Linearity Beginner's Guide — Slope, Intercept, r & R² Explained

📐 Linearity Calculator 📘 Beginner Tips & Guide

📘 Beginner Tips & Guide

Everything You Need to Know About Linearity

1. What Is Linearity, and Why Does It Matter?

Linearity is the ability of an analytical method to produce results that are directly proportional to the concentration of the analyte, across a defined range. In plain terms: if you double the concentration, the instrument response (peak area, absorbance, etc.) should also double — not wobble, curve, or plateau.

It is one of the core parameters required by ICH Q2(R2) when validating an analytical method (assay, dissolution, related substances, etc.). A method that isn't linear cannot be trusted to report an accurate result at concentrations you didn't specifically test.

📐

The Idea

Plot concentration (x-axis) against instrument response (y-axis) for several levels. If the points fall on (or very close to) a straight line, the method is linear over that range.

🎯

Why It's Required

Regulators need proof that your method won't quietly give a wrong answer just because a sample happens to be a bit stronger or weaker than expected.

2. How This Calculator Builds Your Levels

This tool uses the Standard Weight & Dilution Method — the most common way linearity is demonstrated for assay-type methods:

You weigh out your working standard and dissolve it in a known Final Volume — this gives your 100% target concentration (Weight ÷ Volume).
You then prepare several dilutions/levels around that target — commonly 80%, 90%, 100%, 110%, 120%, but you can use any levels your protocol calls for (e.g. 50–150%, or just three levels for a quick check).
Each level is injected/measured in replicate (usually triplicate), and the average response is plotted against that level's concentration.

Concentration formula: Conc.(Level) = (Level % ÷ 100) × (Weight of Std ÷ Final Volume)

The calculator's table is deliberately built like a spreadsheet: Add Level inserts a new row, and Add Cell inserts one new reading column across every row at once — so you can go from triplicate to five or six replicates per level in one click, exactly like adding a column in Excel. You type the % value for each level yourself, so the scheme is entirely yours.

3. Slope, Intercept, r & R² — In Plain Language

Term	What It Tells You
Slope (m)	How much the response increases per unit of concentration. This is your method's sensitivity — steeper is more sensitive.
Y-Intercept (c)	The response the line predicts at zero concentration. Ideally close to zero — a large intercept can signal background interference or a blank/baseline problem.
Correlation Coefficient (r)	How tightly the data points cluster around the straight line, from −1 to +1. Closer to 1 means a near-perfect straight-line fit.
R² (Coefficient of Determination)	Simply r squared. R² = 0.998 means 99.8% of the variability in response is explained by concentration — the rest is scatter/noise.
%Y-Intercept	The intercept expressed as a percentage of the response at 100% concentration — a common way SOPs express "how small is small enough" for the intercept.

4. Typical Acceptance Criteria

Exact limits always come from your own validation protocol/SOP, but these are the ranges most commonly seen in industry:

Correlation Coefficient (r)

≥ 0.999 is a common specification for assay methods; ≥ 0.995–0.998 is sometimes used for trace-level or related-substance/impurity methods, where more scatter is expected.

%Y-Intercept

Often required to be within ±2% of the response at 100% concentration — though this varies by SOP and is not always a formal requirement.

Reference

ICH Q2(R2) — Validation of Analytical Procedures is the internationally recognized guideline that defines linearity as a required validation characteristic and describes how it should be evaluated.

5. Common Mistakes That Fail a Linearity Study

⚠️ Too few levels

Two points always form a "perfect" line — that proves nothing. Most SOPs require a minimum of 5 levels to properly demonstrate linearity.

⚠️ Levels bunched too closely together

If every level is within a narrow 95–105% band, you're not really testing the working range. Spread levels across the full range your method will actually see in routine use (e.g. 80–120%).

⚠️ One bad replicate dragging the average

A single outlier reading at one level can visibly bend the regression line and tank your r value. Always sanity-check individual replicates, not just the final average, before accepting a result.

⚠️ Wrong Weight or Final Volume entered

Since every level's concentration is derived from Weight of Standard ÷ Final Volume, a typo here shifts every single point on the x-axis — always double check these two numbers first if your r looks unexpectedly low.

6. Quick Checklist Before You Calculate

Weight of Standard and Final Volume entered correctly (this sets your 100% concentration).
At least 5 levels, spread across your intended working range.
Each level has enough replicate readings (commonly triplicate) filled in.
Correct acceptance limit (minimum r) selected for your method type.
Reviewed the regression chart for any obvious outlier before trusting the r/R² values.

Ready to calculate?

Back to the Linearity Calculator

Build your levels, add replicate readings in the Excel-style table, and get your slope, intercept, r and R² instantly.

Open Calculator

Understanding Linearity — A Beginner's Guide