Discussion: View Thread

Dear Members of the Section on Statistical Consulting (CNSL):

Two months ago I sent out a survey to see what topics you thought should be taught in a course on statistical consulting: https://virginiatech.qualtrics.com/SE/?SID=SV_3wPA11EM7dvIrop

Thank you to all who responded!

From a list of 41 randomly ordered topics to teach in "Statistical Consulting," 57 complete responses ranked topics on a 5-3-2-1-0 point scale as to their importance to teach in a 2-semester course. Then considering the importance of teaching the topics ranked "5-Must Have!" or "3-Nice to Have" in a 1-semester course, respondents re-ranked the topics on a 10-5-1-0-0 point scale.

I sent the same survey to the ASA Section on Statistical Education (StatEd), from which 45 respondents completed the rankings. The scores for each topic were surprisingly similar for CNSL and StatEd, with correlation greater than 0.99. I have combined the results from both groups of respondents. See the attached document for the ranks and point totals of all 41 topics.

The top seven topics to teach in a course on "Statistical Consulting" were:

1. 1. Presenting statistical results and interpretations in writing
2. 2. Explaining statistics to non-statisticians
3. 3. Presenting statistical results and interpretations orally
4. 4. Asking the client good questions
5. 5. How to communicate statistics to different audiences
6. 6. Statistical ethics
7. 7. Practical experience with a real client to solve a research or business problem

I have taught "Communication in Statistical Collaboration" four times before this semester and never focused on statistical ethics except in the context of negotiating co-authorships with clients. This semester, because of this survey, I have added "Statistical ethics" to my syllabus and will teach it in a few weeks.

The bottom nine topics were:

1. 33. Using video to improve statistical consulting skills
2. 34. Study skills and time management skills
3. 35. Basics of accounting, client management, billing, etc.
4. 36. Finding consulting clients
5. 37. Job interview skills
6. 38. Identifying personal and professional goals
7. 39. Networking and job seeking skills
8. 40. CV/resume writing
9. 41. Creating a webpage to advertise your statistical services

In my class, "Communication in Statistical Collaborations," we use video of real statistical collaboration meetings (as well as Janice Derr's fake meeting videos) to improve our statistical collaboration skills. Even though this topic was poorly rated, it's one of the best things that we do. Students love watching, commenting on, and learning from 1-3 minute clips of real meetings. Even better would be to video record the students themselves in real meetings, but I have 52 students this semester and haven't yet figured out the logistics to do that. However, because it is such a valuable experience to see oneself (and collect and analyze data on yourself) performing one's job on video, we record-and review in a small group-at least one meeting every semester for every LISA (Virginia Tech's Laboratory for Interdisciplinary Statistical Analysis) statistical collaborator who regularly meets with clients. Since fall of 2010 we have done this 138 times and review 1-2 new videos every week.

Let me know if you have any questions about the survey or its results, and feel free to comment on this thread about the attached results.

Sincerely,
Eric Vance

Dear Members of the Section on Statistical Consulting (CNSL):

Two months ago I sent out a survey to see what topics you thought should be taught in a course on statistical consulting: https://virginiatech.qualtrics.com/SE/?SID=SV_3wPA11EM7dvIrop

Thank you to all who responded!

From a list of 41 randomly ordered topics to teach in "Statistical Consulting," 57 complete responses ranked topics on a 5-3-2-1-0 point scale as to their importance to teach in a 2-semester course. Then considering the importance of teaching the topics ranked "5-Must Have!" or "3-Nice to Have" in a 1-semester course, respondents re-ranked the topics on a 10-5-1-0-0 point scale.

I sent the same survey to the ASA Section on Statistical Education (StatEd), from which 45 respondents completed the rankings. The scores for each topic were surprisingly similar for CNSL and StatEd, with correlation greater than 0.99. I have combined the results from both groups of respondents. See the attached document for the ranks and point totals of all 41 topics.

The top seven topics to teach in a course on "Statistical Consulting" were:

1. 1. Presenting statistical results and interpretations in writing
2. 2. Explaining statistics to non-statisticians
3. 3. Presenting statistical results and interpretations orally
4. 4. Asking the client good questions
5. 5. How to communicate statistics to different audiences
6. 6. Statistical ethics
7. 7. Practical experience with a real client to solve a research or business problem

I have taught "Communication in Statistical Collaboration" four times before this semester and never focused on statistical ethics except in the context of negotiating co-authorships with clients. This semester, because of this survey, I have added "Statistical ethics" to my syllabus and will teach it in a few weeks.

The bottom nine topics were:

1. 33. Using video to improve statistical consulting skills
2. 34. Study skills and time management skills
3. 35. Basics of accounting, client management, billing, etc.
4. 36. Finding consulting clients
5. 37. Job interview skills
6. 38. Identifying personal and professional goals
7. 39. Networking and job seeking skills
8. 40. CV/resume writing
9. 41. Creating a webpage to advertise your statistical services

In my class, "Communication in Statistical Collaborations," we use video of real statistical collaboration meetings (as well as Janice Derr's fake meeting videos) to improve our statistical collaboration skills. Even though this topic was poorly rated, it's one of the best things that we do. Students love watching, commenting on, and learning from 1-3 minute clips of real meetings. Even better would be to video record the students themselves in real meetings, but I have 52 students this semester and haven't yet figured out the logistics to do that. However, because it is such a valuable experience to see oneself (and collect and analyze data on yourself) performing one's job on video, we record-and review in a small group-at least one meeting every semester for every LISA (Virginia Tech's Laboratory for Interdisciplinary Statistical Analysis) statistical collaborator who regularly meets with clients. Since fall of 2010 we have done this 138 times and review 1-2 new videos every week.

Let me know if you have any questions about the survey or its results, and feel free to comment on this thread about the attached results.

Sincerely,
Eric Vance