Business Statistics Quiz: Prediction And Forecasting
20 questions · exam conditions
0:00
Prediction And ForecastingQuestion 1 of 20

A real estate company models the price of houses (in thousands of dollars) based on their size (in thousands of square feet). A partial regression output is provided:

PredictorCoefficientStd. Error
Intercept50.510.2
Size ('000 sq ft)150.28.5

According to the model, what is the predicted price for a house that is 2,200 square feet?

$330,440
$330,490.50
$380,940
$380.94
← Back to quizzes

Business Statistics Quiz

Business Statistics Quiz: Prediction And Forecasting

Practice Prediction And Forecasting in Business Statistics with focused quiz questions that help you check what you know, review explanations, and build confidence with test-style prompts.

What this quiz covers

This quiz focuses on Prediction And Forecasting, giving you a quick way to practice the rules, question types, and explanations that matter most for Business Statistics.

How to use this quiz

Try each quiz question before looking at the correct answer. Use the explanations to review missed ideas, then come back to similar questions until the pattern feels familiar.

All questions

Question 1

A real estate company models the price of houses (in thousands of dollars) based on their size (in thousands of square feet). A partial regression output is provided:

PredictorCoefficientStd. Error
Intercept50.510.2
Size ('000 sq ft)150.28.5

According to the model, what is the predicted price for a house that is 2,200 square feet?

  1. $330,440
  2. $330,490.50
  3. $380,940 (correct answer)
  4. $380.94
Explanation: The model requires the 'Size' variable to be in thousands of square feet. Therefore, 2,200 square feet must be converted to X=2.2X = 2.2. The regression equation is Y^=50.5+150.2X\hat{Y} = 50.5 + 150.2X. Plugging in X=2.2X = 2.2 gives Y^=50.5+150.2(2.2)=50.5+330.44=380.94\hat{Y} = 50.5 + 150.2(2.2) = 50.5 + 330.44 = 380.94. Since the output variable Y^\hat{Y} is in thousands of dollars, the predicted price is 380.94 \times \1,000 = $380,940$.

Question 2

A supply chain manager uses regression analysis on 36 months of data to model delivery times: y^=12.5+0.3x1+1.8x2−0.5x3\hat{y} = 12.5 + 0.3x_1 + 1.8x_2 - 0.5x_3 where x1x_1 = distance, x2x_2 = package weight, x3x_3 = warehouse efficiency score. The model performs well (R2=0.79R^2 = 0.79), but recent global events have disrupted normal operations. For next quarter's planning, what statistical approach would best address forecast reliability concerns?

  1. Increase the prediction interval width to account for the additional uncertainty from operational disruptions
  2. Validate the model using only the most recent 6 months of data to ensure parameter stability under current conditions
  3. Apply scenario analysis with different parameter values to assess forecast sensitivity to potential relationship changes (correct answer)
  4. Use bootstrap resampling on the original 36-month dataset to generate more robust confidence intervals for the forecasts
Explanation: Given that global events have disrupted normal operations, the fundamental relationships in the model may have changed. Scenario analysis allows testing how forecasts change if the coefficients shift due to new conditions (e.g., different distance effects due to route changes, altered weight impacts due to handling changes). This directly addresses the concern that the model's relationships may no longer hold. Option A only addresses uncertainty quantification, not relationship changes. Option B may have too little data for reliable validation. Option D uses historical data that may not reflect current disrupted conditions.

Question 3

An econometric model predicting housing prices shows strong performance with training data R² = 0.89 and validation data R² = 0.83. However, when tested on data from a different geographic region, R² drops to 0.34. What does this suggest about using the model for forecasting in new markets?

  1. The model is overfitted to the original region and needs regularization techniques before applying elsewhere
  2. The model lacks external validity and may not capture universal relationships necessary for cross-regional forecasting (correct answer)
  3. The validation approach was inadequate and a proper geographic cross-validation should have been performed initially
  4. The model's predictive variables are region-specific and require local calibration to maintain forecasting accuracy
Explanation: The dramatic performance drop when applied to a different region (R² from 0.83 to 0.34) indicates the model lacks external validity - it doesn't capture universal relationships that would apply across different markets. The model learned region-specific patterns that don't generalize. This is different from overfitting (A), which would show poor validation performance in the original region. While C and D identify process issues, they don't address the fundamental problem that the model may not represent generalizable relationships needed for reliable cross-regional forecasting.

Question 4

A retail chain develops separate regression models for each store location, all showing similar R² values (0.75-0.82) and using identical predictor variables. When the corporate team creates a pooled model combining all stores' data, the R² increases to 0.91. For forecasting new store performance in untested markets, which model approach presents the greater risk?

  1. The individual store models lack sufficient sample size to provide stable parameter estimates for new locations
  2. The individual store models capture idiosyncratic effects that are irrelevant for predicting new store performance
  3. The pooled model's higher R² likely reflects spurious correlations that won't generalize to new market conditions
  4. The pooled model may obscure important location-specific relationships that affect forecasting accuracy in new markets (correct answer)
Explanation: When you encounter regression modeling questions involving pooled versus individual models, focus on the trade-off between capturing general patterns and preserving important variations across subgroups. The pooled model's higher R² (0.91 vs. 0.75-0.82) might seem appealing, but it creates significant forecasting risk. By combining all stores' data, the pooled model likely captures broad retail patterns while smoothing over crucial location-specific relationships—like how demographic differences, local competition, or regional preferences affect sales. When forecasting for new stores in untested markets, these location-specific factors could be critical, but the pooled model has essentially averaged them away. Answer D correctly identifies this as the greater risk. Looking at the wrong answers: A misses the point—individual models with R² values around 0.75-0.82 actually show reasonably stable relationships, not sample size problems. B incorrectly assumes location-specific effects are "idiosyncratic" noise, when they're likely meaningful market variations that new stores will also experience. C suggests the pooled model's higher R² comes from spurious correlations, but the improvement more likely stems from capturing overall retail patterns—the issue isn't spuriousness, but rather the loss of important local variations. Study tip: In model selection questions, remember that higher R² doesn't automatically mean better forecasting performance. Always consider whether your model preserves the relationships that matter most for your specific prediction task—sometimes more complex or specialized models outperform simpler, more general ones.

Question 5

A time series regression model using 60 months of data produces the forecast equation y^t+h=125+1.4h+0.6yt\hat{y}_{t+h} = 125 + 1.4h + 0.6y_t where hh is the forecast horizon and yty_t is the current period value. For long-term strategic planning, management requests forecasts for h=24h = 24 months when yt=200y_t = 200. What is the primary statistical concern with this approach?

  1. The forecast error variance increases quadratically with the horizon, making 24-month forecasts highly uncertain
  2. The model assumes the trend and autoregressive parameters remain constant over the entire 24-month forecast period (correct answer)
  3. The linear trend component becomes the dominant factor in long-horizon forecasts, potentially overriding realistic bounds
  4. The autoregressive component's influence diminishes over time, reducing the model's ability to capture cyclical patterns
Explanation: The primary concern is the assumption of parameter stability over 24 months. The model assumes the trend (1.4 per period) and autoregressive coefficient (0.6) remain constant, which is often unrealistic for long horizons where economic conditions, business cycles, or structural changes may alter these relationships. While A (increasing uncertainty), C (trend dominance), and D (diminishing autoregressive influence) are valid technical concerns, the fundamental issue is assuming the estimated parameters will remain stable over such a long forecasting horizon.

Question 6

A forecasting model trained on pre-2020 data shows excellent performance metrics (MAPE = 3.2%, MAD = 12.5). When applied to 2020-2022 data, the same model shows MAPE = 18.7% and MAD = 67.3. The model is now being considered for 2024 forecasts. What does this performance degradation most likely indicate about future forecasting reliability?

  1. The model requires recalibration with recent data before making reliable 2024 forecasts
  2. The underlying data relationships have shifted, suggesting the model structure may be fundamentally inappropriate (correct answer)
  3. The model's parameters have become outdated but the underlying relationships remain stable for forecasting
  4. The performance degradation is temporary and the model will return to historical accuracy levels
Explanation: The dramatic performance degradation (MAPE increased from 3.2% to 18.7%) suggests a structural break in the underlying relationships, not just parameter drift. This indicates the fundamental model structure may no longer be appropriate for the changed environment. Simply recalibrating parameters (A) may not address structural changes. Option C contradicts the evidence of relationship changes. Option D assumes the degradation is temporary without evidence, which is dangerous for forecasting decisions.

Question 7

A logistics company developed a simple linear regression model to predict fuel consumption (in gallons) based on the total weight of cargo (in tons). The model was built using data from 200 delivery trips, with cargo weights ranging from 5 to 25 tons. The model is: Predicted Gallons = 10 + 2.5 * Weight. The model has an R-squared of 0.92. A manager uses the model to estimate the fuel needed for a special shipment weighing 35 tons.

Which statement is the most accurate assessment of the manager's fuel estimate for the 35-ton shipment?

  1. The estimate is likely to be highly accurate because the model's R-squared value is very high, indicating a strong predictive relationship.
  2. The estimate is unreliable because it requires extrapolating far beyond the observed range of cargo weights in the original data. (correct answer)
  3. The fuel estimate should be adjusted upwards by at least 8% to account for the increased uncertainty of predicting outside the data range.
  4. The prediction interval for this estimate will be narrower than for a prediction at the mean cargo weight of 15 tons due to the high R-squared.
Explanation: The model was built on data with weights from 5 to 25 tons. Predicting for a 35-ton shipment is an extrapolation. Even if a model has a very high R-squared value within the data range, its accuracy outside that range is not guaranteed. The relationship might change (e.g., fuel consumption might increase at a different rate for very heavy loads), making the extrapolated forecast unreliable.

Question 8

A market analyst develops a linear regression model to predict a product's sales based on its price. The data used for the model includes prices between $10 and $20. The model shows a strong negative linear relationship. When is a forecast generated from this model most likely to be misleading?

  1. When predicting sales for a price of $15, because this is the midpoint of the data range where variance is often highest.
  2. When predicting sales for a competitor's product that has a similar price point but different features.
  3. When predicting sales for a new promotional price of $5, as the demand curve may become non-linear at such a low price point. (correct answer)
  4. When the standard error of the estimate is large, indicating significant variability in sales at any given price.
Explanation: Predicting sales at a price of $5 is an act of extrapolation, as it falls outside the $10-$20 range of the data used to build the model. The primary danger here is that the relationship between price and demand might not be linear outside the observed range. At a very low price, demand might increase exponentially, or it might hit a saturation point, making the linear forecast misleading.

Question 9

A manufacturing firm uses a regression model to predict the shear strength (in psi) of a new adhesive based on its curing time (in hours). The model is y^=1200+50x\hat{y} = 1200 + 50x. For a specific curing time of x=10x=10 hours, the standard error of the mean response is s{μ^10}=5s\{\hat{\mu}_{10}\} = 5 psi and the residual standard error is se=20s_e = 20 psi. The critical t-value for a 95% interval is approximately tcrit=2t_{crit} = 2.

The firm needs to provide a 95% interval forecast for the strength of a single, specific batch of adhesive cured for 10 hours. Which calculation most accurately represents this required interval?

  1. 1700±2×51700 \pm 2 \times 5
  2. 1700±2×201700 \pm 2 \times 20
  3. 1700±2×52+2021700 \pm 2 \times \sqrt{5^2 + 20^2} (correct answer)
  4. 1200±2×52+2021200 \pm 2 \times \sqrt{5^2 + 20^2}
Explanation: First, calculate the point prediction for x=10: y^=1200+50(10)=1700\hat{y} = 1200 + 50(10) = 1700. For a prediction interval for a single observation, the standard error of prediction is s{y^pred}=se2+s{μ^x0}2s\{\hat{y}_{pred}\} = \sqrt{s_e^2 + s\{\hat{\mu}_{x_0}\}^2}. Plugging in the values: 202+52=400+25=425\sqrt{20^2 + 5^2} = \sqrt{400 + 25} = \sqrt{425}. The interval is the point estimate ±\pm (t-critical value) ×\times (standard error of prediction). This gives 1700±2×52+2021700 \pm 2 \times \sqrt{5^2 + 20^2}. Choice A calculates the confidence interval for the mean, a common error.

Question 10

A marketing firm created a regression model to forecast weekly sales based on TV advertising spend. For an upcoming campaign with a planned spend of $10,000, the model provides a point forecast of 150,000insalesanda90150,000 in sales and a 90% prediction interval of [120,000, $180,000]. The firm's break-even sales level for the campaign is $125,000.

Based on this forecast, which of the following is the most prudent conclusion for the marketing manager?

  1. The campaign is a safe investment because the point forecast of $150,000 is well above the break-even level of $125,000.
  2. The campaign should be cancelled because the prediction interval includes values below the break-even point, indicating a certain loss.
  3. The campaign carries a non-trivial risk, as the lower bound of the 90% prediction interval ($120,000) is below the break-even point. (correct answer)
  4. There is a 90% chance that actual sales will be some value between $120,000 and $180,000 for this specific campaign.
Explanation: The prediction interval provides a range of plausible outcomes for sales. Since the interval [$120,000, $180,000] contains values below the break-even point of $125,000, there is a statistically significant possibility that the campaign will not be profitable. A prudent manager must acknowledge this risk, which is revealed by the lower bound of the interval, rather than focusing solely on the optimistic point forecast.

Question 11

A data scientist is building two different regression models (Model A and Model B) to predict the same dependent variable, Y. Both models are based on the same dataset. The standard error of the estimate (residual standard error, ses_e) for Model A is 15. The standard error of the estimate for Model B is 25. Assuming all other factors are equal, what is the primary implication of this difference for creating prediction intervals?

  1. Prediction intervals from Model A will be wider than those from Model B, indicating less precision.
  2. Prediction intervals from Model A will be narrower than those from Model B, indicating greater precision for individual forecasts. (correct answer)
  3. The width of the prediction intervals will be the same for both models, as it depends on the confidence level and sample size, not the model fit.
  4. Model B is preferable for forecasting because a higher standard error of the estimate implies it captures more of the total variation in Y.
Explanation: The standard error of the estimate (ses_e) measures the typical size of the prediction errors (residuals). A smaller ses_e means that the data points lie closer to the regression line, on average. Since ses_e is a key component in the formula for the width of a prediction interval, a smaller ses_e (Model A) will result in narrower, more precise prediction intervals compared to a model with a larger ses_e (Model B), all else being equal.

Question 12

A manager uses a simple linear regression of sales on a time index to forecast sales for the next four quarters. A critical, often unstated, assumption required for these forecasts to be considered reliable is that:

  1. the residuals of the model are normally distributed with a mean of zero.
  2. the underlying business and economic processes that generated the historical sales data will remain stable and continue into the future. (correct answer)
  3. the R-squared of the model is statistically significant and greater than a predetermined threshold, such as 0.70.
  4. the sales data has been deseasonalized before the linear trend model was applied.
Explanation: All forecasting models based on historical data fundamentally assume that the future will, in some sense, resemble the past. For a time-series regression, this means assuming that the structural factors driving the observed trend (e.g., market growth, customer behavior, competitive environment) will continue to operate in the same way during the forecast period. If this assumption is violated (e.g., by a new competitor or an economic recession), the forecast will likely be inaccurate, regardless of how well the model fit the historical data.

Question 13

An analyst is choosing between two models to forecast next year's demand for a luxury good.

  • Model 1: Regresses demand on a time trend variable. R-squared = 0.90.
  • Model 2: Regresses demand on a national disposable income index. R-squared = 0.75.

Historical data shows the time trend has been very stable, but economists are predicting a major economic slowdown next year.

Which statement provides the best guidance for the analyst?

  1. Use Model 1 because its higher R-squared indicates a superior predictive fit to the historical data, which is the best guide for the future.
  2. Use Model 2, as it allows the analyst to incorporate the external forecast of an economic slowdown into the demand prediction. (correct answer)
  3. Neither model should be used; the prediction of an economic slowdown invalidates any forecast based on historical data.
  4. Average the forecasts from both models to balance the high historical fit of Model 1 with the economic logic of Model 2.
Explanation: Model 1, a simple time trend, assumes the future will follow the same path as the past. The forecast of an economic slowdown directly contradicts this assumption. Model 2, a structural model, links demand to a relevant economic driver (disposable income). This allows the analyst to use forecasts for disposable income (which would be low during a slowdown) to generate a more realistic and context-aware forecast for demand. In this case, adaptability to new information is more valuable than a higher historical R-squared.

Question 14

A regression analysis of home prices versus square footage reveals that the variability of prices increases as the square footage increases. A plot of the residuals versus the fitted values shows a distinct fan or cone shape, opening to the right.

What is the primary implication of this pattern for a real estate agent using the model to create 95% prediction intervals for home prices?

  1. The prediction intervals will be unreliable, likely being too narrow for large homes and too wide for small homes. (correct answer)
  2. The model's point predictions are biased, consistently over-predicting the price of large homes and under-predicting for small homes.
  3. The linear relationship is inappropriate; a logarithmic transformation of price must be used to achieve accurate point forecasts.
  4. The prediction intervals are still valid in their calculation, but the confidence level should be stated as approximate rather than exact.
Explanation: The described pattern in the residuals is a classic sign of heteroscedasticity, a violation of the constant variance assumption in linear regression. Standard prediction interval formulas assume constant variance (homoscedasticity). When heteroscedasticity is present, the model uses an 'average' level of variance. This results in prediction intervals that are too wide where the actual variance is small (for small homes) and dangerously too narrow where the actual variance is large (for large homes), making them unreliable for risk assessment.

Question 15

An HR department uses a linear regression model to predict job performance scores (Y, scale 0-100) from an applicant's score on a problem-solving test (X, scale 0-50). The model was developed using data from 200 current employees whose test scores (X) ranged from 25 to 45. A new applicant, a math prodigy, scores a 49 on the test.

What is the most significant concern when using the regression model to forecast this applicant's job performance?

  1. The applicant's unique talent may not conform to the linear trend established by the general employee population, making the forecast an unreliable extrapolation. (correct answer)
  2. The model's intercept is not meaningful because no applicant scored zero, which invalidates predictions for high-scoring applicants.
  3. The prediction for this applicant will have the widest possible prediction interval, but the point estimate remains the best guess of performance.
  4. The model is only valid for predicting the average score of a group of applicants, not the performance of a single individual.
Explanation: The applicant's score of 49 is outside the range of the data [25, 45] used to build the model. This is an extrapolation. The primary risk of extrapolation is that the relationship observed in the data may not hold outside that range. For example, the effect of problem-solving skills on job performance might plateau at very high levels, which a linear model would not capture. The forecast's reliability is questionable not just because it's less precise, but because the model itself may be wrong for this type of applicant.

Question 16

A car dealership uses the simple linear regression model Price = 25000 - 800 * Age to estimate the value of used cars, where Age is in years. The model was built using data on cars between 1 and 8 years old. A manager attempts to use the model to price a 35-year-old vintage car and gets a prediction of -$3000.

Statistically, what is the best explanation for this nonsensical negative price prediction?

  1. The sample data used to build the model must have contained measurement errors, leading to a flawed intercept or slope.
  2. The model incorrectly assumes that price is the dependent variable when it should be the independent variable for pricing decisions.
  3. The R-squared of the model must be very low, meaning the predictions are not reliable for any age of car, new or old.
  4. The true relationship between a car's age and its price is curvilinear, and the linear model is an invalid approximation when extrapolated far outside the original data range. (correct answer)
Explanation: When you encounter regression predictions that produce impossible or nonsensical results, you should immediately think about the limits and assumptions of linear models, particularly the dangers of extrapolation. The correct answer is D because linear regression models are only reliable within or very close to the range of data used to build them. This model was created using cars aged 1-8 years, but the manager tried to predict the value of a 35-year-old car—far outside this range. While the relationship between age and price might be reasonably linear for relatively new used cars, this relationship almost certainly becomes curvilinear when extended to vintage cars. Very old cars can actually increase in value due to rarity and collector interest, making a simple linear decline inappropriate. Option A is incorrect because measurement errors would affect the quality of predictions within the valid range, but wouldn't specifically cause problems with extrapolation. The negative result stems from using the model outside its intended scope, not from flawed data. Option B misunderstands regression modeling—price is correctly specified as the dependent variable since we want to predict price based on age. Option C incorrectly assumes the model is generally unreliable. A low R-squared would mean poor predictions across all ages, but this model might work perfectly well for cars in the 1-8 year range it was designed for. Study tip: Always check whether you're extrapolating beyond the original data range when using regression models. Linear relationships rarely hold indefinitely, and dramatic extrapolation often produces meaningless results even from good models.

Question 17

A consulting firm builds a client retention model using logistic regression on 2019-2021 data. The model includes variables like contract value, service satisfaction scores, and industry type. Key findings show that satisfaction scores above 8.0 strongly predict retention (OR = 4.2), and technology clients have higher retention rates than manufacturing clients (OR = 2.1). The model achieves 78% classification accuracy on holdout data from the same time period.

The firm wants to use this model to forecast 2024 client retention rates for budget planning. Given the passage information, what represents the most critical limitation for this forecasting application?

  1. The logistic regression coefficients may have shifted due to changing business relationships and market conditions over the 3-year gap (correct answer)
  2. The 78% classification accuracy is insufficient for reliable financial planning and budget forecasting applications
  3. The model's industry effects may no longer be valid due to technological disruption affecting client retention patterns differently across sectors
  4. The satisfaction score threshold of 8.0 for strong retention may have changed as client expectations evolved over time
Explanation: The most critical limitation is that the fundamental relationships captured by the logistic regression coefficients (like the odds ratios for satisfaction and industry type) may have changed over the 3-year gap from training data (2019-2021) to application (2024). Business relationships, competitive landscapes, and client expectations can shift significantly over this timeframe, making the model's learned parameters unreliable. While C and D identify specific ways relationships might change, A captures the broader issue of parameter instability. Option B incorrectly suggests 78% accuracy is inadequate - this is reasonable for business applications, and the issue is temporal validity, not accuracy level.

Question 18

Two separate research teams are building regression models to predict crop yield (Y) based on fertilizer amount (X). Team A collects data where fertilizer amounts range from 10 to 30 kg/hectare. Team B collects data where fertilizer amounts range from 0 to 50 kg/hectare. Both teams find a similar linear relationship and have a similar sample size and residual standard error. How will the precision of their models likely differ?

  1. Team A will produce more precise predictions because their data is more focused on a specific range of fertilizer application.
  2. Team B's model will be less precise because the wider range of X values is likely to introduce more random variation into the model.
  3. Both teams will have identical prediction precision because their residual standard errors and sample sizes are the same.
  4. Team B's model will likely have a more precisely estimated slope coefficient, leading to more reliable predictions within its observation range. (correct answer)
Explanation: When you encounter regression questions about data ranges and precision, focus on how the spread of X values affects the reliability of your coefficient estimates and predictions. The key insight here is that wider ranges of X values lead to more precise slope estimates. When Team B collects data across 0-50 kg/hectare (range = 50) versus Team A's 10-30 kg/hectare (range = 20), Team B's regression line is anchored by points that are much further apart. This greater spread in X values reduces the standard error of the slope coefficient, making the estimated relationship more reliable. Since both teams have similar residual standard errors and sample sizes, the primary difference comes from this range effect. Option A incorrectly suggests that focusing on a narrow range improves precision. While this might reduce some sources of variation, it actually hurts the precision of slope estimation because you're fitting a line through points that are closer together horizontally. Option B misunderstands the relationship between X-value range and model precision. A wider range of X values doesn't introduce more random variation - the residual standard error captures that, and it's similar for both teams. Option C ignores the crucial role that X-value spread plays in coefficient precision. Even with identical residual standard errors and sample sizes, the range of X values significantly affects how precisely you can estimate the slope. Remember: In regression analysis, spreading your X values over a wider range generally improves the precision of your slope estimate, assuming the linear relationship holds throughout that range.

Question 19

A factory manager uses a regression model to predict daily output (in units) based on the number of machine hours scheduled. The model was developed on data where machine hours ranged from 20 to 80 per day. The manager needs forecasts for two upcoming days: Day A with 75 scheduled hours, and Day B with 95 scheduled hours.

Which statement provides the most accurate comparison of the two forecasts?

  1. The forecasts for Day A and Day B will be equally reliable because they are generated from the same regression model.
  2. The prediction interval for Day B will be narrower than for Day A because the model is more certain about predictions far from the mean.
  3. The forecast for Day B is likely to be more accurate because more machine hours should logically lead to a more stable and predictable output level.
  4. The forecast for Day A is an interpolation, while the forecast for Day B is an extrapolation, making the forecast for Day A more reliable. (correct answer)
Explanation: When you encounter regression forecasting questions, the critical distinction is between interpolation (predicting within the data range) and extrapolation (predicting outside the data range). This fundamentally affects forecast reliability. The regression model was built using machine hours ranging from 20 to 80 per day. Day A's forecast uses 75 hours, which falls within this range (interpolation), while Day B's forecast uses 95 hours, which exceeds the upper bound of 80 (extrapolation). Interpolation is inherently more reliable because you're predicting within the established relationship pattern. Extrapolation assumes the same linear relationship continues beyond the observed data, which may not hold true. Choice A incorrectly assumes that using the same model guarantees equal reliability, ignoring the crucial difference between predicting within versus outside the data range. Choice B gets prediction intervals backwards – they actually widen as you move further from the center of the data and especially when extrapolating, reflecting increased uncertainty. Choice C commits a logical fallacy by assuming more machine hours automatically means more predictable output, when the real issue is that we have no observed data at 95 hours to validate this assumption. The correct answer is D because it properly identifies that Day A involves interpolation (more reliable) while Day B involves extrapolation (less reliable). Study tip: Always check whether forecast values fall within the original data range. Interpolation beats extrapolation every time in terms of reliability, regardless of the specific numbers involved.

Question 20

A software company has tracked its quarterly subscription revenue (in millions) for the past five years (20 quarters, from Q1 2019 to Q4 2023). They fit a linear regression model: Revenue = 5.2 + 0.8 * Quarter, where Quarter is a time index from 1 to 20. The model has an R-squared of 0.95. The CFO asks for a revenue forecast for Q4 2028.

What is the most significant statistical issue with using this model to provide the requested forecast?

  1. The model fails to account for seasonality, which is nearly always present in quarterly revenue data and will bias the forecast.
  2. The forecast for Q4 2028 (Quarter 40) is a major extrapolation, assuming a linear trend will continue unchanged for 5 more years. (correct answer)
  3. The R-squared value is likely inflated due to serial correlation in the residuals, making the model's fit appear better than it is.
  4. A time index is not a true causal variable, so the model is fundamentally invalid for forecasting future revenue streams.
Explanation: The data runs from Quarter 1 to 20. Q4 2028 would be Quarter 40 (20 quarters after the end of the data). This is a significant extrapolation. The primary risk is that the business environment, market growth, or competitive landscape will change over the next five years, making the simple linear trend observed in the past an unreliable guide to the future. While seasonality (A) and serial correlation (C) are potential issues, the sheer length of the extrapolation is the most severe forecasting problem.