It’s a snowy weekend as I draft this with a cup of hot cocoa in hand I went over my survey of 385 young librarians before, as well as critiqued my own methods. In this post, I’ll try working with Excel data. Please bear in mind that a) I’m blogging my own learning curve, and would love your feedback, and b) I got responses from only 1% of the 40,000+ MLS graduates in the past ten years alone. This all points to a need for more discussion, and more people exploring the data
Working in Excel: Cleaning the Data
I’m going to work with the 2014 Recent MLS Grads survey, so I hope you’ve reviewed the basic results and survey methods. If you haven’t run numbers in Excel before, do pause and read School of Data’s intro to data in Excel, or better yet, the Data Journalism Handbook.
Should we be running data in Excel? Well, it’s clunky compared to other statistical tools, as Eva Goldwater at UMass points out: it takes a lot of time, doesn’t like missing values, only lets you do simple stuff, and limits how you can re-arrange and visualize data. But on the plus side, you’ve probably used it, and can find a computer that runs it easily.
Descriptive Statistics in Excel: Librarian Salaries
You’ll see below that I asked individuals to report their salary a) before library school, b) during library school, c) 6 months after, d) 2 years after, and e) 5 years after MLS graduation. To get these numbers, I cleaned up punctation from the numeric data. By looking at the raw data, I noted that some responses seem to be typing errors (an annual salary of $430,000? Was that $43,000… or am I working in the wrong library?).
For descriptive statistics, I copied non-$0 salary data on a new page, then ran descriptive statistics on each column (data tab –> descriptive stats). What you see took a half hour of fiddling to produce. I’m suspicious of how outliers may have impacted the results, even using the median, but I like the detailed statistics produced:
Salary before MLS, during studies, and 6 months/2 years/5 years after.
Above, then, we see a median wage of $22,900 before grad school, $18,000 during school, $35,000 six months after graduation, $40,000 two years after, and $47,000 five years after. Note that the ‘count’ of responses drops after two years out, due to the recent grads, and also that these statistics only include those who had a non-zero salary. Including unemployed folk in the totals would make it a bit lower! Continue reading →