Statistical Investigation: Steps & Laws of Statistics
1. What Is a Statistical Investigation?
A Statistical Investigation (or statistical inquiry) is a systematic search for truth conducted through quantitative measurements and data collection to solve a specific management or research problem.
Examples of Business Investigations:
- An automobile brand investigating why electric two-wheeler sales in Madhya Pradesh dipped by 12% in Q3.
- An FMCG firm surveying 2,000 college students in Indore to test consumer preference for a new sugar-free energy drink.
2. The 7 Steps of a Statistical Investigation (Activity 2 Flowchart)
Conducting a successful statistical investigation follows 7 disciplined, sequential steps:
7-Step Roadmap of Statistical Investigation
Step 1
Define Objectives
Specify exact research question & scope of inquiry
➔
Step 2
Plan the Inquiry
Choose source, statistical unit, and Census vs Sample
➔
Step 3
Data Collection
Deploy questionnaires, interviews, or secondary records
➔
Step 4
Data Editing
Filter incomplete responses, correct errors & outliers
➔
Step 5
Classification & Table
Group data systematically and create tables/graphs
➔
Step 6
Statistical Analysis
Calculate Mean, SD, Correlation, and Trends
➔
Step 7
Report & Action
Draw managerial conclusions & present final findings
Step-by-Step Breakdown:
- Clarifying Problem & Objective:
- What is the exact goal? (e.g., “Determine the price elasticity of consumer demand for organic milk in Vijay Nagar, Indore”).
- Planning the Inquiry:
- Define Statistical Unit: Is the unit an individual customer, a household, or a retail store?
- Census vs. Sample: Surveying all 50,000 households (Census - costly, slow) vs. a representative sample of 500 households (Sample - fast, cost-effective).
- Collection of Data:
- Administering structured questionnaires via Google Forms, telephonic interviews, or field visits.
- Scrutiny & Editing of Data:
- Checking returned questionnaires for omissions, extreme inconsistencies (e.g., age listed as 180), and falsified entries.
- Classification and Tabulation:
- Condensing hundreds of raw questionnaire responses into neat frequency distribution tables and bar charts.
- Statistical Analysis & Measurement:
- Applying mathematical formulas: calculating Average (Mean), Spread (Standard Deviation), and Relationships (Correlation Coefficient).
- Interpretation & Report Writing:
- Translating statistical figures into actionable managerial English with executive summaries, visual charts, and strategic recommendations.
3. The Two Fundamental Laws of Statistics
Why can we trust a survey of 1,000 voters to predict an election of 10 crore citizens? Because modern sampling is grounded in two scientific laws of statistics:
Foundational Laws of Statistical Sampling
1. Law of Statistical Regularity
Core Principle: A moderately large sample chosen at random from a population will, on average, possess the exact characteristics of that entire population.
Condition: The sample MUST be selected purely at random (every item having an equal chance of selection).
Example: Testing a spoonful of cooked rice from a giant pot to check if all the rice is cooked.
2. Law of Inertia of Large Numbers
Core Principle: Large aggregates are much more stable and constant over time than small groups. Variations in one direction cancel out variations in the opposite direction.
Condition: The larger the size of the sample, the greater the statistical stability and accuracy.
Example: Predicting how long 1 specific person will live is impossible; predicting average life expectancy of 10 lakh citizens is highly accurate!
In-Depth Analysis of the Two Laws:
A. The Law of Statistical Regularity
- Origin: Derived directly from the mathematical Theory of Probability.
- Statement: If a sufficiently large sample is selected at random from a population, the sample is representative of the whole universe.
- Why It Works: Nature operates with underlying regularity. If we test 500 light bulbs at random out of a production batch of 50,000, and find that 2% are defective, we can safely conclude that approximately 2% of the entire 50,000 lot is defective!
B. The Law of Inertia of Large Numbers
- Origin: A direct corollary of the Law of Statistical Regularity.
- Meaning of “Inertia”: Just as in Physics a heavy flywheel resists sudden erratic changes, a large body of numbers resists sudden wild fluctuations.
- Why Compensating Errors Occur: In a huge sample of consumers, some will buy more this month, while others will buy less. Some people will earn higher bonuses, others will face wage deductions. Individual variations offset and neutralize each other, leaving the grand total remarkably stable!
- Business Application: Life insurance companies like LIC thrive on this law. While they cannot predict which specific policyholder will pass away tomorrow, they can accurately predict how many people out of 10,00,000 policyholders will pass away in a year, allowing them to price premiums safely.