STATISTICS QNS

StatisticsQuestion1.docx

Home >Mathematics homework help >Statistics homework help >STATISTICS QNS

Question 1

Suppose a researcher is interested in assessing how the efficacy of a firm's advertisements measured by two outcome variables varies by four groups created by two factors with two levels each. The two dependent variables are ability to gain attention and ability to obtain a purchase, both measured on a 10-point scale. The first factor measures product type, categorized as product 1 and product 2. The second factor measures whether a customer is current, categorized as current customer and ex-customer. Research data was used to construct the following graphA visual examination of the results suggests that

		the research should run multiple separate one-way ANOVAs.
		the differences for customer type are less than the differences for product type.
		the differences for product type are are less than the differences for customer type.
		the differences for product type and customer type are statistically insignificant.

Question 2

Instead of grouping variables based on their correlations, cluster analysis groups _____ based on a group of _____ selected by the _____.

		variables, objects, clustering algorithm
		objects, variables, researcher
		variables, correlations, researcher
		correlations, objects, clustering algorithm

Question 3

Interdependence techniques studied so far this semester include

		Cluster Analysis
		Multidimensional Scaling
		Factor Analysis
		all of the above.

Question 4

Which technique is most appropriate if researchers are interested in studying whether multiple outcome variables vary by groups defined by one or more categorical independent varaibles?

		discriminant analysis
		cluster analysis
		manova
		all of the above

Question 5

The univariate procedures of ANOVA described in chapter 7 are valid if it is assumed that

		the dependent variable is normally distributed.
		groups are statistically the same.
		that the dependent variable is normally distributed, the groups are independent in their responses on the dependent variable, and the variances are unequal for all treatment groups.
		that the dependent variable is normally distributed, the groups are independent in their responses on the dependent variable, and the variances are equal for all treatment groups

Question 6

If the composite dependent variable is simply the average of the sum of the outcome variables by group, its values for each group are

		2.25, 4.25, 5, and 6.75
		8, 11.125, 6.25, 12.875
		4.25, 8.25, 11.75, and 14
		2, 4, 5, and 7

Question 7

What are some tools that researchers can use to determine how many clusters to retain in a Cluster Analysis?

		Stopping rules
		The Dendrogram
		Heterogeneity measures
		all of the above

Question 8

To achieve the suggested power of 0.80 in MANOVA when assessing medium effect sizes in a 5 group design comprised of six dependent variables, the researcher must include at least _______ subjects per group?

		82
		74
		90
		105

Question 9

MANOVA assumptions include

		observations are independent.
		multivariate normality.
		equal variance-covariance matrices for all treatment groups
		all of the above.

Question 10

Given the multidimensional scaling results below, how many dimensions should be retained? What percentage of the total variation did eigenvalues account for?

		2, 44
		2,54
		2, 43
		3, 64

Question 11

What are the two biggest assumptions in performing a Cluster Analysis?

		Sample representativeness and no multicollinearity
		Sample representativeness and multicollinearity.
		Sample representativeness and the number of cluster chosen.
		Selecting the clustering algorithm and selecting the similarity measure.

Question 12

Once clusters have been created, researchers can validate the results by

		running a MANOVA using criterion validity outcome variables as the dependent variables and cluster membership as the categorical independent variable.
		profile the final cluster solution on a set of additional variables not included in the clustering variate.
		assess cluster stability by resorting the data on an arbitrarily chosen variable, rerunning your cluster analysis, and examining the results for consistency.
		doing all of the above.

Question 13

Suppose you are interested in studying whether two dependent variables are statistically significantly different for multiple groups formed by one or more categorical variables. What is an important consideration when running consecutive one-way ANOVAs?

		If the dependent variables are statistically independent, then running multiple one-way ANOVAs results in bias.
		If the independent variables are statistically dependent, then running multiple one-way ANOVAs results in bias.
		If the independent variables are statistically independent, then running multiple one-way ANOVAs results in bias.
		If the dependent variables are statistically dependent, then the researcher runs the risk of increasing the probability of a Type 1 error.

Question 14

How many clusters should be retained based on the dendrogram below

		2
		3
		4
		5

Question 15

What are the biggest criticisms of Cluster Analysis?

		Cluster Analysis has no statistical basis on which to draw inferences from a sample to a population.
		Cluster Analysis will always create clusters, regardless of whether there exists any structure in the data.
		The cluster solution is not generalizable since it is totally dependent upon the variables used as the basis for the similarity measure.
		All over the above.

Question 16

In deciding how many clusters to include in a Cluster Analysis, the research must balance the trade-off between

		the number of clusters chosen and the level of statistical significance.
		the number of clusters chosen and the similarity measure.
		the number of clusters chosen and within-cluster homogeneity.
		a small number of clusters and more homogeneity within clusters.

Question 17

Suppose that we are interested in assessing whether two groups formed by variable X5 exhibit statistically significant differences for three outcome variables measured by X19, X20, and X21. After running the command "manova X19 X20 X21 = X5" in Stata, we get the following results:

Based on the results above

		a statistically significant difference exits for X19 by X5.
		MANOVA results are inconclusive.
		a statistically insignificant difference exists for X19 by X5.
		a statistically significant difference exists for a variate comprised of X19, X20, and X21 by X5.

Question 18

How does Multidimensional Scaling differ from Cluster Analysis and Factor Analysis?

		Factor analysis defines structure by grouping variables into variates that represent underlying dimensions in the original set of variables.
		Cluster analysis defines structure by grouping objects according to their profile on a set of variables in which objects in close proximity to each other are grouped together.
		both (a) and (b)
		Multidimensional scaling does not use a variate. Also, a solution can be obtained for each respondent.