form of social practice. Umbrella term:
Rhetoric, Text/Narrative analysis. Focus on Language as Social Action, HOW they are said
How do you construct scientific arguments?
Claim - A statement that something is true or false, “it’s dark outside”
Reasons - Something to support that claim, “when i look outside, i don’t see any light”
Evidence - accepted as a fact
Convincing → Representing → linking claims to evidence
Claim that… because of reasons… which I base on this evidence…
What is evidence in qualitative research?
Your own or others research; analyses, experiments, interviews etc. Authoritative statements and “accepted facts”.
How do you make qualitative research trustworthy?
Trustworthy: Research should be credibility and transferability.
The text and language is central and should be interpreted as truthful.
Whats the difference between autocorrelation and heteroskedasticity?
Autocorrelation deals with correlations between residuals (temporal or spatial relationships), while heteroskedasticity focuses on unequal variability of residuals.
What is research design and how do you reach a good design?
Overall plan for answering your research problem/ question. A framework reflecting decisions regarding priority in relation to several aspects of the research process. Design Match RQ. Research design turns a research question and objectives into a project.
Good if it considers strategies and choices for what data to collect, how to collect it, and how to access it
External validity (generalized to a broader population) → Internal validity (research design answer RQ) → Reliability
What does it mean to code qualitative data and how can it be done?
Finding your way from “raw data” (rough and unsorted) to “making a statement”. Connect “claims” to “evidence”. Categorize data, organizing into similar “chunks”.
Inductive, Deductive, Grounden, Gioia etc.
Data (read/hear/see/feel)
→ Interpretation (make sense of data) → Statement (Claim/Argument)
Get to know your data → Mark the text → Code → Relate to theory
How do we theorize from data?
We Add interpretation! Relate codes to research question, to existing theory
Theorize - can we use this to make sense “at a higher level”?
Read! → Code! → Analyze!
What are the assumptions needed to fit a linear regression to data?
- Linearity
Linear relationship between X and Y - Normality of errors
The residuals (errors) are normally distributed. - Homoskedasticity
The residuals (errors) are evenly distributed. Constant error variance assumption. - No autocorrelation (independence)
Residuals (errors) not correlated across time.
Is panel data and event study methods both quasi-experimental?
Yes. Both rely on observed data and assumptions (e.g., no confounding trends) to infer causality without randomization. Often, in event study, the “control group” is just the counterfactual.
What is the benefit to using portfolio sorts over a (panel) linear regression method?
Portfolio sorts make it easy to see how average returns change across groups (e.g., quintiles of size or value), while regressions give coefficients that are harder to interpret. Sorts also handle extreme values better and don’t rely on strict linear assumptions.
What is the estimation window?
Estimation Window: Provides a benchmark for “normal” returns, used to calculate abnormal returns. Usually 120 days.
Do NOT confuse with pre-event window: directly examines what happens before the event.
What are the draw downs of cross-sectional regression?
Can only handle linear relationship.
Difficult to handle inclusion/exclusion.
Does not allow for a hedge portfolio.
What are the positives of portfolio sorts?
Allows for non-linear relationship.
Easy to handle inclusion/exclusion.
Allows for hedge portfolio.
What is the difference between correlation and causality?
a) Correlation is a statistical term, causality is not.
b) Only causality shows the direction of the relationship between variables
c) Only causality shows if the relationship between variables is positive or negative
d) Causality can be due to third variables, correlation cannot
Only causality shows the direction of the relationship between variables
What is Cartesian dualism?
Cartesian DUAlism is the idea that the mind and body are two completely separate things. The mind is non-physical (thinking, consciousness), while the body is physical (material, extended in space).
Knowledge comes from rational thinking and innate ideas that are certain and beyond doubt.
Innate ideas are concepts or knowledge that are believed to be present in the mind from birth, without being learned through experience. Descartes argued that some truths, like the existence of God or basic logical principles, are built into the human mind naturally.
What is explanans and explanadum?
- Explanandum: The thing you want to explain (the phenomenon or question). Why is the sky blue?
- Explanans: The explanation or reason that answers it. Because sunlight scatters blue light more in the atmosphere.
What is Lockeian view?
The Lockean view sees knowledge as coming from sensory experience and reflection. According to John Locke, the mind starts as a blank slate and gains knowledge through perceptions and experiences, rejecting the idea of innate ideas.
What is a cross-sectional study?
Different groups compared at the same time.
Quantitative research is often (always) aimed at being..
generalizable to a larger population.
What is important in quantitative research?
Credibility.
External validity: can the result be generalized to a broader population?
Internal validity: is the research design appropriate to answer the research question?
When do you use a research case? (harvard case)
When asking “how” and “why” questions
When the investigator has little or no control over events
Focus on a phenomenon in its real-life context and creates context-dependent knowledge
For theory development, not statistical generalization
especially appropriate in new topic areas.
What is coding?
A way of working.
A set in the analytical process, but no everything.
Coding is not necessary. Coding is most often inductive, but can also be deductive.
What type of analysis is this?
Grounded
Content
Discourse
Aim: understanding the role of business media setting the corporate CSR agenda by exploring how CSR is presented in two
English-language business newspapers with an international readership, Financial Times and The Guardian, between 2000-2009.
Content analysis
What is time-demeaning fixed effect?
focuses on deviations from entity-specific means and isolates within-entity variations while controlling for constant unobserved factors.
- avoids autocorrelation in the error term.
- preserves the degrees of freedom
- It does not rely on T-consistency (large time periods).
- the standard method for fixed-effects regressions due to its simplicity and effectiveness.
First-differencing focuses on changes between consecutive periods, while time-demeaning focuses on deviations from the individual’s average.
How can you choose between Pooled OLS or Fixed Effects?
Use the F-test.
How can you choose between one-way vs two-way fixed effect?
Use the F-test.
