Exam Notes: Research Methods and Integrity

Both assess the relationship between two variables.
Main Difference: Experimental method manipulates the presumed causal variable; correlational method measures it.
Reasons for not knowing causal direction in correlational studies: Third-variable problem, unknown direction of cause.
Complications with experiments: Uncertainty about exact manipulation, creation of unlikely variable levels, often requires deception, not always possible.
Takeaway: Experiments are not always better; ideal research includes both designs.

Statistical significance: A result that would occur by chance less than $5\%$ of the time.
Null-hypothesis significance testing (NHST): Determines the chance of getting a result if no true effect exists.
$p$ -level: Probability of obtaining a result if there is no difference between groups or no relationship between variables.
Problems with NHST: Difficult logic, arbitrary significance criterion, doesn't indicate practical significance, non-significant results often misinterpreted as "no result."

Effect size: An index of the magnitude or strength of the relationship between variables.
Quantifies practical significance.
Correlation coefficient: Most commonly used measure of effect size, ranging between $-1$ and $+1$ . If $0$ , there is no relationship.

Provides the range of values within which the true population correlation is likely to be found (e.g., $95\%$ confidence level).

Replication: Finding the same result repeatedly with different participants and labs.
Publication bias: Studies with strong results are more likely to be published; small/weak effects often unreported.
Questionable research practices (p-hacking): Manipulating data to achieve statistical significance for publication.
- Examples: Deleting unusual responses, adjusting results, not reporting failed experiments.
Making research dependable: Use larger participant numbers, disclose all methods, share data, report all studies (successful and unsuccessful).
Scientific conclusions are always subject to change; no single study is conclusive proof.

WEIRD samples: Common participant groups (Western, Educated, Industrial, Rich, Democratic).
Lack of diversity among researchers leads to limited diversity in research topics/perspectives.
Increasing efforts towards Diversity, Equity, and Inclusion (DEI).

Open Science: Practices promoting transparency and rigor in research.
Key principles: Avoid plagiarism and data fabrication, report data completely, fully describe all study aspects, report both successful and failed studies, freely share data.