FDSC 515: Sensometrics - Applied Multivariate Analysis in Sensory and Food Science Syllabus

FDSC 515: Sensometrics - Applied Multivariate Analysis in Sensory & Food Science (3 credits). The main objective of this course is to allow each student to develop the necessary data analysis skills needed for analyzing and interpreting sensory and consumer data.

Prerequisite: STAT 500 or BBH 505

Instructor

Helene Hopfer, Ph.D.
Associate Professor of Food Science
Program Chair, Graduate Certificate Sensory & Consumer Science

Department of Food Science
221 Rodney A. Erickson Food Science Building
University Park, PA 16802

Phone (Office): 814-863-5572
E-mail: Use Canvas Inbox or hopfer@psu.edu

Office Hours: Wednesday, 4:30–5:30 p.m. ET on Zoom

Course Objectives

Upon completion of this course, you will be able to:

1. use statistical software to analyze sensory data sets with uni- and multivariate methods
2. create code to import, analyze, and plot data
3. select an appropriate data analysis method to answer a research question
4. evaluate the quality and interpret the results of statistical analyses to answer a research question
5. explain and write about the results of the data analysis in a scientific report
6. critically evaluate scientific literature for their experimental design and used statistical methods
7. lead and guide a discussion on scientific literature
8. experience how statistical methods are used for sensory data sets

Course Outline

Module 1: Intro to R

Module 2: Working With Data

Module 3: Experimental Design & (M)ANOVA

Module 4: Power Analysis & Missing Data

Module 5: Principal Component Analysis (PCA)

Module 6: Correspondence Analysis

Module 7: Multidimensional Scaling (MDS) & DISTATIS

Module 8: Multi-Factor Analysis (MFA) & Generalized Procrustes Analysis (GPA)

Module 9: Cluster Analysis

Module 10: Discriminant Analysis (CVA)

Module 11: Modeling (PLS)

Module 12: Preference Mapping (IPM, EPM)

Module 13: Conjoint Analysis & Consumer Segmentation

Course Schedule

For due dates, refer to the Course Summary on the Syllabus page in Canvas.

Course Materials

No required text, but useful books can be found on the Further Readings page in Canvas.

Assignments

Your grade in this course consists of 6 parts:

1. Pre-lecture code quiz (10% of your grade)
2. Homework assignments (20% of your grade)
3. Peer reviews of homework assignments (5% of your grade)
4. Milestone assignments (25% of your grade)
5. Literature discussions (20% of your grade)
6. Take-home final (20% of your grade)
   • You will analyze a data set using the techniques studied in the course (see below). I encourage you to analyze your own data. However, if you would like to analyze your own data for the final, I need to approve the data set by at the latest April 1st, in writing!

Code Quiz (10% of your grade)

To help you with the coding for the homework assignments there will be a coding review and interpretation every week. To prepare for this code review class I will provide a data set and a R code script file on Canvas, together with an online quiz (graded).

Homework Assignments (20% of your grade)

To prepare you for your milestone assignments and your take-home final, there will be weekly homework assignments that are peer-reviewed. You are expected to create code to analyze and interpret the results for data sets assigned for each topic. The code and interpretation will be reviewed by one of your peers, using a grading rubric.

The purpose is to provide you with the opportunity to apply your knowledge of the selected data analysis method(s) and to help you master written communication of statistical analysis results. You are expected to submit your R code and a 1-page summary of your interpretation for peer review — often, you will be asked to also include answers to specific questions. Your code needs to run on somebody's computer, be understandable to that person (so use the comment function extensively), and your interpretation and reporting need to make sense and be supported by the data. You may include a graph or table in support of your interpretation, but you also need to ensure that your report covers the major findings of your analysis.

Please name all uploaded files in the following way:

• Report: LastName_HW#.pdf where # is the number
• R script file: LastName_HW#.r

Peer Reviews of Homework Assignments (5% of your grade)

Each week when there is a homework assignment due, you will be assigned the assignment of another student to peer review. The idea is to provide you the opportunity to see how others have solved the homework and to improve your critical evaluation skills in a professional and courteous manner. Think about how you would receive feedback on your work. To help you in your assessment use the provided rubric to evaluate your peer's submission. The evaluation rubrics are specific to the homework assignments but typically include some criteria tied to the code script and some regarding the submitted report. I urge you to make extensive use of the comments, especially when you deduct points to assist your colleague in understanding where they made a mistake in your opinion.

Peer reviews are due Wednesday after homework assignments are due.

Milestone Assignments (25% of your grade)

There will be three graded milestone assignments, where you are expected to demonstrate the skills you learned up to then. You are expected to create code, analyze, interpret, and report your findings in a report for the assigned data set. These assignments will build upon your weekly homework assignments.

The purpose of the milestone assignments is to provide you with the opportunity to apply your knowledge of the selected data analysis technique(s) and to help you master written communication of data analysis results using these methods.

Unless otherwise noted, your report should:

• Include an abstract, introduction, materials & methods, followed by results & discussion, and conclusions.
• Include the research questions at the end of the introduction.
• Describe your dataset, the experimental design of the dataset, and all statistical methods used in the Materials & Methods section.
• Place all necessary tables and graphs directly into the document. Use only graphs and/or tables that support your results and discussion, do not include any tables/graphs just because R provided them!
• Provide a full narrative interpretation of the results & discussion using the GEE principles.
• In your conclusion section suggest further analyses if needed.
• Include your R script as an .r file as an appendix. If you make any changes to the data set you need to upload that file as well.
• Acceptable file formats are .pdf for your final report, .csv, .txt, or .xlsx for your raw data set if you made any changes to the provided data set or use your own data set, and .r for your R script.

Please name all uploaded files in the following way:

• Report: LastName_MS#.pdf where # is MS1, MS2, or MS3
• R script file: LastName_MS#.r

Literature Discussions (20% of your grade)

In order to further your critical thinking skills, we are using online discussions for the weekly paper and general bigger-picture questions. Participation is a requirement for this course, and the Canvas discussion platform will be used for online discussion.

The goals are:

1. to critically evaluate the discussion papers with regards to the study, the methods used, the results, and conclusions (e.g., used statistical methods, experimental design, data analysis plan, general "usefulness").
2. to look at the bigger picture, e.g., how does the material covered in the paper relate to the methods we discussed, how could you apply these methods for your own data, what could be some of the advantages and limitations, which areas you are confused about, etc.

In order to receive your grade every week, you must post 2 new questions and 2 responses to other people's questions per week. There will be a Sunday 11:59 p.m. deadline for submissions each week.

Take-Home Final (20% of your grade)

For the final, you will analyze a larger data set with different techniques that we discussed in class. You can analyze your own data set if you have obtained prior authorization (see above), otherwise, I will provide a data set.

• You have to develop 2–3 research questions for the provided data set. Come up with a background story that allows you to develop those research questions for the provided data set.
• You have to use at least three different methods to analyze the data and you have to justify why you are using each method. That means, tell me how these methods allow you to answer your research questions.
• You need to hand in a fully written research paper in the form of a journal article. Use the provided word template to write your final paper. You are limited to a maximum of 3 pages.

Unless otherwise noted, your final report should:

• Include an abstract, introduction, materials & methods, followed by results & discussion, and conclusions.
• Include the research questions at the end of the introduction.
• Describe your dataset, the experimental design of the dataset, and all statistical methods used in the Materials & Methods section.
• Place all necessary tables and graphs directly into the document. Use only graphs and/or tables that support your results and discussion, do not include any tables/graphs just because R provided them!
• Provide a full narrative interpretation of the results & discussion using the GEE principles.
• In your conclusion section suggest further analyses if needed.
• Include your R script as an .r file as an appendix. If you make any changes to the data set you need to upload that file as well.
• Acceptable file formats are .pdf for your final report, .csv, .txt, or .xlsx for your raw data set if you made any changes to the provided data set or use your own data set, and .r for your R script.

Please name all uploaded files in the following way:

• Report: LastName_Final.pdf
• R script file: LastName_Final.r

Only material uploaded to Canvas by the submission deadline will be used for your final grade.

Late Submission Policy

Based on previous experiences, I developed a late submission policy as follows:

• Homework assignments build upon each other, therefore, submission of your homework assignments is due as indicated (typically the Wednesday after the module ends). Peer reviews of the homework assignments are due 1 week after the submission due date.
• There are 3 milestone submission deadlines throughout the semester: You are expected to submit your report by the listed due date, past that deadline no late submissions will be accepted!
• The same holds true for the take-home final.

Scheduling Conflicts

I'd be happy to make arrangements for scheduling conflicts but need to be advised ahead of time. I am happy to discuss how to best accommodate scheduling conflicts but strongly encourage you to contact me as early as possible, rather than waiting until a deadline to ask for an accommodation.

Grading Policy

The following table is the grading criteria for the course.

The following table is the grading scheme for the course.

Please refer to the University Grading Policy for Graduate Courses for additional information.

NOTE: If you are planning to graduate this semester, please communicate your intent to graduate to your instructor. This will alert your instructor to the need to submit your final grade in time to meet the published graduation deadlines. For more information about graduation policies and deadlines, please see "Graduation" under World Campus Student Resources.

Online Students Use of the Library

As Penn State students, you have access to many of the materials that the library offers to students. The library website has a lot to offer but can be overwhelming. A guide has been created to serve as your introduction to important library resources, services, and important pages within the library. The Online Student Library Guide is updated regularly by the online librarian. It is intended to provide a level of comfort through an introduction to help you feel comfortable navigating the library website to find valuable information for your coursework.