- A manager is visually analyzing a scatterplot of cost versus activity. To improve visibility, the manager adjusts the graph's axes. The original plot had axes from zero to the maximum observed values. The new plot's axes are 'zoomed in,' showing only the narrow range where the data actually lie (e.g., Y-axis from $95,000 to $105,000 instead of $0 to $120,000). How does this change in scaling typically affect the visual interpretation of the cost relationship's strength?
Answer: It makes the relationship appear weaker, as the data points will seem more widely scattered and less linear.
This is a common perceptual trick in data visualization. When you 'zoom in' on the data by narrowing the range of the axes, the absolute distance between the points and a potential trendline is magnified. This makes the random variation appear much larger relative to the total area of the plot. Consequently, a relationship that looks very strong and tight on a full-range graph can look weak and scattered on a zoomed-in graph. The underlying statistical relationship is unchanged, but the visual perception is that the relationship is weaker.
- A scatterplot of total costs versus production output for a chemical processing plant shows a distinct 'kink' or inflection point. Below an output of 1,000 tons, the points follow a steep linear trend. Above 1,000 tons, the points follow a second, much flatter linear trend. Fitting a single trendline through all the data would be inappropriate. What is the most plausible business reason for this observed pattern?
Answer: Economies of scale are realized at higher production levels, reducing the variable cost per ton for output beyond 1,000 tons.
A 'kink' where the slope decreases indicates a change in the variable cost per unit. The steep slope below 1,000 tons represents a higher variable cost, and the flatter slope above 1,000 tons represents a lower variable cost. This is consistent with economies of scale, such as receiving bulk discounts on raw materials or using more efficient large-batch processes only available at higher output levels. A step-up in fixed costs (A) would cause a vertical jump, not a change in slope.
- A company plots monthly advertising costs against monthly sales revenue for the past 24 months. The scatterplot shows that in one particular month, advertising cost was five times the monthly average due to a major product launch, while sales revenue for that month was only moderately higher than average. This data point is a significant vertical outlier. What is the most likely effect of including this outlier when visually fitting a trendline to estimate the relationship?
Answer: It will dramatically increase the slope of the trendline, suggesting a strong but spurious relationship between advertising and sales.
A vertical outlier (high on the y-axis, but not extreme on the x-axis) pulls the trendline towards itself. Since this point is far above the general trend, it will pull the line up. This vertical pull will cause the line to pivot, increasing both its slope (implying advertising is more effective than it is) and its y-intercept. The most dramatic effect, however, is on the slope, creating a misleadingly strong apparent relationship. Note: this differs from a high-leverage point, which is extreme on the x-axis.
- A trainee accountant creates a scatterplot but inadvertently reverses the axes, plotting total costs on the horizontal x-axis and activity level (e.g., machine hours) on the vertical y-axis. The trainee then fits a linear trendline to this inverted plot, obtaining the equation Y = 500 + 0.5X. What is the correct interpretation of the 0.5 slope value?
Answer: The value represents the machine hours required for each dollar of total cost, which is not a standard managerial cost parameter.
The slope of a trendline is the change in the y-axis variable for a one-unit change in the x-axis variable (ΔY/ΔX). In this incorrectly specified plot, Y is machine hours and X is total cost. Therefore, the slope of 0.5 represents ΔHours / ΔCost, or 0.5 machine hours per dollar of cost. This is not the variable cost rate (which is ΔCost / ΔHours). While related (it's the reciprocal of the variable rate if there were no fixed costs), it's not a standard cost parameter itself and cannot be used directly for cost estimation in the usual C = F + V(x) formula.
- A cost analyst creates a scatterplot of monthly shipping costs versus number of shipments and calculates a correlation coefficient of r=0.92. The analyst then adds three data points from peak holiday seasons that show unusually high costs per shipment due to expedited shipping premiums. After adding these points, the correlation coefficient drops to r=0.76. Which conclusion is most appropriate?
Answer: A separate cost estimation model should be developed for holiday periods due to the different cost structure during peak seasons
The correct answer is C. When certain periods have fundamentally different cost structures (like holiday expedited shipping), separate models should be developed rather than forcing all data into one model. The drop in correlation indicates the holiday data represents a different cost behavior pattern. Choice A is incorrect because higher correlation doesn't automatically mean better estimation if it excludes relevant scenarios. Choice B misses that including structurally different data in one model reduces its predictive power for normal operations. Choice D is wrong because the correlation decrease simply indicates different cost behaviors, not unreliable data.
- If the resulting trendline from the scatterplot is Cost = $50,000 + $25 * (Billable Hours), what would be the estimated total cost for a month with 1,200 billable hours, assuming this activity level is within the relevant range?
Answer: $80,000
This question requires applying the cost formula derived from a trendline. The formula is Total Cost = Fixed Costs + (Variable Cost per Unit Activity Level). Plugging in the given values: Total Cost = $50,000 + ($25 1,200). First, calculate the total variable cost: $25 * 1,200 = $30,000. Then, add the fixed cost: $50,000 + $30,000 = $80,000. The distractors represent common errors: $30,000 is only the variable portion, $50,000 is only the fixed portion, and $25,000 confuses the variable rate with a total cost.
- What is the most significant logical flaw in the city manager's conclusion based on the scatterplot data?
Answer: The analysis fails to consider a lurking variable, such as the severity of the winter, that influences both road salt usage and pothole formation.
This is a classic case of correlation does not imply causation. A third, or lurking, variable is the most likely explanation. A severe winter (heavy snow and ice) would cause the city to use more road salt AND would cause more potholes to form. The two variables are correlated because they are both driven by the same cause. The assignment of independent/dependent variables does not change the logical flaw. A positive correlation means the variables move in the same direction, which is what the manager observed. The location of the y-intercept is irrelevant to the question of causality.
- A company plots total factory overhead against machine hours using monthly data for the past year. Visually, the scatterplot shows two distinct, parallel clusters of points. The cluster for January through June is positioned noticeably lower than the cluster for July through December. What is the most logical conclusion an analyst can draw before fitting any trendline?
Answer: There was a structural change in the company's monthly fixed overhead costs starting in July.
Two parallel clusters indicate that for any given level of machine hours, the cost is systematically higher in the second period. Since the clusters are parallel, their slopes (variable rate) are the same. The upward shift of the entire group of points signifies an increase in the fixed cost component (a change in the y-intercept) that occurred mid-year, such as a rent increase or hiring of new salaried supervisors.
- An analyst at a software company plots total development costs against the number of product features for several completed projects. The scatterplot reveals a distinct curvilinear pattern: costs increase with each additional feature, but the incremental cost for each subsequent feature is progressively smaller. This suggests a learning curve or economies of scale. If a manager ignores this curve and fits a single linear trendline to the data, what will be the characteristic pattern of prediction errors?
Answer: The model will systematically underestimate costs for projects with few or many features, and overestimate costs for projects with a mid-range number of features.
Fitting a straight line through a concave-down curve (like a typical total cost learning curve) will cause the line to be below the curve at the beginning and end, and above the curve in the middle. Therefore, the linear model will underestimate costs at low and high activity levels (few/many features) and overestimate costs for mid-range activity levels.
- What is the primary reason the analyst should have created and reviewed a scatterplot before accepting the regression formula?
Answer: To visually confirm that the relationship is approximately linear and identify any significant outliers.
Regression analysis relies on several assumptions, including a linear relationship between variables. A high R-squared value can be misleading if the underlying relationship is, for example, strongly curvilinear. A scatterplot provides a crucial visual check to ensure the linear model is appropriate and to spot outliers, seasonal patterns, or other anomalies that are not apparent from the numerical output alone. The other options describe actions that are either less important or incorrect; visual inspection is for validation, not calculation, and the R-squared value is an output of the analysis, not something confirmed by the plot itself.
- A company's cost analyst plots monthly utility costs against production volume and observes that the scatterplot shows a positive correlation with R2=0.68. However, the analyst notices that the residuals (differences between actual and predicted values) are larger in summer months compared to winter months, even at similar production volumes. What does this pattern suggest about the cost estimation model?
Answer: Seasonal factors affect utility costs beyond production volume, suggesting a multiple regression model may be more appropriate
The correct answer is B. The pattern of larger residuals in summer months indicates that factors other than production volume (likely air conditioning costs) significantly affect utility costs. This suggests that a simple linear regression with only production volume is insufficient, and additional variables (like seasonal factors) should be included. Choice A is incorrect because the R2 threshold alone doesn't address the systematic pattern in residuals. Choice C is wrong because seasonal variations are legitimate business costs, not outliers to be excluded. Choice D is incorrect because there is still a positive correlation; the issue is that other factors also matter. - An analyst plots a company's monthly shipping costs against the weight of goods shipped over three years. A strong positive linear relationship is apparent. However, the data points, when traced chronologically, spiral around the trendline. That is, points for consecutive months are located near each other, forming a cyclical pattern above and below the line. This pattern indicates serial correlation. What is the most likely cause and consequence of this pattern?
Answer: The cause is likely an unmodeled seasonal factor affecting costs; the consequence is that short-term cost predictions will have systematic, predictable errors.
A cyclical or spiraling pattern of residuals (the points around the line) where consecutive points are related is a classic sign of serial correlation, often caused by seasonality. For example, higher fuel surcharges in winter months might cause costs to be systematically above the trendline, while lower surcharges in summer put them below. The consequence is that the model fails to capture this predictable pattern, leading to systematic errors in monthly forecasts (e.g., consistently underestimating in winter and overestimating in summer).
- A scatterplot shows the cost of a key raw material versus units produced. The pattern is perfectly linear up to 10,000 units, at which point the slope of the trend line abruptly and significantly increases. The trend is again linear, but steeper, for all production levels above 10,000 units. Which of the following scenarios provides the most plausible economic explanation for this observed cost behavior?
Answer: The company must pay a significant premium for expedited shipping or purchases from a higher-cost supplier after exhausting its primary inventory.
An abrupt increase in the slope of the total cost line signifies an increase in the variable cost per unit. Scenario B provides a perfect reason for this: after a certain point (10,000 units worth of material), the low-cost source is depleted, and the company must turn to a higher-cost source, increasing the variable cost for all subsequent units. A quantity discount (A) would decrease the slope. A step-fixed cost (C) would cause a vertical jump, not a change in slope. Inefficiency (D) is plausible but less specific than the common business practice of tiered supplier pricing or spot market purchases (B).
- A manufacturing firm's controller creates a scatterplot of monthly maintenance costs versus machine hours. The data points show a strong positive linear relationship for the observed activity range of 2,000 to 8,000 machine hours. However, one data point corresponding to a new plant expansion project is included at 15,000 machine hours, with a cost significantly below the trend line suggested by all other data points. If a single linear trendline is visually fitted to all data points including this high-leverage observation, what is the most likely distortion to the resulting cost formula, C = F + V(x), compared to a formula based only on the original 2,000-8,000 hour range?
Answer: The variable cost rate (V) will be understated, and the fixed cost component (F) will be overstated.
This describes a high-leverage influential point. A data point at a high activity level (x-axis) that falls significantly below the existing trend will 'pull' the right side of the trendline down. This action pivots the line, causing two effects: 1) the slope (variable cost rate, V) will decrease (become understated), and 2) to compensate, the y-intercept (fixed cost component, F) will increase (become overstated).
- A scatterplot of total maintenance costs versus machine hours contains a data point for a month where a major, unscheduled machine breakdown occurred, requiring exceptionally expensive emergency repairs. This data point is located at an average level of machine hours but at a cost level that is three times higher than any other point. If this point is not removed, how does it most likely bias the interpretation of the routine cost structure?
Answer: It will shift the entire trendline upward without significantly changing the slope, primarily inflating the estimate of fixed costs.
This point is a vertical outlier at an average activity level. Its primary effect is to pull the entire trendline upward. Because it is not at an extreme x-value (a high-leverage point), it has less power to pivot the line (change the slope) and more power to cause a parallel shift. Therefore, it will disproportionately increase the y-intercept, leading to an overestimation of the fixed cost component of the routine cost structure. The slope (variable rate) would be less affected than the intercept.
- A manager plots total overhead costs versus direct labor hours. The scatterplot shows a cloud of data points with no discernible upward or downward trend. A trendline fitted to the data is nearly horizontal. What is the correct interpretation of the y-intercept and slope of this trendline?
Answer: The slope is effectively zero, indicating no correlation between the two variables, and the y-intercept represents the average overhead cost across the observed activity levels.
A horizontal trendline indicates a slope of (or near) zero. This means that as the independent variable (labor hours) changes, the dependent variable (overhead cost) does not systematically change. Thus, there is no correlation. In this case, the best estimate for the cost at any activity level is simply the overall average cost. The y-intercept of a horizontal line fitted by least-squares regression will be this average cost. Option A is incorrect because the costs are not truly fixed (they vary, just not with labor hours), but rather the chosen driver is irrelevant.
- A manufacturing company is analyzing the relationship between production volume and total manufacturing costs using a scatterplot. The data shows a strong positive correlation with an R2 value of 0.85. However, when management examines the individual data points more closely, they notice that three observations corresponding to months with significant equipment breakdowns fall well above the trendline. What is the most appropriate next step in developing a reliable cost estimation model?
Answer: Remove the three outlier observations and recalculate the trendline to improve the predictive accuracy of the model
The correct answer is A. Outliers caused by abnormal circumstances (equipment breakdowns) should be removed from the dataset used for cost estimation because they don't represent normal operating conditions. These outliers would skew the cost function and reduce its predictive value for normal operations. Choice B is incorrect because even with a high R² value, outliers from abnormal events compromise the model's reliability. Choice C is wrong because artificially adjusting data points violates data integrity principles. Choice D is incorrect because keeping outliers in the model and adding arbitrary safety margins doesn't address the underlying data quality issue.
- An analyst is presented with two scatterplots. Plot 1 shows monthly cost vs. activity, with the cost axis ranging from $0 to $1,000,000. The data points form a tight, steep line. Plot 2 uses the exact same data, but with the cost axis 'zoomed in' to a range of $400,000 to $500,000. In Plot 2, the data points appear much more widely dispersed around the trendline. Which statement correctly synthesizes this information?
Answer: Plot 2 is misleading; the relationship is actually very strong, as indicated by the high proportion of variation explained when viewed in its full context in Plot 1.
This question tests the understanding of visual perception versus statistical reality. A large fixed cost component (a high y-intercept) means that the total variation in cost is large, but the variation around the trendline (the error) can be very small in comparison. Plot 1, showing the full range from $0, correctly visualizes that the activity driver explains a very high proportion of the total cost variation. Plot 2, by zooming in, magnifies the small random error, making it look large and the relationship appear weak. Plot 1 provides the better strategic overview of the cost structure's predictability.
- An analyst creates a scatterplot of monthly advertising expenses versus sales revenue and finds a correlation coefficient of r=0.45. Management questions whether this relationship is strong enough to use advertising expense as a predictor for sales revenue in their cost-volume-profit analysis. Which evaluation of this correlation is most accurate?
Answer: The correlation strength is acceptable, but management should consider that other factors likely have significant influence on sales revenue
The correct answer is C. A correlation of 0.45 indicates a moderate positive relationship (R² = 0.20, meaning 20% of variance explained), which suggests advertising has some influence on sales but other factors are clearly important. This level of correlation can be useful but requires acknowledging its limitations. Choice A is incorrect because 20% explained variance, while not strong, still indicates some predictive value. Choice B is wrong because weak correlations shouldn't be used just because they're positive. Choice D is incorrect because there's no standard 0.40 threshold that makes correlations automatically suitable for precise estimation.
- How should the variables be plotted on the graph's axes?
Answer: The independent variable (cost driver) on the x-axis and the dependent variable (total cost) on the y-axis.
By convention in mathematics and statistics, the independent variable (the one that is presumed to influence the other) is plotted on the horizontal x-axis, and the dependent variable (the one being influenced) is plotted on the vertical y-axis. In cost analysis, the activity or cost driver (X) is considered independent, and the total cost (Y) is dependent on that activity. This convention is essential for correctly interpreting the y-intercept as fixed cost and the slope as variable cost per unit.
