{"id":2072,"date":"2012-08-02T14:18:33","date_gmt":"2012-08-02T14:18:33","guid":{"rendered":"http:\/\/occamstypewriter.org\/vwxynot\/?p=2072"},"modified":"2012-08-02T14:28:50","modified_gmt":"2012-08-02T14:28:50","slug":"whats-in-a-name-towards-the-development-of-a-novel-analysis-of-grant-titles","status":"publish","type":"post","link":"https:\/\/occamstypewriter.org\/vwxynot\/2012\/08\/02\/whats-in-a-name-towards-the-development-of-a-novel-analysis-of-grant-titles\/","title":{"rendered":"What’s in a name?: towards the development of a novel analysis of grant titles"},"content":{"rendered":"

A couple of weeks ago, I blogged about how I’m going to be responsible for helping all trainees in my new department identify, apply for, and manage their awards (comment on my post, you bastards<\/del> please!<\/a>). I’ve spent some time since then exporting all trainee award entries in our comprehensive grants database – which contains every funding application of any kind submitted by anyone affiliated with our department since before 2007 – into an Excel spreadsheet for ease of analysis. While I was engaged in this thrilling<\/del> necessary exercise, I spotted some patterns in the first few words of the grant titles (the field by which all database entries are sorted by default). I therefore spent some time last night (hooray for VPN!) doing a quick-and-dirty analysis* of all database entries (i.e. not just the trainee awards) by status (successful, unsuccessful, pending), and pulled out all title first words that showed up more than once in any category.<\/p>\n

I ignored excessively field-specific words (a lot of our grants start with Genomic \/ Genetic \/ Molecular \/ Personalised etc.), and lumped different forms of the same word together – so, for example, all grants starting with “The evaluation of”, “Evaluation of”, and “Evaluating” are grouped together as “Evaluation”.<\/p>\n

Here<\/p>\n

is<\/p>\n

some<\/p>\n

padding<\/p>\n

text<\/p>\n

so<\/p>\n

people<\/p>\n

who<\/p>\n

want<\/p>\n

to<\/p>\n

guess<\/p>\n

which<\/p>\n

title<\/p>\n

first<\/p>\n

words<\/p>\n

are<\/p>\n

the<\/p>\n

most<\/p>\n

common<\/p>\n

don’t<\/p>\n

accidentally<\/p>\n

see<\/p>\n

the<\/p>\n

graph<\/p>\n

before<\/p>\n

they’re<\/p>\n

ready.<\/p>\n

NB there are ~25% more unsuccessful than successful grants in the database**, with a much smaller number of pending applications.<\/p>\n

 <\/p>\n

\"grant<\/a><\/p>\n

Some unique entries that caught my eye:<\/p>\n