Lunch & Learn: Jeffrey Himpele on Coursera

Princeton courses on Coursera

Prince­ton courses on Coursera

Jef­frey Him­pele, who directs the McGraw Center’s Grad­u­ate Teach­ing Tran­script Pro­gram, leads ped­a­gogy work­shops and indi­vid­ual teach­ing con­sul­ta­tions for grad­u­ate stu­dents & fac­ulty (http://www.princeton.edu/mcgraw/about/staff/), and directs Princeton’s pro­gram of online courses offered on Coursera.org. He spoke to the Lunch & Learn audi­ence on Novem­ber 7th, 2012 about the Cours­era Project and Princeton’s involvement.

One of the most notable things about Cours­era is its large adop­tion num­bers, despite less than a year in use. Almost two mil­lion stu­dents have enrolled in Cours­era, which offers free online access to classes, includ­ing lec­tures, read­ings, course con­tent and assess­ments. At the time of this writ­ing, in Novem­ber of 2012, there are 1,822,099 par­tic­i­pants, called Cours­e­ri­ans, doc­u­mented in a num­ber that appears at the top of the home page at http://coursera.org and updated every few sec­onds. Prince­ton began offer­ing 9 courses in Cours­era taught by 7 fac­ulty in May, 2012. Stan­ford pro­fes­sors Daphne Koller and Andrew Ng launched the ser­vice in 2012 and now there are 34 insti­tu­tions (https://www.coursera.org/universities). There are 203 courses as of today in Cours­era (https://www.coursera.org/courses).

Him­pele explained the acronym MOOC, which stands for Mas­sive Online Open Course, as one of the ways peo­ple define Cours­era. It is mas­sive indeed. In Princeton’s courses offered in Cours­era, There are 94, 835 stu­dents enrolled in the most pop­u­lar Prince­ton offer­ing to date, a Sta­tis­tics I course (https://www.coursera.org/course/stats1). Him­pele noted that the num­ber of peo­ple who enroll is often much dif­fer­ent from the num­ber of peo­ple who actively par­tic­i­pate, e.g. take exams, write papers, watch videos, and so on. For exam­ple, in a Soci­ol­ogy course taught by Princeton’s Mitchell Duneier (https://www.coursera.org/course/soc101), over 37,000 peo­ple enrolled, but only about 2,000 took the midterm exam, and about 1,000 took the final exam.

Him­pele stressed that although the num­bers were inter­est­ing, that they were not the most impor­tant part of why Prince­ton uses the plat­form. Him­pele, in his role at the McGraw Cen­ter (http://www.princeton.edu/mcgraw/) in help­ing fac­ulty to explore new ways of teach­ing, is using Cours­era to change the way teach­ers teach and think about their courses and stu­dents, e.g. with more of a student-centered approach, and more inter­ac­tion between stu­dents, their peers, and fac­ulty. One ques­tion that Him­pele raised with Cours­era fac­ulty was what the impli­ca­tions might be for using this plat­form to extend teach­ing and learn­ing. He calls the project (and Princeton’s involve­ment in it) an exper­i­ment. He also stressed that key prin­ci­ples of ped­a­gogy and schol­ar­ship (includ­ing evi­dence, logic, and rigor) are employed in this plat­form as in any other set­ting. He also asked fac­ulty what kind of teach­ing is most effec­tive for dif­fer­ent sit­u­a­tions, with dif­fer­ent audi­ences, and at dif­fer­ent lev­els of stu­dent understanding.

Him­pele talked about the ways in which fac­ulty used Cours­era to flip lec­tures, one kind of active learn­ing, in which stu­dents use asyn­chro­nous time (where fac­ulty and stu­dents are apart) to watch pre-recorded lec­tures and other learn­ing objects, and then use syn­chro­nous time (where the class comes together) to go deeper into the inquiry and explo­ration of the lec­ture top­ics. In other words, flip­ping sug­gests that fac­ulty show con­tent out­side of class, and have active learn­ing, think­ing and dis­cus­sion hap­pen inside the class­room. Cours­era talks about the poten­tial ben­e­fits of active learn­ing over tra­di­tional lecture:

“(A)n exper­i­ment in an intro­duc­tory physics class (com­pares) a tra­di­tional lec­ture set­ting to one that uses active learn­ing. In the active learn­ing group, stu­dent engage­ment nearly dou­bled, atten­dance increased by 20%, and aver­age scores on the same test increased from 41% to 74% (where ran­dom guess­ing would give a score of 23%).” (https://www.coursera.org/about/pedagogy)

The video lec­tures for Prince­ton based Cours­era courses are often done in the Broad­cast Cen­ter, which has resulted in an espe­cially high level of con­tent qual­ity. In Cours­era, videos of lec­tures are fol­lowed by a quiz to assess stu­dent learn­ing and out­comes.  Cours­era assess­ments allow for feed­back that can assess stu­dent under­stand­ing, then quickly cor­rect mis­un­der­stand­ing and  rein­force what is actu­ally cor­rect. Com­pare this to a typ­i­cal exam where a stu­dent leaves the test think­ing that they are cor­rect in their answers for a time between meet­ings, just to find out that they were wrong some time later, pos­si­bly let­ting the wrong idea incu­bate and take hold. According to Him­pele, Cours­era allows many forms of exams and assign­ments, includ­ing self-graded prob­lem sets and machine-graded assignments.

One ques­tion that peo­ple have about courses with 80,000 stu­dents: When they all sub­mit a paper, who grades it? In Cours­era, one answer is peer grad­ing. Stu­dents sub­mit assign­ments, then the assign­ments are re-distributed and graded by fel­low stu­dents. Rubrics are often used to align the val­ues and cri­te­ria of graders. Him­pele says that peer grad­ing helps to reduce the poten­tial bias that fac­ulty might oth­er­wise have for typ­i­cally high-performance stu­dents. Cours­era says the fol­low­ing about peer assessment:

“This tech­nol­ogy draws on two bod­ies of lit­er­a­ture: First, the edu­ca­tion lit­er­a­ture on peer assess­ments. Fol­low­ing the lit­er­a­ture on stu­dent peer reviews, we have devel­oped a process in which stu­dents are first trained using a grad­ing rubric to grade other assess­ments. This has been shown to result in accu­rate feed­back to other stu­dents, and also pro­vide a valu­able learn­ing expe­ri­ence for the stu­dents doing the grad­ing. Sec­ond, we draw on ideas from the lit­er­a­ture on crowd-sourcing, which stud­ies how one can take many rat­ings (of vary­ing degrees of reli­a­bil­ity) and com­bine them to obtain a highly accu­rate score. Using such algo­rithms, we expect that by hav­ing mul­ti­ple stu­dents grade each home­work, we will be able to obtain grad­ing accu­racy com­pa­ra­ble or even supe­rior to that pro­vided by a sin­gle teach­ing assis­tant.” (https://www.coursera.org/about/pedagogy)

Forums exist on the Cours­era course web­sites to allow for exchanges of ideas, ques­tions, answers, assess­ment, and feed­back. Study groups often spon­ta­neously emerge using the forums, and stu­dents use the forums to answer ques­tions of other stu­dents. Some fac­ulty cre­ate com­mu­nity teach­ing assis­tants, often a reward for more active, pro­gres­sive stu­dents, accord­ing to Himpele.

For more infor­ma­tion about Princeton’s involve­ment with the Cours­era project, please con­tact Jef­frey Him­pele at 

This entry was posted in Educational Technology, Lunch & Learn. Bookmark the permalink.