Exam II

All of the following are true about Robust OLS except?

Can be applied to a broad range of regression models including both linear and non-linear OLS (e.g., logistic) regression models

Can be used to find undetected specification errors in the tests of the OLS assumptions

Therefore, can be used even if a violation of OLS assumption is not detected, in order to enhance the statistical rigor of the analysis

One of the statistical analysis techniques to perform a linear regression

True or False: Generalized Least Squares (GLS) One of the statistical analysis techniques to perform a linear regression Used when significant violations of the OLS assumptions are found (actually, OLS is a special case of GLS that does not have the problems)

True

False

True or False: Strengths and advantages of OLS with robust standard errors Can be used when OLS assumptions are violated. Even if there are no signs of violation, cannot be used

True

False

True or False: Logistic regression is a type of linear regression analysis technique that has a dichotomous variable (e.g., 1 or 0) as a dependent variable

True

False

True or False: Logistic regression is also subject to the violation of the assumptions that can be applied to the OLS assumptions (c.f., If one of the violations are detected, logistic regression with robust standard errors should be used)

True

False

Which of the following are examples of Business Questions for Logistic Regression

What factors are associated with the offer acceptance of credit card customers?

What factors are associated with daily downloads?

What factors are related with the employees who moved to another company in a year?

What factors are associated with the companies that have experienced a hacking incident?

All of the following are conditions for ANOVA except?

A statistical technique used to compare the means of two or more groups of observations to evaluate the impact of a treatment

Requiring a continuous dependent variable and a discrete independent variable in the analysis

Can have more than one independent varaible

Generally used in science experiments since other conditions should be same, except one condition, called treatment

Should be carefully used in the context of business since it is almost impossible to have such a controlled environment

Which of the following could be questions to analyze with ANOVA?

Do accountants, on average, earn more than teachers?

Do the owners of credit card A spend more than those of credit card B?

Do the online game user who experienced account stealing spend less than those who did not?

Which of the following determines the statistical significance of each independent varaible

Pvalue

Rsquared

Coefficient

Tvalue

Which of the following determines the strength of relationship with the dependent variable?

Pvalue

Rswuared

Coefficient

Tvalue

Which of the following determines the relative strength of relationship with the dependent variable?

Pvalue

Rsquared

Coefficient

Tvalue

True or False: If a Rsquared value is 0.36 this means that 36% of the variance of the dependent variable is explained by the independent variables in the model.

True

False

True or False: 0.1 means that the independent variable is statistically significant related to the dependent variable

True

False

If you had a coefficient of 0.1 this means that when the independent variable increases by one unit, the dependent variable differs by 0.1, assuming all other independent variables remain constant.

True

False

When looking at a linear regression model the only time that the intercept is meaningful is when all the varaibles in the model can have a value of zero.

True

False

Which of the following is the correct interpretation: daily_download = 22.949181 - 0.031739*age – 0.008061*rank

When age increases by one unit, daily downloads differs by -0.031739 assuming all other variables remain constant

When age increases by one unit, daily downloads differs by 0.031739 assuming all other independent varaibles remain constant

When age increases by one unit, daily downloads differs by -0.031739.

Which of the following is the correct interpretation: average_income = 31.45 + 0.031739*age + 156.789214*gender

When gender is 1 (1=female) average income differs by 156.789214 when all other independent variables remain constant

When gender increases by one unit, average income increases b 156.789214 assuming all independent variables remain constant

When gender is 1 average income varies by 156.789214.

Which of the following is the correct interpretation: everyone_dummy = -6.8572 + 0.0977*price + 1.4940*rating + 0.00216*rank

When rank differs by one unit, the log odds of “everyone_dummy” differs by 0.00216

When rank differs by one unit, everyone dummy differs by 0.00216 assuming all other independent varaibles remain constant

If A happened as a result of B Without B, A would have not happened If B had happened, A inevitably happened

Causality

Correlation

True or False: One of the factors that should be considered in the interpretation is endogeneity Particularly, endogeneity should be checked when translating association to causality

True

False

True or False: business data analysts need to avoid using such interpretations as “A impacts on B”, “A causes B”, and “A induces B” that imply a causality between variables

True

False

True or False: Instead, it would be better to use “A is associated with B”, “A is related to B”, and “A is one of the relevant factors to B”

True

False

True or False: When there are evident, multiple previous evidences that can support such a causality, analysts may carefully argue it (e.g., event study concerning the impact of hacking incidents on stock price)

True

False

What are the three basic assumptions in applying statistical results to business?

Association can be interpreted as causality

Association cannot be interpreted as causality

The variables in the analysis are all the important factors that explain the case

The analysis results can accurately predict the future in a similar circumstance

Which of the following are basic rules for making an effective table chart

Follow (strictly) a standard table format of your organization

Use a variety of font syles

Align texts in rows and columns in a consistent way

Work directly with the documentation software that will be used for reporting

Alignment doesn't really matter

Highlight important figures that should be emphasized

Locate the tables properly within a page

Which of the following is not true concerning using Visual Elements

Using images such as icons and images to emphasize or highlight the information in charts

Attract readers’ attention

Avoiding to use too many, complicated visual elements that distract readers (c.f., may not be a good idea to use visual elements for conservative readers)

All of the above are true

Value and usages of diagrams in analytics reports Used for visualizing ideas, conclusions, or summary derived from analysis results Should be simple, straightforward, and insightful MS PowerPoint, offering basic forms of diagrams

True

False

Which of the following are guidelines for making an effective diagram

Use different sizes and colors of shapes to present the relative magnitude of numeric features

Presenting a story from multiple analysis results in a diagram

Avoid using online copyright free pcitures

Make diagrams readable

Use free PowerPoint Diagrams

Drawing your own diagrams despite the time it takes

What type of learning can be defined as: Predefined target variable (i.e., dependent variable) Used for prediction and classification E.g., OLS, GLS, logistic regression, ANOVA, decision tree etc.

Supervised

Unsupervised

What type of learning can be defined as: No predefined target variable to predict or classify E.g., cluster segmentation, market basket analysis, principle component analysis (PCA), etc.

Supervised

Unsupervised

What type of dataset is used to reevaluate effectiveness of the model when applied to unused data

Test data

Train data

Validation data

What type of tree: when the dependent variable is categorical similar to logistic regression

Classification tree

Regression tree

What type of tree: when the dependent is continuous similar to linear regressions

Classification tree

Regression tree

Classification tree: misclassification rate Regression tree: root mean square error (RMSE)

Accuracy

Stability

Simplicity

Model behaves similarly when applied to different datasets (e.g., difference between the results from training & validation datasets)

Accuracy

Stability

Simplicity

Unlock and Upgrade

Remove all limits from your Quizzes

Exam II

More Quizzes