How quickly (if ever) do students read email?

[Mark Stellmack]

Seemingly, the easiest way for me to contact all of my students outside of class is to send them an email.  I send them emails occasionally during the semester to give them information that cannot wait until the next lecture.  The emails will contain important information about assignments or changes to the course schedule.  After I send an email, I proceed as if all of the students have read the message and, accordingly, I may not repeat the content of the message in lectures.  However, the assumption that all students read emails from me in a timely fashion, if ever, may be incorrect.

To test this assumption, I sent an email to all of the students in one of my classes asking them simply to reply (to click on "reply" and "send") as soon as they received the message.  I merely said that I was testing a new way of contacting students.

On Wednesday, November 13, 2013 at 2:45 p.m. (about 2 weeks before Thanksgiving), I sent the email to the 350 students in my undergraduate statistics class.  Below is the number (and percentage) of students that replied within different time intervals.  The numbers in each row from left to right are:

- end time of current time window (or time elapsed after 1 day)
- number of students that replied since end of previous time window
- (percent of total)
- cumulative number of students that replied up to that time
- (percent of total)

message sent at 2:45 p.m.
3:00 p.m.     18 (5%)         18 (5%)
4:00             86 (25)         104 (30)
5:00             51 (14)         155 (44)
6:00             34 (10)         189 (54)
7:00             22 (6)           211 (60)
8:00             10 (3)           221 (63)
9:00             19 (5)           240 (68)
10:00           14 (5)           254 (73)
11:00             8 (2)           262 (75)
12:00 a.m.     7 (2)           269 (77)
9:00 a.m.       4 (1)           273 (78)
10:00             3 (0.8)        276 (79)
11:00             1 (0.3)        277 (79)
12:00 p.m.     3 (0.8)        280 (80)
1:00              2 (0.6)        282 (81)
2:00              3 (0.8)        285 (81)
3:00 (1 day)   0 (0)           285 (81)

2 days         12 (3)          297 (85)
3 days           2 (0.6)       299 (85)
4 days           2 (0.6)       301 (86)
5 days           4 (1)          305 (87)

Thus, 60% of the students had responded within about 4 hours of the message being sent.  81% responded within one day.

The next lecture was about 2 days after the message was sent, at which time 85% of the students had responded.

I received the last reply 5 days after the message was sent.  By that time, two lectures had occurred since the message was sent.  87% of students (305) had replied.

13% of the students (45) never replied.

Did students who replied quickest do the best in the course?  I computed the correlation between "time between original message and reply (in minutes)" and "total points accumulated in the course" for the 296 students who replied and who took all the exams in the course.  Pearson r(294) = -0.11, p = .061, indicating that although students who replied quicker tended to do better in the course, the correlation was not significant.  The same was true for the Spearman rank-order correlation.

Did students who replied at any time do better than students who never replied?  The mean total points in the course for the "replied" group was M = 557.89, SD = 71.26, N = 296.  For the "never replied" group, M = 511.87, SD = 76.71, N = 34.  The sample means were significantly different, t(328) = 3.54, p < .001.

Thus, although there appears to be little or no relationship between how long it took a student to reply to an email and performance in the course, there does appear to be a difference between the performance of students who replied to the email and those who did not.  Perhaps more importantly, 15% of students did not reply to the email before the next lecture and 13% never replied, indicating the extent to which email falls short as a means of communicating with students.

(Thanks to Sarah Bentley for helping me compile the data for this analysis.)

[Patricia Frazier]

Very interesting data, Mark!