Study guide

5.2 Upper Confidence Limits (UCL95) and Compliance Decision Rules

CIH Certified Industrial Hygienist Exam

Track your progress as you read

Finished sections get a check mark.

1 Chapter 1: The CIH Exam and How to Use This Guide

1.1 CIH Exam Format, Scoring, and Test-Day Logistics 1.2 Eligibility, Application Windows, Fees, and Certification Maintenance 1.3 The BGC Blueprint, Subject Areas, and a Calculation-First Study Plan

2 Chapter 2: Basic Science & Calculations

2.1 Gas Laws and Vapor Calculations 2.2 Unit Conversions: ppm, mg/m³, and Standard Temperature/Pressure 2.3 Fluid Mechanics and Airflow Fundamentals

3 Chapter 3: Air Sampling & Instrumentation

3.1 Active and Passive Air Sampling Strategies 3.2 Particulate and Aerosol Sampling 3.3 Direct-Reading Instruments and Calibration

4 Chapter 4: Analytical Chemistry & Lab Methods

4.1 Gas Chromatography and Mass Spectrometry 4.2 High-Performance Liquid Chromatography and Ion Chromatography 4.3 Spectroscopy and Elemental Analysis 4.4 Gravimetric, Microscopy, and Wet Chemical Methods 4.5 Laboratory Quality Assurance, Chain of Custody, and Desorption Efficiency

5 Chapter 5: Biostatistics & Epidemiology

5.1 Lognormal Distributions and Exposure Data Statistics 5.2 Upper Confidence Limits (UCL95) and Compliance Decision Rules 5.3 Epidemiological Study Designs and Measures of Association 5.4 Sampling Strategy: Sample Size, Power, and Worker Selection

6 Chapter 6: Toxicology & Biological Monitoring

6.1 Dose-Response Relationships and Points of Departure 6.2 Toxicokinetics: Absorption, Distribution, Metabolism, and Elimination 6.3 Aerosol Deposition, Clearance, and Respiratory Tract Dosimetry 6.4 Hepatic, Renal, Neurologic, and Hematologic Target Organ Toxicity 6.5 Pulmonary Toxicology, Carcinogenicity, and Reproductive Hazards 6.6 OEL Frameworks, Classifications, and Excursion Limits 6.7 Chemical Mixtures and Extended Work-Shift OEL Adjustments 6.8 Biological Exposure Indices (BEIs) and Biomonitoring

7 Chapter 7: Biohazards & Indoor Environmental Quality

7.1 Mold, Fungi, and Microbial Contaminants 7.2 Infectious Agents, Bloodborne Pathogens, and Biosafety Levels 7.3 Indoor Environmental Quality (IEQ) Investigations 7.4 Sanitation, Potable Water, and Personal Hygiene Programs

8 Chapter 8: Community Exposure & Environmental Health

8.1 Clean Air Act Framework, NAAQS, and New Source Review 8.2 Hazardous Air Pollutants, CEMS, and EPA Reference Methods 8.3 Atmospheric Stability, Lapse Rates, and Plume Morphology 8.4 Gaussian Plume Modeling, Plume Rise, and Environmental Fate 8.5 HAZWOPER Scope, Mandates, and Training Tiers 8.6 Protection Levels, Site Zoning, Incident Command, and Decontamination 8.7 Water Pollution, Hazardous Waste, and Community Right-to-Know

9 Chapter 9: Health Risk Analysis & Hazard Communication

9.1 GHS Hazard Classification: Health and Physical Hazard Criteria 9.2 GHS Labels, Safety Data Sheets, and Trade Secret Rules 9.3 The AIHA Exposure Assessment Strategy and Similar Exposure Groups 9.4 Qualitative Exposure Profiling and the AIHA Risk Matrix 9.5 Quantitative Health Risk Characterization: HQ, HI, and ELCR 9.6 NIOSH Occupational Exposure Banding 9.7 Risk Communication and Reporting to Affected Parties

10 Chapter 10: Industrial Ventilation Principles & LEV Design

10.1 Elimination, Substitution, Isolation, Containment, and Process Change 10.2 LEV Hood Taxonomy and Capture Velocity Selection 10.3 Hood Airflow Equations, Entry Losses, and Hood Static Pressure 10.4 Particulate Transport Aerodynamics and Duct Sizing 10.5 Duct Friction, Fitting Losses, and System Balancing 10.6 Fan Categories and Pressure Dynamics Across the Fan 10.7 System Curves, the Fan Laws, and Density Corrections

11 Chapter 11: General Ventilation, Air Cleaning & Diagnostics

11.1 Dilution Ventilation Principles and Steady-State Modeling 11.2 Transient Purging and Air Distribution Geometry 11.3 Particulate Air Cleaning: Cyclones, Baghouses, Scrubbers, and ESPs 11.4 Gaseous Emission Control and Air Cleaner Selection 11.5 Ventilation System Troubleshooting and Testing

12 Chapter 12: Noise Measurement & Hearing Conservation

12.1 Acoustic Physics and Decibel Mathematics 12.2 Sound Propagation, Octave Bands, and Frequency Weighting 12.3 Noise Dose Metrics and OSHA versus ACGIH Criteria 12.4 Exchange Rates, Extended Shifts, and Dosimetry Field Protocols 12.5 Engineering Noise Controls and Hearing Protector Attenuation 12.6 Hearing Conservation Programs and Standard Threshold Shifts 12.7 Hand-Arm and Whole-Body Vibration

13 Chapter 13: Thermal Stress Assessment & Control

13.1 Wet Bulb Globe Temperature (WBGT) and Heat Stress Indices 13.2 Heat-Related Disorders, Acclimatization, and Work/Rest Regimens 13.3 Cold Stress Evaluation and Wind Chill

14 Chapter 14: Ionizing Radiation Protection

14.1 Ionizing Radiation Physics and Radioactive Decay Kinetics 14.2 Radiation Units, Radiobiology, and Health Effects 14.3 ALARA, Stay-Time, and the Inverse Square Law 14.4 Shielding Physics, Material Selection, and Gamma Constants 14.5 Radiation Detection Instruments and Spectrometry 14.6 Personnel Dosimetry, Dose Limits, and Contamination Surveys

15 Chapter 15: Non-Ionizing Radiation & Laser Safety

15.1 The Non-Ionizing Spectrum and Static/ELF Field Exposures 15.2 Radiofrequency and Microwave Exposure: SAR, Standards, and Surveys 15.3 Ultraviolet, Visible, and Infrared Radiation Hazards 15.4 Laser Safety Classifications and Optical Density Calculations 15.5 Illumination, Lighting Design, and Visual Performance

16 Chapter 16: Non-Engineering Controls & Personal Protective Equipment

16.1 Respiratory Protection Programs and Assigned Protection Factors 16.2 Maximum Use Concentration, Oxygen Deficiency, and Breathing Air 16.3 Particulate Filters, Sorbent Cartridges, and Change Schedules 16.4 Respirator Fit Testing Protocols 16.5 Chemical Protective Clothing and Dermal Exposure Control

17 Chapter 17: Ergonomics & Workstation Design

17.1 Soft-Tissue Biomechanics and WMSD Pathophysiology 17.2 Ergonomic Risk Factors and Their Interaction Effects 17.3 The Revised NIOSH Lifting Equation 17.4 RULA, REBA, the Strain Index, and the ACGIH HAL TLV 17.5 Anthropometry and Workstation Geometry 17.6 VDT Workstations, Hand Tool Design, and Administrative Ergonomics

18 Chapter 18: Work Environments & Industrial Processes

18.1 Welding Fume Emissions, Metal Fume Fever, and Cr(VI)/Manganese Toxicology 18.2 Foundry Operations and Hot Work Fire Prevention 18.3 Vapor Degreasing, Pickling, Electroplating, and Cyanide Baths 18.4 Spray Coatings, Isocyanates, and Semiconductor Manufacturing 18.5 Confined Spaces and High-Hazard Industrial Operations 18.6 Asbestos, Lead, and Silica: Regulated Substance Programs 18.7 Abrasive Blasting and Surface Preparation

19 Chapter 19: IH Program Management & Professional Ethics

19.1 The BGC Code of Ethics: Canons and Enforcement 19.2 Professional Liability, Standard of Care, and Expert Testimony 19.3 Occupational Health & Safety Management Systems (ISO 45001, ANSI Z10) 19.4 Exposure Records Access and Medical Surveillance 19.5 OSHA Injury and Illness Recordkeeping and Enforcement

5.2 Upper Confidence Limits (UCL95) and Compliance Decision Rules

Key Takeaways Explain This Topic AI

The 95% Upper Confidence Limit of the Arithmetic Mean (UCL95) provides 95% statistical confidence that the true long-term population average exposure is below the Occupational Exposure Limit (OEL).

Applying Student's t-distribution UCL formulas to right-skewed lognormal datasets severely underestimates true upper confidence bounds; exact methods such as Land's exact H-statistic, Cox's method, or bootstrap UCLs must be utilized.

The 95% Upper Tolerance Limit of the 95th percentile (UTL95%,95%) accounts for small sample sizes to ensure with 95% confidence that at least 95% of all daily shift exposures fall below the limit.

The AIHA exposure control framework categorizes Similar Exposure Groups into 5 standardized tiers: Category 0 (<1% OEL), Category 1 (<10% OEL), Category 2 (10-50% OEL), Category 3 (50-100% OEL), and Category 4 (>100% OEL).

Under OSHA compliance rules, issuing a citation requires proving non-compliance beyond reasonable statistical doubt (>=95% confidence), established when the Lower Confidence Limit exceeds unity: LCL1,95% = (Measured Conc / PEL) - 1.645 · CV_T > 1.0.

Last updated: August 2026

Upper Confidence Limits (UCL95) and Compliance Decision Rules

In occupational hygiene practice, sample datasets are inherently limited in size—often consisting of only 3 to 10 personal air samples per Similar Exposure Group (SEG). Because of random sampling variability, the sample mean (x̄) or sample 95th percentile (X95) is merely a point estimate and does not guarantee that the true workplace population distribution complies with an Occupational Exposure Limit (OEL).

To make legally defensible, health-protective compliance decisions, industrial hygienists use Upper Confidence Limits (UCL) , Upper Tolerance Limits (UTL) , and formal compliance decision rules established by the American Industrial Hygiene Association (AIHA), NIOSH, and OSHA.

1. Statistical Confidence in Exposure Assessment

When evaluating a group of workers, two distinct statistical parameters are evaluated depending on whether the chemical produces chronic cumulative toxicity or acute peak toxicity:

Long-Term Average Exposure (Population Arithmetic Mean, µ): Evaluated against chronic 8-hour TWA exposure limits to prevent cumulative systemic diseases (e.g., lead plumbism, pneumoconiosis, solvent encephalopathy). Evaluated via the 95% Upper Confidence Limit of the Mean (UCL(1-α)) .

Peak Upper-Tail Exposure (Population 95th Percentile, X95): Evaluated against acute limits, ceiling limits, STELs, and modern AIHA exposure control strategies. Evaluated via the 95% Upper Tolerance Limit of the 95th Percentile (UTL(95%, 95%)) .

Confidence Limit Definitions

One-Sided Upper Confidence Limit (UCL(95%)): A calculated upper numerical bound such that there is a 95% statistical probability (1 - α = 0.95) that the true population parameter (e.g., mean µ) is less than or equal to the UCL.

One-Sided Lower Confidence Limit (LCL(95%)): A calculated lower numerical bound such that there is a 95% statistical probability that the true population parameter exceeds the LCL.

  • 2. Methods for Calculating Upper Confidence Limits of the Mean
  • The Student's t-Distribution (Normal Data Only)

For normally distributed data, the one-sided (1 - α) Upper Confidence Limit of the arithmetic mean is calculated as: U C L 1 − α = x ˉ + t 1 − α , n − 1 ⋅ s n UCL_{1-\alpha} = \bar{x} + t_{1-\alpha, \, n-1} \cdot \frac{s}{\sqrt{n}} U C L 1 − α ​ = x ˉ + t 1 − α , n − 1 ​ ⋅ n ​ s ​

Where:

  • x̄ = sample arithmetic mean
  • s = sample standard deviation
  • n = sample size

t(1-α, n-1) = critical value from the Student's t-distribution with n - 1 degrees of freedom at significance level α (e.g., t(0.95, 5) = 2.015 for n = 6)

Fatal Flaw for Exposure Data: Applying the Student's t-interval to lognormally distributed exposure data results in severe under-coverage . Because lognormal distributions are right-skewed, standard t-intervals substantially underestimate the true upper confidence limit of the arithmetic mean, falsely certifying hazardous workplaces as compliant.

Land's Exact Method (H-Statistic)

C.E. Land (1971) developed the mathematically exact parametric method for constructing two-sided and one-sided confidence limits on the arithmetic mean of a lognormal distribution (E[X] = exp(µy + 0.5σy²)).

The one-sided (1 - α) Upper Confidence Limit using Land's method is: U C L 1 − α = exp ⁡ ( y ˉ + 1 2 s y 2 + s y ⋅ H 1 − α , n , s y n − 1 ) \mathbf{UCL_{1-\alpha} = \exp\left( \bar{y} + \frac{1}{2} s_y^2 + \frac{s_y \cdot H_{1-\alpha, \, n, \, s_y}}{\sqrt{n - 1}} \right)} UC L 1 − α ​ = exp ( y ˉ ​ + 2 1 ​ s y 2 ​ + n − 1 ​ s y ​ ⋅ H 1 − α , n , s y ​ ​ ​ )

Where:

ȳ and sy are the sample mean and sample standard deviation of the log-transformed data ln(x)

H(1-α, n, sy) is Land's tabulated H-statistic, which depends on the confidence level (1-α), sample size (n), and log-scale standard deviation (sy)

Limitations of Land's Method: When sample sizes are small (n < 10) and variability is high (sy > 1.0 or GSD > 2.7), Land's H-factor increases exponentially, often yielding an astronomical and practically uninformative UCL value that exceeds the physical saturation concentration of the chemical.

Cox's Direct Approximation Method

Cox proposed a direct approximation for lognormal UCLs based on the delta method. When sample sizes exceed 15 to 20, Cox's modified UCL provides a robust estimate: U C L 1 − α = exp ⁡ ( y ˉ + s y 2 2 + Z 1 − α ⋅ s y 2 n + s y 4 2 ( n − 1 ) ) UCL_{1-\alpha} = \exp\left( \bar{y} + \frac{s_y^2}{2} + Z_{1-\alpha} \cdot \sqrt{\frac{s_y^2}{n} + \frac{s_y^4}{2(n - 1)}} \right) U C L 1 − α ​ = exp ​ y ˉ ​ + 2 s y 2 ​ ​ + Z 1 − α ​ ⋅ n s y 2 ​ ​ + 2 ( n − 1 ) s y 4 ​ ​ ​ ​

Non-Parametric Bootstrap & Chebyshev UCLs

When exposure data cannot be proven to fit a lognormal distribution or when dealing with highly skewed datasets containing non-detects (censored values):

Chebyshev Inequality UCL: Calculates a conservative, distribution-free upper bound: UCL(95%) = x̄ + √(1/α - 1) · s/√n = x̄ + 4.359 · s/√n (for α = 0.05).

Bias-Corrected and Accelerated (BCa) Bootstrap UCL: Resamples the dataset with replacement 10,000 times to generate an empirical sampling distribution of the mean, adjusting for bias and skewness.

3. Upper Tolerance Limits (UTL(95%, 95%)) for the 95th Percentile

While the point estimate formula X95 = GM · GSD(1.645) assumes the population parameters µy and σy are known exactly, in practice they are estimated from small samples (n). Sampling error introduces uncertainty in both ȳ and sy.

An Upper Tolerance Limit (UTL(p, 1-α)) provides (1 - α) confidence that at least proportion p of the entire population distribution falls below the limit. In occupational hygiene, the standard benchmark is the UTL(95%, 95%) (95% confidence that 95% of daily shifts are below the limit). ln ⁡ ( U T L p , 1 − α ) = y ˉ + K n , p , 1 − α ⋅ s y ⟹ U T L p , 1 − α = exp ⁡ ( y ˉ + K ⋅ s y ) \mathbf{\ln(UTL_{p, \, 1-\alpha}) = \bar{y} + K_{n, \, p, \, 1-\alpha} \cdot s_y} \implies \mathbf{UTL_{p, \, 1-\alpha} = \exp\left( \bar{y} + K \cdot s_y \right)} ln ( UT L p , 1 − α ​ ) = y ˉ ​ + K n , p , 1 − α ​ ⋅ s y ​ ⟹ UT L p , 1 − α ​ = exp ( y ˉ ​ + K ⋅ s y ​ )

Where K(n, p, 1-α) is the one-sided tolerance factor derived from the non-central t-distribution:

One-Sided Tolerance Factors (K) for p = 0.95 and 1 - α = 0.95

Sample Size (n) Tolerance Factor (K(0.95, 0.95)) Standard Normal Z(0.95) (Infinite n) Impact of Sample Size Uncertainty

  • n = 3 7.656 1.645 Extreme penalty for tiny sample size (K / Z = 4.65)
  • n = 5 4.203 1.645 High uncertainty factor (K / Z = 2.55)
  • n = 6 3.708 1.645 Common minimum sample size in IH surveys
  • n = 10 2.911 1.645 Moderate uncertainty penalty
  • n = 20 2.396 1.645 Stable estimate
  • n = 50 2.065 1.645 Approaching asymptotic Z-score
  • n = ∞ 1.645 1.645 Exact population parameter value

Key Principle: As sample size n increases, K decreases toward 1.645. Collecting more samples directly reduces statistical uncertainty, lowering the calculated UTL(95%, 95%) without altering the physical environment.

4. AIHA Exposure Decision Categories (5-Tier Framework)

The AIHA Exposure Assessment Strategy establishes a standardized 5-tier classification framework based on the estimated 95th percentile (X95) or UTL(95%, 95%) expressed as a fraction of the Occupational Exposure Limit (OEL):

Exposure Category Exposure Level (% OEL) Statistical Definition (X95) Recommended Industrial Hygiene Action & Control Strategy

Category 0 < 1% of OEL X95 < 0.01 · OEL Trivial / Negligible Exposure. No action required; routine re-evaluation only upon major process or chemical change.

Category 1 < 10% of OEL 0.01 · OEL ≤ X95 < 0.10 · OEL Highly Controlled. Maintain existing general controls; basic hazard communication and chemical awareness training.

Category 2 10% - 50% of OEL 0.10 · OEL ≤ X95 < 0.50 · OEL Controlled. Well-controlled below Action Level (50% OEL). Establish baseline monitoring; re-characterize every 3 years.

Category 3 50% - 100% of OEL 0.50 · OEL ≤ X95 ≤ 1.00 · OEL Poorly Controlled. Exceeds Action Level. Implement medical surveillance, routine periodic monitoring (annual/semi-annual), and explore engineering upgrades.

Category 4 > 100% of OEL X95 > 1.00 · OEL Unacceptable Exposure. Violation of health standard. Immediate implementation of engineering/administrative controls; mandatory respiratory protection and PPE interim controls. AIHA EXPOSURE PROFILE SPECTRUM

  • [Cat 0: <1%] [Cat 1: 1-10%] [Cat 2: 10-50%] | [Cat 3: 50-100%] | [Cat 4: >100%]
  • <----------------- ACCEPTABLE -----------------> <-- ACTION LEVEL --> <-- UNACCEPTABLE -->
  • (50% OEL) (100% OEL)
  • Statistical Decision Rule for Acceptability

An exposure profile is statistically classified as Acceptable if there is at least 95% statistical confidence that the true 95th percentile is below the OEL : U T L 95 % , 95 % ≤ OEL ⟺ P ( X 95 ≤ OEL ) ≥ 0.95 \mathbf{UTL_{95\%, \, 95\%} \le \text{OEL}} \quad \Longleftrightarrow \quad P(X_{95} \le \text{OEL}) \ge 0.95 UT L 95% , 95% ​ ≤ OEL ⟺ P ( X 95 ​ ≤ OEL ) ≥ 0.95

5. OSHA Compliance Decision Rules and Sampling Error (CVT)

While industrial hygienists design control programs to protect workers on 95% of all shifts (UTL(95%, 95%) ≤ OEL), regulatory compliance officers from OSHA enforce standards under a different legal framework. Under the OSHA Technical Manual (OTM) and Field Operations Manual, OSHA must prove in court beyond reasonable statistical doubt (≥ 95% confidence) that an employer was out of compliance on the specific day sampled.

Total Coefficient of Variation (CVT / CVtotal)

Every occupational air measurement contains physical sampling and chemical analytical uncertainty:

Sampling Pump Error (CVS): Flow rate calibration drift, pulsation, and timing variations (typically assumed to be CVS = 0.05 or 5% for personal sampling pumps).

Analytical Laboratory Error (CVA): Extraction recovery, gas chromatography integration, and instrumental calibration variation (typically CVA = 0.04 to 0.09 depending on OSHA method validation).

Under error propagation physics, the Total Coefficient of Variation (CVT) is the root sum of squares: C V T = C V S 2 + C V A 2 \mathbf{CV_T = \sqrt{CV_S^2 + CV_A^2}} C V T ​ = C V S 2 ​ + C V A 2 ​ ​

Standardized Exposure Ratio (Y) and Confidence Limits

For a full-shift single sample or time-weighted average concentration C measured against the Permissible Exposure Limit (PEL): Y = C PEL Y = \frac{C}{\text{PEL}} Y = PEL C ​

Assuming measurement errors are normally distributed around the true exposure with standard deviation σerror = Y · CVT, the one-sided 95% confidence limits are: L C L 1 , 95 % = Y − 1.645 ⋅ C V T = C PEL − 1.645 ⋅ C V T \mathbf{LCL_{1, \, 95\%} = Y - 1.645 \cdot CV_T = \frac{C}{\text{PEL}} - 1.645 \cdot CV_T} LC L 1 , 95% ​ = Y − 1.645 ⋅ C V T ​ = PEL C ​ − 1.645 ⋅ C V T ​ U C L 1 , 95 % = Y + 1.645 ⋅ C V T = C PEL + 1.645 ⋅ C V T \mathbf{UCL_{1, \, 95\%} = Y + 1.645 \cdot CV_T = \frac{C}{\text{PEL}} + 1.645 \cdot CV_T} UC L 1 , 95% ​ = Y + 1.645 ⋅ C V T ​ = PEL C ​ + 1.645 ⋅ C V T ​

The OSHA Decision Classification Framework

Calculated Bounds Statistical Determination OSHA Regulatory Action

LCL(1, 95%) > 1.00 Established Non-Compliance (≥95% confidence) Citation Issued. Legal burden of proof met; exposure conclusively exceeded the PEL.

UCL(1, 95%) ≤ 1.00 Established Compliance (≥95% confidence) No Citation. Process is compliant within statistical margin of measurement error.

LCL ≤ 1.00 ≤ UCL Inconclusive / Possible Overexposure No Citation Issued. Measured value C may exceed the PEL (i.e., Y > 1.0), but sampling/analytical error prevents proving violation with ≥95% confidence. Employer receives letter of warning/recommendations. OSHA COMPLIANCE DECISION CONTINUUM

  • [ Established Compliance ] | [ Inconclusive / Possible Overexposure ] | [ Non-Compliance Citation ]
  • <--------------------------------- 1.0 --------------------------------------------->
  • PEL
  • UCL <= 1.00 LCL <= 1.00 <= UCL LCL > 1.00
  • (No Citation) (No Violation Cited) (Citation Issued)
  • Consecutive Sampling Across a Single Shift

When an 8-hour shift is sampled using m consecutive samples of durations T1, T2, s, Tm with individual concentrations C1, C2, s, Cm: T W A = ∑ i = 1 m C i ⋅ T i ∑ i = 1 m T i , Y = T W A PEL TWA = \frac{\sum_{i=1}^{m} C_i \cdot T_i}{\sum_{i=1}^{m} T_i}, \quad Y = \frac{TWA}{\text{PEL}} T W A = ∑ i = 1 m ​ T i ​ ∑ i = 1 m ​ C i ​ ⋅ T i ​ ​ , Y = PEL T W A ​ C V T = ∑ i = 1 m ( T i 2 ⋅ C V T , i 2 ) ∑ i = 1 m T i CV_T = \frac{\sqrt{\sum_{i=1}^{m} (T_i^2 \cdot CV_{T, i}^2)}}{\sum_{i=1}^{m} T_i} C V T ​ = ∑ i = 1 m ​ T i ​ ∑ i = 1 m ​ ( T i 2 ​ ⋅ C V T , i 2 ​ ) ​ ​

For m equal-duration consecutive samples with identical method error CV(T, 0): C V T = C V T , 0 m CV_T = \frac{CV_{T, 0}}{\sqrt{m}} C V T ​ = m ​ C V T , 0 ​ ​

Exam Tip: Taking multiple consecutive short-term samples throughout an 8-hour shift reduces overall sampling error by a factor of √m, narrowing the confidence interval (LCL moves up, UCL moves down) and providing greater statistical power to establish compliance or non-compliance.

  • 6. Worked Step-by-Step Calculation Examples
  • Worked Example 4.2.1: OSHA Single-Sample Compliance Determination

Problem: An OSHA compliance officer conducts a full-period 8-hour personal air sample for methylene chloride (OSHA PEL = 25 ppm). The analytical chemistry report indicates an 8-hour TWA concentration of 28.5 ppm. The sampling and analytical method documentation specifies a sampling pump coefficient of variation CVS = 0.05 and an analytical laboratory coefficient of variation CVA = 0.06.

Calculate the total coefficient of variation (CVT).

Calculate the standardized exposure ratio (Y), LCL(1, 95%), and UCL(1, 95%).

Determine whether OSHA will issue a citation for non-compliance.

Solution Steps:

Calculate total coefficient of variation (CVT): C V T = C V S 2 + C V A 2 = ( 0.05 ) 2 + ( 0.06 ) 2 = 0.0025 + 0.0036 = 0.0061 = 0.0781 ( 7.81 % CV_T = \sqrt{CV_S^2 + CV_A^2} = \sqrt{(0.05)^2 + (0.06)^2} = \sqrt{0.0025 + 0.0036} = \sqrt{0.0061} = \mathbf{0.0781} \quad (7.81\% C V T ​ = C V S 2 ​ + C V A 2 ​ ​ = ( 0.05 ) 2 + ( 0.06 ) 2 ​ = 0.0025 + 0.0036 ​ = 0.0061 ​ = 0.0781 ( 7.81%

Calculate standardized exposure ratio (Y): Y = C PEL = 28.5 ppm 25.0 ppm = 1.140 Y = \frac{C}{\text{PEL}} = \frac{28.5\text{ ppm}}{25.0\text{ ppm}} = \mathbf{1.140} Y = PEL C ​ = 25.0 ppm 28.5 ppm ​ = 1.140

Calculate Lower and Upper Confidence Limits (LCL and UCL): L C L 1 , 95 % = Y − 1.645 ⋅ C V T = 1.140 − ( 1.645 × 0.0781 ) = 1.140 − 0.1285 = 1.0115 LCL_{1, 95\%} = Y - 1.645 \cdot CV_T = 1.140 - (1.645 \times 0.0781) = 1.140 - 0.1285 = \mathbf{1.0115} L C L 1 , 95% ​ = Y − 1.645 ⋅ C V T ​ = 1.140 − ( 1.645 × 0.0781 ) = 1.140 − 0.1285 = 1.0115 U C L 1 , 95 % = Y + 1.645 ⋅ C V T = 1.140 + 0.1285 = 1.2685 UCL_{1, 95\%} = Y + 1.645 \cdot CV_T = 1.140 + 0.1285 = \mathbf{1.2685} U C L 1 , 95% ​ = Y + 1.645 ⋅ C V T ​ = 1.140 + 0.1285 = 1.2685

Compliance Decision:

Because LCL(1, 95%) = 1.0115 > 1.00, OSHA has established with > 95% statistical confidence that the true airborne concentration exceeded the PEL.

Determination: A citation for non-compliance will be issued.

Worked Example 4.2.2: AIHA Upper Tolerance Limit (UTL) & Category Rating

Problem: A facility industrial hygienist collects six (n = 6) random 8-hour samples for ethylbenzene (OEL = 100 ppm) from a chemical reactor operator SEG. Statistical analysis reveals a log-scale mean ȳ = 3.20 and a log-scale standard deviation sy = 0.42.

Calculate the point estimate of the Geometric Mean (GM), Geometric Standard Deviation (GSD), and 95th percentile (X95).

Calculate the 95% Upper Tolerance Limit of the 95th percentile (UTL(95%, 95%)) using K(6, 0.95, 0.95) = 3.708.

Determine the AIHA Exposure Category and state whether the profile is statistically acceptable.

Solution Steps:

Calculate point estimates:

  • GM = exp(3.20) = 24.53 ppm
  • GSD = exp(0.42) = 1.522
  • Point estimate X95 = GM · GSD(1.645) = exp(3.20 + 1.645 × 0.42) = exp(3.20 + 0.6909) = exp(3.8909) = 48.95 ppm
  • Expressed as percentage of OEL: (48.95 ppm)/(100 ppm) = 48.95%

Point estimate rating: AIHA Category 2 (10% - 50% of OEL) .

Calculate UTL(95%, 95%) to account for sample size (n = 6): ln ⁡ ( U T L 95 % , 95 % ) = y ˉ + K ⋅ s y = 3.20 + ( 3.708 × 0.42 ) = 3.20 + 1.5574 = 4.7574 \ln(UTL_{95\%, 95\%}) = \bar{y} + K \cdot s_y = 3.20 + (3.708 \times 0.42) = 3.20 + 1.5574 = 4.7574 ln ( U T L 95% , 95% ​ ) = y ˉ ​ + K ⋅ s y ​ = 3.20 + ( 3.708 × 0.42 ) = 3.20 + 1.5574 = 4.7574 U T L 95 % , 95 % = exp ⁡ ( 4.7574 ) = 116.44 ppm UTL_{95\%, 95\%} = \exp(4.7574) = \mathbf{116.44\text{ ppm}} U T L 95% , 95% ​ = exp ( 4.7574 ) = 116.44 ppm

Acceptability Evaluation:

Although the point estimate X95 = 48.95 ppm is below the 100 ppm OEL, the UTL(95%, 95%) = 116.44 ppm > 100 ppm.

Because UTL > OEL, the industrial hygienist does not yet have 95% statistical confidence that the 95th percentile is below the limit . The CIH must either collect additional samples to reduce the K-factor penalty or implement precautionary administrative/engineering controls.