- (Topic 1)
An analytics team has completed some initial data analysis but is considering revising their research question based on their analysis findings. The team was concerned the original question was too broad. What outcome would lead the team to have this concern?
Correct Answer:
C
A research question is a clear and focused question that guides the data analytics process and defines the expected outcome or value of the analysis1. A research question that is too broad may lead to the concern of being difficult to identify the key performance indicators (KPIs) to measure, as KPIs are specific, quantifiable, and relevant metrics that indicate the progress and success of the analysis in relation to the research question23. A broad research question may also result in too much or too little data, unclear or conflicting objectives, or irrelevant or ambiguous results4. References: 1: Guide to Business Data Analytics, IIBA, 2020, p. 202: Guide to Business Data Analytics, IIBA, 2020, p. 233: Key Performance Indicators: Developing, Implementing, and Using Winning KPIs, David Parmenter, 2015, p. 34: How to Write a Good Research Question, ThoughtCo, 2021, 1.
- (Topic 2)
A movie production company wants to use analytics to decide which customers would choose to watch or not watch a particular movie after seeing a promotional teaser. The business analysis professional suggests they could make that prediction by identifying characteristics of the new movie and determining if the customer has watched other movies with similar characteristics.Thisis an example of using the following technique:
Correct Answer:
A
Logistic regression is a technique that can be used to model the probability of a binary outcome, such as choosing to watch or not watch a movie, based on one or more predictor variables, such as the characteristics of the movie and the customer??s viewing history. Logistic regression can help the business analysis professional to identify the factors that influence the customer??s decision and to estimate the likelihood of each customer??s preference. Logistic regression can also be used to test hypotheses and to evaluate the performance of the predictive model. References: [Guide to Business Data Analytics], page 55; [Business Data Analytics: A Practical Guide], page 93; [Introduction to Business Data Analytics: A Practitioner View], page 14.
- (Topic 2)
An analytics team is sourcing data for a new analytics initiative and is deciding between two comparable data sources. One source being considered is a very large dataset and another consists of three smaller sources. What advantage will the larger dataset provide over the three smaller sources?
Correct Answer:
A
A larger dataset may provide more significant results than three smaller sources, as it may have more statistical power to detect differences or relationships among variables1. Statistical power is the probability of finding a statistically significant result when there is a true effect in the population2. A larger dataset may have more power because it may have more variability, less sampling error, and higher precision than smaller datasets3. More significant results may lead to more confident and valid conclusions and recommendations for the analytics initiative.
Higher validity, more reproducibility, and higher reliability are not necessarily advantages of a larger dataset over three smaller sources, as they depend on other factors besides the size of the data. Validity is the degree to which the data and the analysis measure what they are intended to measure4. Reproducibility is the degree to which the data and the analysis can be replicated by another analyst using the same methods and data sources. Reliability is the degree to which the data and the analysis produce consistent results under the same conditions. These qualities may be affected by the quality, accuracy, completeness, and relevance of the data, as well as the appropriateness, transparency, and rigor of the analysis methods. A larger dataset may not be valid, reproducible, or reliable if it has errors, biases, missing values, or irrelevant variables, or if the analysis methods are not suitable, documented, or verified.
References:1: Guide to Business Data Analytics, IIBA, 2020, p. 542: Introduction to Business Data Analytics: A Practitioner View, IIBA, 2019, p. 233: Data Analysis: The Definitive Guide, Tableau, 4: Guide to Business Data Analytics, IIBA, 2020, p. 26. : Introduction to Business Data Analytics: A Practitioner View, IIBA, 2019, p. 25. : Guide to Business Data Analytics, IIBA, 2020, p. 26. : Introduction to Business Data Analytics: An Organizational View, IIBA, 2019, p. 13.
- (Topic 2)
An analyst is working through data on comparing performance scores in different schools across the state, for ranking purposes. Since there is a lot of data and some extreme outliers, the analyst is trying to determine which type of statistical average would best represent the results. Which of the following is a concern when relying too heavily on summary statistics during data analysis?
Correct Answer:
A
Summary statistics are numerical measures that describe certain characteristics of a data set, such as the mean, median, mode, standard deviation, range, or quartiles. Summary statistics can help simplify and communicate complex data, but they can also obscure or distort important information, such as the distribution, shape, outliers, or trends of the data. Contextualization is the process of providing relevant background information, assumptions, limitations, or explanations for the data analysis and its results. Contextualization can help avoid misinterpretation, confusion, or bias when using summary statistics. Contextualization can also help connect the data analysis to the business problem, objectives, and stakeholders.
References:Guide to Business Data Analytics, page 43; Introduction to Business Data Analytics: A Practitioner View, page 13.
- (Topic 1)
After analyzing sales data, the analytics team finds that the older the customer, the more expensive the neckties purchased. The team felt this was a breakthrough insight but on closer analysis realized that other factors could account for this relationship. This is a clear indication that:
Correct Answer:
D
The analytics team found a correlation between the age of the customer and the price of the neckties purchased, meaning that as one variable changes, the other tends to change in the same direction. However, this correlation does not imply causation, meaning that one variable does not necessarily cause the other to change. There could be other factors, such as income, preference, or quality, that affect both variables and create a spurious relationship. Therefore, the team realized that they need to investigate further to determine if there is a causal link between the variables, or if the correlation is coincidental12 References: 1: Correlation vs. Causation | Difference, Designs & Examples - Scribbr 2: Correlation vs Causation: Understanding the Differences - Statistics By Jim
