Econ Statistics & Data Lab

NODE 01

Data

101010111

• Observations or facts collected for analysis.
• Income, prices, population, employment.

NODE 02

Primary Data

👤

• Collected directly by investigator for specific purpose.
• Surveys, interviews, questionnaires.

NODE 03

Secondary Data

→

• Data already collected by someone else.
• Census, RBI publications, World Bank databases.

NODE 04

Primary vs Secondary

PRI

vs

SEC

• Primary: Tailored, first-hand.
• Secondary: Cheaper/faster, but requires evaluation.

NODE 05

Data Sources

RBI

NSO

WB

• Census, NSO, RBI, Govt ministries, SEBI.
• Exact source should always be recorded.

NODE 06

Census

• Attempts to collect info from EVERY unit of population.
• E.g., Population census.

NODE 07

Sample

• Subset selected from population for study.
• Representative fraction instead of entire group.

NODE 08

Population

• Complete set of units relevant to investigation.
• E.g., All relevant households in a city.

NODE 09

Census vs Sample

ALL

vs

SUB

• Sampling reduces time, cost, collection burden.
• But introduces sampling-error considerations.

NODE 10

Sampling Methods

• Random, systematic, stratified, cluster.
• Objective: obtain representative sample.

NODE 11

Random Sampling

• Each unit has specified probability of selection.
• Reduces selection bias.

NODE 12

Sampling Error

GAP

• Diff between sample estimate and population value.
• Larger sample can reduce this uncertainty.

NODE 13

Non-Sampling Error

X

• Incorrect responses, data-entry mistakes.
• Large sample ≠ automatically high-quality data.

NODE 14

Variable

• Characteristic taking different values.
• Income, age, price, employment status.

NODE 15

Qual vs Quant Data

vs123

• Qualitative: Categories (gender, occupation).
• Quantitative: Numerical (income, age, output).

NODE 16

Discrete Variable

1, 2, 3

• Takes distinct, countable values.
• E.g., Number of children, number of employees.

NODE 17

Continuous Variable

• Can take values over a continuous range.
• E.g., Height, weight, temperature, income.

NODE 18

Raw Data

12 18 25 19 18 31 12

• Data in original collected form.
• Before systematic organisation.

NODE 19

Classification of Data

• By qualitative/quantitative traits.
• By time, location, or other criteria.

NODE 20

Frequency

• How many times a value/category occurs.
• E.g., 25 households in the 20k-30k income group.

NODE 21

Frequency Distribution

VAL

FREQ

0-10

12

10-20

24

• Organizes observations into classes/values.
• Shows their respective frequencies.

NODE 22

Grouped Data

• Observations organized into intervals/classes.
• Useful for large numbers of observations.

NODE 23

Class Interval

• Interval between lower and upper boundaries.
• E.g., 20–30 is a class interval.

NODE 24

Class Width

• Size of the class interval.
• E.g., 20–30 → width = 10.

NODE 25

Cumulative Frequency

• Adds frequencies progressively.
• Less-than vs More-than Ogives.

NODE 26

Tabulation

T

I

T

L

E

!

• Presents data systematically in rows/columns.
• Should identify title, units, source, time period.

NODE 27

Diagrams

• Visual presentation of economic data.
• Bar, pie, histograms, line graphs.

NODE 28

Bar Graph

• For discrete or categorical data.
• Bars are generally separated. E.g., GDP by sector.

NODE 29

Histogram

• For frequency distribution of continuous data.
• Bars touch each other (no gaps).

NODE 30

Pie Chart

• Represents components of a whole (360°).
• Angle = (Component / Total) × 360°.

NODE 31

Line Graph

• Useful for showing changes over time.
• E.g., Inflation over years 2020 to 2023.

NODE 32

Arithmetic Mean

• Sum of observations ÷ Number of obs.
• Uses every observation in the dataset.

NODE 33

Weighted Mean

Σwx/Σw

• Used when observations have different importance.
• Crucial for economic indexes.

NODE 34

Median

• Middle value when data is ordered.
• Better for highly skewed income distributions.

NODE 35

Mode

• Value occurring most frequently.
• Depends purely on frequency peaks.

NODE 36

Mean vs Median vs Mode

MEAN (All)

MED (Mid)

MOD (Freq)

• Mean uses all data, sensitive to outliers.
• Median ignores extremes, good for income.

NODE 37

Dispersion

• How spread out are the observations?
• Range, Quartiles, Mean Dev, Std Dev, Variance.

NODE 38

Range

• Maximum value − Minimum value.
• Simple but highly sensitive to extreme observations.

NODE 39

Quartiles

• Divide ordered data into four parts.
• Q1 (25%), Q2 (Median), Q3 (75%).

NODE 40

Interquartile Range

• IQR = Q3 − Q1.
• Measures spread of the middle 50% of data.

NODE 41

Standard Deviation

• Dispersion of observations around their mean.
• Higher SD = greater dispersion.

NODE 42

Variance

SD²

• Squared measure of dispersion.
• Variance = (Standard Deviation)².

NODE 43

Coefficient of Variation

CV = (SD/Mean)

• Relative measure of dispersion.
• Useful for comparing datasets with different scales.

NODE 44

Skewness

• Asymmetry in a distribution.
• Right skew: long tail to higher values (Income).

NODE 45

Correlation

• Degree and direction of association.
• E.g., Income ↔ Consumption.

NODE 46

Positive Correlation

• Variables move in the same direction.
• X ↑ → Y ↑.

NODE 47

Negative Correlation

• Variables move in opposite directions.
• X ↑ → Y ↓.

NODE 48

Zero Correlation

• Little or no linear association.
• Scattered randomly.

NODE 49

Correlation Coeff (r)

-1≤ r ≤+1

• Value lies between -1 and +1.
• +1 = Perfect positive linear correlation.

NODE 50

Corr ≠ Causation

🍦☀️≠

• Most important data literacy rule.
• X & Y moving together doesn't mean X causes Y.

NODE 51

Time Series

• Observations across time (GDP by year).
• Trend, seasonality, cycles, irregular.

NODE 52

Trend

• Underlying long-term direction.
• E.g., Real GDP increasing over decades.

NODE 53

Seasonality

• Regular patterns repeating during periods.
• E.g., Festival demand, agri cycles.

NODE 54

Index Numbers

100

• Relative change compared to base period.
• Base is assigned 100.

NODE 55

Simple Price Index

120

/

100

* 100

• Price Relative = (Current / Base) × 100.

NODE 56

Price Index

• Tracks changes across a basket (CPI, WPI).
• Connects Realm 17 with Realm 11.

NODE 57

Quantity Index

• Measures changes in quantities of goods.

NODE 58

Value Index

• Changes in total monetary value of a basket.

NODE 59

Weighted Index

• Items get weights based on importance.
• More meaningful than equal weighting.

NODE 60

CPI as Econ Index

• Major example of index number.
• Tracks cost of consumer basket over time.

NODE 61

Nominal vs Real

N

vs

R

• Nominal: Current prices.
• Real: Adjusted for price-level changes.

NODE 62

Percentage Change

• (New − Old) ÷ Old × 100.

NODE 63

Growth Rate

• Rate at which a variable changes over time.

NODE 64

Compound Growth

• FV = PV × (1 + g)ⁿ.
• Useful for long-term economic projections.

NODE 65

Data Comparability

• Check definition, time period, units, base year.
• Otherwise comparisons are misleading.

NODE 66

Absolute vs Relative

10M

vs

6%

• 10M unemployed vs 6% unemployment rate.
• Population size dictates interpretation.

NODE 67

Rate vs Level

RATE5%

LEVEL10M

• Level: Absolute quantity.
• Rate: Quantity relative to denominator.

NODE 68

Per Capita Data

• Adjusts an aggregate by population.
• Improves comparisons but ignores inequality.

NODE 69

Real PC Income

• Accounts for both population AND price changes.
• Better for long-term living standards.

NODE 70

Base Effect

• Growth rates highly influenced by previous value.
• 100→200 (100%), then 200→220 (10%).

NODE 71

Index Base Effect

• Applies to inflation and indices.
• High/low previous base distorts current growth rate.

NODE 72

Data Bias

• Data ≠ automatically truth.
• Biased by sampling, selection, reporting.

NODE 73

Sample Selection Bias

• Systematically excluding certain groups.
• Sample ≠ representative population.

NODE 74

Survivorship Bias

X

• Analyzing only entities that survive.
• E.g., Studying successful startups, ignoring failed.

NODE 75

Cherry-Picking

• Selecting only data supporting a conclusion.
• Good analysis uses the full dataset.

NODE 76

Data Vis Literacy

• Check axes, units, time period, base.
• Is axis truncated? Scale consistent?

NODE 77

Meaningful Source

• Identify source, methodology, definition, units.
• Makes data reproducible and interpretable.

NODE 78

Official Statistics

• Published by govt statistical agencies.
• E.g., MoSPI/NSO in India.

NODE 79

Data & Policy

DATA↓POLICY

• Diagnosis → Design → Monitor → Evaluate.
• Data → Evidence → Policy.

NODE 80

Data → Theory

DATA↓THEORY

• Reality → Data → Stats → Model → Theory.
• Built into the Measurement Engine.

NODE 81

Statistical Misinterp

CORRCAUS

PCTPCT PT

NOMREAL

• Confusing correlation with causation.
• Confusing nominal with real, level with rate.

NODE 82

Pct vs Pct Point

5%→6%= 1PP

5%→6%= 20%

• Change of 5% to 6% is 1 percentage point.
• But it is a 20% increase. Not the same.

NODE 83

Real Data Lab

NOM - PR = REAL

• If Nom GDP ↑ 10% and Infl = 5%...
• Real GDP growth ≠ 10%. Adjust for prices.

NODE 84

Stats & Decision

• Question → Data → Clean → Measure.
• Visualize → Analyze → Interpret → Decide.

NODE 85

Master Data Lab Chain

• Question → Sample → Raw Data → Tables/Graphs.
• Mean/Dispersion/Correl → Econ Conclusion.

NODE 86

Master Literacy Chain

• Number → Unit? → Denominator? → Base?
• Collection Method? → Bias? → Actual Proof?

NODE 87

Formula Bank

x̄ = Σx/n

• Mean = Σx/n. Range = Max - Min. CV = SD/Mean*100.
• Correlation: -1 ≤ r ≤ +1.
• Pct Change = (New-Old)/Old*100.