Lecture 1 – Introduction
Lecture 2 – September 4th
What is theory? A causal explanation of why things happen, a cause that leads to an effect.
Why is theory important? Can theories be wrong? Theories can be wrong depending on the
circumstances of different environments where the theory is applied to.
Induction: starting for an observation, trying to find patterns in the observations, proposing
a causal explanation, leading to a theory
Deduction: starting with a theory, deriving a hypothesis from this theory, making
observations and testing if the theory is matched, finally falsifying, or supporting the theory.
Descriptive theory: a theory that describes how things really are as opposed how they
should be.
Normative theory: a theory involving moral standards that regulate right and wrong.
Looking for the boundary conditions of a theory: looking for when a theory works and when
it will not work. Boundaries conditions are needed to avoid making mistakes or bad
decisions.
What is a construct? A theoretical concept based on empirical observations. Can you
observe a contruct? Example: diversity and team performance. It is an abstract thing which
we cannot measure directly. How can you measure a construct? We can measure it
indirectly through variables.
Failure points in research:
- Internal validity: the extent to which the observed results represent the truth in the
population we are studying. Are you actually doing that correctly?
- Construct validity: the extent to which your test or measure accurately assesses what
it is supposed to. Are we measuring the construct really well?
- External validity: if causal relationship exist, how generalizable is this relationship?
, Interpretation of a linear regression:
1. Check the statistical significance (rule of thumb: p-values less than 5% are considered
statistically significant). Theoretically you fail to reject your hypothesis if the
coefficients are statistically significant.
2. Check the sign of the coefficient (positive/negative)
3. Interpret the coefficient
Lecture 3 – September 11th
Characteristics of a strong theory:
1) Studied in many different context
2) Robust to alternative explanations
3) Boundary conditions (moderation)
4) Mechanisms (mediation)
Moderation (or interaction effect): Does the hypothesized relationship depend on another
variable? What is the boundary condition?
Statistical equivalent: outcome = alpha (intercept) + beta 1 * cause 1 + beta 2 * cause 2 +
beta 3 * cause 1 * cause 2 + e
To interpret the red marked part of the equation, we need to visualize it. A highly motivated
employee with years of experience has a way higher work productivity compared to a low
motivated employee with not many years of experience.
Mediation: How does X cause Y through M?
Lecture 2 – September 4th
What is theory? A causal explanation of why things happen, a cause that leads to an effect.
Why is theory important? Can theories be wrong? Theories can be wrong depending on the
circumstances of different environments where the theory is applied to.
Induction: starting for an observation, trying to find patterns in the observations, proposing
a causal explanation, leading to a theory
Deduction: starting with a theory, deriving a hypothesis from this theory, making
observations and testing if the theory is matched, finally falsifying, or supporting the theory.
Descriptive theory: a theory that describes how things really are as opposed how they
should be.
Normative theory: a theory involving moral standards that regulate right and wrong.
Looking for the boundary conditions of a theory: looking for when a theory works and when
it will not work. Boundaries conditions are needed to avoid making mistakes or bad
decisions.
What is a construct? A theoretical concept based on empirical observations. Can you
observe a contruct? Example: diversity and team performance. It is an abstract thing which
we cannot measure directly. How can you measure a construct? We can measure it
indirectly through variables.
Failure points in research:
- Internal validity: the extent to which the observed results represent the truth in the
population we are studying. Are you actually doing that correctly?
- Construct validity: the extent to which your test or measure accurately assesses what
it is supposed to. Are we measuring the construct really well?
- External validity: if causal relationship exist, how generalizable is this relationship?
, Interpretation of a linear regression:
1. Check the statistical significance (rule of thumb: p-values less than 5% are considered
statistically significant). Theoretically you fail to reject your hypothesis if the
coefficients are statistically significant.
2. Check the sign of the coefficient (positive/negative)
3. Interpret the coefficient
Lecture 3 – September 11th
Characteristics of a strong theory:
1) Studied in many different context
2) Robust to alternative explanations
3) Boundary conditions (moderation)
4) Mechanisms (mediation)
Moderation (or interaction effect): Does the hypothesized relationship depend on another
variable? What is the boundary condition?
Statistical equivalent: outcome = alpha (intercept) + beta 1 * cause 1 + beta 2 * cause 2 +
beta 3 * cause 1 * cause 2 + e
To interpret the red marked part of the equation, we need to visualize it. A highly motivated
employee with years of experience has a way higher work productivity compared to a low
motivated employee with not many years of experience.
Mediation: How does X cause Y through M?
