Study guide
1.2 Sampling, Hypothesis Testing & Regression Analysis
Certified Investment Management Analyst CIMA
Track your progress as you read
Finished sections get a check mark.
1 Chapter 1: Statistical Methods & Quantitative Foundations
1.1 Descriptive Statistics, Probability Distributions & Return Metrics 1.2 Sampling, Hypothesis Testing & Regression Analysis 1.3 Time Series Analysis, Monte Carlo Simulation & Machine Learning 1.4 Time Value of Money, Cash Flow Discounting & Quantitative Decision Tools
2 Chapter 2: Applied Economics: Theory, Policy & the Business Cycle
2.1 Schools of Economic Thought & Core Economic Principles 2.2 Macroeconomic Indicators, Business Cycles & Central Bank Monetary Policy
3 Chapter 3: Global Economy, Trade & Currency Dynamics
3.1 Global Trade, Balance of Payments & International Institutions 3.2 Demographic Forces & Long-Run Economic Growth 3.3 Foreign Exchange Dynamics, Purchasing Power Parity & Currency Risk Management 3.4 Exchange Rate Regimes, Currency Pegs, Dollarization & Special Drawing Rights
4 Chapter 4: Global Capital Markets History & Valuation
4.1 Global Capital Markets History, Market Regimes & Long-Run Asset Class Returns 4.2 Sovereign Default History, Inflation Regimes & Emerging/Frontier Market Valuation
5 Chapter 5: Investment Vehicles & Managed Structures
5.1 Open-End Mutual Funds & Closed-End Funds: Structures and Trading Mechanics 5.2 Exchange-Traded Funds, Semi-Transparent Structures & Unit Investment Trusts 5.3 Separately Managed Accounts (SMAs), Unified Managed Accounts (UMAs) & Collective Investment Trusts
6 Chapter 6: Equity Analysis & Valuation
6.1 Intrinsic Equity Valuation: Dividend Discount & Discounted Cash Flow Models 6.2 Relative Valuation Multiples, Residual Income & Economic Value Added
7 Chapter 7: Equity Indices, Factor Strategies & Global Equity Allocation
7.1 Equity Index Construction Methodologies 7.2 Equity Factors & Smart Beta Index Construction 7.3 Factor Implementation Challenges & Multi-Factor Portfolio Integration 7.4 International Equity Diversification & Correlation Dynamics
8 Chapter 8: Fixed Income Fundamentals, Yield & Term Structure
8.1 Fixed Income Instruments, Pricing, Yield Measures & Term Structure of Interest Rates 8.2 Yield Curve Strategies, Spreads & Credit Analysis
9 Chapter 9: Fixed Income Risk, Duration & Structured Products
9.1 Duration, Convexity & Immunization Strategies 9.2 Mortgage-Backed Securities, Prepayment Risk & CMO Tranche Architecture 9.3 Asset-Backed Securities, Credit Enhancement & Collateralized Loan Obligations
10 Chapter 10: Alternative Investments: Hedge Funds, Private Markets & Investor Eligibility
10.1 Hedge Fund Strategies, Fund-of-Funds & Due Diligence Considerations 10.2 Private Equity Lifecycle, Leveraged Buyouts & Venture Capital 10.3 Private Credit, Direct Lending & Private Market Performance Measurement 10.4 Investor Eligibility Standards: Accredited Investors & Qualified Purchasers
11 Chapter 11: Derivatives & Risk Management Strategies
11.1 Option Fundamentals, Put-Call Parity & Black-Scholes-Merton Pricing 11.2 The Option Greeks & Institutional Option Overlay Strategies 11.3 Forwards, Futures Pricing, Basis & Institutional Hedging Applications 11.4 Swaps, Swaptions, Credit Default Swaps & Derivatives Execution Risk
12 Chapter 12: Real Assets, Commodities & Digital Assets
12.1 Real Estate Investment Trusts (REITs), Direct Real Estate & Infrastructure 12.2 Commodities, Natural Resources & Inflation Hedging 12.3 Digital Assets: Definition, Major Assets, Trends, Benefits & Risks
13 Chapter 13: Portfolio Theory & Asset Pricing Models
13.1 Markowitz Portfolio Theory, Mean-Variance Optimization & the Efficient Frontier 13.2 Capital Allocation, Investor Utility & the Practical Limits of Mean-Variance Optimization 13.3 The Capital Asset Pricing Model, CML versus SML & Beta Estimation 13.4 Arbitrage Pricing Theory, Multifactor & Statistical Factor Models
14 Chapter 14: Behavioral Finance & Decision Architecture
14.1 Heuristics, Cognitive Biases & Emotional Fallacies 14.2 Prospect Theory, Loss Aversion & Overcoming Behavioral Pitfalls in Portfolio Management
15 Chapter 15: Investment Philosophies, Styles & Allocation Strategies
15.1 Strategic Asset Allocation (SAA) vs. Tactical Asset Allocation (TAA) 15.2 Active vs. Passive Management, Core-Satellite & Factor Allocation Frameworks 15.3 Technical Analysis: Trend, Supply and Demand & Momentum Indicators 15.4 Sustainable Investing: SRI, ESG Integration & Impact Investing
16 Chapter 16: Portfolio Construction Methodologies & LDI
16.1 Liability-Driven Investing (LDI) & Goals-Based Wealth Management 16.2 Portfolio Rebalancing Triggers, Tax-Aware Construction & Liquidity Management
17 Chapter 17: Risk Measurement & Downside Analytics
17.1 Downside Risk Metrics: Semivariance, Maximum Drawdown & Sortino Ratio 17.2 Value at Risk & Conditional Value at Risk 17.3 Stress Testing, Scenario Analysis & Reverse Stress Testing
18 Chapter 18: Performance Measurement, Attribution & GIPS Standards
18.1 Risk-Adjusted Return Measures: Sharpe, M-Squared, Treynor & Jensen's Alpha 18.2 The Information Ratio & Selecting the Right Risk-Adjusted Metric 18.3 Brinson Attribution, Sector/Security Selection & Global Investment Performance Standards (GIPS)
19 Chapter 19: Professional Ethics & the IWI Code of Professional Responsibility
19.1 The IWI Code of Professional Responsibility: The Nine Principles 19.2 IWI Disciplinary Rules, the Professional Review Board & Professional Practice Standards
20 Chapter 20: Fiduciary Duties & Regulatory Frameworks
20.1 The Fiduciary Standard, Suitability, Regulation Best Interest & the Advisers Act 20.2 ERISA Fiduciary Standards & Conflict of Interest Management
21 Chapter 21: Client Discovery, Constraints & IPS Development
21.1 Client Profile Discovery: Risk Tolerance, Capacity & Time Horizon 21.2 Investment Policy Statement (IPS) Formulation & Constraint Analysis (RRTTLLU)
22 Chapter 22: Implementation, Manager Due Diligence & Portfolio Governance
22.1 Investment Implementation Approaches: In-House, Sub-Advisory & Outsourced CIO (OCIO) 22.2 Manager Search, Selection & Operational Due Diligence 22.3 Portfolio Monitoring, Performance Reporting & Governance Reviews
1.2 Sampling, Hypothesis Testing & Regression Analysis
Key Takeaways Explain This Topic AI
The standard error of a sample mean is the sample standard deviation divided by the square root of the observation count, so quadrupling the sample only halves estimation error.
Three years of monthly returns (n = 36) is almost never enough to establish that a manager's alpha is statistically different from zero at the 95% confidence level.
In a single-factor regression of fund returns on benchmark returns, the slope is beta, the intercept is Jensen's alpha, and R-squared is the fraction of return variation explained by the benchmark.
Multicollinearity inflates coefficient standard errors without biasing the coefficients, producing a high R-squared alongside individually insignificant factor loadings.
Serial correlation in reported returns — common in appraisal-priced real estate and private funds — biases standard errors downward and makes risk look artificially low.
Last updated: August 2026
- 1.2 Sampling, Hypothesis Testing & Regression Analysis
- 1. From Population to Sample: Sampling Distributions and Standard Error
Every statistic an investment consultant quotes — a manager's average excess return, an asset class's volatility, a factor loading — is a sample estimate drawn from an unobservable population . The CIMA exam repeatedly tests whether a candidate understands that an estimate is a random variable with its own distribution, not a fact.
The sampling distribution of the sample mean has a standard deviation called the standard error of the mean : S E x ˉ = s n SE_{\bar{x}} = \frac{s}{\sqrt{n}} S E x ˉ = n s
where s s s is the sample standard deviation and n n n is the number of observations. Because n n n enters under a square root, precision improves slowly: to halve the standard error you must quadruple the sample. This single fact explains why performance track records are so weak as evidence.
The Central Limit Theorem (CLT) states that the sampling distribution of the mean approaches a normal distribution as n n n grows, regardless of the shape of the underlying population. In practice, n ≥ 30 n \geq 30 n ≥ 30 is the conventional threshold. The CLT is what licenses the use of normal-based confidence intervals on mean returns even when the return series itself is skewed and fat-tailed.
When the population standard deviation is unknown — always the case in practice — the correct reference distribution is Student's t with n − 1 n - 1 n − 1 degrees of freedom , not the standard normal. The t-distribution has fatter tails, so it produces wider (more honest) confidence intervals. The two converge as n n n rises above roughly 120.
- Sample size (n) Degrees of freedom Two-tailed 95% critical value
- 10 9 2.262
- 20 19 2.093
- 36 35 2.030
- 60 59 2.001
- 120 119 1.980
- ∞ \infty ∞ (normal) — 1.960
A confidence interval for the mean is: x ˉ ± t α / 2 , n − 1 × s n \bar{x} \pm t_{\alpha/2,\, n-1} \times \frac{s}{\sqrt{n}} x ˉ ± t α /2 , n − 1 × n s
2. The Hypothesis-Testing Framework
A hypothesis test formalizes the question "could this result plausibly be luck?" The structure is fixed:
State the null hypothesis H 0 H_0 H 0 — the position of no effect, e.g. "the manager's true alpha is zero."
State the alternative H a H_a H a — e.g. "true alpha is not zero" (two-tailed) or "true alpha is positive" (one-tailed).
Choose a significance level α \alpha α , conventionally 0.05.
Compute the test statistic , e.g. t = ( x ˉ − μ 0 ) / S E t = (\bar{x} - \mu_0) / SE t = ( x ˉ − μ 0 ) / SE .
Compare the statistic to the critical value, or compare the p-value to α \alpha α .
The p-value is the probability of observing a result at least as extreme as the one obtained if the null were true . It is not the probability that the null is true — a distinction the exam tests directly.
Two errors are possible:
H 0 H_0 H 0 actually true H 0 H_0 H 0 actually false
Reject H 0 H_0 H 0 Type I error (false positive), probability = α \alpha α Correct decision ( power = 1 − β 1 - \beta 1 − β )
Fail to reject H 0 H_0 H 0 Correct decision Type II error (false negative), probability = β \beta β
Lowering α \alpha α to reduce false positives mechanically raises β \beta β and reduces power. Only a larger sample improves both simultaneously.
Worked Example: Is This Manager's Alpha Real?
A consultant reviews a large-cap manager with 36 monthly excess returns . The sample mean monthly excess return is +0.90% with a sample standard deviation of 4.20% .
Standard error: S E = 4.20 % / 36 = 4.20 % / 6 = 0.70 % SE = 4.20\% / \sqrt{36} = 4.20\% / 6 = 0.70\% SE = 4.20%/ 36 = 4.20%/6 = 0.70%
Test statistic: t = ( 0.90 % − 0 % ) / 0.70 % = 1.286 t = (0.90\% - 0\%) / 0.70\% = 1.286 t = ( 0.90% − 0% ) /0.70% = 1.286
Critical value: two-tailed, α = 0.05 \alpha = 0.05 α = 0.05 , df = 35 → \rightarrow → 2.030
Decision: 1.286 < 2.030 1.286 < 2.030 1.286 < 2.030 , so fail to reject H 0 H_0 H 0 .
Despite an annualized excess return of roughly 10.8%, three years of data cannot distinguish this manager from luck. Solving for the sample size that would be needed at this signal-to-noise ratio requires t ≥ 2 t \geq 2 t ≥ 2 , i.e. n ≥ ( 2 × 4.20 / 0.90 ) 2 ≈ 87 n \geq (2 \times 4.20 / 0.90)^2 \approx 87 n ≥ ( 2 × 4.20/0.90 ) 2 ≈ 87 months — more than seven years. This is the quantitative justification for the consulting convention of requiring full market-cycle track records.
3. Simple Linear Regression: Estimating Beta and Alpha
The single-index model regresses a fund's excess return on the benchmark's excess return: R p , t − R f , t = α + β ( R m , t − R f , t ) + ε t R_{p,t} - R_{f,t} = \alpha + \beta (R_{m,t} - R_{f,t}) + \varepsilon_t R p , t − R f , t = α + β ( R m , t − R f , t ) + ε t
Ordinary least squares (OLS) chooses α \alpha α and β \beta β to minimize the sum of squared residuals. The slope has a closed form: β = C o v ( R p , R m ) V a r ( R m ) = ρ p , m × σ p σ m \beta = \frac{Cov(R_p, R_m)}{Var(R_m)} = \rho_{p,m} \times \frac{\sigma_p}{\sigma_m} β = Va r ( R m ) C o v ( R p , R m ) = ρ p , m × σ m σ p
The intercept α \alpha α is Jensen's alpha — return unexplained by benchmark exposure.
The slope β \beta β is systematic risk relative to the benchmark.
R-squared is the proportion of the fund's return variance explained by the benchmark; in a single-factor regression R 2 = ρ 2 R^2 = \rho^2 R 2 = ρ 2 .
The standard error of the estimate measures typical residual size; residual (non-systematic) risk is the volatility of ε \varepsilon ε .
Worked Example: Beta and R-Squared
A fund has σ p = 18.0 % \sigma_p = 18.0\% σ p = 18.0% , the benchmark has σ m = 16.0 % \sigma_m = 16.0\% σ m = 16.0% , and their correlation is ρ = 0.90 \rho = 0.90 ρ = 0.90 . β = 0.90 × 18.0 16.0 = 0.90 × 1.125 = 1.0125 \beta = 0.90 \times \frac{18.0}{16.0} = 0.90 \times 1.125 = 1.0125 β = 0.90 × 16.0 18.0 = 0.90 × 1.125 = 1.0125 R 2 = 0.90 2 = 0.81 R^2 = 0.90^2 = 0.81 R 2 = 0.9 0 2 = 0.81
81% of the fund's return variation is benchmark-driven; the remaining 19% is idiosyncratic. An R 2 R^2 R 2 this high indicates the benchmark is appropriate — a critical input to the benchmark-selection process covered in the manager-monitoring material.
4. Multiple Regression and Model Diagnostics
Multi-factor attribution extends the model to several explanatory variables: R p − R f = α + β 1 F 1 + β 2 F 2 + ⋯ + β k F k + ε R_p - R_f = \alpha + \beta_1 F_1 + \beta_2 F_2 + \dots + \beta_k F_k + \varepsilon R p − R f = α + β 1 F 1 + β 2 F 2 + ⋯ + β k F k + ε
Because adding any regressor mechanically raises R 2 R^2 R 2 , multiple regression is judged on adjusted R-squared , which penalizes the addition of variables that do not earn their degrees of freedom: R a d j 2 = 1 − ( 1 − R 2 ) n − 1 n − k − 1 R^2_{adj} = 1 - (1 - R^2)\frac{n-1}{n-k-1} R a d j 2 = 1 − ( 1 − R 2 ) n − k − 1 n − 1
Adjusted R 2 R^2 R 2 can fall when a useless factor is added — that decline is the signal to drop it.
OLS inference relies on assumptions, and violations are a favorite exam target because each one has a different consequence:
Violation Definition Effect on coefficients Effect on inference Typical diagnostic
Multicollinearity Regressors highly correlated with one another (e.g. value and quality factors) Unbiased Standard errors inflated ; high overall R 2 R^2 R 2 with individually insignificant t-stats Variance inflation factor (VIF)
Heteroskedasticity Residual variance changes with the level of a regressor (volatility clustering) Unbiased Standard errors biased; t-stats unreliable Breusch–Pagan test
Serial correlation Residuals correlated across time (appraisal-based or stale pricing) Unbiased Standard errors biased downward ; significance overstated, risk understated Durbin–Watson statistic
The practical consulting warning: serial correlation is pervasive in private real estate, private credit, and fund-of-funds returns because valuations are appraisal-based and smoothed. Smoothed returns understate true volatility, understate beta, and inflate Sharpe ratios. A Durbin–Watson statistic materially below 2.0 (positive autocorrelation) should trigger an unsmoothing adjustment before those returns enter a mean-variance optimizer.