Here are my notes and thoughts on economic methods.
A model is a simplified description of reality. Economists use data to evaluate the accuracy of models and understand how the world works. Correlation does not imply causality. Experiments help economists measure cause and effect. Economic research focuses on questions that are important to society and can be answered with models and data.
Empiricism uses data to analyze the world. It is at the heart of all scientific analysis. The scientific method is the name for the ongoing process that economists, other social scientists, and natural scientists use to develop models of the world and evaluate those models by testing them with data.
Testing models with data help separate good models from bad models. A model is a simplified description of reality. Sometimes a model is also called a theory. Because models are simplified, they are not perfect replicas of reality.
Scientific models are used to make predictions that can be checked with empirical evidence. Empirical evidence is a fact that can be obtained through observation and measurement. Empirical evidence is also known as data. A model’s prediction is known as an hypothesis. A model is only an approximation and accordingly, understand that the model is not exactly correct. A model also makes predictions that can be tested with data.
The man, or average, is the sum of all the different values divided by the number of values and is a commonly used technique for summarizing data.
Causation occurs when one thing directly affects another. Scientists refer to a changing factor, like temperature, as a variable. Scientists say that causation occurs when one variable causes another variable to change. Correlation means that two variables tend to change at the same time. As one variable changes, the other changes as well. When two variables are correlated, it suggests that causation may be possible and further investigation is warranted.
Correlations are divided into three categories, positive, negative, and zero correlation. Positive correlation implies that two variables tend to move in the same direction. Negative correlation implies that the two variables tend to move in opposite directions. When two variables are not related, we say they have zero correlation.
An omitted variable is something that has been left out of a study. Reverse causality is another problem that plagues our effort to distinguish correlation and causation. It occurs when we mix up the direction of cause and effect.
One method of determining cause and effect is to run an experiment. An experiment is a controlled method of investigating causal relationships among variables. To run an experiment, researchers usually create a treatment group and a control group. Randomization is the assignment of subjects by chance, rather than by choice to a group. The treatment group and the control group are treated identically except along one dimension that they are testing for.
A natural experiment is an empirical study in which some process has assigned subjects to control and treatment groups in a random or nearly random way. Natural experiments are a useful source of data in empirical economics. In many problems, they help us separate correlation from causation.
Good economic questions address topics that are important to individual economic agents and to our society. Good economic questions can be answered.
To say that economists use the scientific method means that they are using
- An ongoing process to develop models of the world and then test and evaluate those models
How do economists distinguish between models that work and those that don’t?
- They test their models against real world data
How does the sample size affect the validity of an empirical argument?
- The larger the sample size the better
Suppose country A had 5 families. Their incomes are 10000,19000,30000,39000,and 49000.
Country A’s median income is 30000
Its mean income is 29400
Suppose country B also has 5 families. Their incomes are 10000,19000,30000,39000, and 151000.
Country B’s median income is 30000
Its mean income is 49800
Country B has greater income inequality
Suppose you thought income inequality in the US had increased over time. Would you expect the ratio of the mean income in the US to the median income has risen or fallen?
- Risen, because means change more with extreme values
Suppose you have just been hired as a management consultant by a major oil company to help it optimally price gasoline at its service stations. During a meeting with your client, the ceo asks if your economic models include all factors that impact gasoline prices. What is your response to this question?
- No, the model is a simplified representation of reality
Your client becomes critical of your sloppy technique of using a model that does not include all factors. What is the most appropriate reply to this criticism?
- Economic models are meant to be approximations that predict what happens in most circumstances.
He is still unconvinced about the reliability of using economic models to make business decisions. You can answer this concern by sharing that you will confirm the accuracy of the model by:
- Testing its predictions with empirical data
What is meant by randomization in the context of an economic experiment?
- Subjects are assigned by chance, rather than by choice, to a group
Causation occurs when there is:
- A logical cause and effect relationship
Demonstrate causation or correlation with the following examples:
- More police officers and lower crime rates is likely to be causation
- More economic growth and higher employment levels is likely to be causation
- The length of women’s skirts and stock market performance is likely to be correlation
In general, people with more education earn higher salaries. Economists have offered two explanations of this relationship.
- The signaling argument implies that a college student who drops out of school one month before graduation should earn much less than a student who graduates.
- The human capital argument implies that a college student who drops out of school one month before graduation should earn almost the same as a student who graduates
Some people go to a new gym because it just happened to open up right next to where they live. People who live farther away do not go to this gym.
Using the concept of a natural experiment, which of the following statements is true?
- This is a good natural experiment
A simple economic model predicts that a fall in the price of bus tickets means that more people will take the bus. This isn’t always the case. Is the model incorrect?
- No, because it predicts the outcome of increased bus ridership on average
How would you test this model?
- You should run a natural experiment by analyzing bus ridership and price changes
Your client has asked you to plot crude oil prices and gasoline prices on a graph. The cause or independent variable should be plotted on the:
The effect or dependent variable should be plotted on the:
There is a positive correlation between oil prices and gasoline prices
Your client has asked you to plot gdp and gasoline prices on a graph.
- There is a negative correlation between gdp and gasoline prices
Suppose you have been hired as a management consultant by a major oil company to help it optimally price gasoline at its service stations. Your client would like your team to perform a study on customers’ gasoline purchasing habits when they notice price increases. You suggest that the team:
- Design and execute an experiment
Debbie’s method is a controlled experiment and Troy’s method is a natural experiment
Group A is the treatment group and Group B is the control group
Suppose that you are on a date with an economics major, and you want to impress them by talking about economics. Your date challenges you to state your knowledge of positive and normative questions.
- You say that positive questions ask what is and normative questions ask what ought to be
Which of the following examples do you provide as a normative question?
- Should welfare be repealed
Which of the following do you provide as a positive question?
- How much is the national debt?