Moments Skewness Kurtosis Questions for SSC CGL

What Are Moments?

Moments are numerical measures that capture different properties of a distribution. Think of them as 'snapshots' of data: the first moment tells you where the data is centered (mean), the second tells you how spread out it is (variance), and higher moments reveal shape. There are two types: raw moments (calculated from origin) and central moments (calculated from mean). Central moments are more useful because they ignore location and focus on shape.

• Raw moment of order r: Mr = (1/n) × sum of (Xi)^r
• Central moment of order r: μr = (1/n) × sum of (Xi - mean)^r
• First raw moment = mean; second central moment = variance
• Third and fourth central moments define skewness and kurtosis
• Moments are building blocks for understanding distribution shape

Skewness: Measuring Asymmetry

Skewness measures whether a distribution leans left or right. A symmetric distribution (like normal curve) has zero skewness. Positive skewness means the tail stretches to the right (mean > median > mode); negative skewness means tail stretches left (mean < median < mode). In SSC, you'll see questions asking you to identify skewness from histograms or calculate it from data. The formula uses the third central moment divided by the cube of standard deviation.

• Positive skew: right tail longer, mean pulled right, data clusters on left
• Negative skew: left tail longer, mean pulled left, data clusters on right
• Zero skew: perfectly symmetric (normal distribution, uniform distribution)
• Skewness range: −3 to +3 (extreme values rare)
• Interpretation: |skewness| < 0.5 = fairly symmetric; 0.5−1 = moderately skewed; > 1 = highly skewed

Kurtosis: Measuring Tail Heaviness

Kurtosis measures how heavy or light the tails of a distribution are compared to a normal distribution. High kurtosis means fat tails (extreme values more likely) and sharp peak; low kurtosis means thin tails and flat peak. SSC questions test whether you can identify leptokurtic (peaked), mesokurtic (normal), or platykurtic (flat) distributions from graphs or data. The formula uses the fourth central moment divided by the fourth power of standard deviation, minus 3 (the '−3' makes normal distribution have kurtosis = 0).

• Excess kurtosis = kurtosis − 3 (normal distribution has excess kurtosis = 0)
• Leptokurtic (excess > 0): sharp peak, fat tails, extreme values more common
• Mesokurtic (excess = 0): normal distribution shape
• Platykurtic (excess < 0): flat peak, thin tails, data more uniform
• High kurtosis = risk of outliers; low kurtosis = data more concentrated near center

Relationship Between Moments, Skewness, and Kurtosis

These three concepts form a hierarchy. Moments are the foundation; skewness and kurtosis are derived from moments. Understanding this chain helps you solve complex problems. First moment = location (mean). Second moment = spread (variance). Third moment = skewness (asymmetry). Fourth moment = kurtosis (tail behavior). In SSC, you may see questions that ask you to calculate multiple moments from the same dataset, or to interpret how a change in data affects all three properties.

• Moments are raw data; skewness and kurtosis are interpretations of moments
• All three describe distribution shape; together they give a complete picture
• Normal distribution: all odd moments = 0, skewness = 0, excess kurtosis = 0
• Changing outliers affects higher moments more than lower moments
• Skewness and kurtosis are scale-invariant (dividing all data by constant doesn't change them)

• Confusing raw moments with central moments. Raw moments depend on origin; central moments don't. Always use central moments for skewness/kurtosis unless explicitly told otherwise.
• Forgetting the '−3' in excess kurtosis formula. Many students calculate μ4/σ^4 and call it kurtosis, but SSC expects excess kurtosis (subtract 3). Normal distribution should give 0, not 3.
• Misinterpreting positive vs negative skewness. Remember: positive skew = right tail (mean > median), NOT 'more data on right'. The tail points right, but bulk of data is on left.
• Assuming high kurtosis = bad. High kurtosis just means fat tails; whether that's bad depends on context. In finance, it's risky. In quality control, it might be acceptable.
• Calculating skewness from grouped data without adjusting for class width. If data is grouped, use class midpoints and frequency, not raw values.

• MOMENTS = M.O.M.E.N.T.S: Mean (1st), Outliers (2nd via variance), More outliers (3rd = skewness), Even more (4th = kurtosis), Never forget central, Third and fourth matter most.
• SKEW = Stretch: Positive skew = tail stretches right (like a right arrow →). Negative skew = tail stretches left (like a left arrow ←). Mean chases the tail.
• KURT = Kurtosis = 'Curve' shape: Lepto = Lean/Lofty (peaked), Meso = Medium (normal), Platy = Platypus (flat, spread out).
• FORMULA CHAIN: Raw moment M'r → Central moment μr (subtract mean) → Divide by σ^r → Skewness (r=3) or Kurtosis (r=4, minus 3).

• SSC Tier-II typically asks 1-2 direct calculation questions: 'Calculate skewness/kurtosis from given data.' Memorize the formulas and practice with 5-10 datasets. Speed matters; use Pearson's formula when possible (faster than calculating all moments).
• Graph interpretation questions are common: 'Which histogram shows positive skewness?' or 'Identify the distribution with highest kurtosis.' Train your eye to spot right tails (positive skew), left tails (negative skew), and peaked vs flat shapes.
• Conceptual questions appear in Tier-I: 'If mean > median, the distribution is ___.' Answer: positively skewed. These are quick 1-mark wins if you remember the relationship.
• Watch for 'excess kurtosis' vs 'kurtosis' in question wording. If it says 'excess,' subtract 3 from your calculation. If it just says 'kurtosis,' check the answer options to infer which definition they're using.
• Recent papers (2022-2024) show more emphasis on interpreting real-world data (income, test scores, stock returns) rather than abstract calculations. Focus on understanding what skewness/kurtosis mean in context, not just formulas.