Biostatistics Quiz: Time Series Plots
20 questions · exam conditions
0:00
Time Series PlotsQuestion 1 of 20

A researcher presents a time series plot of weekly COVID-19 cases that appears to show exponential growth. However, the y-axis uses a logarithmic scale. On this log-scale plot, the data points form an approximately straight line. What does this pattern indicate about the underlying growth rate?

The growth rate is constant over time, indicating a stable exponential increase with consistent doubling time throughout the period
The growth rate is decreasing over time, showing that the epidemic is slowing down despite continued increases in case numbers
The growth rate is accelerating over time, indicating that the exponential growth is becoming more rapid in later periods
The growth pattern is linear rather than exponential, with the logarithmic scale creating an illusion of exponential growth
The growth rate varies unpredictably over time, with the straight line representing an average across periods of varying growth
← Back to quizzes

Biostatistics Quiz

Biostatistics Quiz: Time Series Plots

Practice Time Series Plots in Biostatistics 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 Time Series Plots, giving you a quick way to practice the rules, question types, and explanations that matter most for Biostatistics.

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 researcher presents a time series plot of weekly COVID-19 cases that appears to show exponential growth. However, the y-axis uses a logarithmic scale. On this log-scale plot, the data points form an approximately straight line. What does this pattern indicate about the underlying growth rate?

  1. The growth rate is constant over time, indicating a stable exponential increase with consistent doubling time throughout the period (correct answer)
  2. The growth rate is decreasing over time, showing that the epidemic is slowing down despite continued increases in case numbers
  3. The growth rate is accelerating over time, indicating that the exponential growth is becoming more rapid in later periods
  4. The growth pattern is linear rather than exponential, with the logarithmic scale creating an illusion of exponential growth
  5. The growth rate varies unpredictably over time, with the straight line representing an average across periods of varying growth
Explanation: When interpreting time series data on logarithmic scales, you need to understand the fundamental relationship between exponential growth and linear patterns on log plots. The key insight is that exponential functions become linear when plotted on a logarithmic scale. If data shows exponential growth of the form y=aebty = ae^{bt}, taking the natural logarithm gives ln⁡(y)=ln⁡(a)+bt\ln(y) = \ln(a) + bt. This is a linear equation where the slope represents the growth rate parameter bb. When points form a straight line on a log-scale plot, it confirms true exponential growth with a constant rate parameter. Answer A is correct because a straight line on the log scale indicates the growth rate coefficient remains constant throughout the time period, meaning consistent exponential growth with the same doubling time. Answer B is wrong because if the growth rate were decreasing, you'd see a curved line on the log plot that flattens over time (decreasing slope), not a straight line. Answer C is incorrect because accelerating growth would appear as an upward-curving line on the log scale (increasing slope), not a straight line. Answer D misunderstands the mathematics entirely. Linear growth would appear curved (concave down) on a log scale, not straight. The logarithmic scale doesn't create an "illusion" of exponential growth—it reveals whether growth is truly exponential. Study tip: Remember that straight lines on log scales = constant exponential growth rates. Curved lines indicate changing growth rates. Practice identifying these patterns, as they're crucial for interpreting epidemiological data.

Question 2

In a time series plot of daily step counts from a fitness tracker over 10 weeks, the data shows a repeating pattern where 5 consecutive days have high values (8000-10000 steps) followed by 2 days with low values (3000-5000 steps). This pattern repeats consistently throughout the observation period. What does this most likely represent?

  1. Weekly behavioral pattern reflecting workday versus weekend activity levels, with consistent lifestyle habits maintained over time (correct answer)
  2. Equipment malfunction occurring on a predictable schedule, causing systematic underreporting of steps on certain days
  3. Seasonal variation with 7-day periodicity, driven by environmental factors that influence physical activity levels
  4. Autoregressive behavior where high activity days lead to fatigue, requiring recovery periods before returning to high activity
  5. Data collection artifact where the device fails to sync properly on weekends, resulting in incomplete step recording
Explanation: When analyzing time series data with regular patterns, you need to consider what real-world processes could generate the observed periodicity. A 7-day repeating cycle (5 high days + 2 low days) strongly suggests weekly behavioral patterns rather than biological or technical causes. Option A correctly identifies this as a weekly behavioral pattern. The 5 consecutive high-activity days (8000-10000 steps) align perfectly with a typical Monday-Friday work schedule, where people walk more during commutes, lunch breaks, and daily routines. The 2 lower-activity days (3000-5000 steps) correspond to weekends when people are more sedentary at home. The consistency over 10 weeks indicates stable lifestyle habits. Option B (equipment malfunction) is unlikely because malfunctions typically occur randomly or show progressive deterioration, not predictable 7-day cycles. Modern fitness trackers don't fail on scheduled intervals. Option C (seasonal variation) is incorrect because seasonal changes occur over months, not weekly cycles. A 7-day periodicity isn't driven by environmental factors like weather or daylight. Option D (autoregressive fatigue behavior) doesn't fit the data pattern. True exercise-induced fatigue would create irregular recovery periods based on activity intensity, not a rigid 5-day-on, 2-day-off schedule that repeats identically each week. Study tip: In biostatistics, when you see regular periodicity in behavioral data, first consider whether the cycle length matches common human social patterns (daily, weekly, monthly) before looking at biological or technical explanations. Weekly patterns almost always reflect work-life schedules rather than physiological cycles.

Question 3

A time series plot displays monthly mortality rates over 5 years. The plot shows regular peaks occurring every 12 months, with the peaks getting progressively higher each year while the baseline level remains constant. What pattern does this represent and what might it suggest?

  1. Annual seasonality with increasing amplitude over time, suggesting that seasonal mortality factors are becoming more pronounced (correct answer)
  2. Linear trend with additive seasonal components, indicating both overall mortality increase and consistent seasonal variation
  3. Cyclical pattern with random amplitude variation, where the increasing peaks represent normal fluctuation around seasonal norms
  4. Multiplicative seasonal pattern with stable baseline, where seasonal effects scale proportionally with underlying mortality risk
  5. Autoregressive pattern with annual lag structure, where each year's peak depends on the magnitude of the previous year's peak
Explanation: Time series analysis requires you to identify distinct patterns: trends (long-term direction), seasonality (regular periodic patterns), and amplitude changes (how the magnitude of variations evolves over time). The key details here point to a specific pattern: regular 12-month peaks indicate clear annual seasonality, while "progressively higher peaks" with a "constant baseline" describes increasing amplitude over time. This means the seasonal effect itself is intensifying while the underlying mortality rate stays stable. Option A correctly identifies this as annual seasonality with increasing amplitude, suggesting seasonal mortality factors (like winter flu, summer heat waves, or holiday stress) are becoming more pronounced over time. This is a common public health phenomenon where seasonal vulnerabilities in populations can worsen due to factors like aging demographics or climate change. Option B is incorrect because there's no linear trend - the baseline remains constant, so overall mortality isn't increasing. Option C wrongly characterizes this as random variation when the pattern shows systematic amplitude increases, not random fluctuation. Option D describes multiplicative seasonality where effects scale with baseline levels, but here the baseline is stable while only the seasonal peaks grow. The critical distinction is between different types of time series components. Additive models add seasonal effects to trends, multiplicative models scale seasonal effects by baseline levels, but this pattern shows seasonality with evolving amplitude independent of baseline changes. Remember: When analyzing time series patterns, separately examine the baseline trend, the seasonal pattern, and how the seasonal amplitude behaves over time.

Question 4

When examining a time series plot of hourly emergency room admissions over one week, a pattern emerges where admissions are consistently low from midnight to 6 AM, increase gradually through the day, peak in early evening, then decline. This 24-hour pattern repeats each day with minor variations. How should this pattern be interpreted for staffing decisions?

  1. Daily cyclical pattern with predictable timing, allowing for evidence-based staffing adjustments to match anticipated demand fluctuations (correct answer)
  2. Random variation in admission timing, requiring constant high staffing levels to accommodate unpredictable patient arrival patterns
  3. Seasonal pattern requiring long-term trend analysis, with current data insufficient for operational planning purposes
  4. Artifact of data collection timing, where apparent patterns reflect administrative schedules rather than true admission patterns
  5. Autoregressive dependency in admissions, where each hour's volume depends primarily on the previous hour's census levels
Explanation: When analyzing time series data in healthcare settings, you're looking for patterns that can inform operational decisions. The key is distinguishing between meaningful patterns and random noise, then determining what type of pattern you're observing. The described pattern shows a clear daily cycle that repeats consistently over the week-long observation period. This is a classic example of circadian variation in healthcare utilization, where human behavior and biological rhythms create predictable 24-hour cycles. The pattern makes intuitive sense: fewer people seek emergency care during sleeping hours, activity increases during the day, and early evening represents peak activity before people settle in for the night. Answer A correctly identifies this as a daily cyclical pattern with predictable timing. Since the pattern repeats reliably, administrators can confidently adjust staffing levels to match anticipated demand, placing more staff during peak hours and reducing staff during predictably quiet periods. Answer B incorrectly characterizes this clear pattern as random variation. The consistency across multiple days demonstrates this is definitely not random, making constant high staffing both unnecessary and economically wasteful. Answer C confuses daily cycles with seasonal patterns. Seasonal patterns occur over months or years, not hours. One week provides sufficient data to identify and act on daily patterns. Answer D suggests the pattern reflects administrative schedules rather than true patient behavior. However, emergency admissions are driven by patient need, not administrative convenience, making this explanation implausible. Remember: In biostatistics, distinguishing pattern types (daily, seasonal, random) is crucial for making appropriate operational decisions based on your data.

Question 5

A researcher analyzes a time series plot of monthly vaccination rates and observes that the data appears to oscillate around a horizontal line with no apparent trend, but the oscillations seem to have a period of approximately 24 months. What type of pattern does this represent and what might it indicate?

  1. Cyclical pattern with 2-year periodicity, possibly reflecting policy cycles, funding cycles, or public health campaign schedules (correct answer)
  2. Seasonal variation with extended period, indicating that vaccination behavior follows a predictable annual cycle with lag effects
  3. Autoregressive pattern with long memory, where current vaccination rates depend on rates from 24 months in the past
  4. Random walk around a stable mean, with the apparent 24-month cycle representing spurious pattern detection in noise
  5. Measurement error with systematic bias that varies on a 2-year schedule due to changes in data collection methodology
Explanation: When analyzing time series data in biostatistics, you need to distinguish between different types of temporal patterns to understand what drives the observed variation. A pattern that oscillates around a horizontal line with no trend but shows regular periodicity over 24 months suggests systematic, non-seasonal forces at work. Answer A correctly identifies this as a cyclical pattern with 2-year periodicity. Unlike seasonal patterns that repeat annually due to weather or calendar effects, cyclical patterns have longer periods and typically reflect institutional, policy, or economic cycles. In public health, 2-year cycles often correspond to budget cycles, policy implementation schedules, or the timing of major public health campaigns and their follow-up phases. Answer B incorrectly suggests seasonal variation with extended period. Seasonal patterns are specifically tied to annual cycles (12 months or divisions thereof) related to weather, holidays, or school years. A 24-month period exceeds seasonal variation by definition. Answer C misidentifies this as an autoregressive pattern. While autoregressive processes can show dependencies over time, the regular 24-month oscillation described suggests external cyclical forces rather than internal statistical dependence structures. Answer D dismisses the pattern as spurious noise, but a consistent 24-month periodicity across multiple cycles is unlikely to be random and warrants investigation for underlying systematic causes. Study tip: Remember that in public health time series, look for institutional explanations for non-seasonal cycles. Budget periods, election cycles, and policy implementation schedules often create predictable 2-3 year patterns in health data.

Question 6

A time series plot of weekly medication errors in a hospital shows a stable baseline level for most weeks, but every 6-8 weeks there are dramatic spikes that are 3-4 times higher than the baseline. These spikes last for only one week before returning to baseline. What pattern does this represent?

  1. Intermittent systematic events creating temporary level shifts, possibly related to staff rotations, training periods, or system updates (correct answer)
  2. Seasonal pattern with irregular periodicity, reflecting external factors that influence medication error rates on predictable schedules
  3. Random outliers occurring at coincidentally regular intervals, representing normal variation in a complex healthcare environment
  4. Measurement error due to periodic changes in reporting sensitivity, where certain weeks capture errors missed in other periods
  5. Autoregressive volatility clustering, where periods of stable error rates alternate with periods of increased variability
Explanation: When analyzing time series data in healthcare quality improvement, you need to distinguish between different types of variation patterns. Regular, dramatic spikes that consistently return to baseline suggest systematic causes rather than random variation or measurement issues. The pattern described here—stable baseline with predictable 3-4x spikes every 6-8 weeks lasting only one week—points to intermittent systematic events. Answer A correctly identifies this as temporary level shifts likely caused by operational changes like staff rotations, new employee orientation periods, or system updates. These events create temporary disruptions in normal processes, leading to higher error rates until staff adapt or systems stabilize. Answer B is incorrect because true seasonal patterns typically show more gradual changes and longer-duration effects, not sharp one-week spikes. The 6-8 week interval also doesn't align with typical seasonal cycles. Answer C misinterprets the pattern as random outliers, but the regularity and predictable timing indicate systematic causes. Random variation wouldn't produce such consistent intervals and spike magnitudes. Answer D suggests measurement error due to periodic reporting changes, but this would more likely create sustained differences in reporting levels rather than dramatic one-week spikes followed by immediate returns to baseline. Study tip: In quality improvement contexts, look for operational explanations when you see regular patterns in performance data. Predictable timing often reveals systematic causes like staffing cycles, training schedules, or policy implementations rather than random variation or measurement artifacts.

Question 7

A time series plot shows hourly blood glucose readings from a continuous monitor over 72 hours. The data reveals a pattern where values are typically lowest around 3-4 AM each day, gradually increase to peaks around 8-9 PM, then decline overnight. This 24-hour cycle repeats for all three days, but the overall average level appears to be increasing across the 3-day period. What does this suggest about the patient's condition?

  1. Normal circadian rhythm in glucose metabolism combined with deteriorating glycemic control, requiring medical attention for the underlying trend (correct answer)
  2. Medication timing effects creating artificial daily cycles, with the increasing trend indicating need for dosage adjustment
  3. Measurement device calibration drift causing apparent daily patterns, with the trend reflecting progressive sensor degradation
  4. Dietary pattern effects where meal timing creates predictable glucose cycles, with weekend eating causing the overall increase
  5. Normal glucose variation around a stable mean, with apparent patterns and trends representing random fluctuation in short-term data
Explanation: When analyzing time series data in biostatistics, you need to distinguish between normal physiological patterns and pathological trends. This question tests your ability to recognize circadian rhythms versus concerning clinical developments. The 24-hour glucose cycle described here reflects normal circadian physiology. Blood glucose naturally follows a diurnal pattern, with lowest levels during early morning hours (3-4 AM corresponds to the deepest sleep phase) and higher levels in evening hours due to cortisol fluctuations, activity patterns, and metabolic changes throughout the day. This cyclical pattern repeating consistently over three days indicates the patient's circadian rhythm is intact. However, the key concern is the increasing overall average level across the 72-hour period. This upward trend, superimposed on normal daily cycles, suggests deteriorating glycemic control that requires immediate medical evaluation. Answer A correctly identifies both components: normal circadian rhythm plus a worrisome clinical trend. Answer B incorrectly assumes medication timing creates the cycles, but proper diabetes medications typically flatten glucose curves rather than create pronounced daily peaks. Answer C suggests device malfunction, but calibration drift wouldn't create such precise 24-hour repetitive patterns—it would show random or progressive error patterns. Answer D focuses on dietary effects, but normal meal timing doesn't typically create such dramatic overnight lows, and the weekend explanation doesn't account for the consistent daily pattern. Remember: In time series analysis, always separate predictable physiological cycles from concerning trends. Normal patterns can coexist with pathological developments—don't let one mask the other.

Question 8

In a time series plot of daily step counts over 8 weeks, the data shows a consistent pattern where step counts gradually decrease over 6-day periods, then jump back to high levels on the 7th day, and the pattern repeats. The overall level of this saw-tooth pattern is gradually increasing over the 8-week period. What is the most likely interpretation?

  1. Weekly activity cycle with improving fitness trend, where activity accumulates fatigue during the week but overall capacity increases over time (correct answer)
  2. Device battery life pattern affecting step recording accuracy, with weekly charging cycles creating apparent activity patterns
  3. Behavioral pattern where weekend activity compensates for declining weekday activity, with seasonal effects increasing overall activity
  4. Measurement artifact where step counting algorithm sensitivity varies with device wear time, creating false weekly patterns
  5. Random variation in daily activity with coincidental weekly clustering, superimposed on gradual lifestyle improvement
Explanation: When analyzing time series data in biostatistics, you need to distinguish between biological patterns and technical artifacts by considering which explanation best fits the observed data characteristics and real-world plausibility. The described pattern shows two key features: a repeating 6-day decline followed by a 7th-day spike (weekly cycle), plus an overall upward trend over 8 weeks. Answer A correctly identifies this as a weekly activity cycle with improving fitness. This makes biological sense—people often reduce activity during weekdays due to work fatigue, then increase activity on weekends when they have more time and energy. The gradual overall increase suggests improving fitness capacity or motivation over the 8-week period, which is consistent with exercise adaptation. Answer B incorrectly attributes the pattern to device charging cycles. Battery depletion would cause step counts to drop to zero or show erratic readings, not a gradual 6-day decline with precise weekly timing. Answer C misinterprets the pattern direction—it describes weekend compensation for declining weekday activity, but the data shows weekday decline followed by weekend peaks, which is normal behavior, not compensation. "Seasonal effects" over 8 weeks is also implausible. Answer D suggests measurement artifacts from varying device sensitivity. However, technical artifacts typically create random or irregular patterns, not consistent weekly cycles that align perfectly with calendar days. Study tip: In time series analysis, always consider whether patterns align with known biological or behavioral cycles (daily, weekly, seasonal) versus technical issues, which usually create irregular artifacts.

Question 9

A biostatistician examines a time series plot of monthly blood glucose measurements from a diabetic patient over 2 years. The plot shows values fluctuating around 150 mg/dL for the first year, then around 120 mg/dL for the second year, with a sharp transition between the two periods. What is the most appropriate interpretation?

  1. Permanent level shift occurred between years, likely due to treatment intervention, with stable variance in both periods (correct answer)
  2. Gradual improvement in glucose control over time, with the apparent sharp transition being an artifact of monthly averaging
  3. Seasonal pattern with annual periodicity, where glucose levels naturally cycle between higher and lower values each year
  4. Measurement bias changed between years, with the second year showing systematically lower values due to calibration differences
  5. Random variation around a constant mean, with the apparent level difference representing normal fluctuation in a small sample
Explanation: When analyzing time series data in biostatistics, you need to distinguish between different types of patterns: gradual trends, periodic cycles, sudden shifts, and measurement artifacts. The key is examining both the timing and nature of changes in the data. The description reveals a sharp transition from one stable level (150 mg/dL) to another (120 mg/dL) between years, with fluctuations around each mean remaining consistent. This pattern is classic for a permanent level shift caused by an intervention like medication adjustment, dietary changes, or improved treatment compliance. Answer A correctly identifies this as a structural change with maintained variance, which is exactly what you'd expect when treatment improves glucose control but doesn't eliminate normal biological variation. Answer B is wrong because gradual improvement would show a smooth downward trend, not a sharp transition. Monthly averaging wouldn't create an artificial sharp break—it would smooth out any abrupt changes. Answer C misinterprets the pattern as seasonal cycling. True seasonality would show glucose levels returning to higher values in year two's corresponding months, but the description indicates sustained lower levels throughout the second year. Answer D assumes measurement bias, but this would typically affect data quality randomly or show calibration drift over time. The clean transition between stable periods, combined with the clinical context of diabetes management, strongly suggests a real physiological change rather than instrument error. Study tip: In time series questions, always consider the clinical context first. Sharp transitions in biomedical data often indicate interventions or policy changes, while gradual changes suggest natural progression or environmental factors.

Question 10

A biostatistician examines a time series plot of monthly flu vaccination rates over 5 years. The plot shows sharp increases every 12 months (always in the same month), followed by gradual declines over the subsequent 11 months. However, the peak values are getting progressively higher each year. What does this pattern indicate?

  1. Annual seasonal pattern with increasing trend in peak vaccination coverage, suggesting improved public health outreach over time (correct answer)
  2. Cyclical pattern with multiplicative amplitude, where seasonal effects grow proportionally with baseline vaccination rates
  3. Autoregressive seasonal model, where each year's peak depends on the magnitude of previous years' vaccination campaigns
  4. Random variation in seasonal timing, with apparent peak increases representing normal fluctuation around constant seasonal norms
  5. Measurement bias increasing over time, where later years show artificially inflated vaccination rates due to improved record keeping
Explanation: When analyzing time series data in biostatistics, you need to distinguish between different types of patterns: seasonal effects (predictable, recurring patterns), trends (long-term directional changes), and cyclical variations (irregular fluctuations). This vaccination data shows a clear seasonal pattern - sharp increases every 12 months in the same month (likely fall when flu vaccination campaigns begin), followed by gradual declines. The key observation is that while the timing remains constant, the peak values are progressively increasing each year. This indicates two simultaneous phenomena: a predictable seasonal effect combined with an upward trend in vaccination coverage. Answer A correctly identifies this as an annual seasonal pattern with an increasing trend, logically attributable to improved public health outreach over time. This interpretation makes biological and public health sense - vaccination campaigns occur seasonally but become more effective over years. Answer B incorrectly suggests multiplicative amplitude effects, which would imply the seasonal variation grows proportionally with baseline rates rather than the observed consistent timing with increasing peaks. Answer C misidentifies this as an autoregressive model where each peak depends on previous years' magnitudes, but the data shows a steady upward trend, not dependency relationships. Answer D dismisses the clear pattern as random variation, ignoring the obvious systematic seasonal timing and progressive increase. Study tip: In time series analysis, always separate the components you observe: identify seasonal patterns first (consistent timing), then look for trends (directional changes over time). Don't overcomplicate straightforward patterns with complex statistical models when simpler explanations fit the data.

Question 11

When examining a time series plot of weekly infection rates in a hospital over two years, a biostatistician notices that the variance appears to increase proportionally with the mean level. Which transformation would be most appropriate before conducting further time series analysis?

  1. Square root transformation to stabilize variance when the mean and variance are proportionally related (correct answer)
  2. Logarithmic transformation to address exponential growth patterns while maintaining interpretability of relative changes
  3. Linear detrending to remove the systematic increase in both mean and variance over time
  4. Standardization using z-scores to normalize the distribution and eliminate variance heterogeneity across time periods
  5. Differencing transformation to convert the series to stationary form by removing trend and variance components
Explanation: When you encounter time series data where variance increases proportionally with the mean, you're dealing with a classic variance stabilization problem. This pattern is common in count data like infection rates, where larger values naturally have more variability. The square root transformation is specifically designed for this situation. When variance is proportional to the mean (Var(Y)∝Mean(Y)\text{Var}(Y) \propto \text{Mean}(Y)), applying Y\sqrt{Y} creates a new variable where the variance becomes approximately constant. This mathematical property makes the square root transformation ideal for count data following Poisson-like distributions, where the variance equals the mean. Answer A correctly identifies this relationship and the appropriate solution. Looking at the incorrect options: Answer B suggests logarithmic transformation, which is better suited when variance increases with the square of the mean (Var(Y)∝Mean(Y)2\text{Var}(Y) \propto \text{Mean}(Y)^2), not proportionally. While log transformation can help with exponential growth, it's not the optimal choice here. Answer C proposes linear detrending, but this only removes systematic trends in the mean—it doesn't address the variance structure problem. Detrending assumes constant variance after removing the trend, which isn't the case here. Answer D suggests standardization, but z-scores don't actually stabilize variance patterns; they just rescale the data to have mean zero and variance one at each time point without addressing the underlying variance-mean relationship. Study tip: Remember the transformation hierarchy: square root when variance is proportional to mean, logarithmic when variance is proportional to mean squared. This pattern appears frequently in biostatistics with count and rate data.

Question 12

A time series plot displays daily hospital admissions over 18 months. The plot shows three distinct periods: months 1-6 have stable admissions around 50 per day, months 7-12 show a sudden increase to around 80 per day with high variability, and months 13-18 return to 50 per day with low variability. What type of pattern does this represent?

  1. Structural break with temporary level shift, indicating an external intervention or event affecting the middle period only (correct answer)
  2. Cyclical pattern with increasing amplitude, suggesting seasonal effects becoming more pronounced over the observation period
  3. Random walk with drift, demonstrating permanent changes in the underlying admission rate with increasing uncertainty
  4. Autoregressive pattern with lag effects, where previous admission levels influence subsequent periods through feedback mechanisms
  5. Measurement error in the middle period, with the true underlying process remaining constant throughout all periods
Explanation: When analyzing time series patterns in biostatistics, you need to identify whether changes are permanent, temporary, cyclical, or follow specific statistical models. The key is matching the observed pattern to the underlying data-generating process. The described pattern shows a clear temporary disruption: stable baseline (months 1-6), sudden increase with high variability (months 7-12), then return to original stable levels (months 13-18). This is the textbook definition of a structural break with temporary level shift, making A correct. The "structural break" refers to the sudden change points at months 7 and 13, while "temporary level shift" describes how the series returns to its original level after the intervention period. This pattern typically indicates an external event (like a disease outbreak, policy change, or facility closure) that temporarily affected hospital admissions. B is wrong because cyclical patterns repeat regularly over time with predictable periods, not a one-time middle disruption. C is incorrect because random walks with drift show permanent changes that don't revert to previous levels—here, admissions returned to baseline. D is wrong because autoregressive patterns show gradual transitions where past values smoothly influence future ones, not the sharp breaks and sudden return seen here. Study tip: For time series questions, always ask: "Did the data return to baseline?" If yes, look for temporary effects or interventions. If no, consider permanent changes like trends or random walks. Sharp transitions usually indicate structural breaks rather than smooth autoregressive processes.

Question 13

A time series plot shows daily patient visits to an emergency department over 4 weeks. The plot reveals a clear weekly pattern where values are consistently lowest on certain days and highest on others, but this weekly pattern appears to be gradually shifting upward over the month. How should this pattern be characterized?

  1. Additive seasonal pattern with linear trend, where weekly cycles maintain constant amplitude while the overall level increases systematically (correct answer)
  2. Multiplicative seasonal pattern with exponential trend, where weekly cycles increase proportionally with the rising baseline level
  3. Autoregressive pattern with weekly lag structure, where patient visits depend on values from exactly seven days previous
  4. Random variation around a constant weekly pattern, with the apparent upward shift representing sampling fluctuation rather than true trend
  5. Cyclical pattern with increasing amplitude, where the weekly fluctuations become more pronounced over time while maintaining constant mean
Explanation: When analyzing time series data in biostatistics, you need to identify three key components: trend (long-term direction), seasonality (regular patterns), and the relationship between them. This question tests your ability to distinguish between additive and multiplicative seasonal patterns. The scenario describes a weekly pattern that maintains consistent amplitude (the difference between peak and trough values stays the same) while the entire pattern shifts upward uniformly. This is the hallmark of an additive seasonal model, where seasonal effects are added to the trend component rather than multiplied by it. The "gradually shifting upward" indicates a linear trend superimposed on the weekly cycles. Answer A correctly identifies this as an additive seasonal pattern with linear trend, where the weekly amplitude remains constant while the baseline level increases systematically. Answer B is wrong because a multiplicative pattern would show the weekly cycles getting larger as the baseline increases - the peaks and valleys would become more extreme over time, not maintain constant amplitude. Answer C incorrectly focuses on autoregression, which describes how current values depend on past values. While emergency visits might show some autocorrelation, the question emphasizes the systematic weekly pattern and upward trend, not the predictive relationship between consecutive observations. Answer D misses the systematic nature of both components. The description clearly indicates both a consistent weekly pattern AND a gradual upward shift - neither represents random variation. Study tip: In time series questions, always ask: "Are the seasonal fluctuations getting bigger (multiplicative) or staying the same size (additive) as the trend changes?" This distinction is crucial for proper model selection.

Question 14

A time series plot shows quarterly hospital readmission rates over 4 years. The data displays an overall declining trend, but with notable upward spikes occurring every 4th quarter (Q4 of each year). The Q4 spikes appear to be getting smaller relative to the declining baseline. What does this pattern suggest?

  1. Multiplicative seasonal effect combined with improving baseline performance, where seasonal factors have proportionally less impact as overall rates improve (correct answer)
  2. Additive seasonal effect with linear trend, where consistent Q4 increases are superimposed on steadily declining baseline rates
  3. Cyclical pattern with decreasing amplitude, indicating that seasonal influences are becoming less important over time
  4. Measurement bias in Q4 reporting that is being gradually corrected, leading to smaller apparent spikes in later years
  5. Random variation around a declining trend, where Q4 spikes represent coincidental clustering rather than systematic seasonal effects
Explanation: When analyzing time series data with both trend and seasonal components, you need to distinguish between additive and multiplicative seasonal effects. In additive models, seasonal variations remain constant in absolute terms regardless of the baseline level. In multiplicative models, seasonal effects change proportionally with the baseline—as the baseline decreases, the seasonal spikes become smaller in absolute terms. The pattern described here—Q4 spikes getting smaller as the overall readmission rate declines—is the hallmark of a multiplicative seasonal effect. This makes biological sense: if winter conditions historically cause a 20% increase in readmissions, that 20% represents a smaller absolute number when applied to an improved (lower) baseline rate. Answer A correctly identifies this multiplicative relationship where seasonal factors have proportionally less impact as overall performance improves. Answer B describes an additive model where Q4 increases would remain constant in absolute terms regardless of baseline changes—but that's not what we observe here. Answer C mentions "cyclical pattern with decreasing amplitude," but cyclical patterns are longer-term fluctuations (typically several years), not the regular annual seasonality described. Answer D suggests measurement bias, but there's no indication of systematic reporting errors—the pattern is more consistent with genuine seasonal healthcare phenomena. For time series questions, always ask yourself: "Are the seasonal variations staying the same size (additive) or changing proportionally with the trend (multiplicative)?" The answer will guide you to the correct model interpretation.

Question 15

In the time series plot showing monthly medication adherence rates, the data points alternate between high values (around 85%) and low values (around 65%) in a regular pattern. What is the most likely explanation for this alternating pattern?

  1. Measurement artifact due to alternating data collection methods between months, creating systematic differences in recorded adherence
  2. True behavioral pattern where patients cycle between periods of good and poor adherence, possibly related to medication supply cycles (correct answer)
  3. Seasonal variation with 2-month periodicity, reflecting external factors that influence medication-taking behavior on a regular schedule
  4. Statistical regression to the mean, where extreme values in one month are naturally followed by more moderate values
  5. Data entry error causing systematic misclassification of adherence rates in alternating time periods during database construction
Explanation: Regular alternating patterns in medication adherence often reflect real behavioral cycles, such as patients receiving monthly medication supplies, taking them consistently early in the cycle, then becoming less adherent as supplies run low, or psychological patterns of motivation and fatigue. Choice A assumes measurement problems without evidence. Choice C uses incorrect terminology (2-month cycles aren't typically called seasonal). Choice D misapplies regression to the mean. Choice E assumes data errors rather than recognizing plausible behavioral patterns.

Question 16

When interpreting the time series plot shown for quarterly pharmaceutical sales, an analyst notices that the data shows both an increasing trend and increasing variability over time. The later quarters show much larger fluctuations than earlier quarters. What does this suggest about the appropriate analytical approach?

  1. The data should be log-transformed before analysis to stabilize variance and make the trend more linear for modeling purposes (correct answer)
  2. The increasing variability indicates measurement error that increases over time, requiring adjustment for heteroscedastic errors
  3. A linear trend model is appropriate since the increasing variability is a natural consequence of larger absolute values
  4. The data should be differenced to remove both the trend and the increasing variability simultaneously
  5. Seasonal decomposition should be applied first to separate trend from variability before determining transformation needs
Explanation: When both the mean and variance increase together over time (often proportionally), log transformation is typically appropriate. This stabilizes variance and often linearizes exponential trends, making the data more suitable for standard analytical methods. Choice B assumes measurement error rather than recognizing a common pattern in growth data. Choice C ignores the problematic variance structure. Choice D may not address the variance issue. Choice E addresses seasonality but doesn't directly tackle the variance problem.

Question 17

A researcher is examining the time series plot shown for monthly blood pressure measurements in a patient over 24 months. The plot shows an overall upward trend with regular fluctuations. What is the most appropriate interpretation of the pattern observed?

  1. The measurements show seasonal variation superimposed on a linear increasing trend, suggesting both progressive hypertension and cyclical factors (correct answer)
  2. The fluctuations represent measurement error around a stable mean, indicating no clinically significant change in blood pressure
  3. The upward trend is an artifact of the time scale used, and the actual pattern shows random variation around a constant value
  4. The regular fluctuations indicate equipment malfunction, while the trend reflects accurate measurement of stable blood pressure over time
  5. The pattern demonstrates measurement bias increasing over time, with the fluctuations representing normal biological variation around true values
Explanation: A time series showing both an overall upward trend and regular fluctuations suggests two components: a systematic change over time (progressive hypertension) and cyclical variation (possibly seasonal, stress-related, or medication compliance patterns). Choice B ignores the clear trend. Choice C incorrectly attributes the trend to scale artifacts. Choice D misinterprets fluctuations as equipment issues. Choice E incorrectly identifies bias rather than recognizing legitimate trend and cyclical components.

Question 18

The time series plot shows weekly antibiotic prescriptions at a clinic over 18 months. The data exhibits a sawtooth pattern where values gradually increase over 4-6 weeks, then drop sharply, and the pattern repeats. The overall level of the sawtooth pattern appears stable over time. What is the most likely explanation?

  1. Periodic inventory and restocking cycles, where prescriptions accumulate until supply constraints force temporary reduction in prescribing (correct answer)
  2. Seasonal illness patterns with rapid onset and gradual recovery, creating natural cycles in antibiotic demand over time
  3. Measurement error in data collection, where accumulated recording errors are periodically corrected causing apparent drops
  4. Regulatory compliance cycles, where prescribing increases until audit periods require temporary reduction to meet guidelines
  5. Random variation around a constant mean, with the sawtooth pattern representing coincidental alignment of natural fluctuations
Explanation: A repeating sawtooth pattern with gradual increases followed by sharp drops typically indicates accumulation processes with periodic resets. In healthcare, this often reflects inventory or supply chain patterns where prescribing gradually increases until supply constraints or restocking cycles cause temporary reductions. Choice B would show more irregular patterns. Choice C assumes measurement error. Choice D is possible but less common than supply issues. Choice E ignores the clear systematic pattern.

Question 19

In the time series plot shown below, daily patient satisfaction scores over 6 months show a gradual upward trend. However, every 7th data point consistently falls below the trend line, while the other 6 points cluster around or above it. What is the most appropriate interpretation of this pattern?

  1. Weekly systematic variation superimposed on improving trend, likely reflecting day-of-week effects in patient experience or staffing
  2. Outlier pattern indicating data quality issues occurring on a regular schedule, requiring investigation of collection procedures
  3. Autoregressive structure with 7-day memory, where satisfaction depends on scores from exactly one week previous
  4. Measurement artifact from different survey administration methods used on alternating days throughout the week
Explanation: A

Question 20

The time series plot shows daily calorie intake recorded by a patient over 12 weeks. The data shows 5 days of stable intake around 2000 calories, followed by 2 days of much higher intake around 3000 calories, and this 7-day pattern repeats consistently. Additionally, there is a gradual downward trend in both the baseline and peak values. How should this pattern be characterized?

  1. Weekly seasonal pattern with multiplicative trend, where both weekday and weekend behaviors show proportional improvement over time
  2. Additive weekly pattern with linear trend, where consistent day-of-week effects are combined with systematic behavior change
  3. Cyclical eating disorder pattern, where regular binge episodes occur on a predictable schedule with increasing severity
  4. Autoregressive pattern with weekly structure, where each week's eating pattern depends on the previous week's behavior
Explanation: B