Behavior Change Via the Internet

1
Behavior Change Via the Internet

• Matthew L. Porritt
• Alan D. Poling
• Mark Alavosius
• The Agility Lab
• Western Michigan University

2
The Internet

• Has revolutionized communication
• Allows us new ways to communicate
• Most of these are facilitated by permanent
  products of verbal behavior created by internet
  verbal behavior.

3
The Internet

• Listener behavior can occur with multiple people.
• Time and geography do not matter.

4
The Internet

• We can apply
• Social contingencies
• Conditioned reinforcers
• A wide range of stimuli
• Essentially any thing that can be represented on
  the computer screen.

5
Interventions Over The Internet

• Should adhere to our standards of research
• Particularly objective recording of behaviors.
• This can be difficult when behavior is not under
  direct observation.

6
Introduction

• Writing (typing) produces permanent products of
  behavior.
• Behavioral treatments for increasing writing
  productivity have been successful.
• Can we successfully apply contingency management
  techniques previously proven successful to an
  online format to increase writing productivity?

7
Methods

• Participants
• 10 hobby and professional fiction writers from an
  online community.
• Community based around critiquing each others
  work.

8
Methods

• Participants downloaded packets containing
  program information and consent form.
• Packet contained goal-setting worksheet and
  performance contract.
• Goals were of the format x words for x days
  each week.
• Participants mailed or faxed me consent form and
  performance contract.

9
Methods

• Then emailed me with goal and initial manuscript.
  
• I ran a word count and check for relevancy of
  material (not just cut and pasted from
  somewhere).
• Continued emailing me manuscript for one or two
  weeks, allowing for multiple baselines (Phase I).
• Email sent three days prior, and on the day
  contingency management started (Phase II).

10
Methods

• Phase II, contingency management.
• Intervention package
• Crit points awarded within community for
  critiquing others works
• Web site with feedback (click)
• Weekly emails outlining progress
• Emails went out to individuals and group
• Social recognition
• In email to group if all goals met for the week
• In email to entire community (4000) if all goals
  met for three weeks
• Bonus Allowed to submit a manuscript for
  immediate critique rather than wait in a
  four-week queue.
• Penalty Removed crit points for goals missed

11
Method

• Changing goals.
• I suggested goal increases every two weeks with
  all goals met.
• Participants sometimes asked for goal increases.
• Increase in goal also increased crits earned.
• Phase II lasted six weeks.

12
Participant One
Started writing to goal level before phase II.
400/4
300/4
300/3
Off
300/4
250/4
Words written per submission by treatment day.
Data points are connected by weeks. First phase
change line represents change from phase 1 to
phase 2, successive phase change lines represent
changes in goal. Horizontal lines reflect goal
within each phase. Goals are listed as number of
words per submission over number of days per week.
13
Participant Two
740/4
625/4
500/4
400/4
14
Participant Three
500/4
350/4
15
Participant Four
Professional writer. Wrote a novel during course
of program Publisher interest before last goal
change. Goal was changed per his request, in
order to finish novel for the publisher.
1000/5
1200/5
2500/5
16
Participant Five
Two goals, 4 days during the week and each day on
the weekend. Missed some goals during holidays.
200/4 900/2
200/4 800/2
17
Participant Six
Performed above goal Should have received a goal
change after week four of phase 2. Regression of
words written per submission. Reported lower
satisfaction with his writing productivity.
500/4
300/4
18
Participant Seven
350/4
250/4
500/4
19
Participant Eight
330/4
230/4
20
Participant Nine
Did not visit web page No word count function in
his word processor Rated someone checking up on
him as most valuable treatment component. Rated
goals and feedback not as effective.
1500/3
300/3
21
Participant Ten
Dropped over holidays Came back after
penalty Reported penalty had effect
500/4
400/4
300/4
22
SurveyResults
23
SurveyResults
24
SurveyResults
25
Limitations

• No real baseline phase.
• Was not direct observation cannot tell for sure
  if participants did not cheat.
• I had to program often software would be nice
• More rigorous continued evaluation software
  would be nice

26
Discussion

• Power in computers.
• Participants in seven states, U.K., and Ireland.
• Easy data collection.
• Automatic feedback.
• All interactions were saved.

27
Discussion

• Participants rated social recognition low.
• Little prescribed contact with each other.
• Slim chance of community recognition from email.
• Possibility of contact with me was high.
• Goal statement only shared with me.
• Rated Someone (me) checking up on you as highly
  effective.

28
Future research

• How do social interactions over the Internet
  become reinforcing?
• Find digital permanent products, then apply the
  same methods.
• Make methods available.

29
Thanks

• Dr. Alavosius and the Agility Lab
• Dr. Richard Malott

30
Questions?
Paper is available at www.mattporritt.com Email
me s8porrit_at_wmich.edu