Research methodology, statistical best practices, and updates from the GraphHelix team.
Log transformation and back-transformation worked end to end: why the geometric mean is not the mean, and why the back-transformed CI is asymmetric.
Shapiro-Wilk failed? When to use the Mann-Whitney U test instead of a t-test turns on four inputs, and the normality test is the weakest of them.
A VIF of 8.4 inflates that standard error 2.9x. Which multicollinearity threshold should worry you in a regression, and the three checks that settle it.
Chi-square test of independence: when Cochran's expected-count rule applies, when to switch to Fisher's exact, and what to report alongside p.
Welch's t-test handles unequal variance and unequal n correctly; Student's only works in a narrow case. Why R defaults to Welch's and SPSS doesn't.
Bootstrap the indirect effect a×b with 5,000 BCa resamples. R workflow, reporting template, and why Sobel and Baron & Kenny are out of date.
Gene expression heatmaps in R: log2 counts before row-scaling, Pearson + Ward.D2 clustering, colorblind-safe palettes, pheatmap and ComplexHeatmap.
Statistics methods section template for biology papers: five required elements, APA formatting rules, by-test templates, and reviewer-catch list.
Volcano plots for RNA-seq in R: choose adjusted p-values, set log2FC thresholds, and label crowded plots with ggplot2 or EnhancedVolcano.
How to visualize longitudinal data: spaghetti plots for individual trajectories, lasagna plots for high density, and summary plots with proper error bars.
Compare free biostatistics software for researchers: JASP, jamovi, R, Python, GraphHelix, and online calculators by analysis type and learning curve.
Decide when normality matters for ANOVA: Shapiro-Wilk on residuals, Q-Q plot patterns, sample-size rules, and when to switch to Kruskal-Wallis.
Configure matplotlib for journal-ready figures: column widths in mm, pinned Helvetica/Arial fonts, vector output, colorblind-safe palettes.
Compare Tukey HSD, Bonferroni, Holm, and Benjamini-Hochberg FDR for ANOVA post-hoc tests with a worked 10-comparison example and selection guide.
Worked example of a Cox proportional hazards model: hazard ratios, confidence intervals, Schoenfeld residuals, and a publication-ready write-up.
A four-decision framework for choosing the right statistical test for experimental data: variable type, number of groups, pairing, and assumptions.
When to use SD, SEM, or 95% confidence intervals as error bars in scientific figures. Covers formulas, the overlap trap, R and Python code, figure legend conventions, and common reviewer requests.
Side-by-side comparison of GraphPad Prism alternatives for scientific graphing and statistics. Covers R, JASP, Jamovi, Julius AI, Python, and GraphHelix with feature tables and workflow recommendations.
Step-by-step guide to creating publication-ready Kaplan-Meier survival curves in R with ggsurvplot. Covers log-rank tests, risk tables, censoring marks, median survival lines, and manuscript reporting conventions.
Three methods for justifying animal numbers: formal power analysis, resource equation, and pilot study justification. Covers IACUC documentation requirements, ARRIVE 2.0, the 3Rs, and attrition adjustments.
Calculate sample size for an independent two-sample t-test. Covers Cohen d conventions, the per-group formula, a quick reference table, worked examples, and common mistakes in power analysis.
Learn how to calculate sample size for clinical trials. Covers the four core parameters - alpha, power, effect size, and variability - with formulas, worked examples, and common pitfalls.
Join the beta waitlist and be the first to try GraphHelix.