Practical Multivariate Analysis 5th Edition Abdelmonem Afifi, Susanne May, Virginia A. Clark - Solutions

Discover comprehensive resources for "Practical Multivariate Analysis 5th Edition" by Abdelmonem Afifi, Susanne May, and Virginia A. Clark. Access the online answers key and detailed solutions PDF to enhance your understanding of the textbook. Benefit from the solution manual and solved problems for a clear grasp of complex concepts. Explore the test bank for practice with chapter solutions and step-by-step answers. Our instructor manual offers invaluable insights, while the questions and answers section ensures thorough preparation. Enjoy free download options to facilitate your learning journey with this essential academic resource.

What are key differences in operators between SAS and Stata?
What are some instances in which we would need to use expressions?
How can you add a comment to SAS and to Stata?
How do you tell SAS and Stata where a line ends, by default?
Where in the batch program can you create new variables in SAS and Stata?
Why is a skip map useful?
What are some conventions you might use to name variables?
What are some conventions you might use to name a project folder and to name your analytic data fi le, your batch program, and your results fi les?
What are the conventional extensions for SAS and for Stata data fi les, batch program fi les, and results fi les?
Answer the following questions to help your instructor get to know you better and help you think about your goals for the course. Some questions may be diffi cult to answer, especially if your research interests and empirical approach are still developing, so answer as best you can. In addition to
Locate a journal article using regression analyses, preferably on a topic of interest to you. You can use an article you already have or draw something from the syllabus of one of your substantive classes or search a bibliographic database.Read (or reread) the article, paying particular attention
For the school example and GEE marginal approach we mentioned that the most appropriate correlation structure is exchangeable. Give reasons why the AR(1) or the unstructured options for correlation structure might not be appropriate for this setting.
For the mice data, instead of modeling weight linearly over time, model weight as a quadratic function of time by including a term for days as well as for days2. Which model (random intercept, random slope) and correlation structure(s) seem most appropriate for these data?Describe your reasoning
There might be siblings represented in the school data set. The factor parental education could potentially be considered another level in the hierarchical model. Give reasons for when and why this might or should (not) be addressed.
For the school data set (again, using math score as the outcome variable) start with fitting a model with the following fixed effects: SES centered around its respective school mean, school mean SES, public school (vs. private), hours spent on homework, race, gender, and the following random
For the school data set (again, using math score as the outcome variable) fit the following models:a) A random intercept model with random intercepts for schools and both the raw SES scores and the respective school means as fixed effects.b) The same model as in (a) except using SES centered around
For the school data set (again, using math score as the outcome variable) fit a random intercept model with random intercepts for schools and SES centered around its respective school mean as a fixed effect. What is the estimated within school slope for SES? Compute the estimated between school
For the school data set using math score as the outcome variable, fit the following models:a) A fixed effects model with schools and SES as fixed effects.b) The same as (a), but with each SES centered around its respective school mean.c) Compare the models in (a) and (b). What effect did centering
For the school data set generate box plots of SES by schools (similar to Figure 18.2). Interpret the graph with respect to potential differences in mean SES and differences in variability of SES across schools.
Three students have the following scores on their math, english and social science classes.Student 1: 87.3, 91.2 and 86.0, Student 2: 75.4, 81.3 and 79.6, Student 3: 98.8, 95.0 and 94.7.a) Using a spreadsheet program or a calculator, calculate the intra-class correlation coefficient.b) Change three
(Problem 17.9 continued) Compare the results you get from Problem 17.9 and Problem 12.30.
This problem uses the Northridge earthquake data set. Look for significant associations in homeowner status (V449), home evacuation status (V173), and status of home damage (V127)using a log-linear analysis.
Look for significant associations among the variables gender, currently living with(LIVWITH), the financial situation (FINSIT), and having enough money for food (MONFOOD)in the Parental HIV data using log-linear analyses. For the final model assess goodness-of-fit. Interpret the findings.
In Problem 17.5, eliminate the four-way interaction, then compare the two models: the one with and the one without Smokfu. Interpret the results.
Rerun the analyses given in Problem 17.4 adding a delta of 0.01, 0.1, and 0.5 to see if it changes any of the results.
Run a four-way log-linear analysis using the variables Stagen, Hist, Smokfu and Death in the lung cancer data set. Report on the significant associations that you find.
Perform a log-linear analysis using data from the lung cancer data set described in Section 13.3 and Appendix A. Check if there are any significant associations among Staget, Stagen and Hist. Which variable is resulting in a small number of observations in the cells?
Split the variable EDUCAT in the depression data set into two groups, those who completed high school and those who did not. Also split INCOME at 18 or less and greater than 18 (this is thousands of dollars). Run a log-linear analysis of SEX versus grouped EDUCAT versus grouped INCOME and describe
Run a log-linear analysis using the variables DRINK, CESD (split at ten or less), and TREAT from the depression data set to see if there is any significant association among these variables.
Using the variables AGE and CESD from the depression data, make two new categorical variables called CAGE and CCESD. For CAGE, group together all respondents that are 35 years old or less, 36 up to and including 55, and over 55 into three groups. Also group the CESD into two groups; those with a
For the Parental HIV data consider the subgroup of adolescents who have used marijuana. Perform a hierarchical cluster analysis separately for females on the following variables from the Parental HIV data: the sum of the variables describing the neighborhood (NGHB1–NGHB11), enough money for food
Repeat Problem 16.1, using the variables age, income, and education instead of the last seven variables.
Create a data set from the family lung function data described in Appendix A as follows. It will have three times the number of observations that the original data set has—the first third of the observations will be the mothers’ measurements, the second third those of the fathers, and the final
Repeat Problem 16.6, using the K-means method for K = 4. Compare the results with the four clusters produced in Problem 16.6.
Repeat Problem 16.6, using AREA as an additional clustering variable. Comment on your results.
Describe how you would expect guards, forwards, and centers in basketball to cluster on the basis of size or other variables. Which variables should be measured?16.6 For the family lung function data described in Appendix A, perform a hierarchical cluster analysis using the variables FEV1 and FVC
Perform a cluster analysis on the chemical company data in Table 9.1, using the K-means method for K = 2;3; 4.16.4 For the accompanying small, hypothetical data set, plot the data by using methods given in this chapter, and perform both hierarchical and K-means clustering with K = 2. Cases X1 X2 1
For the situation described in Problem 8.7, modify the data for X1, X2,. . . , X9 as follows. For the first 25 cases, add 10 to X1, X2, X3. For the next 25 cases, add 10 to X4, X5, X6. For the next 25 cases, add 10 to X7, X8. For the last 25 cases, add 10 to X9.Now perform a cluster analysis to
For the depression data set, use the last seven variables to perform a cluster analysis producing two groups. Compare the distribution of CESD and cases in the groups. Compare also the distribution of sex in the groups. Try two different programs with different options. Comment on your results.
This problem uses the Northridge earthquake data set. Perform a factor analysis on the following items in the Brief Symptom Inventory (BSI): V346, V357, V364, V383, V390, V394, V351, V358, V385, V386, V391. Note that the first six items relate to the anxiety subscale of BSI and the next five items
Repeat Problem 15.8 using an oblique rotation. Do the substantive conclusions change?
Perform a factor analysis on all of the items of the Parental Bonding scale for the Parental HIV data (see Appendix A and the codebook). Retain two factors. Rotate the factors using an orthogonal rotation. Do the items with the highest loadings for each of the factors correspond to the items of the
For the depression data set, perform four factor analyses on the last seven variables DRINK–CHRONILL (Table 3.4 or 3.5). Use two different initial extraction methods and both orthogonal and oblique rotations. Interpret the results. Compare the results to those from Problem 14.1.
Separate the depression data set into two subgroups, men and women. Using four factors, repeat the factor analysis in Table 15.7. Compare the results of your two factor analyses to each other and to the results in Table 15.7.
For the data generated in Problem 8.7, perform four factor analyses, using two different initial extraction methods and both orthogonal and oblique rotations. Interpret the results.
Perform a factor analysis on the data in Table 9.1 and explain any insights this factor analysis gives you.
Another method of factor extraction, maximum likelihood, was mentioned in Section 15.6 but not discussed in detail. Use one of the packages which offers this option to analyze the data along with an oblique and orthogonal rotation. Compare the results with those in Problems 15.1 and 15.2, and
Repeat the analysis of Problem 15.1 and Table 15.7, but use an iterated principal factor solution instead of the principal components method. Compare the results.
The CESD scale items (C1–C20) from the depression data set in Chapter 3 were used to obtain the factor loadings listed in Table 15.7. The initial factor solution was obtained from the principal components method, and a varimax rotation was performed. Analyze this same data set by using an oblique
Perform a principal components analysis on all the items of the Parental Bonding scale for the Parental HIV data (see Appendix A and the codebook). How many principal components would you expect to retain for this scale? How many components should be retained based on the rules of thumb mentioned
Perform a principal components analysis on AGE and INCOME using the depression data set.Include all the additional data points listed in Problem 7.9(b). Plot the original variables and the principal components. Indicate the added points on the graph, and discuss the results.
Using the family lung function data, perform a principal components analysis on mother’s height, weight, age, FEV1, and FVC. Use the covariance matrix, then repeat using the correlation matrix. Compare the results and comment.
(Continuation of Problem 14.7.) Perform a regression of FEV1 for the oldest child on the principal components found in Problem 14.7. Compare the results to those from Problem 8.15.
Using the family lung function data, perform a principal components analysis on age, height, and weight for the oldest child.
Using the family lung function data described in Appendix A define a new variable RATIO= FEV1/FVC for the fathers. What is the correlation between RATIO and FEV1? Between RATIO and FVC? Perform a principal components analysis on FEV1 and FVC, plotting the results. Perform a principal components
Perform a principal components analysis on the data in Table 9.1 (not including the variable P/E). Interpret the components. Then perform a regression analysis with P/E as the dependent variable, using the relevant principal components. Compare the results with those in Chapter 9.
(Continuation of Problem 14.3.) Perform the regression of Y on the principal components.Compare the results with the multiple regression of Y on X1 to X9.
For the data generated in Problem 8.7, perform a principal components analysis on X1;X2; : : : ;X9. Compare the results with what is known about the population.
(Continuation of Problem 14.1.) Perform a regression analysis of CASES on the last seven variables as well as on the principal components. What does the regression represent? Interpret the results.
For the depression data set described in Appendix A, perform a principal components analysis on the last seven variables DRINK–CHRONILL (Table 3.5). Interpret the results.
Perform a proportional hazards regression using the variables Staget, Treat, and Perfbl and stratify the model by Poinf. Compare the estimated hazard ratios to the ones obtained in Problem 13.9.
Evaluate graphically and statistically the proportional hazards assumption for the variables Perfbl, Poinf, and Treat in the model presented in Table 13.4 using the methods described in Section 13.8.
Define a variable Smokchng that measures change in smoking status between baseline and follow-up, so that Smokchng equals 1 if a person changes from being a smoker to being an ex-smoker and equals 0 otherwise.(a) Is there an association between Staget and Smokchng?(b) What effects does Smokchng
Repeat Problem 13.6 assuming a proportional hazards model.
Assuming a log-linear model, do the effects of smoking status upon survival change depending on the tumor size at diagnosis?
Repeat Problem 13.4 assuming a proportional hazards model.
Assuming a log-linear model for survival, does smoking status (i.e., the variables Smokbl and Smokfu) significantly affect survival?
Do the patterns of censoring appear to be the same for smokers at baseline, ex-smokers at baseline, and nonsmokers at baseline? What about for those who are smokers, ex-smokers, and nonsmokers at follow-up?
Repeat Problem 13.1, using a Cox proportional hazards model instead of a log-linear. Compare the results.
(a) Find the effect of Stagen and Hist upon survival by fitting a log-linear model. Check any assumptions and evaluate the fit using the graphical methods described in this chapter.(b) What happens in part (a) if you include Staget in your model along with Stagen and Hist as predictors of survival?
(Problem 12.30 continued) Using an appropriate method in your software package, obtain confidence intervals for the odds ratios you computed in partsa, b and c of Problem 12.30.
(Problem 12.29 continued) Fit a logistic regression model (again using as the outcome “evacuate home” (V173)) which includes as the only covariates home owner status (rent/own, V449)and status of home damage (V127) and an appropriate interaction term. Is there a statistically significant
This problem and the following ones also use the Northridge earthquake data set. Perform an appropriate regression analysis using variable selection techniques for the following outcome:evacuate home (use V173 of the questionnaire). Choose from the following predictor variables:MMI (which is a
(Problem 12.22 continued) Perform diagnostic procedures to identify influential observations.Remove the four (4) most influential observations using the delta chi-square method. Rerun the analysis and compare the results. What is your conclusion?
(Problem 12.22 continued) Is there an interaction effect between age and home ownership, controlling for gender and ethnicity?
(Problem 12.22 continued) Is there an interaction effect between gender and home ownership?That is, are the estimated effects of home ownership upon reporting emotional injuries different for men and women, controlling for age and ethnicity?
(Problem 12.22 continued) Are the effects of ethnicity upon reporting emotional injuries statistically significant, controlling for home ownership status, age, and gender? Use a likelihood ratio test to answer this question.
(Problem 12.22 continued) Based on your results, what is the estimated probability of reporting emotional injuries for a 30-year-old white female renter? For a 50-year-old Latino home owner?
(Problem 12.22 continued) Fit a logistic regression model using emotional injury (yes/no, W238) as an outcome and using home ownership status (rent/own, V449), age (RAGE), gender(RSEX), and ethnicity (NEWETHN) as independent variables. Give parameter estimates and their standard errors. What are
This problem and the following ones use the Northridge earthquake data set. We wish to answer the questions: Were homeowners more likely than renters to report emotional injuries as a result of the Northridge earthquake, controlling for age (RAGE), gender (RSEX), and ethnicity (NEWETHN)? Use V449,
For the model in 12.20 use the referent time as given by the variable HMONTH to define the offset. Run the same model as in 12.20. Present rate ratios, including confidence intervals, and interpret the results. Describe any discrepancies compared to the results obtained in 12.20.
For the Parental HIV data perform a Poisson regression on the number of days the adolescents were absent from school without a reason. For this analysis assume that the referent time period was one month for all adolescents. As independent variables consider gender, age, and to what degree the
For the model in 12.18 find an appropriate cutoff point to discriminate between adolescents who were absent without a reason and those who were not. Assess how well the model predicts the outcome using sensitivity, specificity, and the ROC curve.
Perform a binary logistic regression analysis using the Parental HIV data to model the probability of having been absent from school without a reason (variable HOOKEY). Find the variables that best predict whether an adolescent had been absent without a reason or not.Assess goodness-of-fit for the
Perform a nominal and ordinal logistic regression analysis using the health scale as the outcome variable and age and income as independent variables. Present and interpret the results for an increase in age of 10 years and an increase in income of $5,000. How do the results from the ordinal
For the family lung function data perform a nominal logistic regression where the outcome variable is place of residence and the predictors are those defined in Problem 12.11. Test the hypothesis that each of the variables can be dropped, while keeping all the others in the model.Compare your
For the depression data perform an ordinal logistic regression analysis using the same outcome categories as in Section 12.9 which reverses the order and hence compares more severe depression to less severe depression. Use age, income, and sex as the independent variables.Compare your results to
Generate a graph for income similar to Figure 12.2 to assess whether modeling income linearly seems appropriate.
Perform a logistic regression analysis for the depression data which includes income and sex and models age as ana) quadratic orb) cubic function. Use likelihood ratio test statistics to determine whether these models are significantly better than a model which includes income and sex and models
Assume a logistic regression model includes a continuous variable like age and a categorical variable like gender and an interaction term for these. Is the P value for any of the main effects helpful in determining whethera) the interaction is significant,b) whether the main effect should be
For the family lung function data set, define a new variable VALLEY (residence in San Gabriel or San Fernando Valley) to be one if the family lives in Burbank or Glendora and zero otherwise.Using variable selection techniques, perform a logistic regression of VALLEY on mother’s age and FEV1,
Using the definition of low FEV1 given in Problem 12.9, perform a logistic regression of low FEV1 on area for the fathers. Include all four areas and use a dummy variable for each area.What is the intercept term? Is it what you expected (or could have expected)?
Define low FEV1 to be an FEV1 measurement below the median FEV1 of the fathers in the family lung function data set given in Appendix A. What are the odds that a father in this data set has low FEV1? What are the odds that a father from Glendora has low FEV1? What are the odds that a father from
Repeat Problem 12.7, but for chronic rather than acute illness.
Using the depression data set, appropriate variable selection techniques, and logistic regression, describe the probability of an acute illness as a function of age, education, income, depression, and regular drinking.
(a) Using the depression data set, fill in the following table:What are the odds that a woman is a regular drinker? That a man is a regular drinker? What is the odds ratio?(b) Repeat the tabulation and calculations for part (a) separately for people who are depressed and those who are not. Compare
Perform a logistic regression analysis on the data described in Problem 11.2.
Perform a logistic regression analysis with the same variables and data used in the example in Problem 11.13.
The accompanying table presents the number of individuals by smoking and disease status.What are the odds that a smoker will get disease A? That a nonsmoker will get disease A?What is the odds ratio? Disease A Smoking Yes No Total Yes 80 120 200 No 20 280 300 Total 100 400 500
Using the formula odds = P=(1????P), fill in the accompanying table. Odds P 0.25 0.20 0.5 1.0 0.5 1.5 2.0 2.5 3.0 0.75 5.0

Showing 200 - 300 of 2391

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Last

Practical Multivariate Analysis 5th Edition Abdelmonem Afifi, Susanne May, Virginia A. Clark - Solutions

Step by Step Answers