Hypothesis-Testing Demands Trustworthy Data-A Simulation Approach to Inferential Statistics Advocating the Research...

Area of Science:

• Psychological research methodology
• Statistical inference

Background:

• Null-hypothesis significance testing (NHST) is the primary statistical strategy in psychology.
• Recent replication failures indicate NHST results often lack statistical power and have high error rates.

Purpose of the Study:

• To propose the research program strategy (RPS) as a superior alternative to NHST.
• To address key deficits in both Frequentist and Bayesian inference methods.
• To provide a tool for restoring trust in psychological research.

Main Methods:

• Data-simulation was used to estimate error rates of NHST results.
• The research program strategy (RPS) was developed, integrating Frequentist and Bayesian inference elements.
• RPS guides from preliminary discovery (H0) to statistical verification (H1).

Main Results:

• RPS demonstrates significantly lower error rates compared to NHST.
• RPS effectively aggregates underpowered research findings.
• RPS addresses limitations of pure Frequentist and standard Bayesian approaches.

Conclusions:

• The research program strategy (RPS) offers a more robust methodology for empirical research.
• RPS can help mitigate the impact of the ongoing replicability crisis in psychology.
• Adopting RPS can enhance the reliability and trustworthiness of scientific findings.