Question 1

Which of the following represents the correlation between an original variable and a derived factor?

Accepted Answer

Factor loading represents the correlation between an original variable and a derived factor.

Question 2

Two of the most widely used methods to decide the number of factors to retain are the _____.

Accepted Answer

The two most widely used methods to decide the number of factors to retain are the latent root criterion and the percentage of variance criterion. Latent root criterion involves retaining all factors with eigenvalues greater than 1, while percentage of variance criterion involves retaining factors that explain a certain amount of total variance, such as those that explain 70% of the total variance.

Question 3

The latent root criterion for determining the number of factors to retain is that with principle components analysis, factors that have a latent root of _____ or higher are retained.

Accepted Answer

Factors with a latent root of 1 or higher are retained in principle components analysis.

Question 4

The rule of thumb for the percentage of variance criterion for retaining factors in factor analysis is that a factor solution should account for a minimum of _____ percent of the total variance.

Accepted Answer

The answer of The rule of thumb for the percentage...

Question 5

The _____ is the total amount of variance in a factor analysis.

Accepted Answer

The trace is the total amount of variance in a factor analysis, as it is the sum of the eigenvalues of the correlation matrix.

Question 6

_____ represents the row sum of squared factor loadings and tells us how much of the variance in a particular variable is accounted for by the factor solution.

Accepted Answer

Communality represents the proportion of variance in a variable that can be accounted for by the factor solution. It is calculated as the sum of the squared factor loadings for a variable and represents its common variance with the other variables in the factor solution.

Question 7

_____ refers to a variable loading on more than one factor.

Accepted Answer

Cross loading refers to a variable that loads significantly on more than one factor. This can cause difficulty in interpreting the specific factor associated with the variable.

Question 8

A factor analysis solution that does not show the smaller, less significant factor loadings is called a(n) _____.

Accepted Answer

The answer of A factor analysis solution that does not...

Question 9

_____ is an interdependence multivariate technique that enables us to combine objects into groups so that objects within each of the groups are similar to each other and different from objects in all other groups.

Accepted Answer

Cluster analysis is the technique used to group objects into clusters based on similarities between multiple variables. It is often used in market segmentation, image recognition, and social network analysis.

Question 10

Which is NOT a phase to be performed in cluster analysis?

Accepted Answer

Developing a linear function to predict cluster membership is not a phase in cluster analysis. The other options (A, B, C, and E) are all phases commonly used in cluster analysis.

Question 11

The distances between the objects we are clustering can be measured in several ways. Which measure is most commonly used and has the fewest weaknesses?

Accepted Answer

Squared Euclidean distance is the most commonly used measure for calculating distance between objects in clustering due to its simplicity and efficiency in computation. It minimizes the effect of outliers and is suitable for various types of data, making it a versatile choice with relatively fewer weaknesses compared to other distance measures.

Question 12

Procedures used to group objects that are similar in cluster analysis all can be classified into two types, which are ____.

Accepted Answer

The answer of Procedures used to group objects that are...

Question 13

Agglomerative and divisive approaches to cluster analysis are types of _____ clustering procedures.

Accepted Answer

Agglomerative and divisive approaches are both methods of hierarchical clustering, which involves creating a hierarchy of clusters by either merging smaller clusters into larger ones (agglomerative) or splitting larger clusters into smaller ones (divisive). Nonhierarchical clustering, on the other hand, involves directly assigning individual data points to clusters without creating a hierarchy. Orthogonal, oblique, and dependence clustering are not commonly used terms in the context of cluster analysis.

Question 14

A tree-graph for hierarchical cluster analysis is referred to as a(n) _____.

Accepted Answer

A dendogram is a type of tree-graph commonly used in hierarchical cluster analysis to visually display the relationships between clusters of data. The other answer choices are not applicable to dendograms or clustering analysis.

Question 15

In _____ clustering procedures, one or more cluster seeds (initial starting points) are selected and objects within a prespecified distance from the cluster seeds are considered to be in particular cluster.

Accepted Answer

The description given pertains to nonhierarchical clustering procedures, where cluster seeds are chosen and objects within a distance from the seeds are assigned to particular clusters. Factor, orthogonal, and oblique clustering procedures are not defined or distinguished by how cluster seeds are selected or how distances are used to assign objects. Hierarchical clustering procedures construct a hierarchy of nested clusters based on distances between objects, without requiring the selection of cluster seeds.

Question 16

What is an advantage of the nonhierarchical clustering over the hierarchical clustering procedure?

Accepted Answer

Nonhierarchical clustering allows for flexibility in cluster assignments where objects can be removed from an early-formed cluster and reassigned to another cluster formed later in the sequence. This can lead to more accurate and meaningful clusters as the algorithm adjusts to the data in real-time.

Question 17

Harriet is examining the demographic and psychographic characteristics of the individuals in each cluster in an attempt to explain why the individuals were grouped in the manner they were. Harriet is _________ clusters.

Accepted Answer

Profiling involves examining the characteristics of individuals in each cluster, which is exactly what Harriet is doing. The other choices do not accurately describe Harriet's actions.

Question 18

_____ is a statistical tool that develops linear combinations of metric independent variables to predict group membership as defined by the nonmetric dependent variable.

Accepted Answer

Discriminant analysis is the statistical tool that develops linear combinations of metric independent variables to predict group membership as defined by the nonmetric dependent variable. Regression analysis is a statistical tool that predicts a metric dependent variable based on the values of one or more metric independent variables. Cluster analysis is a statistical tool that groups similar cases together based on their attributes. Factor analysis is a statistical tool that identifies underlying factors that explain the correlations among a set of variables. ANOVA is a statistical tool that tests for significant differences in means among groups.

Question 19

Jose wants to predict employees likely to leave versus those who are not likely leave the company based on the number of years an individual has worked for the company and their level of satisfaction with their job. The most appropriate type of analysis for Jose to use would be _____.

Accepted Answer

Discriminant analysis is appropriate for predicting group membership based on predictor variables. In this case, the groups would be employees likely to leave versus those who are not likely to leave, while the predictor variables are the number of years an individual has worked for the company and their level of job satisfaction.

Question 20

Estimates of the predictive power of each independent variable in discriminant analysis are called _____.

Accepted Answer

Estimates of the predictive power of each independent variable in discriminant analysis are called discriminant weights.

Quiz 15: Other Multivariate Techiques