1) What do you mean by 'innovative'?
2) What do you mean by 'interactive'?
3) What is meant by 'above and beyond'?
4) How much data analysis does the display need to incorporate?
5) Can we combine data from multiple tables/graphs in the motivating report?
6) Do we have to use everything contained in the table?
7) There’s a variable we’d like, but it isn’t in the data set!
8) What if we change our mind about the data set?
9) What are the weights in the data sets?
10) Are we required to use R, Tableau, SAS, ArcGIS, Data-Driven Documents JavaScript Library - D3, or Python?
11) Why do the teams need to be cross-disciplinary?
12) What if I'm having trouble with my team or it needs to change members?
13) So what exactly do we need to submit?
14) How do we submit it?
15) What if I have a different question?
1) What do you mean by 'innovative'?
Your interactive graphic doesn't need to be something of a type that's never been seen before, but the display needs to move well beyond the static histogram or table of numbers that have been a mainstay of classic statistics texts and educational reports. The big questions are, will seeing the graphic push the viewer to: (i) Interact More and (ii) Click on the Article. (top)
2) What do you mean by 'interactive'?
The graphics in the current report are static. Even if a variety of questions suggest themselves, there is no way to pursue them. Your graphic should allow the user to view what happens when some other variables (predictor and/or response) are taken into account. It also needs to either show the measures of uncertainty or make it possible to see them. (top)
3) What is meant by 'above and beyond'?
“Above” means “better than a 2-D histogram,” certainly more eye-catching, preferably more informative as well. “Beyond” means “interacting displays new information” to follow a logically sequenced inquiry that yields a new insight of substantive interest and relevance (see requirement for explanation of reasoning). Think about subsets, about relationships among variables and/or covariates, about higher dimensional representations, and definitely about uncertainty (variation, random or otherwise).
4) How much data analysis does the display need to incorporate?
The goal of the graphic is to convey more information about the substantive area of the report and give more insghts. It is not to re-analyze the data (in terms of model fitting, estimation, or testing). The goal is to help the reader explore the data to discover what important stories it has to tell. (top)
5) Can we combine data from multiple tables/graphs in the motivating report?
Yes! In many cases the interesting questions the report is trying to study are found across multiple tables or graphs, or involve variables collected that weren't included in the published displays. (top)
6) Do we have to use everything contained in the table?
Use as much or as little as needed for insight. Some of the reports involve tables that contain a huge number of variables; choose what you need to look more in depth, graphically and subject matter question-wise. (top)
7) There’s a variable we’d like, but it isn’t in the data set!
If you would like to demonstrate something graphically, it is ok to simulate a synthetic variable based on the real data set (maybe converting something discrete to continuous). Be sure to include that you have done so in your report. (top)
8) What if we change our mind about the data set?
If you've decided your first choice isn't what you wanted, or see a new addition that will work better, then
contact Randy Freret at RFreret@NISS.org. We will replace the previous data files in your DropBox. (Rember that a team can propose up to two entries!). (top)
9) What are the weights in the data sets?
The samples used in these studies are generally fairly complex; raw unweighted data will give biased estimates unless adjusted. The weights are used to make estimates from the sample represent the target population, and adjust for the probability individuals are selected and the patterns of non-response. The data sets provided here are accompanied by a weighting variable. (top)
Survey weights can be used in R (https://cran.r-project.org/web/packages/pricesensitivitymeter/vignettes/using-weighted-data.html) , Tableau (https://www.datarevelations.com/resources/working-with-weighted-survey-data/) , SAS, and others. (These links open at sites outside of NISS). (top)
10) Are we required to use R, Tableau, SAS, ArcGIS, Data-Driven Documents JavaScript Library - D3, or Python?
No, entries are not limited to these. If you have another free or commercial software package or platform you are experienced with, that will work too. If you're not sure about a particular one, please contact us. (top)
11) Why do the teams need to be cross-disciplinary?
Statistical accuracy must go hand in hand with subject matter relevance; so partnership is essential. Also, teams benefit - combining expertise in thinking about educational questions and expertise in statistical graphics and computing is the basis for a strong entry! (top)
12) What if I'm having trouble with my team or it needs to change members?
Contact Brian Habing at bhabing@niss.org (top)
13) So what exactly do we need to submit?
* The signed release granting NISS sole rights to your submission.
* Color images (8.5" x 11" in .pdf format) showing the graphic and how it would be used.
* An explanatory text (8.5" x 11" in .pdf format) in the style of what would be found in the motivating report.
* A file with the code used to generate the graphic with any needed instructions
You may also include a url (such as to a Shiny) and/or an .mp4 video of you (or your team) demonstrating the graphic. (top)
14) How do we submit it?
The final submission must be placed in the “Submission” folder inside your team's DropBox. If you’re having trouble with DropBox, contact Randy Freret at RFreret@NISS.org. (top)
15) What if I have a different question?
If this FAQ did not help, then contact Brian Habing at bhabing@niss.org. (top)