Psych/Soc Deck 8: Research Methods, Statistics & Experimental Design

Experimen

A study in which the researcher manipulates an independent variable to determine its effect on a dependent variable.

Observational Study

The researcher observes subjects without manipulating variables.

Can identify associations but cannot prove causation.

Cross-Sectional Study

Measures different individuals at one point in time.

Fast and inexpensive.

Cannot determine cause and effect.

Longitudinal Study

Follows the same participants over time.

Can examine changes over time.

Expensive and time-consuming.

Case-Control Study

Starts with participants who already have a disease and compares them to controls.

Looks backward for possible exposures.

Cohort Study

Starts with exposure status and follows participants to see who develops the outcome.

Can be prospective or retrospective.

Case-Control vs Cohort

Case-Control
Start with the disease → look backward.

Cohort
Start with exposure → follow forward.

Independent Variable (IV)

The variable the researcher manipulates.

Dependent Variable (DV)

The variable that is measured as the outcome.

IV vs DV

Independent

"I change it."

Dependent

"I measure it."

Control Group

Does not receive the experimental treatment.

Serves as a comparison.

Experimental Group

Receives the treatment or intervention.

Confounding Variable

An outside variable associated with both the independent and dependent variables that may distort the results.

Random Assignment

Randomly assigning participants to experimental groups.

Helps reduce confounding.

Random Sampling

Randomly selecting participants from the population.

Improves generalizability.

Random Sampling vs Random Assignment

Sampling→ Who enters the study?

Assignment→ Which group do they enter?

Selection Bias

Study participants are not representative of the target population.

Recall Bias

Participants inaccurately remember past events.

Common in retrospective studies.

Observer Bias

Researchers' expectations influence observations.

Response Bias

Participants answer questions inaccurately.

Social Desirability Bias

Participants give answers they believe are socially acceptable rather than truthful.

Reliability

The consistency or reproducibility of a measurement.

Validity

Whether a test actually measures what it is intended to measure.

Reliability vs Validity

Reliability→ Consistent.

Validity→ Accurate.

Internal Validity

The degree to which changes in the dependent variable are caused by the independent variable rather than other factors.

External Validity

The extent to which study findings can be generalized to other populations or settings.

Positive Correlation

Both variables increase or decrease together.

Negative Correlation

As one variable increases, the other decreases.

Correlation Coefficient (r)

Ranges from -1 to +1.

• +1: Perfect positive correlation

• 0: No correlation

• -1: Perfect negative correlation

Does correlation imply causation?

No.

Correlation indicates an association, not a cause-and-effect relationship.

Mode

Most frequently occurring value.

Standard Deviation (SD)

Measures how spread out data are around the mean.

Smaller SD = less variability.

Normal Distribution

Bell-shaped curve.

Approximately:

• 68% within 1 SD

• 95% within 2 SD

  • 99.7% within 3 SD

68–95–99.7 Rule

1 SD = 68%

2 SD = 95%

3 SD = 99.7%

p-value

The probability of obtaining results at least as extreme as those observed if the null hypothesis is true.

Typically:
p < 0.05 = statistically significant.

Confidence Interval (CI)

A range of values likely to contain the true population parameter.

Narrower CI = more precise estimate

Null Hypothesis (H₀)

Assumes no difference or no relationship between variables.

Alternative Hypothesis (H₁)

States that a difference or relationship exists.

Type I Error

Rejecting a true null hypothesis.

False Positive

Type II Error

Failing to reject a false null hypothesis.

False Negative

Type I vs Type II Error

Type I

False Positive

"I found an effect that isn't really there."

Type II

False Negative

"I missed a real effect."

Sensitivity

Ability of a test to correctly identify people who have the disease.

True Positive Rate

Specificity

Ability of a test to correctly identify people who do not have the disease.

True Negative Rate

Sensitivity vs Specificity

Sensitivity

Finds disease.

Few false negatives.

Specificity

Rules out disease.

Few false positives.

Informed Consent

Participants voluntarily agree to participate after receiving adequate information about the study.

Confidentiality

Protecting participants' private information

Institutional Review Board (IRB)

Committee that reviews research involving human participants to ensure ethical standards are met.

Placebo Effect

Participants improve because they believe they are receiving treatment.

Double-Blind Study

Neither participants nor researchers know who receives the treatment.

Reduces placebo and observer bias.

Researchers assign volunteers to receive either a new antidepressant or a placebo by flipping a coin. What process is this?

Random assignment.

A study recruits only college students to represent all U.S. adults. What bias is most likely?

Selection bias

Researchers ask participants to remember how many sugary drinks they consumed over the last five years. What bias is most likely?

Recall bias

A screening test correctly identifies 98% of people with a disease but incorrectly labels many healthy people as positive. What property is high?

Sensitivity

Researchers find that ice cream sales and drowning deaths increase together during the summer. Does this prove ice cream causes drowning?

No. Correlation does not imply causation. A confounding variable (summer weather) likely explains the association.

A study follows smokers and nonsmokers for 20 years to compare lung cancer rates. What type of study is this?

Cohort study

A researcher begins with patients who have lung cancer and compares their past smoking history to healthy controls. What study design is this?

Case-control study

A statistically significant study reports p = 0.02. What does this mean?

If the null hypothesis were true, there would be a 2% probability of observing results this extreme (or more extreme) by chance alone.