Statistics Certification Exam — Questions and Answers

Question 1: A statistician is presenting regression results to a non-technical board. Which approach best communicates a significant predictor?

Show the full covariance matrix
Explain degrees of freedom in detail
Report the p-value and t-statistic only
State 'for every 1-unit increase in X, Y increases by B units on average' (Correct answer)

Correct answer: State 'for every 1-unit increase in X, Y increases by B units on average'

Plain-language interpretation of the slope coefficient makes regression results actionable for non-technical audiences.

Question 2: A researcher finds r = 0.85 between two raters scoring the same essays. This is an example of:

Internal consistency
Inter-rater reliability (Correct answer)
Predictive validity
Construct validity

Correct answer: Inter-rater reliability

Inter-rater reliability quantifies agreement between different scorers or observers measuring the same thing.

Question 3: A risk analyst finds that the loss distribution has a kurtosis of 6. Compared to a normal distribution (kurtosis = 3), this indicates:

Lower variance than a normal distribution
Heavier tails and higher chance of extreme losses (Correct answer)
A symmetric distribution with no skew
Fewer extreme events than normal

Correct answer: Heavier tails and higher chance of extreme losses

Excess kurtosis (kurtosis > 3) indicates leptokurtic (fat-tailed) distributions where extreme events are more probable than Gaussian models predict.

Question 4: Which of the following best defines 'selective reporting' in statistical practice?

Publishing only statistically significant outcomes while suppressing null results (Correct answer)
Reporting results separately for different demographic groups
Using stratified sampling to target specific subgroups
Focusing a report on the most policy-relevant findings

Correct answer: Publishing only statistically significant outcomes while suppressing null results

Selective reporting occurs when null or unfavorable results are suppressed, distorting the evidence base.

Question 5: What does a frequency polygon display?

Cumulative relative frequency
Distribution shape using connected midpoint frequencies (Correct answer)
Proportion of each category
Relationship between two variables

Correct answer: Distribution shape using connected midpoint frequencies

A frequency polygon connects midpoints of histogram bars with line segments to show the distribution's shape.

Question 6: Which of the following is an example of confounding in an observational study?

Coffee drinkers have higher rates of lung cancer because they also smoke more (Correct answer)
The outcome measure has low test-retest reliability
Researchers know which group participants are in
Participants who drop out have worse outcomes than completers

Correct answer: Coffee drinkers have higher rates of lung cancer because they also smoke more

Smoking is a confounder because it is associated with both coffee drinking and lung cancer, creating a spurious association.

Question 7: In the context of professional statistical practice, 'pre-registration' of a study primarily addresses which concern?

Ensuring IRB approval is obtained before data collection
Complying with federal funding agency reporting requirements
Preventing outcome switching and p-hacking by committing to analysis plans in advance (Correct answer)
Registering intellectual property rights to statistical methods

Correct answer: Preventing outcome switching and p-hacking by committing to analysis plans in advance

Pre-registration locks in hypotheses and analysis plans before data collection, preventing post-hoc flexibility that inflates false positives.

Question 8: What happens to the width of a confidence interval as sample size increases?

It remains the same
It depends on the population mean
It increases
It decreases (Correct answer)

Correct answer: It decreases

Larger samples reduce the standard error, which narrows the confidence interval.

Question 9: In a regression study, multicollinearity refers to:

Non-linear relationships between predictors and the outcome
Heteroscedasticity in the residuals
Measurement error in the outcome variable
High correlation among predictor variables, making it difficult to isolate their individual effects (Correct answer)

Correct answer: High correlation among predictor variables, making it difficult to isolate their individual effects

Multicollinearity occurs when predictors are highly correlated, inflating standard errors and destabilizing coefficient estimates.

Question 10: What does a leverage point in regression mean?

An outlier in the response variable Y
A data point with a very large residual
A predictor with near-zero variance
An observation with an extreme X value that has potential to strongly influence the regression line (Correct answer)

Correct answer: An observation with an extreme X value that has potential to strongly influence the regression line

High-leverage points have unusual predictor values and can disproportionately pull the regression line toward them.

Question 11: Which of the following is the correct range for the Pearson correlation coefficient r?

0 to 1
−∞ to +∞
−1 to +1 (Correct answer)
0 to +∞

Correct answer: −1 to +1

Pearson's r is bounded between −1 (perfect negative) and +1 (perfect positive linear relationship).

Question 12: What is the variance of the dataset {2, 4, 4, 4, 5, 5, 7, 9}?

4 (Correct answer)
3
2
16

Correct answer: 4

The mean is 5 and the average squared deviation from the mean equals 4.

Question 13: Which SQL aggregate function computes the arithmetic mean of a numeric column named 'score'?

COUNT(score)
AVG(score) (Correct answer)
SUM(score)
MEDIAN(score)

Correct answer: AVG(score)

AVG() is the standard SQL aggregate function that calculates the arithmetic mean of non-null values in a column.

Question 14: Which regression diagnostic plot is used to check the linearity and homoscedasticity assumptions?

Normal Q-Q plot
Histogram of the response variable
Residuals vs. fitted values plot (Correct answer)
Scatter plot of predictors

Correct answer: Residuals vs. fitted values plot

Plotting residuals against fitted values reveals non-linear patterns and changing variance (heteroscedasticity).

Question 15: A sports analyst examines whether home-field advantage exists by testing if home teams win more than 50% of games across 500 games. 275 home wins are observed. What test is appropriate?

Chi-square goodness-of-fit test comparing wins across seasons
Paired t-test for home vs away scores
Independent samples t-test comparing two team means
One-sample z-test for a proportion against H₀: p=0.50 (Correct answer)

Correct answer: One-sample z-test for a proportion against H₀: p=0.50

Testing a single observed proportion against a known hypothesized value (0.50) calls for a one-sample z-test for proportions.

Question 16: A p-value of 0.03 means:

There is a 3% probability the null hypothesis is true
The effect size is clinically meaningful
The observed data (or more extreme) would occur 3% of the time if the null hypothesis were true (Correct answer)
There is a 97% probability the alternative hypothesis is true

Correct answer: The observed data (or more extreme) would occur 3% of the time if the null hypothesis were true

The p-value is the probability of obtaining results at least as extreme as observed, assuming the null hypothesis is true.

Question 17: In Design of Experiments (DOE) applied to quality improvement, a '2^k factorial design' refers to an experiment with:

Two response variables and k factors
Two blocks and k treatment combinations
k replications of a two-factor experiment
k factors each tested at exactly two levels (Correct answer)

Correct answer: k factors each tested at exactly two levels

A 2^k factorial design tests k factors, each at two levels (high and low), resulting in 2^k total treatment combinations when fully replicated.

Question 18: Which format is most effective for communicating uncertainty to a general audience in a written report?

Phrases like 'we estimate X, with a plausible range of Y to Z' (Correct answer)
Variance-covariance matrices
Standard error tables
Z-score distributions

Correct answer: Phrases like 'we estimate X, with a plausible range of Y to Z'

Pairing a point estimate with a plain-language range communicates uncertainty accessibly without requiring statistical literacy.

Question 19: When reporting a confidence interval, professional standards require that the statistician clearly communicate:

That the true parameter has a 95% probability of falling within the interval
Only the width of the interval, not the procedure
The procedure used, confidence level, and what the interval estimates (Correct answer)
That the interval will contain the sample statistic 95% of the time

Correct answer: The procedure used, confidence level, and what the interval estimates

Professional reporting requires specifying the estimation procedure, confidence level, and the quantity being estimated to avoid misinterpretation.

Question 20: In environmental regulation, which statistical approach does the EPA use to set National Ambient Air Quality Standards?

Geometric mean of all monitoring station readings
Simple arithmetic mean of annual measurements
Maximum likelihood estimation of the 50th percentile
Risk-based standards using specified percentile exposures (Correct answer)

Correct answer: Risk-based standards using specified percentile exposures

EPA NAAQS standards are typically set at specific percentiles (e.g., 98th percentile for 24-hour PM2.5) based on health risk assessments.

Question 21: In the context of regression, what is overfitting?

A model where all coefficients are statistically significant
A model with too few predictors to explain the outcome
A model that fits the training data too closely and performs poorly on new data (Correct answer)
A model that underestimates the true population effect

Correct answer: A model that fits the training data too closely and performs poorly on new data

Overfitting occurs when a model captures noise rather than the true signal, leading to poor generalization.

Question 22: ISO 9001 is best described as a standard for:

Defining statistical methods for process control
Setting environmental management requirements
Establishing a quality management system framework focused on customer satisfaction and continual improvement (Correct answer)
Specifying product performance requirements for manufactured goods

Correct answer: Establishing a quality management system framework focused on customer satisfaction and continual improvement

ISO 9001 is a process-based Quality Management System (QMS) standard requiring organizations to demonstrate the ability to consistently provide products/services that meet customer and regulatory requirements.

Question 23: In the context of federal statistical agencies, 'respondent burden' refers to:

The time and effort required of survey participants to respond (Correct answer)
The statistical weight assigned to rare population segments
Legal liability for inaccurate survey responses
The cost to agencies of processing survey data

Correct answer: The time and effort required of survey participants to respond

Respondent burden is the time, effort, and resources required of individuals or organizations to complete surveys, which agencies must minimize under the Paperwork Reduction Act.

Question 24: What does the standard error of the estimate (SEE) measure in regression?

The correlation between observed and predicted values
The standard deviation of the slope coefficient
The variance of the predictor variable
The typical size of the prediction error (residual) (Correct answer)

Correct answer: The typical size of the prediction error (residual)

The SEE (also called RMSE) measures the average magnitude of residuals, indicating how accurate predictions are.

Question 25: Which data visualization mistake most often misleads stakeholders in statistical reports?

Labeling axes clearly
Using a line chart for time series data
Using color in bar charts
Truncating the y-axis to exaggerate small differences (Correct answer)

Correct answer: Truncating the y-axis to exaggerate small differences

A y-axis that doesn't start at zero visually amplifies differences, making small changes appear dramatic and misleading non-expert readers.

Question 26: When a pharmaceutical company submits a New Drug Application, what statistical significance threshold is conventionally required for pivotal trial results?

p < 0.01
p < 0.10
p < 0.001
p < 0.05 (Correct answer)

Correct answer: p < 0.05

FDA guidance and industry convention require p < 0.05 in two adequate and well-controlled clinical trials for NDA approval.

Question 27: What role does peer review play in Statistics practice?

It creates unnecessary competition
It is only for beginners
It replaces formal certification
It provides quality assurance and professional development through collegial evaluation (Correct answer)

Correct answer: It provides quality assurance and professional development through collegial evaluation

This is fundamental to Statistics practice. It provides quality assurance and professional development through collegial evaluation represents the professional standard for professional standards in the Statistics certification framework.

Question 28: In the context of Measurement System Analysis, 'reproducibility' refers to:

The variation caused by different operators measuring the same part (Correct answer)
The variation when the same operator measures the same part repeatedly
The agreement between the measured value and the true value
The ability to detect small differences between parts

Correct answer: The variation caused by different operators measuring the same part

Reproducibility is the variation in measurement averages when different appraisers (operators) measure the same set of parts under the same conditions.

Question 29: Which measure of central tendency is most affected by extreme outliers in a dataset?

Mean (Correct answer)
Midrange
Mode
Median

Correct answer: Mean

The mean is most affected by outliers because it incorporates every value in its calculation.

Question 30: In reliability engineering within quality assurance, the 'bathtub curve' describes the failure rate of a product over time. The three regions of this curve are:

Infant mortality, useful life, and obsolescence
Design, testing, and deployment
Inspection, production, and warranty
Burn-in, random failure, and wear-out (Correct answer)

Correct answer: Burn-in, random failure, and wear-out

The bathtub curve shows high early failures (burn-in/infant mortality), a low constant rate during useful life (random failure), and increasing failures at end of life (wear-out).

Question 31: The 'replication crisis' in science primarily implicates which statistical practice?

Overfitting models to small training datasets
Using Bayesian instead of frequentist methods
Excessive use of nonparametric tests
Routine underpowered studies combined with p-hacking and selective reporting (Correct answer)

Correct answer: Routine underpowered studies combined with p-hacking and selective reporting

The replication crisis is largely attributed to underpowered studies, flexible analysis choices, and suppression of null results inflating the false positive rate.

Question 32: A data scientist notices that a predictive model performs well overall but poorly for a minority subgroup. Professional standards suggest they should:

Report only overall performance metrics since they are statistically more stable
Disclose the differential performance and recommend targeted model improvements (Correct answer)
Exclude the subgroup from the evaluation sample to improve reported accuracy
Weight the subgroup's outcomes to match majority group performance

Correct answer: Disclose the differential performance and recommend targeted model improvements

Algorithmic fairness and professional ethics require disclosing subgroup performance disparities and actively working to address them.

Question 33: Which scenario represents a violation of the 'independence' standard for statisticians serving as peer reviewers?

Recommending rejection of a paper using a method you have criticized in print
Reviewing a paper co-authored by your current doctoral student (Correct answer)
Requesting additional data before recommending acceptance
Reviewing a paper in a methodology area outside your primary specialty

Correct answer: Reviewing a paper co-authored by your current doctoral student

Reviewing work by someone in a direct supervisory or mentorship relationship constitutes a conflict of interest that compromises independence.

Question 34: What assumption about residuals is required for ordinary least squares regression to be valid?

Residuals must be positively skewed
Residuals must increase as X increases
Residuals are approximately normally distributed with constant variance (Correct answer)
Residuals must be zero for all observations

Correct answer: Residuals are approximately normally distributed with constant variance

OLS regression assumes residuals are normally distributed, have constant variance (homoscedasticity), and are independent.

Question 35: A statistician conducting a meta-analysis should be most concerned about which threat to the validity of conclusions?

Differences in statistical software used by original authors
Publication bias favoring positive results (Correct answer)
Heterogeneity of effect sizes across studies
Variation in sample sizes across included studies

Correct answer: Publication bias favoring positive results

Publication bias systematically excludes null results, causing meta-analyses to overestimate true effect sizes.

Question 36: What does the adjusted R² penalize for compared to R²?

Having a small sample size
Using non-normal predictors
Having correlated residuals
Adding predictors that don't improve the model (Correct answer)

Correct answer: Adding predictors that don't improve the model

Adjusted R² decreases when a predictor is added that doesn't sufficiently improve model fit, preventing overfitting.

Question 37: Which statistical approach is used to control for multiple potential confounders simultaneously in an observational study?

Multivariable regression (Correct answer)
Wilcoxon signed-rank test
Pearson correlation
Chi-square test

Correct answer: Multivariable regression

Multivariable regression models allow researchers to adjust for multiple confounders and isolate the effect of a primary predictor.

Question 38: The coefficient of variation (CV) is used in risk analysis to:

Measure absolute risk regardless of scale
Estimate the maximum loss at a confidence level
Compare relative risk across variables with different means (Correct answer)
Calculate the probability of ruin

Correct answer: Compare relative risk across variables with different means

CV = standard deviation / mean, allowing apples-to-apples risk comparison across projects or assets with different expected values.

Question 39: Why is evidence-based practice important in Statistics?

It only applies to academic settings
It replaces experience
It integrates best available evidence with professional expertise for optimal outcomes (Correct answer)
It is a theoretical concept only

Correct answer: It integrates best available evidence with professional expertise for optimal outcomes

This is fundamental to Statistics practice. It integrates best available evidence with professional expertise for optimal outcomes represents the professional standard for research in the Statistics certification framework.

Question 40: What does a box plot's whisker typically represent?

One standard deviation from the mean
1.5 × IQR beyond the quartiles (Correct answer)
The minimum and maximum values
The 10th and 90th percentiles

Correct answer: 1.5 × IQR beyond the quartiles

By convention, box plot whiskers extend to 1.5 times the IQR beyond Q1 and Q3.

Question 41: A positively skewed distribution has its tail extending toward which direction?

Right (Correct answer)
Both directions equally
Left
Neither direction

Correct answer: Right

A positively skewed (right-skewed) distribution has a long tail on the right side.

Question 42: An odds ratio of 1.0 in a case-control study indicates:

A strong positive association
A statistically significant finding
A strong negative association
No association between exposure and outcome (Correct answer)

Correct answer: No association between exposure and outcome

An OR of 1.0 means the odds of the outcome are equal in exposed and unexposed groups, indicating no association.

Question 43: What is the primary purpose of blinding participants in a clinical trial?

To reduce sample size requirements
To improve external validity
To prevent selection bias
To reduce placebo and reporting bias (Correct answer)

Correct answer: To reduce placebo and reporting bias

Blinding prevents participants (and sometimes assessors) from knowing treatment assignments, reducing placebo and reporting biases.

Question 44: Which graphical display is best suited for showing the distribution of a single quantitative variable?

Bar chart
Scatter plot
Histogram (Correct answer)
Pie chart

Correct answer: Histogram

A histogram displays the frequency distribution of a continuous quantitative variable using bins.

Question 45: What does the correlation coefficient r = 0 indicate?

No linear relationship between the two variables (Correct answer)
A non-linear relationship exists
A perfect negative linear relationship
A perfect positive linear relationship

Correct answer: No linear relationship between the two variables

A correlation of 0 means there is no linear association; a non-linear relationship may still exist.

Question 46: A dataset has values: 2, 4, 4, 6, 8. What is the median?

5
4 (Correct answer)
4.8
6

Correct answer: 4

With 5 values sorted in order, the median is the 3rd value, which is 4.

Question 47: The empirical rule states that approximately what percentage of data falls within 2 standard deviations of the mean in a normal distribution?

50%
68%
95% (Correct answer)
99.7%

Correct answer: 95%

The empirical rule states approximately 68%, 95%, and 99.7% of data fall within 1, 2, and 3 standard deviations respectively.

Question 48: What is the complement rule in probability?

P(A') = 1 − P(A) (Correct answer)
P(A') = P(A) + P(B)
P(A') = P(A)
P(A') = P(A) × P(B)

Correct answer: P(A') = 1 − P(A)

The complement of an event A contains all outcomes not in A, and their probabilities sum to 1.

Question 49: Which type of data display would be most appropriate for comparing the distribution of test scores across five different classrooms?

Side-by-side box plots (Correct answer)
Line graph
Pie chart
Single histogram

Correct answer: Side-by-side box plots

Side-by-side box plots allow direct visual comparison of distribution shape, center, and spread across groups.

Question 50: A researcher conducts 20 separate hypothesis tests simultaneously at α=0.05 and finds 2 significant results. What is the primary statistical concern?

Two significant results out of 20 is a strong finding
The familywise Type I error rate is inflated, increasing false-positive risk (Correct answer)
A paired test should have been used instead
The sample size is too small for 20 tests

Correct answer: The familywise Type I error rate is inflated, increasing false-positive risk

Running multiple tests without correction inflates the probability of at least one false positive (familywise error); corrections like Bonferroni should be applied.

Question 51: How do Statistics professionals maintain digital competency?

Through ongoing training, practice with new tools, and staying current with technological advances (Correct answer)
By hiring IT support for all tasks
Digital skills are not required
Skills from initial training are sufficient

Correct answer: Through ongoing training, practice with new tools, and staying current with technological advances

This is fundamental to Statistics practice. Through ongoing training, practice with new tools, and staying current with technological advances represents the professional standard for technology in the Statistics certification framework.

Question 52: Which probability distribution is appropriate for modeling the number of successes in a fixed number of independent trials?

Binomial distribution (Correct answer)
Uniform distribution
Normal distribution
Poisson distribution

Correct answer: Binomial distribution

The binomial distribution models successes in a fixed number of independent Bernoulli trials with constant probability.

Question 53: In a fishbone (Ishikawa) diagram used in quality analysis, the 'bones' typically represent:

Individual defective items found in inspection
Data points plotted over time
Major categories of potential causes such as Man, Machine, Method, Material (Correct answer)
Steps in the production process flow

Correct answer: Major categories of potential causes such as Man, Machine, Method, Material

The bones of a fishbone diagram represent major cause categories (the 6 M's or 4 P's) branching off the spine leading to the effect (quality problem).

Question 54: In R, the function lm(y ~ x, data=df) fits a linear model. Which function extracts the model's R-squared value from the result object?

summary()$r.squared (Correct answer)
residuals()
coef()
predict()

Correct answer: summary()$r.squared

summary() on an lm object returns a list, and the $r.squared element specifically extracts the coefficient of determination.

Question 55: A systematic review differs from a narrative review primarily because it:

Is always conducted by multiple authors
Uses pre-specified, reproducible methods to identify and synthesize evidence (Correct answer)
Requires a larger number of included studies
Includes only randomized controlled trials

Correct answer: Uses pre-specified, reproducible methods to identify and synthesize evidence

Systematic reviews use transparent, reproducible search and selection criteria that reduce bias in evidence synthesis.

Question 56: What is the significance of a code of conduct for Statistics professionals?

It establishes expected behaviors and ethical standards that protect the public and profession (Correct answer)
It applies only to new practitioners
It limits professional freedom
It is merely symbolic

Correct answer: It establishes expected behaviors and ethical standards that protect the public and profession

This is fundamental to Statistics practice. It establishes expected behaviors and ethical standards that protect the public and profession represents the professional standard for professional standards in the Statistics certification framework.

Question 57: How should an Statistics professional handle a situation outside their scope of competency?

Ignore the situation
Attempt it anyway
Recognize limitations and refer to or consult with appropriate specialists (Correct answer)
Decline all unfamiliar work

Correct answer: Recognize limitations and refer to or consult with appropriate specialists

This is fundamental to Statistics practice. Recognize limitations and refer to or consult with appropriate specialists represents the professional standard for professional standards in the Statistics certification framework.

Question 58: What is the least squares criterion used for in regression?

Maximizing the R² value
Minimizing the number of predictors
Minimizing the sum of squared residuals (Correct answer)
Maximizing the F-statistic

Correct answer: Minimizing the sum of squared residuals

Ordinary least squares (OLS) finds the line that minimizes the total of squared vertical distances from the data points.

Question 59: An IRB (Institutional Review Board) primarily ensures that research:

Obtains a sufficiently large sample
Protects the rights and welfare of human subjects (Correct answer)
Achieves statistical significance
Uses valid and reliable measures

Correct answer: Protects the rights and welfare of human subjects

IRBs evaluate study protocols to protect participants from harm and ensure ethical standards are met.

Question 60: What is the purpose of a cumulative frequency distribution?

To identify outliers in a dataset
To show the number of observations at or below each value (Correct answer)
To display the most frequent value in a dataset
To compare two datasets side by side

Correct answer: To show the number of observations at or below each value

A cumulative frequency distribution shows how many observations fall at or below each data value.

Question 61: What does 'statistical significance' fail to communicate that stakeholders often most need to know?

The practical or economic magnitude of the effect (Correct answer)
Whether normal distribution was assumed
The sample size used
The confidence level chosen

Correct answer: The practical or economic magnitude of the effect

Statistical significance only addresses whether an effect is distinguishable from zero, not whether it is large enough to matter practically.

Question 62: In Tableau, a 'dimension' field versus a 'measure' field differs primarily in that:

Dimensions appear on the y-axis; measures on the x-axis
Dimensions are always numeric; measures are categorical
Dimensions are categorical/qualitative; measures are quantitative and aggregatable (Correct answer)
Dimensions require sorting; measures do not

Correct answer: Dimensions are categorical/qualitative; measures are quantitative and aggregatable

Tableau classifies categorical fields used to slice data as dimensions, while numeric fields that are aggregated (sum, avg, etc.) are measures.

Question 63: What is the role of professional journals in Statistics practice?

They only benefit academics
They are outdated by publication time
They disseminate current research, best practices, and professional developments (Correct answer)
They are optional reading

Correct answer: They disseminate current research, best practices, and professional developments

This is fundamental to Statistics practice. They disseminate current research, best practices, and professional developments represents the professional standard for research in the Statistics certification framework.

Question 64: How should Statistics professionals apply research findings to practice?

Ignore research that contradicts experience
Only follow systematic reviews
Critically evaluate applicability, adapt to context, and monitor outcomes (Correct answer)
Implement immediately without evaluation

Correct answer: Critically evaluate applicability, adapt to context, and monitor outcomes

This is fundamental to Statistics practice. Critically evaluate applicability, adapt to context, and monitor outcomes represents the professional standard for research in the Statistics certification framework.

Question 65: An analyst fits a time-series forecasting model and plots the residuals. The residuals show a clear sinusoidal pattern. What does this indicate?

The model is performing well because residuals are centered at zero
Sinusoidal residuals indicate the model is overfitting
The residuals should always follow a sinusoidal pattern
The model has not captured seasonal or cyclical patterns in the data (Correct answer)

Correct answer: The model has not captured seasonal or cyclical patterns in the data

A systematic pattern in residuals (like a sine wave) indicates the model missed a periodic structure such as seasonality, violating the randomness assumption.

Question 66: Spearman's rank correlation is used instead of Pearson's when?

The sample size is very large
Both variables are normally distributed
The data are ordinal or the relationship is monotonic but not necessarily linear (Correct answer)
The variables have no measurement error

Correct answer: The data are ordinal or the relationship is monotonic but not necessarily linear

Spearman's correlation uses ranked values and is appropriate for ordinal data or non-linear monotonic relationships.

Question 67: Which of the following is a core competency specified in the American Statistical Association's endorsed data science framework?

Mastery of machine learning without classical statistical training
Ability to communicate statistical uncertainty to non-technical audiences (Correct answer)
Specialization in a single domain such as medicine or finance
Proficiency in at least three programming languages

Correct answer: Ability to communicate statistical uncertainty to non-technical audiences

Communicating uncertainty and statistical findings clearly to diverse audiences is a foundational competency in professional statistical practice.

Question 68: In a symmetric bell-shaped distribution, which of the following is true?

Mean > Median > Mode
Mean = Median = Mode (Correct answer)
Mean < Median < Mode
Mode > Median > Mean

Correct answer: Mean = Median = Mode

In a perfectly symmetric unimodal distribution, all three measures of center are equal.

Question 69: Under FDA 21 CFR Part 11, electronic records used in statistical analysis of clinical trials must have what feature to ensure data integrity?

Monthly printed backups
On-site storage only
Audit trails that cannot be disabled (Correct answer)
Approval from two independent statisticians

Correct answer: Audit trails that cannot be disabled

21 CFR Part 11 requires secure, computer-generated audit trails that record who changed data, what was changed, and when.

Question 70: Which measure of spread is expressed in the same units as the original data?

Standard deviation (Correct answer)
Variance
Sum of squares
Mean squared error

Correct answer: Standard deviation

Standard deviation is the square root of variance, returning to the original unit of measurement.

Question 71: What does the coefficient of determination R² represent?

The correlation coefficient squared only for perfect fits
The probability that the regression is significant
The proportion of variance in Y explained by the regression model (Correct answer)
The slope divided by the standard error

Correct answer: The proportion of variance in Y explained by the regression model

R² measures the percentage of total variation in the response variable that is accounted for by the predictor(s).

Question 72: What is the purpose of polynomial regression?

To test for multicollinearity
To compare means across multiple groups
To model a curved (non-linear) relationship between X and Y (Correct answer)
To reduce the number of predictor variables

Correct answer: To model a curved (non-linear) relationship between X and Y

Polynomial regression adds squared or higher-order terms of X to model non-linear relationships within a linear framework.

Question 73: A project sponsor pressures a statistician to rerun the analysis until it yields a significant p-value. This practice is known as:

Bayesian updating
Bootstrapping
Cross-validation
p-hacking or data dredging, which inflates Type I error (Correct answer)

Correct answer: p-hacking or data dredging, which inflates Type I error

Running multiple analyses and reporting only significant results dramatically inflates the false-positive rate beyond the nominal alpha level.

Question 74: A statistician is asked to design a survey that will systematically oversample high-income households to make the results look more favorable to a client. The correct professional response is to:

Agree only if the client signs a liability waiver
Accept if appropriate weighting will be applied post-collection
Refuse, as intentional design bias violates core professional standards (Correct answer)
Comply if the sampling plan is disclosed in the methodology section

Correct answer: Refuse, as intentional design bias violates core professional standards

Deliberately biasing a survey design to produce favorable results is professionally and ethically impermissible regardless of disclosure.

Question 75: A nonprofit director asks why your survey results have a 'margin of error of ±4%.' How do you explain this?

It means the true population value likely falls within 4 percentage points of the sample estimate, based on sampling variability (Correct answer)
It means the survey was 4% inaccurate
It reflects the error rate of the survey software
It measures response bias

Correct answer: It means the true population value likely falls within 4 percentage points of the sample estimate, based on sampling variability

Margin of error quantifies sampling variability: the range within which the true population value is expected to fall with the specified confidence level.

Question 76: When comparing two datasets, one with a standard deviation of 3 and another with a standard deviation of 9, what can be said about their variability?

The second dataset has three times the spread of the first (Correct answer)
No comparison can be made without knowing the means
The datasets have equal variability
The first dataset has more variability

Correct answer: The second dataset has three times the spread of the first

A higher standard deviation directly indicates greater spread; 9 is three times 3.

Question 77: What is the purpose of regular risk reviews in Statistics practice?

To reduce workload
To identify new risks, evaluate control effectiveness, and update mitigation strategies (Correct answer)
To generate reports
To satisfy auditors only

Correct answer: To identify new risks, evaluate control effectiveness, and update mitigation strategies

This is fundamental to Statistics practice. To identify new risks, evaluate control effectiveness, and update mitigation strategies represents the professional standard for risk management in the Statistics certification framework.

Question 78: A data scientist splits a dataset 80/20 into training and test sets, fits a model on training data, and evaluates on the test set. The test set accuracy is 78%. What does this metric measure?

The percentage of correct training predictions
How well the model memorized the training data
The model's ability to generalize to unseen data (Correct answer)
The statistical significance of the model's coefficients

Correct answer: The model's ability to generalize to unseen data

Evaluating on a held-out test set estimates how well the model generalizes to new, unseen observations not used during training.

Question 79: Which statistic is most appropriate for describing the center of a heavily skewed income distribution?

Mean
Standard deviation
Mode
Median (Correct answer)

Correct answer: Median

The median is resistant to skewness and extreme values, making it ideal for skewed distributions.

Question 80: A quality control engineer samples 50 bolts per hour from a production line and plots X-bar and R charts. She notices 3 consecutive points above the upper control limit. What action is most appropriate?

Increase the sample size to reduce variability
Investigate and identify the assignable cause of variation (Correct answer)
Shut down the production line immediately
Recalculate control limits using the new data

Correct answer: Investigate and identify the assignable cause of variation

Points outside control limits signal a special (assignable) cause that must be identified and corrected before resuming normal production.

Question 81: What is a residual in linear regression?

The intercept of the regression equation
The difference between observed and predicted values (Correct answer)
The slope of the regression line
The standard deviation of X

Correct answer: The difference between observed and predicted values

A residual is the observed Y minus the predicted Ŷ, representing unexplained variation.

Question 82: How should an Statistics professional respond to a compliance violation?

Conceal it if minor
Blame the regulatory framework
Report it promptly, investigate the root cause, and implement corrective actions (Correct answer)
Wait for an external audit to find it

Correct answer: Report it promptly, investigate the root cause, and implement corrective actions

This is fundamental to Statistics practice. Report it promptly, investigate the root cause, and implement corrective actions represents the professional standard for regulatory in the Statistics certification framework.

Question 83: On a control chart, a run of 8 consecutive points on the same side of the center line indicates:

A possible process shift requiring investigation (Correct answer)
The control limits need to be recalculated
Normal random variation
The process is in statistical control

Correct answer: A possible process shift requiring investigation

Eight consecutive points on one side of the center line is a non-random pattern (Western Electric Rule) signaling a probable process shift.

Question 84: Under GDPR and U.S. statistical agency standards, which principle governs limiting data use to the purpose for which it was originally collected?

Purpose limitation (Correct answer)
Data minimization
Storage limitation
Data accuracy

Correct answer: Purpose limitation

Purpose limitation prohibits using personal data for purposes incompatible with those disclosed at the time of collection.

Question 85: What does heteroscedasticity mean in the context of regression?

The regression coefficients are not significant
The variance of residuals changes across levels of the predictor (Correct answer)
The residuals are not independent
The predictor and response variable are not correlated

Correct answer: The variance of residuals changes across levels of the predictor

Heteroscedasticity means the spread of residuals is not constant, violating the equal-variance assumption of OLS.

Question 86: Why should double-barreled questions be avoided in survey design?

They ask about two separate issues in one question, making responses ambiguous (Correct answer)
They cannot be processed by standard statistical software
They are inappropriate for online survey formats
They inflate the required sample size

Correct answer: They ask about two separate issues in one question, making responses ambiguous

A double-barreled question combines two distinct issues, so a single answer cannot accurately reflect the respondent's views on both issues separately.

Question 87: In meta-analysis, what does a forest plot primarily display?

Effect sizes and confidence intervals from individual studies alongside a pooled estimate (Correct answer)
The correlation between study quality and effect size
The distribution of raw data across studies
Publication bias detection results

Correct answer: Effect sizes and confidence intervals from individual studies alongside a pooled estimate

A forest plot displays individual study effect sizes with their CIs and a pooled overall estimate, enabling visual comparison.

Question 88: What is logistic regression used for?

Predicting a binary or categorical outcome variable (Correct answer)
Predicting a continuous outcome variable
Comparing means of two groups
Testing the normality of a distribution

Correct answer: Predicting a binary or categorical outcome variable

Logistic regression models the probability of a binary outcome (e.g., yes/no, success/failure) as a function of predictors.

Question 89: In evidence-based practice, the PICO framework helps researchers formulate:

Statistical power calculations
Ethical guidelines for research
A well-built clinical question (Correct answer)
Data analysis procedures

Correct answer: A well-built clinical question

PICO (Population, Intervention, Comparison, Outcome) structures clinical questions to guide systematic literature searches.

Question 90: In multiple regression, what does multicollinearity refer to?

High correlation among predictor variables (Correct answer)
Unequal variance of residuals
Non-linear relationship between predictors and outcome
Outliers in the response variable

Correct answer: High correlation among predictor variables

Multicollinearity occurs when two or more predictor variables are highly correlated, making coefficient estimates unstable.

Question 91: In simple linear regression with the equation Ŷ = 5 + 2X, what is the predicted value when X = 4?

10
18
13 (Correct answer)
9

Correct answer: 13

Substituting X = 4: Ŷ = 5 + 2(4) = 5 + 8 = 13.

Question 92: A researcher divides a population by income level and gender, then randomly selects individuals from each group. This technique is known as:

Stratified sampling (Correct answer)
Multistage sampling
Convenience sampling
Cluster sampling

Correct answer: Stratified sampling

Stratified sampling divides the population into homogeneous subgroups (strata) and then randomly samples from each stratum.

Question 93: Which of the following best describes a stem-and-leaf plot?

A graph of cumulative probabilities
A plot of two quantitative variables
A display that preserves original data values while showing distribution shape (Correct answer)
A chart comparing categorical frequencies

Correct answer: A display that preserves original data values while showing distribution shape

A stem-and-leaf plot retains actual data values while simultaneously showing the shape of the distribution.

Question 94: In simple linear regression, what does the slope coefficient represent?

The standard deviation of the residuals
The correlation between X and Y
The average change in Y for a one-unit increase in X (Correct answer)
The value of Y when X equals zero

Correct answer: The average change in Y for a one-unit increase in X

The slope (b₁) describes how much the predicted response Y changes on average for each one-unit increase in the predictor X.

Question 95: What does a z-score of −2 indicate about a data point?

It is 2 units below the median
It falls in the 2nd percentile
It is 2 standard deviations above the mean
It is 2 standard deviations below the mean (Correct answer)

Correct answer: It is 2 standard deviations below the mean

A z-score measures how many standard deviations a value is from the mean; negative means below.

Question 96: Non-response bias is most likely to distort survey results when:

Survey questions use technical language
The sample size is too small
The population is very large
Those who respond differ systematically from those who do not (Correct answer)

Correct answer: Those who respond differ systematically from those who do not

Non-response bias occurs when individuals who choose not to respond differ in meaningful ways from those who do respond, making the obtained sample unrepresentative.

Question 97: What is the coefficient of variation (CV) used for?

Describing the shape of a distribution
Calculating the range of a dataset
Comparing variability between datasets with different units or means (Correct answer)
Measuring the correlation between two variables

Correct answer: Comparing variability between datasets with different units or means

The CV expresses standard deviation as a percentage of the mean, enabling comparison across different scales.

Question 98: What does the interquartile range (IQR) measure?

The range of the middle 50% of data (Correct answer)
The distance between the mean and median
The total spread of all data
The variance of the dataset

Correct answer: The range of the middle 50% of data

The IQR is Q3 minus Q1, capturing the spread of the central 50% of the data.

Question 99: Under the ASA's Ethical Guidelines, when a statistician is listed as a co-author, what is their primary responsibility?

Defer all methodological decisions to the principal investigator
Verify only the statistical sections of the paper
Take responsibility for the statistical accuracy of the entire work (Correct answer)
Ensure only the data collection methods are sound

Correct answer: Take responsibility for the statistical accuracy of the entire work

Co-authorship implies responsibility for the statistical integrity of the complete work, not just isolated sections.

Question 100: What does statistical power represent?

The probability of making a Type I error
The effect size of a study
The width of a confidence interval
The probability of correctly rejecting a false null hypothesis (Correct answer)

Correct answer: The probability of correctly rejecting a false null hypothesis

Power is the probability of detecting a true effect (1 − β, where β is the Type II error rate).

Question 101: The number needed to treat (NNT) is calculated as:

Relative risk × 100
Relative risk reduction / Baseline risk
1 / Absolute risk reduction (Correct answer)
Odds ratio − 1

Correct answer: 1 / Absolute risk reduction

NNT = 1 / ARR; it tells clinicians how many patients must be treated to prevent one additional adverse outcome.

Statistics Certification Exam

A comprehensive statistics competency exam covering descriptive statistics, inferential methods, regression analysis, research methodology, and professional standards for data analysis practitioners.

100

Questions

120 min

Time Limit

70%

To Pass

Checking for saved progress...

Exam Rules

You can skip questions and return to them later
Flag questions for review before submitting
No feedback shown until you submit the entire exam
Unanswered questions count as wrong — answer everything
10 pretest questions are mixed in and don't affect your score
Timer auto-submits when time runs out
Your progress is auto-saved every 30 seconds