An Early Alert System Will Fail Without an Early Alert Philosophy

Higher Education

Con­grat­u­la­tions to the two schools I spoke to over this week­end in their deci­sion to rec­om­mend again­st invest­ing in an ear­ly alert sys­tem before they were ready.

Ear­ly Alert sys­tems are designed to high­light stu­dents who are at high-risk of neg­a­tive behav­ior, like strug­gling, with­draw­ing from a course, or worse, drop­ping out of col­lege entire­ly. Hun­dreds of com­pa­nies world­wide have been offer­ing sys­tems to track and tar­get at-risk stu­dents in every type of edu­ca­tion­al envi­ron­ment, but for some rea­son high­er edu­ca­tion is the most prof­itable one. Any idea why that is?

May­be it’s because high­er edu­ca­tion is con­stant­ly look­ing for mon­ey to spend, even though what it should spend mon­ey on is right under its nose (hint: it’s stu­dents). Ear­ly Alert sys­tems are often seen as the mag­ic ingre­di­ent that your col­lege has been miss­ing. At last, a piece of soft­ware that will get more of our stu­dents com­plet­ing degrees and cer­tifi­cates faster than ever before!

Eric McIn­tosh of Civ­i­tas Learn­ing Sys­tems sums up how a lot of admin­is­tra­tors feel when it comes to the poten­tial of ear­ly alert sys­tems on their cam­pus:

Knowl­edge of which stu­dents need which sup­ports, and when they need those sup­ports, would be a great way to ensure every stu­dent has access to the sup­port they need, when they need it. Unfor­tu­nate­ly, it is has been dif­fi­cult to know what inter­ven­tion – or inspi­ra­tion – each stu­dent needs, and it has been dif­fi­cult to know exact­ly when they need the out­reach. For­tu­nate­ly, pre­dic­tive mod­el­ing affords a more nuanced and focused per­spec­tive on stu­dent engage­ment and risk, and let’s us be more pre­cise in our efforts to sup­port every student’s jour­ney.

Yes! Data Sci­ence in High­er Edu­ca­tion for the win!

Except there’s a flaw with Ear­ly Alert sys­tems, and that flaw has noth­ing to do with the sys­tems them­selves. EAB, Pharos, Civ­i­tas, Hobson’s Starfish, and more are doing a great job build­ing great soft­ware to do great things. I’ve worked with many tools over the years (even though I’m a huge advo­cate of build­ing your own) and I’m very impressed with what they can do with the right data. The prob­lem is that insti­tu­tions are buy­ing and imple­ment­ing the­se tools before they need them.

It’s the cart before the horse, folks. We’re buy­ing tools to get ear­ly alerts on stu­dent dropout behav­ior before we have a plan to col­lect the data we need to plug into the sys­tem! Why is that? May­be it has some­thing to do with the Law of Triv­i­al­i­ty, where a group of peo­ple will tend to spend the most amount of time on the most triv­ial of prob­lems in a given prob­lem set. That is, may­be we want to spend all our time play­ing with the tools because we are afraid to admit that we’re not ready to use them.

To be fair, an old col­league who now works at Deloit­te said that there are good peo­ple at the­se third-par­ty ven­dors that try to use the demos of their ear­ly alert tools as a way of show­ing col­leges just what is pos­si­ble if you are col­lect­ing the right data. If you’re not, though, it’s very easy to be swept away with what you could be doing (and enter into a huge annu­al con­tract), instead of focus­ing on what you should be doing.

We need an Ear­ly Alert Phi­los­o­phy. Before we spend any mon­ey, let’s spend some time think­ing about what data we need, how we are col­lect­ing it, and most impor­tant­ly, what we will do with the infor­ma­tion about at-risk stu­dents. If we don’t have at least a phi­los­o­phy about what we will be doing with our ear­ly alert data, then we can­not pos­si­bly ben­e­fit from invest­ing in the tools that gen­er­ates that data for us.

The foun­da­tion­al mind­set need­ed to cap­i­tal­ize on an ear­ly alert sys­tem can be forged by fol­low­ing my free check­list for Ana­lyt­ics Readi­ness. Spend some time with your col­lege lead­ers and the pro­fes­sion­als in stu­dent ser­vices to come up with a solid plan of what you will do with ear­ly alert infor­ma­tion. Ear­ly alert sys­tems should aug­ment exist­ing process­es, not cre­ate them. 

In sum­ma­ry:

  • Don’t invest in Ear­ly Alert soft­ware unless you are pre­pared to take action with what it pro­duces
  • Good intel­li­gence is only as use­ful as the knowl­edge it cre­ates
  • Knowl­edge is only as use­ful as the action is it inspires

Leave a Reply