Questionnaire subcommittee meeting 5/22

[rebexxxxxx]

In this meeting we discussed some new graphics that RS proposed for usage on the SHMDB. These graphics can be reviewed in the attached pdf. Below is a combination of notes taken by RS and JP
data_visualization_db.pdf

General Comments

1. Is there a way to select date ranges/version number for the completion of a module? this would allow us to compare completion by module version

• Modules changed at certain dates, so it would be helpful to view the completion times and ratios by module version
• This could either be done by using 2 different plots, or allowing the user to select a date in a filter

2. Where do people drop out of surveys?

• Are there certain questions that people stop on
• Could look at which question participants stopped on by putting the questions in order
• Break off analysis -> look at subset that did not complete the module, why didnt they complete it? The non completers are the most interesting group
• Keep skip logic present in this analysis, look at the core questions or required questions

3. Does the length it takes to complete one survey predict the likelihood to complete another survey or take a long time on the next? (i.e. pr(time to complete module X | module Y completion time)
4. General idea of measuring "willingness" to complete surveys based on the time to complete each and demographic info.
5. General Question idea-- think about how to answer something like this graphically-- Of the individuals who completed module 1 in 10 minutes, what are the number of non-skips/skipped questions. The number of skips should only count the "core questions", since not everyone is presented with the same questions due to their answers to previous questions.

Survival Curves

1. Instead of survival curves for the module time completions, a histogram of completion times would be more helpful

• Want to be able to answer, "how long does it take the majority of participants to complete X module"

Sankey Diagram

1. Very positive overall towards sankey diagrams
2. Split sankey diagram into 2 diagrams-- before and after switch of module order from [1,2,3,4] to [1,4,3,2]

• Will need to obtain the date that this happened