BENJAMIN MITCHELL TTU-ISQS EXAM AND
ANSWERS A+ GRADED WITH EXPERT SOLUTIONS -
120 Questions with Answers
Page 1
,Q1. In a multi-tiered data warehouse architecture, a business wants to implement a
real-time data integration layer that minimizes latency and supports operational
reporting without overburdening the transactional systems. Which approach best
achieves this while maintaining data consistency?
A. Replace the operational data store with an in-memory OLAP cube that directly
queries transactional databases.
B. Use change data capture (CDC) to feed a stream processing engine that updates a
separate operational data store (ODS) in near real-time.
C. Consolidate all data into a single centralized data warehouse and require all
reporting to use that warehouse exclusively.
D. Implement a data lake that stores raw transactional logs and rely on ad-hoc queries
to derive reports.
Correct Answer: B. Use change data capture (CDC) to feed a stream processing
engine that updates a separate operational data store (ODS) in near real-time.
Rationale: CDC captures changes as they occur and streams them to an ODS, enabling
near real-time reporting without impacting OLTP performance. This approach maintains
consistency by applying changes in order and decouples operational reporting from
transactional systems.
Why Wrong:
A - In-memory cubes querying transactional systems directly would introduce
significant load and risk locking, defeating the purpose of isolation.
C - A centralized warehouse is typically batch-oriented and cannot provide real-time
updates without complex incremental loads.
D - Ad-hoc queries on raw logs do not guarantee consistency and are inefficient for
operational reporting.
Reference: Kimball, R., & Ross, M. (2013). The Data Warehouse Toolkit, 3rd Ed., Ch. 1
Q2. A data scientist is building a classifier to predict customer churn. The dataset is
highly imbalanced (5% churn). They use logistic regression with a default threshold
of 0.5 and obtain an ROC-AUC of 0.85. However, the model's precision is very low.
Which of the following changes is most likely to improve precision without
significantly sacrificing recall?
A. Increase the classification threshold to 0.7.
B. Decrease the classification threshold to 0.3.
C. Apply SMOTE to oversample the minority class.
D. Use a random forest with class weights balanced.
Correct Answer: A. Increase the classification threshold to 0.7.
Rationale: Raising the threshold makes the model more conservative in predicting churn,
which typically increases precision (fewer false positives) at the cost of recall. Since the
Page 2
,goal is to improve precision, increasing the threshold is the direct adjustment.
Why Wrong:
B - Decreasing the threshold would increase recall but lower precision.
C - SMOTE can improve recall but may not directly increase precision; it often trades
off.
D - Balanced class weights can improve overall performance but do not specifically
target precision threshold tuning.
Reference: Provost, F., & Fawcett, T. (2013). Data Science for Business, Ch. 6
Q3. During a requirements elicitation session for a new healthcare management
system, stakeholders are divided on the priority of reporting features. The project
manager wants to ensure that the most critical business needs are captured while
avoiding scope creep. Which technique is most effective for prioritizing requirements
and reaching consensus?
A. MoSCoW analysis
B. Focus groups
C. Joint Application Design (JAD) sessions
D. Kano model analysis
Correct Answer: C. Joint Application Design (JAD) sessions
Rationale: JAD sessions bring stakeholders together in a structured workshop to define
and prioritize requirements collaboratively, reducing ambiguity and gaining consensus. It
is specifically designed to resolve conflicts and agree on scope.
Why Wrong:
A - MoSCoW is a prioritization framework but does not inherently facilitate
stakeholder consensus.
B - Focus groups gather opinions but are less structured for prioritization.
D - Kano model categorizes features based on satisfaction but is not a
consensus-building technique.
Reference: Satzinger, J., Jackson, R., & Burd, S. (2016). Systems Analysis and Design in a
Changing World, 7th Ed., Ch. 3
Q4. A logistics company uses a simulation model to optimize warehouse operations.
The model includes stochastic arrival times and service rates. To ensure the
simulation results are statistically reliable, which of the following is the most critical
practice?
A. Run a single long simulation run and discard the initial transient period.
B. Use antithetic variates to reduce variance without increasing run length.
C. Perform multiple independent replications and compute confidence intervals.
D. Set the simulation to terminate after a fixed number of events.
Page 3
, Correct Answer: C. Perform multiple independent replications and compute
confidence intervals.
Rationale: Multiple independent replications allow estimation of the variance and
construction of confidence intervals, providing statistical reliability. This is essential for
stochastic simulations to account for random variation.
Why Wrong:
A - A single run, even with warm-up, does not provide a measure of variability or
confidence.
B - Antithetic variates are a variance reduction technique but do not replace the need
for replications.
D - Fixed termination may introduce bias and does not address statistical reliability.
Reference: Law, A. M. (2015). Simulation Modeling and Analysis, 5th Ed., Ch. 9
Q5. A company is deciding whether to invest in a new ERP system. The CFO argues
for using Net Present Value (NPV) as the primary criterion, while the CTO favors an
options-based approach. Under which condition is the real options approach most
appropriate?
A. When the project has high uncertainty and management can make future decisions
contingent on outcomes.
B. When the project has a fixed scope and deterministic cash flows.
C. When the company has a high cost of capital.
D. When the project is a mandatory regulatory requirement.
Correct Answer: A. When the project has high uncertainty and management can
make future decisions contingent on outcomes.
Rationale: Real options valuation is valuable when there is significant uncertainty and
managerial flexibility to adapt decisions (e.g., expand, defer, abandon). This captures the
value of future choices that NPV ignores.
Why Wrong:
B - Fixed scope and deterministic cash flows make NPV straightforward and real
options unnecessary.
C - High cost of capital affects both methods but does not make real options more
appropriate.
D - Mandatory projects have little flexibility, so real options are less relevant.
Reference: Trigeorgis, L. (1996). Real Options, Ch. 1
Q6. In the context of information systems governance, an organization is aligning its
IT strategy with business strategy. Which framework best enables this alignment by
balancing value delivery, risk optimization, and resource management?
A. COBIT 2019
Page 4
ANSWERS A+ GRADED WITH EXPERT SOLUTIONS -
120 Questions with Answers
Page 1
,Q1. In a multi-tiered data warehouse architecture, a business wants to implement a
real-time data integration layer that minimizes latency and supports operational
reporting without overburdening the transactional systems. Which approach best
achieves this while maintaining data consistency?
A. Replace the operational data store with an in-memory OLAP cube that directly
queries transactional databases.
B. Use change data capture (CDC) to feed a stream processing engine that updates a
separate operational data store (ODS) in near real-time.
C. Consolidate all data into a single centralized data warehouse and require all
reporting to use that warehouse exclusively.
D. Implement a data lake that stores raw transactional logs and rely on ad-hoc queries
to derive reports.
Correct Answer: B. Use change data capture (CDC) to feed a stream processing
engine that updates a separate operational data store (ODS) in near real-time.
Rationale: CDC captures changes as they occur and streams them to an ODS, enabling
near real-time reporting without impacting OLTP performance. This approach maintains
consistency by applying changes in order and decouples operational reporting from
transactional systems.
Why Wrong:
A - In-memory cubes querying transactional systems directly would introduce
significant load and risk locking, defeating the purpose of isolation.
C - A centralized warehouse is typically batch-oriented and cannot provide real-time
updates without complex incremental loads.
D - Ad-hoc queries on raw logs do not guarantee consistency and are inefficient for
operational reporting.
Reference: Kimball, R., & Ross, M. (2013). The Data Warehouse Toolkit, 3rd Ed., Ch. 1
Q2. A data scientist is building a classifier to predict customer churn. The dataset is
highly imbalanced (5% churn). They use logistic regression with a default threshold
of 0.5 and obtain an ROC-AUC of 0.85. However, the model's precision is very low.
Which of the following changes is most likely to improve precision without
significantly sacrificing recall?
A. Increase the classification threshold to 0.7.
B. Decrease the classification threshold to 0.3.
C. Apply SMOTE to oversample the minority class.
D. Use a random forest with class weights balanced.
Correct Answer: A. Increase the classification threshold to 0.7.
Rationale: Raising the threshold makes the model more conservative in predicting churn,
which typically increases precision (fewer false positives) at the cost of recall. Since the
Page 2
,goal is to improve precision, increasing the threshold is the direct adjustment.
Why Wrong:
B - Decreasing the threshold would increase recall but lower precision.
C - SMOTE can improve recall but may not directly increase precision; it often trades
off.
D - Balanced class weights can improve overall performance but do not specifically
target precision threshold tuning.
Reference: Provost, F., & Fawcett, T. (2013). Data Science for Business, Ch. 6
Q3. During a requirements elicitation session for a new healthcare management
system, stakeholders are divided on the priority of reporting features. The project
manager wants to ensure that the most critical business needs are captured while
avoiding scope creep. Which technique is most effective for prioritizing requirements
and reaching consensus?
A. MoSCoW analysis
B. Focus groups
C. Joint Application Design (JAD) sessions
D. Kano model analysis
Correct Answer: C. Joint Application Design (JAD) sessions
Rationale: JAD sessions bring stakeholders together in a structured workshop to define
and prioritize requirements collaboratively, reducing ambiguity and gaining consensus. It
is specifically designed to resolve conflicts and agree on scope.
Why Wrong:
A - MoSCoW is a prioritization framework but does not inherently facilitate
stakeholder consensus.
B - Focus groups gather opinions but are less structured for prioritization.
D - Kano model categorizes features based on satisfaction but is not a
consensus-building technique.
Reference: Satzinger, J., Jackson, R., & Burd, S. (2016). Systems Analysis and Design in a
Changing World, 7th Ed., Ch. 3
Q4. A logistics company uses a simulation model to optimize warehouse operations.
The model includes stochastic arrival times and service rates. To ensure the
simulation results are statistically reliable, which of the following is the most critical
practice?
A. Run a single long simulation run and discard the initial transient period.
B. Use antithetic variates to reduce variance without increasing run length.
C. Perform multiple independent replications and compute confidence intervals.
D. Set the simulation to terminate after a fixed number of events.
Page 3
, Correct Answer: C. Perform multiple independent replications and compute
confidence intervals.
Rationale: Multiple independent replications allow estimation of the variance and
construction of confidence intervals, providing statistical reliability. This is essential for
stochastic simulations to account for random variation.
Why Wrong:
A - A single run, even with warm-up, does not provide a measure of variability or
confidence.
B - Antithetic variates are a variance reduction technique but do not replace the need
for replications.
D - Fixed termination may introduce bias and does not address statistical reliability.
Reference: Law, A. M. (2015). Simulation Modeling and Analysis, 5th Ed., Ch. 9
Q5. A company is deciding whether to invest in a new ERP system. The CFO argues
for using Net Present Value (NPV) as the primary criterion, while the CTO favors an
options-based approach. Under which condition is the real options approach most
appropriate?
A. When the project has high uncertainty and management can make future decisions
contingent on outcomes.
B. When the project has a fixed scope and deterministic cash flows.
C. When the company has a high cost of capital.
D. When the project is a mandatory regulatory requirement.
Correct Answer: A. When the project has high uncertainty and management can
make future decisions contingent on outcomes.
Rationale: Real options valuation is valuable when there is significant uncertainty and
managerial flexibility to adapt decisions (e.g., expand, defer, abandon). This captures the
value of future choices that NPV ignores.
Why Wrong:
B - Fixed scope and deterministic cash flows make NPV straightforward and real
options unnecessary.
C - High cost of capital affects both methods but does not make real options more
appropriate.
D - Mandatory projects have little flexibility, so real options are less relevant.
Reference: Trigeorgis, L. (1996). Real Options, Ch. 1
Q6. In the context of information systems governance, an organization is aligning its
IT strategy with business strategy. Which framework best enables this alignment by
balancing value delivery, risk optimization, and resource management?
A. COBIT 2019
Page 4
