Likelihood ratios: a simple and flexible statistic for empirical psychologists

Area of Science:

• Psychology
• Statistical Inference
• Quantitative Research Methods

Background:

• Empirical psychology commonly utilizes null hypothesis significance testing (NHST) for statistical inferences.
• NHST presents challenges in interpretability and flexibility for researchers.
• A need exists for more straightforward statistical alternatives in psychological research.

Purpose of the Study:

• To propose likelihood ratios as a more direct and interpretable alternative to NHST in empirical psychology.
• To evaluate the utility of likelihood ratios for addressing the statistical needs of psychological researchers.
• To demonstrate the advantages of likelihood ratios over traditional inferential statistics.

Main Methods:

• A survey of empirical articles in psychology was conducted to examine common statistical practices.
• The application of likelihood ratios was assessed for various statistical needs identified in the survey.
• Comparison of interpretability and flexibility between likelihood ratios and NHST.

Main Results:

• Likelihood ratios provide a direct measure of the fit between two competing statistical models.
• The statistic quantifies the relative likelihood of the data under best-fit models.
• Survey results indicate likelihood ratios can fulfill essential statistical requirements in psychology research.

Conclusions:

• Likelihood ratios offer a more intuitive and easily interpretable statistical framework for empirical psychology.
• This approach provides researchers with greater flexibility in constructing empirical arguments.
• Likelihood ratios represent a superior alternative to traditional inferential statistics for psychological research.