Chapter 1: Collecting Data - Experiments and Observational Studies

Flashcards covering key definitions and distinctions between association, causation, confounding variables, experiments, and observational studies, based on Chapter 1.3 of the lecture notes.

When are two variables considered associated?

Two variables are associated if values of one variable tend to be related to the values of the other variable.

What defines a causal association between two variables?

Two variables are causally associated if changing the value of one variable directly influences the value of the other.

Can an association between two variables exist without being causal, and if so, what is a potential reason?

Yes, an association can exist without being causal. It might be due to a confounding variable or purely coincidental.

What is a confounding variable?

A confounding variable is a third variable that is associated with both the explanatory variable and the response variable, offering an alternative explanation for an observed association.

Provide an example of a situation with a confounding variable explaining an association.

The association between increased vehicle registrations and increased life expectancy since 1970 might be confounded by 'Time/Year,' as the population grows and medical improvements occur over time.

What is an experiment in the context of data collection?

An experiment is a study in which one or more of the explanatory variables are actively controlled by the researchers.

How does an observational study differ from an experiment?

An observational study does not control the values of any variable; instead, it only observes and records them without intervention, unlike an experiment where variables are manipulated.

What type of study is it if the value of the explanatory variable is determined randomly before the response variable is measured?

This is a randomized experiment.

What can be concluded about causality if an association is found in a randomized experiment?

If a random experiment yields an association between two variables, there is likely to be a causal relationship (explanatory variable directly influences the response variable).

What are the three main explanations for why an association may be observed in data?

The three explanations are: there is an actual causal association, there is an association due to a third (confounding) variable, or there is no actual association and the observed association is due to random chance (coincidence).

In a study examining the relationship between the amount of iron in soil and the amount of potassium in spinach grown in that soil, what type of study would it be if researchers only recorded these values from existing farms?

It would be an observational study because the researchers are not controlling or manipulating the amount of iron in the soil.

Why can't a causal conclusion typically be drawn from an observational study?

Causal conclusions cannot typically be drawn from observational studies because the absence of controlled manipulation leaves open the possibility of confounding variables influencing the observed association.