The Statistical Applications & Innovation Group (SAIG) offers free short courses to faculty, staff, and graduate students through the VT Professional Development Network. SAIG has supported faculty and student researchers at Virginia Tech for over 75 years!  We provide free courses to members of participating VT Colleges to make you a more effective researcher.  Do you need to analyze your data and aren't sure where to start?  Switching to R from another software package?  Using JMP ® for the first time?  Need a refresher on a particular topic?  Need to learn a new technique?  Faculty - These courses count toward computer refresh credits!

Have questions or need help with a SAIG course? We are here for you! Click the Short Course Questions link to send your inquires.

How can we help you?

Sound Design.  Whether you are working on social  science or rocket science, we can create the study or experimental design to achieve your research objectives. We can maximize information given your resources with sound design.

Appropriate Analytics.  How do I demonstrate statistical significance? What do I do with my non-normal data?  Logistic regression or tree model? Text mining? Bayesian or frequentist? We can sort these issues out by collaborating on appropriate analytics.

Intelligent Interpretation. So, you have statistical results and you are ready to publish or report to your sponsor. You ask: What is the proper way to communicate my conclusions? What visuals support my data analysis? We’ve got your back with intelligent interpretation.

Growth. Onward and upward with SAIG. Sound statistical design, appropriate analytics, and intelligent interpretation are key elements in producing your best proposals and publications. Partner with SAIG and the sky is the limit on your research growth.

Client-Statistical Collaborator Agreement

Services provided by our collaborators are contingent on compliance with the terms below. There is no charge for our services, but clients are limited to 10 hours of assistance per semester. Exceptions to the 10-hour policy may be made for projects in which joint publications are being pursued or for those with potential for future collaborative research.

Client Agreements:

• I will provide 24 hours notice to all parties if I must cancel a meeting or arrive late for any reason. Failure to provide this notice will result in forfeiture of statistical collaboration privileges for the remainder of the semester.
• To make the most of my time with my statistical collaborators I will come prepared to the first meeting with clearly defined research goals and questions, research papers that may assist the collaborators in solving my problem, and the results of any analyses already performed—including plots or graphs.
• I understand that many problems cannot be solved in just one meeting and am prepared to meet several times with the statistical collaboration team if needed.
• I understand that I am limited to 10 hours of assistance per semester. I also understand that this 10-hour time limit is not limited to time in meetings but also includes time that my statistical collaborators spend working on my project outside of meetings. This typically results in approximately 3 meetings. I also understand that if I need more assistance, co-authorship opportunities must be explored. For more information about the 10 hour limit, please see this FAQ.
• I understand that statistical collaborators primarily provide advice and assistance for my project and do not typically clean or format data for clients.
• When appropriate I will give feedback regarding my experience with my statistical collaborators (you will be emailed a link for submitting feedback).
• I understand that I cannot receive for assistance on class projects or homework.

Statistical Collaborator Agreements:

• We will notify all parties involved at least 24 hours in advance if we must cancel a meeting or arrive late.
• We will provide assistance in a timely manner to our clients.
• We will advise clients on appropriate analyses for their data and assist in performing the analyses as time allows.
• We will discuss with the client any statistical assumptions made and the potential drawbacks of the analyses.
• We will attempt to explain the statistical methods used in terms the client understands.

The topic of co-authorship can be a touchy subject. SAIG provides this document to express views on co-authorship. 

Co-authorship is not required for SAIG collaboration projects; however, if collaborators perform the work of a co-author, they should receive the credit of co-authorship. 

It is important to have a discussion about what work is to be completed and if it qualifies for a co-authorship.

SAIG bases co-authorship guidelines on those of the International Committee of Medical Journal Editors (ICMJE) which states that authorship credit should be based on 1) substantial contributions to conception and design, acquisition of data, or analysis and interpretation of data; 2) drafting the article or revising it critically for important intellectual content; and 3) final approval of the version to be published. Authors should meet conditions 1, 2, and 3. 1  

Another authorship guideline is the Galindo-Leal guide for explicit authorship 2. All individuals who work on a paper are assigned points based on their contribution (see Figure 1). If an individual reaches 25 points, he/she deserves to be a co- author. Points are per individual and can total over 100 for the entire paper.

1. http://www.icmje.org/icmje-recommendations.pdf
   http://www.icmje.org/recommendations/browse/roles-and-responsibilities/defining-the-role-of-authors-and-contributors.html
2. http://www.academia.edu/1323050/Explicit_authorship

Within Virginia Tech

• StatCom
• Department of Statistics
• College of Science
• Computer Labs on Campus
• EDRE Research and Consultation Lab

Outside Virginia Tech

• StatCom at Purdue - is a student-run organization that provides pro-bono statistical consulting for local non-profits and governmental organizations.
• UCLA Academic Technology Services: Statistical Computing - this website provides help for using statistical software, downloadable books on statistical topics and software, and data analysis examples in multiple software packages.
• Simon Fraser University: Statistics and Actuarial Science: Datasets; SAS, R, and JMP programs; Output - lots of datasets covering various statistical topics
• JSM (the Joint Statisical Meetings) - the largest gathering of statisticians held in North America. Attended by over 5000 people, activities of the meeting include oral presentations, panel sessions, poster presentations, continuing education courses, exhibit hall (with state-of-the-art statistical products and opportunities), career placement service, society and section business meetings, committee meetings, social activities, and networking opportunities.
• R twotorials: "how to do stuff in r. two minutes or less. for those of us who prefer to learn by watching and listening."

Software

• JMP - low price software for statistical analysis (for VT faculty, staff, and students). [Download]  JMP Webcasts Are available on a variety of topics.
• R - free software environment for statistical computing and graphics. It compiles and runs on a wide variety of UNIX platforms, Windows, and MacOS.
• SPSS - an open source version of SPSS
• SAS - code based software for performing statistics and analytics. Available for purchase through VT Software Distribution. [Download]

General

• How do I request statistical assistance for my research?
• At what stage in my research should I seek statistical advice?
• I want to ask if my experimental design is sound before collecting data. Should I come to Walk-In or Request a Meeting?
• What should I bring to the first collaboration meeting?
• Where do I meet my statistical collaborators when I come for assistance?
• Who are the statistical collaborators?
• What are your policies?
• What about coauthorships and recognition?
• How is the center funded?
• How is the collaboration center evaluated?
• How do I sign up for a website account?

Services

• Who can use the Laboratory's services?
• Can I use the collaboration service if I'm not affiliated with Virginia Tech?
• Do you offer walk-in service?
• Is statistical advice really free?
• What should I expect to happen at the first meeting?
• How do I make the most of my time?
• How does the 10-hour limit work?
• I would like to include statistical collaborators in a grant. What should I know?
• Can I use the center for assistance on a class project?
• Do you offer telephone service?

General

How do I request statistical assistance for my research?

Submit the on-line request form. Please describe your overall research goals and specific questions on the form so that we may serve you better. You should receive an immediate e-mail receipt of your request and a notification in 2-3 days of the statistical collaborators assigned to your project.

At what stage in my research should I seek statistical advice?

The earlier you do so the better. Few things are more frustrating than learning after the fact that your data are inadequate for valid statistical inference and cannot answer your main research questions. We recommend that you meet with us during the pre-proposal phase of your research.

I want to ask if my experimental design is sound before collecting data. Should I come to Walk-In or Request a Meeting?

Since Walk-In Consulting is intended for short questions requiring less than 30 minutes, it would be better for you to fill out the Request for Collaboration and schedule an appointment. This will allow you more time to fully explain the background of your project, your experimental design, and your intended method for analyzing your data.

What should I bring to the first collaboration meeting?

If you seek our advice on experimental design or planning a survey beforegathering your data--and we recommend that you do--bring a clear idea of the purpose of your research: your overall goals, the research questions you would like to answer, and the hypotheses you want to test. Copies or links to articles and earlier work on similar topics in your discipline would also be helpful. If you are seeking our advice on data analysis or report writing after collecting data, in addition to the above bring a copy of the data gathering protocol, a file containing the cleaned data, and the results of any analyses already performed--including any plots or graphs. We strongly encourage everyone to plot their data!

Where do I meet my statistical collaborators when I come for assistance?

Meetings are typically held in 403-J Hutcheson Hall, 101 Old Security Building, or at the client's office. A map of campus can be found here.

Who are the statistical collaborators?

The collaborators are faculty and students in the Department of Statistics. Current collaborators have profiles on our website so you can put a face to the name before your first meeting. In addition, the entire statistics faculty may be available for collaboration on a case-by-case basis.

What are your policies?

We expect clients to come to the meetings prepared, and we expect the statistical collaborators to give good advice and explain it in a way the client understands. Please review our policies here.

What about coauthorships and recognition?

If a statistical collaborator has made a substantial and specific intellectual contribution to your research, co-authorship on papers is appropriate. This may include designing an experiment or analyzing data for the paper. Written acknowledgment of your collaborators in your papers or dissertations is appropriate and encouraged if we have helped you think through your research questions/conclusions or suggested an appropriate statistical analysis that you performed. View student co-authorships here.

How is the center funded?

We are funded jointly by the College of Science, the Provost Office, the Office of Research, the Graduate School, and six additional colleges (Agriculture & Life Sciences, Architecture & Urban Studies, Liberal Arts & Human Sciences, Natural Resources & Environment, Business, and Engineering). The Department of Statistics also provides funding for many of the statistical collaborators and provides other support for our activities.

How is the collaboration center evaluated?

At the conclusion of each project, the client will be asked to fill out a short questionnaire evaluating the quality of service he or she experienced. In addition, we are always happy to receive your comments. Please contact statistics@vt.edu to discuss any issues that arise. Quality feedback including constructive criticism is very important for us!

How do I sign up for a website account?

Registering for a website account is required for clients who wish to use the collaboration services. Signing up for a website account is really simple.

Clicking any request collaboration or request form link will prompt you to sign in or sign up for the website. This link will also take you to the sign up page.

This page presents you with 3 choices:

1. Already Registered? Log in here.
2. I have a Virginia Tech PID, but am not registered. Register here.
3. I do NOT have a Virginia Tech PID and need to register. Register here.

Select the option most appropriate to you.

You will then be asked to complete some personal information about yourself (name, address, college, department, etc). Once this information is completed, complete the math captcha to show that you are really a person and not an automated spammer.

After you have saved your personal information you will be directed to a form to complete for a statistical collaboration project. Please complete this form throughly as it will allow us to match you with an appropriate statistical collaboration team.

If you have any trouble registering for a website account, please contact us.

Services

Who can use the Laboratory's services?

Our services can be used by all Virginia Tech faculty, staff, and students participating in academic research projects.

Can I use the collaboration service if I'm not affiliated with Virginia Tech?

Yes. We provide statistical consulting and collaboration services for non-Virginia Tech clients Interested parties should email or call the directorto discuss arrangements.

Do you offer walk-in service?

Yes, we offer the Walk-in Consulting Service. When classes are in session, a statistical consultant is available to answer your quick questions or to help you with your project. Click here for the current schedule. If more than 30 minutes of assistance is required, you may be referred to meet with a team of statistical collaborators. We do not assist with class projects or homework.

Is statistical advice really free?

Yes. We can offer Virginia Tech faculty, staff, and students in participating Colleges up to 10 hours per semester of statistical advice and assistance for free. If additional assistance is needed, alternative options can be arranged. Users who engage in sponsored research are encouraged to include statistical consulting in research proposals. This can take the form of a direct-cost line item, a full or partial graduate research assistantship, or partial funding of a faculty member’s salary.

What should I expect to happen at the first meeting?

Before the initial meeting, the statistical collaborators review the client's collaboration request form and any additional information they have sent them. We have found that the most successful collaborations occur if the client sees a statistical collaborator before they start collecting data. If the client already has collected data, the collaborators often find it useful to quickly review the data to get an idea of what will be discussed during the first meeting. So we encourage clients to send a copy of their data to the collaborators before the initial meeting.

Collaboration meetings typically begin with the collaborators asking the client what they would like to accomplish during the time available for the meeting. This sets the agenda for the meeting. The collaborators will also ask about your overall research goals to get a better sense of how to help you answer your specific research questions. They want to know what the big picture is: why are you completing this research and what will the results be used for. After looking at the big picture, the collaborators will ask about your specific project. They will want to know exactly how your experiment was designed and executed, what your variables and data are, and what you are most interested in finding out from that data. Having a detailed understanding of the why and how of your project allows the collaborators to decide what is the most appropriate statistical method to use to answer your research questions.

At the end of the meeting, the collaborators will summarize what happened during the meeting including any decisions that were made.

The next steps for the project will be discussed. Sometimes, follow-up meetings are scheduled.

How do I make the most of my time?

Below are a few tips for making the most of your time with our statistical collaborators:

1. Know your project and have thought about ways to explain it to individuals outside of your major field of study.
2. Send informative project information to your collaborators before the first meeting.
   1. When sending technical papers, please consider whether or not people outside of your field will be able to understand them.
3. If you already have data, try to have this data cleaned before coming to the first meeting.
   1. Verify that all data is inputted correctly.
   2. Visually scan the data for things that appear out of place.
      1. Are certain values really outliers or was the value entered incorrectly?
   3. Consider what missing values mean and be prepared to explain this to your collaborators.
      1. Does a 0 mean that there was no measured value? Or does it really mean no data were collected?
         1. If no data were collected, is there a specific reason for this? Is the lack of data somehow informative?
4. Be prepared and on time for your meeting.
   1. If your scheduled meeting is from 1:00-2:00pm and you arrive at 1:15, you should still expect your meeting to end at 2:00pm.
   2. Be sure you bring relevant information and data to your meetings.
   3. If you have preformed any preliminary data analysis, bring it to the first meeting.
   4. It can often be extremely beneficial to simply plot your data before performing any statistical analysis. This allows you to visually see any trends and catch any outliers.
   5. Be prepared to explain your project, especially your overall research goals, specific scientific questions, experimental design, and what your data are.
5. If at any point during a meeting you do not understand something the collaborator has explained or asked, ask them to clarify.
6. At the end of a meeting, there will be a discussion about what the next steps in the project are.
   1. If it is agreed that you will send additional information or update your data, please do so as soon as you can. This will ensure that both you and the collaborators are prepared for your next meeting.

How does the 10-hour limit work?

Our collaborators provides 10 hours total of statistical assistance per project per semester. These hours are calculated based on time spent by each statistical collaborator in meetings as well as time outside of meetings performing research or analysis.

For example, let’s consider a client who requests collaboration and sets up a meeting time with her collaborators. They meet for one hour and agree that the client will clean her data and the collaborators will do some research about a particular method the team agreed would be appropriate for the analysis. They arrange a second meeting. The lead collaborator spends 30 minutes researching the topic and the associate also spends 60 minutes on the topic. Thus far 3.5 hours of the time have been used: 2 hours for the initial meeting and 1.5 hours outside the meeting.

For most projects two collaborators will be assigned to work with the client. It is considered discourteous to ask that a collaborator be removed from a project with the goal of receiving additional meeting or analysis time.

Generally most projects are completed before the 10 hour limit; however, if you find yourself nearing the limit, there are a few options. You can tie up loose ends and close the project for the semester. If you need to go over the limit and a collaborator will be putting a large amount of time into your project, it may be appropriate to consider co-authorships. Co-authorships are appropriate if a collaborator has made a significant intellectual contribution to your project. If co-authorship is not an option, our collabortors can provide additional services for a fee. Additional information about our rates can be found here. If you have specific questions about co-authorships or paying for services, please contact the director.

I would like to include statistical collaborators in a grant. What should I know?

Clients who engage in sponsored research are encouraged to include statistical consulting in research proposals or pay for services from existing grants. This can take the form of a direct-cost line item, a full or partial graduate research assistantship, or partial funding of a faculty member's salary. Interested parties should email or call the director to discuss arrangements.

Can I use the center for assistance on a class project?

No. Assistance from the collaboration and consulting laboratory is for academic research. We cannot assist with class projects or homework.

Do you offer telephone service?

Telephones? No. Skype (VOIP), Yes! If you are outside the Blacksburg area, we can arrange a videoconference meeting with you over Skype. Submit a request for a meeting via Skype. Skype is free to use and download, and is very convenient, especially if your computer has video capabilities.