Preferences for Methodology and Statistics

Introduction

Annals prides itself in maintaining high standards for methodology and statistics. Before any study can be considered for acceptance it must be first reviewed by one of our journal's methodology and statistics editors.

Annals makes requests for a specific style in terms of methods and statistics that may not be routine at all journals, and in this section we provide our guidelines and their rationale. This is not meant to be dogmatic, but an attempt to achieve a consistent style for our journal based on progressive statistical and epidemiological standards. If we suggest modifications to your chosen analytic strategy, we will detail our rationale. If appropriate, we will have a dialog and together identify the soundest way to report your data.

General Philosophies

Our underlying tenets for methodology and statistics are:

• We prefer model development using theory rather than statistical significance.
• We prefer emphasis on the estimation of effect and magnitude of differences rather than hypothesis testing.
• When feasible, we prefer displays of actual data rather than summaries of data.

Annals' Preferences for Methodology and Statistics

Preferences for specific research types and designs:

• Basic science • Chart reviews • Clinical practice guidelines or advisories • Clinical prediction rules • Cost-effectiveness analyses • Decision analyses • Diagnostic test studies • Observational studies • Replication studies • Qualitative research • Meta-analyses • Randomized controlled trials • Surveys

Preferences for multivariable analyses:

• Modeling • Sample size • Stepwise regression • Variable selection

Preferences for other analytic issues:

• Baseline comparisons • Clustering • Missing data • Multiple comparisons • P values • Sensitivity analyses • Size of effect • Subgroup analyses

Preferences for figures and tables.

Return to home page.