A Complete Guide to Sample Size Formulas in Research (With Easy Decision Guide for Students)

By: Quantalpha Algorithms

Introduction: Why Sample Size Matters?

One of the most common mistakes in research is choosing a sample size based on convenience rather than methodology. Your sample size determines:

• The accuracy of your results
• The credibility of your study
• Whether your findings can be generalized

This guide will help you:

• Understand all major sample size approaches
• See their differences clearly
• Use a simple decision guide to pick the right one for your study

PART 1 
The Big Picture: How Sample Size Methods Are Organized?

Instead of memorizing formulas, think this way:

Your research design determines your sample size method.

There are four major families of sample size approaches:

1. Descriptive (Survey) Studies
2. Quantitative (Numerical) Studies
3. Experimental Studies
4. Qualitative Studies

Each uses different logic and formulas.

PART 2 
Sample Size Formulas by Research Type

A. DESCRIPTIVE / SURVEY RESEARCH

Use these when your goal is to describe a population (e.g., opinions, behaviors, percentages).

1) Cochran’s Formula (Large or Unknown Population)

Best for:

• Large populations
• National surveys
• Opinion polls

When to use it:

If you don’t know the population size or it is very large.

Strength: Statistically rigorous

Limitation: Assumes random sampling

2) Cochran’s Correction (Finite Population)

 

Use this after Cochran’s formula when you actually know N.

Best for:

• School-based surveys
• Company surveys
• Barangay/community studies

3) Slovin’s Formula (Simple & Popular in Theses)

 

Best for:

• Undergraduate theses
• Basic descriptive research
• When population is known but variability is unknown

Strength: Easy to compute

Limitation: Less statistically precise than Cochran

4) Yamane’s Formula (Textbook Version of Slovin)

 

Functionally the same idea as Slovin, just presented differently in many books.

5) Krejcie & Morgan Table (No Formula Needed)

Instead of computing, you look up sample size based on population in a standard table.

Best for:

• Education research
• Social science theses

Strength: Simple and widely accepted

Limitation: Less flexible than formulas

Comparison of Survey Methods

Method	Population Known?	Precision	Ease of Use	Typical Use
Cochran	No	High	Moderate	Large surveys
Cochran Corrected	Yes	High	Moderate	Finite populations
Slovin	Yes	Medium	Very easy	Student theses
Yamane	Yes	Medium	Very easy	Textbook problems
Krejcie-Morgan	Yes	Medium	Easiest	Education research

B. QUANTITATIVE (NUMERICAL) RESEARCH

Use these when your data are numbers (scores, time, income, grades, measurements).

6) Sample Size for One Mean

 

Best for:

• Average test scores
• Mean income studies
• Average waiting time studies

Requires an estimate of standard deviation (σ).

7) Sample Size for Correlation Studies

Used when your goal is to see relationships (e.g., study hours vs GPA, etc).

A common power-based formula:

 

Usually computed using G*Power software rather than by hand.

8) Sample Size for Regression Studies

Rule of thumb:

 

Where m = number of independent variables.

Example:

If you have 6 predictors → minimum sample ≈ 98 respondents.

C. EXPERIMENTAL RESEARCH

Use these when you compare groups or test an intervention.

9) Sample Size for Comparing Two Means

 

Best for:

• Pre-test vs post-test
• Control vs experimental group

Often calculated using G*Power software or any other statistical software.

10) Sample Size for Comparing Two Proportions

Used in:

• Clinical trials
• Policy evaluation
• Educational interventions

Also typically computed using software.

11) ANOVA Sample Size (3+ Groups)

Used when you compare three or more groups (e.g., 3 teaching methods).

Almost always done using G*Power or any other statistical software, not a manual formula.

D. QUALITATIVE RESEARCH (NO FORMULA)

Important rule:

Qualitative research does NOT use mathematical formulas.

Instead, sample size is based on data saturation — the point where no new themes emerge.

Typical ranges:

Method	Typical Sample
Phenomenology	8–15 interviews
Grounded Theory	15–30 interviews
Case Study	1–5 cases
Narrative	1–3 participants

PART 3 
EASY DECISION GUIDE (For Researchers)

Step 1: What is your research type?

If your answer is:

• “Survey / descriptive” → Use Cochran, Slovin, Yamane, or Krejcie-Morgan
• “Numerical data” → Use Mean, Correlation, or Regression formulas
• “Experiment” → Use Two Means, Two Proportions, or ANOVA (G*Power)
• “Interviews / qualitative” → Use saturation, not formulas

Step 2: Do you know your population size?

• Yes → Slovin / Yamane / Cochran corrected
• No → Cochran original formula

Step 3: Are you comparing groups?

• Yes → Experimental formulas
• No → Survey or quantitative formulas

PART 4 
One-Page Cheat Sheet

Your Study Type	Use This Method
General survey	Cochran / Slovin
School-based survey	Slovin / Krejcie-Morgan
Averages (scores, time, income)	One-mean formula
Relationship study	Correlation sample size
Multiple predictors	Regression rule
2 groups experiment	Two means formula
3+ groups experiment	ANOVA (G*Power)
Interviews	Saturation (8–30 people)

 



Join our DISCORD to learn more!