Types of Statistical Studies and Experimental Design

The learning goals for types of statistical studies include:
- Understanding the fundamental differences between observational studies and experiments.
- Recognizing key issues within the design of experiments, such as the selection of treatment and control groups.
- Identifying the placebo effect and the importance of blinding in research.

The subjects of a study refer to the people, animals (or other living things), or objects chosen for the sample.
When the subjects of a study are people, they are also frequently referred to as participants.

Observational Study: In this type of study, researchers observe or measure specific characteristics of the subjects. Crucially, they do not attempt to influence or modify these characteristics in any way.
Experiment: In an experiment, researchers actively apply a treatment to the subjects and then observe the effects of that treatment on the subjects.
Example (Identifying Study Types): A poll in which college students are asked if they commute or live on campus is classified as an observational study. This is because the poll seeks only to determine the students' current status without trying to influence their choice.

Statistical studies, whether they are experiments or observations, generally focus on measuring what are known as variables of interest.
Variable: Any item or quantity that can vary or take on different values.
Variables of Interest: Specifically refers to the items or quantities that a statistical study seeks to measure.
Explanatory Variable: When cause and effect are under investigation, this is the variable that may explain or cause the effect.
Response Variable: This is the variable that responds to changes made to (or observed in) the explanatory variable.
Example (Identifying Variables): In a poll asking students if they commute or live on campus, the variables of interest include:
1. The specific responses to the question (commute vs. live on campus).
2. The proportions of responses for the different possible choices (commute, live on campus, or refuse to answer).
- Note: Since this specific poll does not involve cause and effect, there is no need to designate variables as explanatory or response.

While most observational studies collect data all at once, there are two common variations:
- Retrospective (or Case-Control) Study: This study uses data from the past, gathered from sources like official records or past interviews.
- Prospective (or Longitudinal) Study: This study is set up to collect data in the future from specific groups that share common factors.
Example (Selecting Study Type): If you want to determine if teens who try marijuana are more or less likely to abuse hard drugs as adults, a retrospective study is the only real option. Researchers would look at adults who currently abuse drugs and see if they were more likely to have tried marijuana as teens compared to adults who do not abuse drugs.
- Ethical Implication: An experiment would be highly unethical in this case, as researchers cannot give drugs to participants to see the outcome.

Treatment Group: The group of subjects in an experiment who receive the specific treatment being tested.
Control Group: The group of subjects who do not receive the treatment being tested.
Selection Process: In most cases, it is vital to choose members of both the treatment and control groups through random selection from the available pool of subjects.
Example (The Mozart Treatment): A study divided college students into two groups. One group listened to Mozart/classical music before a task (Treatment Group), while the other group was simply assigned the task without music (Control Group).
- Findings: The treatment group performed slightly better, but only if the task was completed within a few minutes of listening. If data were collected later, both groups performed equally.
- The Treatment: The specific treatment in this study was the classical music.

Confounding: A study suffers from confounding if the effects of different variables are mixed to the point where researchers cannot determine the specific effects of the variables of interest.
Confounding Variables: These are the specific variables that lead to confusion regarding the results.
Strategies for Selecting Groups:
- Random Selection: Subjects must be assigned to treatment or control groups at random to ensure each subject has an equal chance of being in either group.
- Large Group Sizes: Groups must be sufficiently large so they are unlikely to differ in a significant way before the treatment is applied.

Placebo: A substance or procedure that lacks the active ingredients of the treatment being tested but looks or feels like the real treatment so participants cannot tell the difference.
Placebo Effect: This occurs when patients show improvement simply because they believe they are receiving a useful treatment.
Ethical Note: While participants should not know which group they are in, they must be informed for ethical reasons that some participants in the study will be given a placebo rather than the real treatment.
Discussion Point: Historically, participants were sometimes told they were all receiving a treatment when some were actually getting a placebo. This raises questions about whether researchers should be allowed to use results from past studies that do not meet modern ethical criteria.

Hawthorne Effect: This occurs when subjects respond differently simply because they know they are part of an experiment, regardless of the treatment.
- Origin: Named after a study of factory workers at Western Electric’s Hawthorne plant.
Experimenter Effect: This occurs when the researcher or experimenter influences subjects through non-verbal cues such as facial expressions, tone of voice, or general attitude.

Single-Blind: The participants do not know if they are in the treatment or control group, but the experimenters do know.
Double-Blind: Neither the participants nor the experimenters interacting with them know who belongs to the treatment group and who belongs to the control group.

In a meta-analysis, researchers review and synthesize many past studies.
The goal is to consider these studies as one combined group to identify trends or findings that were not evident or statistically significant in the individual studies.