# Information about statistics and statisticians

Activity Log

April 2012

This activity log reports the projects that Professional Statistical Services conducted during April of calendar year 2012. Unless otherwise specified, all computerized analyses were conducted using SPSS.

 Project Title Description Methods Used Status Clinical Test of Extended vs. Intermittent Methods of Aminoglycoside Infusion Compute descriptive statistics for each treatment group on all variables; compare means on the 3 dependent variables (i.e., mortality, length of stay and duration of treatment) between each pair of subgroups on 14 categorical variables; compute correlations between 4 continuous independent variables and each of the 3 dependent variables; identify significant predictors of each of the 3 dependent variables. Descriptive statistics Independent sample t-tests Correlation logistic regression completed Review and Edit Methods Chapter of Dissertation Edit wording and grammar; add description and justification of mediation analysis. Writing skills Knowledge of mediation analysis using Preacher & Hayes bootstrap method of computing significance of indirect effect completed Analyze summary data on bleeding vs. non-bleeding events among hospital patients over 3 consecutive years Compare rates of bleeding between each pair of years; determine whether a linear trend in the rate of bleeding existed over the 3 years. Chi-square test of independence Fisher exact test of independence Chi-square test of trend completed Graduate Intro Statistics Problem Set Solve seven problems dealing with multiple regression, residual analysis, assessing fit of data to assumptions of regression, data transformations to correct skewness, fitting nonlinear models, identifying outliers. Multiple regression Data transformations to correct non-normality Use Minitab for all analyses completed Quality Assurance Case Analysis for Graduate Business Course Develop a solution to a case involving defects in an automatic production process. Six Sigma analyses Quality improvement Organizational Change theory and methods completed Graduate Political Science Quantitative Methods Course Problem Set Conduct binary logistic regression on a provided data set, summarize findings, interpret improvement in classification accuracy, interpret odds ratios for 3 predictors. Logistic regression Knowledge of odds ratios completed Data Analysis & Interpretation for Dissertation in Applied Animal Behavior Revise hypotheses to testable form, analyze results of two experiments in canine tracking behavior. Correlation 1-within, 1-between ANOVA Independent samples t-test McNemar and Liddell (exact) tests for paired data in a 2x2 contingency table completed Graduate Nursing Quantitative Methods Course Problem Set Calculate needed sample size needed to infer population proportion with 4% error and 98% confidence, calculate needed sample size to achieve specified error with 90% confidence (estimating variance from range), compute confidence intervals, compute needed sample size for test of difference between independent proportions for specified powers to detect specified differences at specified alphas, and other similar problems. Knowledge of computational formulas for needed sample sizes given specified margins of error, confidence, alpha and beta levels. Compute any of the above based on specified values of the others. completed Graduate Medical Statistics Course Problem Assignment Conduct and report results of survival analysis Kaplan-Meier Cox Regression completed Graduate Biostatistics Course Problem Assignment Compute margin of error values, sample size requirements, test hypotheses regarding proportions. Knowledge of sample size formulas for specified confidence, power, and margin of error. completed Graduate Course in Operations & Supply Chain Management Pr Compute and interpret p and r control charts for a before and after case with data provided; 10 control charts in all. Computation, graphing, and interpretation of p and r control charts. completed Undergraduate Statistics in Business Course Term Project Provide guidance in formulation study and data analysis; analyze data; prepare draft of methods and results sections. Research design Simple 2-group tests (e.g., z-test for differences in proportions) Skills at writing scientific research reports completed Graduate Business Statistics Course Problem Set Use data supplied with case problem to identify and/or construct potential predictors, use alternative methods to identify best regression model, explain procedures and decisions, interpret results. Multiple regression Best subsets procedure Interpretation of Mallow's Cp Writing skills Minitab used for all analyses completed Analysis of Relationship between Weather and Admissions to Hospital Emergency Departments Analyze 6 years of daily weather data, with 4 different weather indexes, for its relationship to hospital emergency admissions. Multiple regression Best subsets procedure Residuals analysis completed Quantitative Methods in Economics Case Problem with Data Analyze the relationship of standardized testing in schools to subsequent indicators of economic well-being; 10 multipart questions; theorizing, analysis, interpretation. Means tests Multiple Regression Using concept of joint significance Scientific reasoning Knowledge of economics completed Biostatistics Course Assignment Compute margin of error, compute needed sample sizes for different margins of error, discuss issues relevant to determining appropriate sample size. Knowledge of formulas relating sample size, confidence, and margin of error Knowledge of sample size determination issues completed Write Journal Article on the Results of a Clinical Research Study on Humans Write a journal article on a controlled study of the effects of a dental product on humans; include lit review, hypotheses, method, results, and discussion sections. Knowledge of structure and appropriate content of scientific journal articles Literature review skills Writing skills Statistical testing (t-tests, normality assumption tests) completed Statistics Course Assignment Follow a 9-step protocol, proceeding from the research question through hypothesis specification, testing, & conclusion) for 5 research questions; extensive writing (10 pages). Hypothesis specification Assumptions of statistical tests z-tests and independent sample t-tests Minitab used for all analyses completed Graduate Business Statistics Course Problem Set Four case problems with data provided; compare focus group ratings of 3 commercials; test for differences in regional preferences for product characteristics; examine data for relationship between age category and political affiliation; examine data for differences between source types on perceived degree to which the source disseminates information. Repeated measures one-way ANOVA Chi-square test of independence Minitab used for all analyses completed Graduate Business Statistics Course Exam Problems Develop answers for 4 multi-part problems: relationship between water quality and watershed areas of streams; relationship between gasoline mileage and weight, displacement, horsepower, length, and width; indications of racial or ethnic differences in rates of being stopped by police; relationship between birth weight and infant survival. Simple regression Multiple regression Stepwise variable selection Chi-square test of independence Binary logistic regression` completed Biostatistics Course Problem Set Compute normal scores of a subset of observations from a sample 130; compute correlation in a normal probability plot; compute correlations between variables. Knowledge of normal scores and how to compute them Pearson correlation completed Graduate Statistics Course Homework Problems Develop answers to 6 multi-part problems: compute predicted values of regression model; compute R2; compute F for difference in R2 between regression models (partial F tests); assess confounding in regression; test for parallelism of 2 straight lines; test for coincidence of two straight lines. Multiple regression Regression predicted values Partial F test t-test for regression coefficients Assessment of confounding in regression Assessment of coincidence of two straight lines completed Graduate Nursing Statistics Course: Problem Set Explore relationship between speed and miles per gallon of fuel, and engine displacement and miles per gallon of fuel; model mpg as a function of speed and engine displacement; evaluate residuals for fulfillment of regression assumptions; model number of state employees as function of state population size; model breast-feeding status against ethnicity; Comparison of fits of alternative functions Multiple regression Logistic regression completed Graduate Nursing Statistics Course: Logistic Regression Problem Set For multiple problems, compute confidence interval for LR predictor coefficient; compute confidence interval for the population odds ratio; compute overall G test for an LR model. Logistic regression Odds ratios Tests of fit of LR models completed Graduate Business Statistics Course: Case problem with 9 sections Compute analysis required in each of the 9 sections; answer questions regarding interpretation of results. Multiple regression Sample size requirements given confidence level, margin of error, and range of responses in a pilot sample Tests for normality One-way ANOVA with post hoc tests t-test assuming unequal variances Excel add-ons used for all analyses completed Graduate Nursing Biostatistics Course: Survival Analysis Problem Set Interpret output from Kaplan-Meier analyses for single group and multiple groups; interpret Cox regression results for a predictor with 2 levels and Injury ; compute hazard ratios for 6 different ISS scores. Kaplan-Meier survival analysis Cox regression completed Graduate Business Statistics Course: Homework problems Answer questions about 5 short case problems; how to measure variables, what analysis to use, justification for analysis choice. Research design Chi-square test of independence ANOVA Correlation completed Graduate Business Statistics Homework Problems Compute logistic regression for provided data; derive equation for outcome probability from the equation for the log odds of the outcome; compare nonlinearity of predicted odds to linearity of predicted log odds; compare least squares and logistic fits. Logistic regression Linear regression Odds, log odds Algebra completed Graduate Statistics Course Project Compute optimal nonlinear regression model for a data set. Identify nonlinear variables Identify best-fitting nonlinear functions Generate scatterplots of raw data, fits, and residuals STATA used for all analyses completed Graduate Business Statistics Course Homework Assignment Compute 10,000 iteration simulation of frequencies of 2 conditions with starting relative frequencies; compute hypergeometric probabilities; chi-square analysis; compute relative risk and odds ratio; Simulation in Excel Computation of hypergeometric p-values Chi-square Relative risk Odds ratio completed Graduate Nursing Statistics Course: Problem Set Compute odds ratio; interpret logistic regression output; compute predicted probability from LR model; compute odds and odds ratios; show equivalence between alternative ways of computing odds and odds ratios. Logistic regression Odds ratios Algebra completed