search

dlab-trainings / social-data-carpentry-2015

Inaugural hackathon for Social Software Carpentry
Other
6 stars 1 forks source link

Data skills training for freshman in social sciences #9

Open tracykteal opened 8 years ago

tracykteal commented 8 years ago

subtitled 'Data with a hug'

From Victoria Cross

" The course that I am designing is aimed far lower than datacarpentry or your data studies course. It is listed as PSC012Y and designed for freshmen. We are imagining that first quarter freshmen will enroll in PSC001 and PSC012Y (or CMN001 and CMN012Y, etc.) It is data with a hug. It is getting their feet wet with hands-on, straight forward summarizing and visualizing. It is aimed at understanding what decisions need to be made when tackling a data set, what the options are and how to decide between them. The hope being that they will then be prepared to understand the data presented to them in future social science courses.

That said, many of the skills are fundamental and relevant to more advanced audiences. Though perhaps too elementary.
We spend a lot of time having the students identify variables and indicate whether they are continuous or categorical (it's amazing how hard they make this!) We cover basic data summarizing - distributions, percent, percent change, means, standard error of mean, standard deviation, correlations We cover basic data visualization - graphs and tables and the decisions that need to be made (when to use lines, when to use bars, deciding which predictor variable goes in the key and which on the axis) We cover basic Excel skills in the context of data sets - I am amazed at how little Excel they know. I'm attaching a draft of one of the first tutorials. Once I have Excel2015, I will be recording this tutorial as a screencast.

The students produce write ups for each data set.  
We introduce the idea of designing research that puts two theories head-to-head and interpreting the data to decide which theory is supported.  Once the students have covered the skills with a practice data set, we provide them with a 'put it into practice' exercise where they need to make the decisions themselves.  I'm attaching the current draft of the one based on Harry Harlow's work.  Note that the Excel file has two sheets - a simple one and then a more complex one.

PSC012Y is mostly using Excel. Though we are also using Marco's SeeIt! for distributions, correlations and sampling. http://seeit.ucdavis.edu/SeeIt3/distributionsv3.html "

tracykteal commented 8 years ago

Thinking of something specifically for freshman is a great idea. The starting place for data skills is learning to think about and evaluate data, and prior to college most students haven't much experience or exposure to this.

Even more than knowing how to do data analysis, knowing how to critically think about data that is presented is, and will increasingly be, a critical skill to be an informed citizen. This type of curriculum seems to make sense from a practical skills perspective and as a part of a liberal arts education.