Scholar’s Datapoints

What can you do with Scholar’s Million Datapoints? Every datapoint is a teachable moment:

  • Feedback is always available to the student.
  • Feedback is focused on improving understanding and extending knowledge.

But, to keep it simple – use only datapoints as needed:

  • Feedback the student needs now to improve, or the teacher to adapt instruction.
  • Feedback the student or teacher asks for.

So, having an answer always at hand to “How am I doing?” through:

  • Examining retrospective progress visualizations that are always accessible.
  • Finding out by drilling down into each datapoint.
  • Teachers and administrators viewing an entire cohort.

Some Data are Just Data ... (Not all feedback is helpful for learning)

Evidence-of-learning in traditional assessments is often not very meaningful:

  • A grade. (But how to improve?)
  • A cohort-normed score. (But how am I progressing?)
  • Item-based tests that only offer a total score. (But what were the answers, what do I need to learn?)
  • Automated essay assessment that provides a rating. (But does it measure things that humans would consider useful?)

Let’s make it our agenda to replace traditional assessments with big educational data!

New Measures of Learning

Quietly in the background, while students do their work, Scholar swings into action with some revolutionary “semantic web” technologies:

  • Rubric-based review – a crowdsourced mashup of self, peer, expert; both qualitative and quantitative; inter-rater reliability weighting

 

  • Natural language processing – coded “change suggestions”; language analyses; short answer analyses; argument and info diagramming.

 

  • Semantic parsing – information/argument; tagging and diagramming with hybrid taxonomy/folksonomy.

  • Data mining student actions and interactions – task application; navigation path analyses; user behavior analyses; network analyses; content analyses; “affect” meter.
  • Survey psychometrics – information and knowledge surveys; computer adaptive testing; diagnostic testing; crowdsourced item evaluation, curation and validation.

 

  • Data analytics – a mix of machine assessment and crowdsourced human assessment, using artificial intelligence methods so the system becomes smarter as more data is collected.

 

 

  • Data presentation – elegant infographics provide an overview and alerts to special issues.