Lecture 2: Collecting Data

Flashcards covering data sources, observational vs. experimental design, confounding, sampling concepts (parameters, statistics, population vs. sample), and common biases and historical examples.

What are the three main data sources discussed in this lecture?

Anecdotal evidence, observational studies, and experiments.

Define anecdotal evidence.

Data based on individual experiences or observations; emotionally persuasive but scientifically weak.

What is an observational study?

A study where variables are measured through surveys or censuses without interference.

What is an experiment?

A study where researchers influence responses by applying a treatment and observing the results.

What distinguishes a designed experiment from an observational study?

A designed experiment assigns individuals to groups and manipulates the explanatory variable.

What is confounding?

Two variables' effects on a response variable cannot be distinguished; lurking variables may be involved.

Give a confounding example involving ice cream and drowning deaths.

Temperature confounds the relationship between ice cream sales and drowning deaths.

Give a confounding example involving cell phone use and brain cancer.

Age, occupation, and place of residence confound the association between cell phone use and brain cancer.

Why are observational studies hampered?

Because of confounding; it is difficult to separate the effects of competing variables.

How do experiments defeat confounding?

Through randomization and control, which help separate effects of variables.

What does the childcare observational study illustrate about confounding?

There could be factors like working parents or stress that influence behavior, making causation unsure.

What is a population?

The entire group of individuals being studied.

What is a sample?

A subset of the population.

What is a parameter?

A numerical summary of a population.

What is a statistic?

A numerical summary based on a sample.

Explain the sampling metaphor used for inference.

Tasting a spoonful of soup; for inference to be valid, the sample must be representative; stirring helps ensure representativeness.

What is a Simple Random Sample (SRS)?

A sample of size n drawn so every set of n individuals has an equal chance of being selected; each individual has an equal chance.

How is a Simple Random Sample practically implemented?

Using random digit tables, computers, or physical random number generators.

What is Stratified Sampling?

Divide the population into strata, take an SRS from each stratum, and combine; strata reflect population shares.

What is Cluster/Multistage Sampling?

Cluster: break population into clusters, randomly select some clusters, and (optionally) perform an SRS within chosen clusters; multistage adds the extra SRS step.

When is cluster sampling typically used?

When the population is too large for an SRS to be practical; often more economical.

What is Undercoverage?

When one or more groups in the population are left out or underrepresented in the sampling process.

What is Nonresponse?

When sampled individuals can’t be contacted or refuse to respond.

What is Response bias?

When respondents lie or misreport to please the interviewer or hide behaviors.

What is the impact of poor wording of questions?

Question wording can bias responses and lead to misleading results.

What is Convenience sampling?

Selecting individuals who are easily accessible, which may not be representative.

What is Voluntary response bias?

Occurs when the sample consists of people who volunteer to respond, often with strong opinions; not representative.

What went wrong in the 1936 Literary Digest election poll?

The sample was biased toward affluent readers (readers, automobile owners, telephone users), leading to an incorrect prediction.

What groups did The Literary Digest sample in 1936?

Its own readers, registered automobile owners, and registered telephone users.