In our final activity (3), we will simulate an actual research task that you might do in grad school. Imagine you have a faculty research advisor, in your future grad school endeavors, with whom you collaborate. (If it's convenient, Maryam or I can play this role in your mind, but of course your advisor may be some other folks in your exciting futures!)

Now, like most advisors, she or he may sometimes email you about some data that you've collected. You're hoping that these data will be part of a research project that you can report at a conference. The advisor might email you about the data and ask "Hey, what did you find about X and Y?" In response, you send a careful email detailing some findings that may be interesting to present at a conference.

So this is what we'll do in this activity: I'll describe some data to you (section 1), I'll then simulate such an email from an advisor (section 2), and you will carry out some analysis using Mode (section 3). Section 4 describes how to write up and hand in this activity.

General point: The relationship of this activity to Modern Everyday Cognition is obvious. The research example itself - in this simulation - might be a kind of crowdsourcing research project we conduct to test some ideas about everyday social experience (some of you are doing this for your project). This activity will also test your hand at building basic SQL on Mode, along with presenting results systematically to communicate to another person (in this case, a "simulated advisor").

So I've put together a simulated data set that will be used in this task. I've tried to make these data distinct from the various projects that you're working on, based on what I've heard. Here is a description of the data:

The data are all in one table.

https://modeanalytics.com/rdale/tables/social_simulated_data

Each row of the table is a survey response from a participant, describing events in their lives on the previous complete day. There are many participants, so there are many rows (so this is a crowdsourced research project). The fields are the following:

ID (the identifier that the participant might enter on GroupSource, distinct for each person)
number of conversations during that day that lasted longer than 1 minute (# of instances)
reported extent of laughter during the day (# of separate instances of audible laughter)
general confidence in one's abilities in academics or work life (1 = low to 7 = high)
general reported emotional happiness (1 = low to 7 = high)
time of day that data was entered as 24-hour (00:00 = 12:00AM to 23:59 = 11:59PM)

Hint: When you are creating a new report, go back to this link and click "Query" again to start a new query (remember: a query is just some SQL to organize and get data out of the table... the report is the graph and other stuff that you use to visualize and present data).

Hi [your name, which Rick now mostly knows for each of you],

Thanks for working on this exciting project for your first-year paper in this exciting master's (and/or Ph.D.) program in social cognition. Woohoo! As we discussed last time, I suggested that we post the data to Mode Analytics for two reasons. One, it makes it really easy to share data with other folks. And second, it lets us do some basic organizing and plotting of our results.

I've given you a couple of SQL demonstrations now and you've got your hands dirty with data, but I'm going to give you a challenge now and let me know what you find. If you find some interesting patterns we can write this up and send it to a conference! I'd be very happy if you led this work as the primary author, and showcased these results to all the attendees.

So, go to this link:

https://modeanalytics.com/rdale/tables/social_simulated_data

You can click "Query" right by the table to get started and building reports.

As you'll remember, we had two main hypotheses. First, we suspect that social interaction episodes (conversations) will predict confidence and emotional well-being. Second, we predict that laughter -- as one part of social interaction -- will also be related to these things, suggesting that laughter is an important component of social support and experience.

First, go to Mode Analytics and make some simple scatterplots with our social variables (interactions, laughter) on the x-axis, and the other variables that we're interested in on the y-axis. Save these figures for me and let me know what you find. This only uses very simple SQL, along with Mode's report and graph tools.

Second, let's do something a bit more complicated. I think it would be interesting to create a bar graph so that we can differentiate more social vs. less social daily experiences more clearly in our graphs. Let's see the average confidence and happiness of people who are high vs. low on social interactions and laughter. To do this, you need to use SQL in a more complicated way, as we discussed. On the x-axis of the bar graph, have bars for low vs. high interactions/laughter, and the height of the bar should show the average happiness/confidence.

That's it! You should just send me a few sentences of description, then I can think about it too. Feel free to send me links to a few Mode reports showing these plots.

Looking forward to seeing what we have!

Rick

Advisors can sometimes be tricky by giving you some important but minimal detail to see what you can do. How cheeky. This happens in grad school all the time. Sometimes your advisor is being sneaky and wants to see what you can do with a bit of basic instructions. Other times your advisor may be really busy (or lacking sleep) and doesn't quite realize that a bit more guidance would help. In any case, you are up for the challenge and you will conquer this!

So, here are some hints to help you, given that this is a simulation and is probably your first time getting an annoying email from a research advisor (maybe not?). Hints:

The first request from the advisor is just a request for some simple plots. All you need is the standard SQL query and then do a report that shows the variables you want on the x and y axes. You could show several of these fields on the same plot (e.g., one field on the x-axis, but several represented on the y-axis with different colors or points). Or, you could do a few reports (several URLs from Mode). This should take you very little time.
The second request is more complicated, and I'm going to give you an "SQL jumble" that essentially contains all the information you need to create a "bin" for laughing (e.g., high/low laughter), as we discussed in class (or, peruse all my reports on Mode to see some examples). This should help for "happiness" and "laughter"... and you can use the same principles for other bar graphs, too:
- avg(confidence) as avgConfidence
  
  GROUP BY laughHighLow
  
  ceiling(laughs/7) as laughHighLow
  
  avg(happiness) as avgHappiness
  
  FROM rdale.social_simulated_data
  
  SELECT

To keep your advisor happy, here are the basic rules for writing it up:

Write 1 sentence of description for each result that the advisor has asked for. For example, the relationship between social interactions and confidence is one result. The relationship between laughter and happiness is a second result. And so on. Include a URL for any such sentence, so your advisor can see the results for her- or himself.

Finally, in one final "text chunk," write at least 5 final sentences of summary. Is there something here? Is there something you'd be happy to report for a conference? Let your research advisor know. She or he will be delighted that you're taking the initiative to suggest something next, such as: "I found something interesting! As you saw above, we find that..."

Yes, yes you may use "I found something interesting" as one of your sentences. :-)

(Bonus: Anytime you have an advisor, remember -- she or he is someone you want to impress. Your advisor recommends you to others. Writes you countless recommendations down the road, and can make you feel awesome by telling you how great you're doing. Why not throw in some magic? Try this: Is there a relationship between the time of day and happiness / confidence? This is not required for this exercise. It's a bonus, a feat of strength, if you will. Your advisor isn't even expecting it... toss it in and they'll smile heartily.)

Hand in a Word document that simulates what your email response might look like to your research advisor. Include links to Mode reports as she or he requests above. Note that you may require several Mode reports (e.g., one for each graph). They are not difficult to create, but each one should have its own URL.

Important note: Make sure you test your URLs to ensure that they show the graph you intend. Be careful not to "write over" one report with your next report (if you want to create a new one, go back and click "Query" on the table again).

Deadline: December 1st, 2015

Final Activity, #3: Simulated Grad School and Mode