Experiment Basics Study Notes Social Psych

Explain what an experiment is.
- Recognize examples of studies that qualify as experiments versus those that do not.
Distinguish between important concepts.
- Manipulation of the independent variable.
- Control of extraneous variables.
- Explain the significance of each.
Identify confounding variables.
- Understand their impact on the internal validity of a study.
Define control conditions.
- Explain their purpose in research regarding treatment effectiveness.
- Describe various types of control conditions.

An experiment is a type of study designed to answer the question:
- Is there a causal relationship between two variables?
- It evaluates if changes in one variable (independent variable) cause a change in another variable (dependent variable).

Manipulation of the independent variable.
- Researchers systematically vary the level of the independent variable.
- Different levels are called conditions.
- Example: In Darley and Latané’s experiment:
  - Independent Variable: Number of witnesses believed to be present.
  - Conditions created: one, two, or five students involved in a discussion.
- Important distinction: There is only one independent variable (number of witnesses) despite having three identified conditions.
Control over extraneous variables.
- Extraneous variables are those not directly measured but can affect the outcome.
- In Darley and Latané’s study:
  - Control measures ensured participants were tested in the same room under identical circumstances.
  - Random assignment was utilized to ensure groups were similar at the start.
- Distinction between "manipulation" (changes to the independent variable) and "control" (holding extraneous variables constant).

Manipulating the Independent Variable:
- Involves changing levels of an independent variable systematically, exposing either different groups or the same group at different times.
- Example: A study on expressive writing may involve groups writing about traumatic experiences versus neutral experiences.
- Conditions can be referred to as the "traumatic condition" and the "neutral condition."
- Researchers must actively intervene rather than simply observe already differing groups, to remove biases from pre-existing conditions.

Some independent variables cannot be manipulated for ethical or practical reasons, e.g., significant early illness experiences.
Alternative non-experimental methodologies may be needed for such studies.

An experiment with one independent variable manipulated across two conditions.

An experiment with one independent variable manipulated across multiple conditions.
- Example:
  - Darley and Latané had three conditions (one, two, five witnesses).

Extraneous Variables Defined:
- Any variable that varies in a study outside of the independent and dependent variables.

Effect on Internal Validity:
- Variation introduces noise, making it hard to detect the effect of the independent variable.
- Example: An experiment on mood (happy vs. sad) affecting memory recall would show variability that obscures clear conclusions.
- Hypothetical data illustrate how extraneous variables can impact perceived outcomes.

Holding Variables Constant:
- Conduct all tests in the same location and under similar instructions.
- Limiting participants to specific characteristics can reduce data noise but may compromise external validity.
To Control or Not to Control:
- Balancing participant diversity with the need to control extraneous variables crucially impacts the research outcome.

Definition:
- An extraneous variable that varies systematically with the independent variable, hence potentially causing confusion.
Example: If one condition consistently has lower IQ scores, it becomes a confounding variable, making it unclear whether independent variable changes or IQ cause the observed differences.
Strategies to limit confounding variables:
- Holding extraneous variables constant.
- Random assignment to conditions.

Definitions: Treatment and Control:
- Treatment: Any intervention intended to change behavior positively.
- Control Condition: Participants do not receive the treatment, enabling comparison.

Ensures that observed improvements or changes can be attributed to the treatment rather than other external influences.

No-Treatment Control:
- Participants receive no treatment.
- Risk of placebo effects wherein participants might expect improvement simply from participation.
- Example: Chicken soup for a cold being a folk remedy with no scientific backing.
Placebo Control Condition:
- Participants receive a placebo, mimicking the treatment without active ingredients.
- Aimed to separate real treatment effects from expectations.
Wait-List Control Condition:
- Participants are told they will receive treatment after others but must wait before their treatment starts, controlling for expectation effects.
Alternative Treatment Comparison:
- Instead of utilizing a control group without treatment, a new treatment can be compared to the best standard treatment available, analyzing efficacy.

Placebos can yield positive outcomes even in physiologically perceived conditions, e.g., asthma or knee pain.
Notable study on arthroscopic surgery showing that sham surgery had similar outcomes to actual surgery.
Ethical implications surround the use of deception in such trials, necessitating thorough IRB consideration of potential risks versus research benefits.

Knecht et al. (2000). Handedness and hemispheric language dominance in healthy humans. Brain: A Journal of Neurology.
Price et al. (2008). A comprehensive review of the placebo effect. Annual Review of Psychology.
Shapiro & Shapiro (1999). The powerful placebo: From ancient priest to modern physician. Johns Hopkins University Press.
Moseley et al. (2002). A controlled trial of arthroscopic surgery for osteoarthritis of the knee. The New England Journal of Medicine.