Dewey
JUL
ALFRED
P.
WORKING PAPER
SLOAN SCHOOL OF MANAGEMENT
Horizontal Diffusion of Innovations:
An Alternative Paradigm to the
Classical Diffusion Model
Dorothy Leonard-Barton
Massachusetts Institute of Technology
WP
1214-81
and
Everett M. Rogers
Stanford University
MASSACHUSETTS
INSTITUTE OF TECHNOLOGY
50 MEMORIAL DRIVE
CAMBRIDGE, MASSACHUSETTS 02139
im mj
Horizontal Diffusion of Innovations:
An Alternative Paradigm to the
Classical Diffusion Model
Dorothy Leonard-Barton
Massachusetts Institute of Technology
and
,
WP
„,
1214-81
Everett M. Rogers
Stanford University
t/i.iX
LIBRARIES
JUL 15
1981
RECEIVilD
Horizontal Diffusion of Innovations:
An Alternative Paradigm to the Classical
Diffusion Model
Dorothy Leonard-Barton
Massachusetts Institute of Technology
and
Everett M. Rogers
Stanford University
Abstract
In this paper
we present a paradigm of diffusion which contrasts along
several key dimensions with the classical diffusion model as exemplified by
the
Department
U.S.
Agriculture
of
Extension
Service.
The
alternative
model, "horizontal" or "decentralized" diffusion, has these characteristics:
(1)
relatively more user innovation and
innovation dissemination among
Several extant, relatively horizontal diffusion systems are
and by peers.
briefly described
and
(2)
order to highlight the differences between centralized
in
decentralized
diffusion
strategies and
to
issues
raise
for
empirical
investigation.
The Alternatives
The Classical Diffusion Model
decades,
For
planners
both
in
innovation
diffusion
one
public
the
originates
from,
model
private
and
and
is
dominated
has
sector.
legitimized
the
this
In
seme
by,
thinking
of
model,
an
expert
source.
This source then diffuses the innovation as a relatively uniform package to
potential
original
model
adopters
form;
accept
reject
or
adopters are not expected
owes much
Agricultural
who
of
its
Extension
popularity
Service
to
the
to
the
innovation,
modify the
success
of
(described belo^i^"
usually
innovation.
the
ahjd.
United
to
in
its
This
States
commercial"
-2-
marketing
especially
strategies,
as
have
latter
the
migrated
into
the
public sector.
The Horizontal Diffusion Model
recent
In
challenges
years,
this
to
classical model
Schon (1971) was among the
several quarters.
have
come
from
scholars to note that
first
theories of diffusion had "characteristically lagged behind the reality of
emerging
systems."
He
particularly
criticized
diffusion
classical
the
theory, which he named the "center-periphery model" because of the tendency
for
innovations
originate
to
social
change
in
out
from,
centralized
a
While recognizing that this model does apply in many
legitimizing source.
cases, Schon noted
radiate
and
in,
that it
which
fails to capture the complexity of system-wide
innovations
originate
numerous
from
sources
and
evolve as they diffuse.
These
two
characteristics of some
innovation
processes
—
multiple
sources of invention and the changing nature of the innovations as they are
disseminated
(1976)
—
have
also
been
noted
by
other
researchers.
Von
Hippel
found in several industries that users invent the product or process
they need and take
production.
the
Rogers
innovation back to manufacturers or
(1977)
recognized
that
the
suppliers for
"re-invention"
of
innovations may occur as adopters adapt an innovation to their particular
needs
and
innovation
situations.
process
at
Thus,
the
whether
level
of
adopters/users
origination
2
or
take
at
part
the
in
level
the
of
"Social marketing," the promotion of goods and services in the public
family planning,
sector for the benefit of society at large (e.g.,
nuitrition, use of seat belts, energy conservation) exemplifies another
form of the classical diffusion model (see Kotler and Zaltman, 1971).
follow here the classification scheme suggested by Pelz and Munson
(1980) in which they divide the innovation process into levels:
(1) origination (invention), (2) adaptation (re-invention), and (3)
borrowing (adoption without change).
vie
^-
adaptation,
they
diffusion model
much
are
more
active
suggest.
would
participants
Diffusion
can
than
the
often
classical
be,
and
the
United States reflects
is,
very
a
interactive process.
The present vogue of the term "networks"
another
—
trend
clusters
individuals
of
organizations.
around
recognition
the
support
who
relatively
These
of
in
the
hunan
resources
each
other
through
unstructured
systems
single political, social or economic issue.
a
embodied
in
self-generated
tend
grow
to
up
Anthropologist Luther
Gerlach and his colleagues, who studied active social movanents such as the
black pov«r
anti-war movements,
leadership
shifting
(1980)
and
centers
characterizes
among
loosely-connected
movements
these
such movements
find
as
characterized
networks,
segmentary
(many
by
Gerlach
groups),
polycentric (many leaders), and integrated in networks (characteristics he
captures in the acronym SPIN).
Such networks foster
the dissemination of
social and political innovations.
These new research trends all
diffusion
which
both
exists
as
an
substantiate an
unintentional
alternative model
of
effect
mass
of
media
concentration on local sites of innovation (the case of Cavis, California
is
described
below)
and
as
a
deliberate diffusion
system built
around
"horizontal" exchange of information about innovations among peers (as
the
case
of
alternative
the
diffusion
"decentralized"
describe
as
computerized
(the
diffusion
strategy,
terms
are
Legitech
which
system
term
we
interchangeable
characterized
described
ei'-her
for
a
in
This
below).
"horizontal"
or
purposes),
we
our
by the dissemination of innovations
from their originators through peer networks
.
There
are
four
crucial
elements in any diffusion strategy whose characteristics distinguish
different types:
-u-
1.
Who innovates.
2.
Who legitimizes the innovation.
3.
Who diffuses the innovation.
4.
How the innovation is diffused.
there
While
system
built
upon
reality no
in
either
strategy,
it
characterized
by
decentralized
extremes,
are
convenient
almost
or
compare
to
opposite
entire
an
centralized
extreme
an
is
forms of
"pure"
emphases
a
two
on
diffusion
completely
hypothetical
sources
the
innovation, their legitimization, and the diffusion of innovations.
al "0
convenient
illustrate
to
the
differences
between
these
of
It
is
systems
by
examining, first, real life examples which approach the theoretical
extremes,
second,
and
other
diffusion
systems
which
are
nybrids of
centralized/decentralized systems, embodying characteristics of both
extremes.
relative
existing
,
In
short, the degree of centralization or decentralization is a
absolute, characteristic in the real
not an
systems
as
sample
the two extremes (Fig ire
points
hypothesized
along a
We
continuun between
Agricultural
Extension
Service
Centralized
Diffusion
Systems.
Decentralized
Diffusion
Systems
Figixe
A Schematic of the
In
the
systems
according
briefly
v*iere
the
a
1
Centralized/Decentralized Continuum of Diffusion Systems
pages which follow,
discussion by
describe
1).
Davis
Legitech
world.
to
the
four
we
compare
crucial
charac*"eristics differ
a
number
of extant
elements
naned
among
systems.
above,
We
diffusion
indicating
follow this
presentation of issues highlighted by the comparisons, and.
-5-
finally, suggest directions for future research on the topic (Figure 2).
Extremely Centralized
Diffusion System
Extremely Decentralized
Diffusion System
Who
Innovates
R&D laboratories
Users
Who
Legitimizes
Experts
Ihe adoption process
Who
Diffuses
Centralized authority
Users
Through a hierarchical.
structured system
using officially
professional
designated
change agents
Face-to-face, usually
through personal contacts;
the users serve as
unofficial change agents
How the innovation
is diffused
,
Figure 2
Major Points of Comparison Among Diffusion Systems
Who innovates.
Agricultural Extension Services
The U.S.
Ihe
cooperatively
Extension
agricultural
50 state
with
the
U.S.
extension
Department
services
of
in
the
Agriculture's
(USDA)
represent the largest public investment in
Service,
operating
U.S.,
Federal
diffusion
a
system in the U.S. and in the world.
The usual flow of agricultural
state
experiment
agricultural
state
stationed
in
extension
agents,
finally
agencies usually have been
provided
the
new
stations
hybrid
specialized fertilizers
—
R&D
seeds,
to
to
individual
specialists,
the
methods
extension
state
through
universities,
agricultural
and
innovations has been from the USDA and
farmers.
Ri.D's
for
The
viiose
crop
them
specialists
to
county
innovation
scientific
rotation,
and
work
the
all of the innovations which have given American
farmers a deservedly high reputation
for
production efficiency.
Once
the
innovation reaches the level of the individual farmer in this predominately
center-periphery model
of diffusion,
lateral
transfer
of
the
information
-6-
does occur, of course,
through peer networks.
the point is that
However,
almost all of the innovations originate from people who never intend to use
the
products
prolific
design
they
primary
whose
and
source of innovations.
3
Thus,
role
is
agricultural
the
function
to
as
a
extension model
serves as an example of a relatively centralized diffusion system.
Legitech on EIES
In
often
used
originate
or
innovations
contrast,
intend
multiple
from
use
to
in
the
relatively horizontal diffusion system
a
sources,
individuals
from
innovation themselves
and
v*io
either
who choose
have
place
to
themselves in the role of expert by providing information to their peers.
Technological
innovations
"craft knowledge"
from
formal
organization
innovations
and
in
such
decentralized
systems
may
be
derive
from
from exemplary practice by practitioners as well
research and development activities.
v*iose
can
formal
role outside the
insinuated
into
a
While
as
individual or
an
network consists of providing
horizontal
diffusion
systan,
the
system is not dependent upon such officially designated "expert" sources.
been
No pure examples of an
intentionally horizontal diffusion network have
identified
authors,
yet
by the
but
the
potential
for
such
a
system
exists through a recently developed form of conputer conferencing called by
its originators "inquiry networking"
3
(Stevens,
1980).
Recently an interesting incident has proven that even the agricultural
Bob Bergland, U.S.
extension service is not inmune to user innovation.
Secretary of Agriculture under the Carter administration, initiated
research into methods of organic farming when he discovered that a
respected neighbor of his in Minnesota had switched to farming his 1,500
from being the
far
that
The research revealed
acres organically.
"crackpots" that agricultural scientists had tended to label them, organic
farmers were producing crops comparable to those raised with chemical
This discovery led to a reversal of
fertilizers
often at lower costs.
The center,
the U.S. D.A.'s former position discouraging organic farming.
However,
in other words, followed the lead of the periphery in this case.
once the idea of organic farming was accepted by the U.S. D.A., it was
diffused in the usual fashion to farmers from the center.
—
-7-
experimental
Several
Electronic
inquiry networks
have been
System
which
Exchange
Information
(EIES),
set
on
up
numerous
supports
other computer conferences and electronic messaging systems as well.
which
inquiry networks
closest
come
horizontal diffusion are the "Legitech"
technical
and
states,
information
"Localtech"
and
municipalities.
Legitech,
In
potential
the
staff
the
of
for
various
information
similar
wishing
legislator
state
a
of
The
exchange of scientific
for
legislative
exchange
the
for
system
the
among
realizing
to
the
among
\
V
solicit
to
suggestions for solving
a
problem
clean
the
(e.g.,
of hazardous waste
up
./
Q^
dumping
sites)
can
network
computer
out
send
how
out
find
to
Any other legislators,
problem.
inquiry
general
a
National
the
Foundation
Science
over
the
to
the
well as a number of resource groups on
as
innovation groups set up
the teleconferencing system, such as the regional
by
responded
have
states
other
topic
this
on
to
aid
local
of
either
goverrments
in
solving
problems, can respond to the inquiry.
response
The
solution,
may
to
access
or
legislator on
answers.
all
innovative
sources
of
provide
as many
designated
print
the
role
of
ideas.
the
(quite
All
In
system
actual
responses as,
apart
from
or
the
through
field, often
who
Legitech
have
tried
originate
the
users
are
than,
computer
from
supply
an
form of reference to
practice,
more
can
technical
in
the response
therefore
the
px^tential
legislative
staff
agencies
whose
those
network)
a
Legitech
Members of the
system.
state legislatures as sources of innovation.
diffused
which
resources
human
specific
a
did not ask the question but who are interested
v*io
also
form
the
Sometimes the response is in the
originated by a
system
take
reference
a
answer, or both.
bill
may
is
to
serve
the
Thus, many of the innovations
users
—
practitioners
in
the
solutions they suggest and whose bills or
\
''
I
\^
*/
/
J^
/
-8-
mandates
serve
can
models
as
legislatures
other
for
system.
the
on
Similarly, municipal authorities facing some local problem vhich they know
before them
faced
peers may have
their
fleets; riot control)
about municipal
seek
can
lease/purchase decision
the
(e.g.
innovative
solutions
from
other municipalities through Localtech.
examples attains the
Neither of these
hypothetical
horizontal
ideal
diffusion system, in which mayors might exchange innovative ideas directly,
Rarely
on-line.
themselves operate
responses
many
before,
noted
usually
computer;
a
the
or
officers
municipal
Moreover,
staff member does.
innovations diffused originate from people on the network.
not all of the
As
the
legislators
state
the
do
innovation
other
references
to
inquiry.
However,
the
to
sources
come
in
the
form
of
experts
in
the
area
of
inquiry
or
to
exemplify many
systems
inquiry networking
these
characteristics v*iich differ from the traditional center-periphery model of
diffusion, and which are therefore worth considering in some detail.
different
Obviously,
sources
in
centralized
system
may
is
centralized
system
of
than
the
products
Given
the
processes
and
extension
fact
to
the
networks.
horizontal
in
agricultural
government-backed.
discredit
credibility accrue
of
types
that
previously
innovation
the
In
service,
diffusion
the
technological
pushed
by
more
advances
government
experts (e.g., chemicals such as DDT), government sources are not regarded
Nevertheless, receivers in such a diffusion system know the
as infallible.
official credentials of the
person or organization offering
advice.
The
sources are technically credible.
However,
innovation
innovation.
can
weight
the
be
In
a
of
detriment
some
cases,
the
as
the
government
well
as
an
centralized
organization
advantage
in
innovation
behind
an
diffusing
the
source
may
be
-9-
perceived as placing more importance on the government's agenda than on the
In
Therefore
such
credibility
An
the
the centralized
short,
adopter's.
.
innovation source in
opposite
is
strength
a
"like
someone
adoption
from
—
me"
rejection
or
decentralized, horizontal system can possess
weakness,
and
innovation s/he suggests and
the
"safety"
not
but
4
S
source of information because,
she
"competence"
possess
may
source
a
source often has "an axe to grind".
often
is
highly
a
point of view of the
the
soneone
who
has
actually
user,
trusted
he
experienced
or
the
does not stand to benefit personally from
viio
of
peer
A
the
innovation.
peer
A
source
therefore
often has "safety credibility."
However,
peer
a
may
lack
technical
expertise,
both
in
judging
the
innovation and in accurately transmitting the information to the adopters.
The
technical
system
credibility of the
derive
can
from
hands-on
innovator
experience
rather than from official credentials.
more credibility,
suggested
would
the
with
horizontal
the
diffusion
innovation
itself,
The longer term the experience, the
since any deficiencies
presumably have
in
surfaced
in
the
over
practice or
time.
product being
However,
obviously
For instance government agencies advocating birth control in developing
nations are sometimes viewed with some suspicion by potential adopters who
believe small families are more to the government's advantage than to
their own
.
^Berlo and others (1970) found these two dimensions of source credibility
Canpetence
through a factor analysis of semantic differential data.
of the
expertise
technical
perceived
the
based
on
credibility is
the
which
to
extent
the
reflects
credibility
information source; safety
the
those
of
to
comparable
needs
with
source is perceived as being a peer,
information recipient's.
powerful influence of
innovation decisions (Leonard-Barton, 1980; Rogers with
A peer could have reason to push an innovation, or
Shoemaker, 1971).
course as for instance to obtain a critical mass of adopters for some
desired community or group innovation.
There
are
peers on
numerous
studies
which demonstrate
the
-10-
this
test
of
but
adopter
time
only apply
can
which
were
invented
fact
in
innovations
to
which
Presumably most innovations by definition do not fall
Technical
credibility
diffusion systems.
therefore
can
be
a
special
is,
"What
the
to
previously.
time
some
at
new
are
that category.
into
problem
horizontal
in
7
Who Legitimates.
Another
diffusion
question
way of stating
system
over
this
quality
the
relates closely to
agricultural
extension
researchers,
workers
legitimating
source
state
legitimate
process
issue
innovations
of
credibility.
The
innovations
presumably
is
that
the
there
is
U.S. D.A.
they
in
a
This
diffuse?"
state
and
the
county
diffuse.
This
specialists,
extension
the
control
and
culmination
of
scientific
evaluations that have been conducted of the innovation.
In
horizontal diffusion system,
the "pure"
of persons officially legitimates the
adoption
process.
actively accepted
challenged
by
An
peers.
innovation
members
In
of
also
sources have earned
the
in
the
is
the
system
innovation is
(and
In
innovation through the
only
Through
system).
for
person or category
cumulative experience.
legitimized
assign varying
Legitech,
—by
whole legitimizes the
by members
users of the system
their
a
one
innovation; the
legitimized only by its success in diffusing
other words, the system as
no
its
in
so
far
as
legitimacy
accumulated
it
is
is
not
experience,
degrees of source credibility to
instance,
certain
respect of others on the
innovation
system
for
information
careful
and
There are systems composed of networks of highly educated and technically
who
cardio-vascular surgeons, for example
competent practitioners
exchange innovative information.
In one sense, technical credibility is
However, the
not a problem in such systems; all participants are experts.
surgeons do not all possess equal credibility as innovation sources.
—
—
-n-
Among professional associations, some of
compietent responses to inquiries.
function
which
surgeon's
horizontal diffusion
as
members
associations)
incompetent innovation
establish them
as
become
systems
known
but
legitimizers,
innovation
time
over
reputations,
Their
sources.
experts,
for
reliable
as
or
other words,
in
role
this
(e.g.,
unofficial.
is
Moreover, there may be no consensus among members of the network regarding
who are the legitimizers.
Who Diffuses the Innovation and How.
centralized
the
In
system,
raison d'etre is to function as
county
the
a
extension
explicit
worker's
change agent, transferring new technology
to farmers.
There is no expectation that the farmers will adapt, modify or
"re-invent"
the
professional
success
at
change
The
innovation.
partially
least
agent,
on
the
measures
who
of
basis
how
his/her
the
well
innovations s/he promotes diffuse through the potential adopter population,
targets
those
farmers
v*iom
s/he
knows
to
be
opinion
leaders
and
those
farmers to whom the innovation offers the greatest relative advantage over
whatever their current practices are.
function
agent's
is
to
select
the
In
other words, part of the change
starting
points
for
the
diffusion
process.
One of the criticisms of the centralized diffusion strategy is that it
has sometimes functioned to favor those adopters who are already advantaged
relative to others in the system.
apt
risk
to
be opinion
anticipate
that
these leaders
farmers are more
leaders and are more likely to have enough resources to
Therefore
innovation.
introduce the
For example, wealthier
it
is
logical
innovations with these
the
innovations
will
for
extension
influential
workers
to
individuals and to
spontaneously diffuse
outward
from
-12-
Since
normal market economy favors those who
a
resources
invest
innovation,
in
always
innovation will
be
it
can
distributed
argued
be
are wealthy enough to
that
regardless
inequitably,
control
system,
over
distribution of innovations
initial
the
of whether
hbwever, in the more
the diffusion system is centralized or decentralized.
centralized
benefits of
the
rests in the hands of change agents.
the most extreme form of horizontal diffusion, there are no change
In
one
following-up
role of
has the
adopters select themselves and no
Potential
agents (as noted previously).
introduction of an
the
innovation
with
technical information and help.
For instance, in
what
innovations
Legitech, the legislators or their assistants decide
they
wish
to
Oi
try.
initiative,
own
their
Theoretic-
further information and technical advice about the innovation.
everyone
ally,
in
the
system
equal
has
access
the
to
seek
they
innovations
being
diffused
Of course
adopters in both centralized and decentralized systems need
technical help in implementing an innovation, or even in making the initial
decision, once they are aware of the innovation.
the
adopter
can
reach
through
back
technical expertise s/he needs
labs,
if
need
be.
However,
—
this
system
the
back to
In
the
the centralized model,
to
whatever
scientists
in
level
national
of
R&D
process of seeking information and help
proceeds through specified channels:
farmer to county extension workers to
state extension specialists to scientists in
R&D laboratories.
It
is rare
that the ultimate adopter of an innovation in the centralized system meets
face-to-face with the originator of that innovation.
In
source
contrast,
are
a
face-to-face
contacts
distinguishing feature of
a
between
adopter
and
innovation
horizontal diffusion system.
In
-13-
Legitech, for example, an adopter can contact directly the other legislator
or
resource
has
v*io
solved
problem
the
how
of
set
to
up
recycling
a
program, or who has originated legislation encouraging energy conservation,
There are no formal, heirarchical channels to go through.
or whatever.
In
many horizontal (or relatively horizontal) diffusion systems, face-
to-face exchange between potential adopter and the innovation originator is
That is,
formalized through site visits.
see
the
their
the
innovation
in
operation,
in
questions directly to their
a
potential adopters travel to
the
context,
peers
have
who
problems they themselves will
implementation
that
so
they may address
actually struggled with
face
if
they decide
to
adopt the innovation
Department of Justice National Institute of Law
For instance, the U.S.
Enforcement
Criminal
and
Justice
which
(NILECJ),
is
part
of
Law
the
Enforcement Assistance Administration (LEAA), funds senior criminal justice
officials
(such
innovations
as
v*iich
For
exanple,
has
developed
the
a
city
police
selected
have been
Street
department
Crime
techniques utilizing
to
New
York City
in
high
decoys
visit
NILECJ committee
by an
Unit of the
chief)
one
as
Police
of
14
models.
9
Department
crime areas,
and
this
unit is a frequent site for visitors.
Another example comes from
Republic of
program
as
China diffused
a
"model"
for
For
China.
several decades, the
Peoples'
innovations by praising some locally developed
the
nation
to
follow.
The
Tachai
Production
F»"ocess occurs v*ien farmers travel to neighboring farms to see a
The difference is that while the owner of the plot is
demonstration plot.
the first adopter of the innovation, s/he is rarely its originator.
A similar
fact that the innovations are selected for diffusion and hence
legitimized by this committee makes this diffusion system somevhat less
However, the innovation is still being
horizontal than Legitech.
transferred laterally, between peers.
^The
-14-
Brigade
Sensi
in
Province,
rocky
of
area
an
erosion
soil,
agriculture, became famous because of the sensational triumph over
made
by
the
brigade
90
terraces and underground
chemical
raised
1949
fertilizers
grain
their
to 5,295
conduits to
other
and
fight
outside
annual
agricultural
new
to 8,220 in
The slogan, "In agriculture, learn
fJ-om
1971
help,
flooding,
In
thousand visitors
a
the
1977,
built
adopted
thereby
and
cultivated
hectare
(Perkins and others,
in
1977).
Tachai" vas ubiquitous in China for
the past decades, with the result that Tachai was flooded anew
with visitors.
nature
they
techniques
1,050 kilograms per
yields from
1965-66 and
in
Without
households.
poor
and
tiny area
received
—
this time
average of over
an
one
day, who came to learn how to grow grain and to develop
self-reliance.
The
United
California,
States
has
which has been
own
its
promoted by the
following Rosalyn Carter's visit in
of
coram unity -lev el
energy
"Tachai"
1978)
in
the
information
about
their
of
Davis,
and by the mass media as a model
conservation.
use
of
Federal government (especially
The
City
of
Development Department is inundated daily by telephone
for
form
energy conserving
Davis'
and
Community
mail
innovations
requests
and
by
visitors who come in person to see the compact police cars, the community
gardens, the passive and active solar-heated homes, the bicycle lanes, and
the energy conserving building codes.
Davis
is
an
example of unintentional
horizontal
diffusion.
Davis'
city planners originated the energy conserving innovations out of concern
During 1980, the grain yields of the Tachai Brigade were questioned
officially by the Peking government, who claimed that they had been
falsified.
Since the demise of the "Gang of Four," the famous Tachai
leader, Chen Yong-Gui
who was associated with that regime, has dropped
from public view.
Ihe slogan has also become much less prominent in the
past two years
,
-15-
for
their own
would
expecting that they
never
future,
be
called
upon
to
diffuse these innovations to countless other community planners and citizen
No change agents were assigned to aid in the diffusion process; no
groups.
government
agency deliberately
travel
Davis.
to
visits
site
up
broadcasts
media
the
Yet
set
funded
or
message
the
people
"in
to
energy
conservation, learn from Davis," and the visitors continue to pour in.
example
This
of
informal
an
diffusion
horizontal
process,
more structured counterparts, has suffered from the lack of
expertise
transfer
about
Overloaded
innovations.
the
like
mechanism to
a
requests
by
its
and
underfunded for response, Davis city officials have served as unpaid change
explaining to visitors and even taking their show on the road with
agents,
Their reward, common to other innovators in a horizontal
slides and talks.
system
such
as
legislators
staff
legislative
or
Legitech,
in
consists
mainly of public recognition.
some
In
innovators.
to
those
horizontal
For
systems,
instance,
individuals or
a
more
a
formal
(they
nodes
communication
information from the system.
the
"free
Such a
problem
rider"
to
future version of Legitech will assign credit
can
be
which contribute to the system, and will deduct points when
extent
accrues
reward
reward
which
organizations)
a
user derives
structure alleviates to some
tends
to
horizontal
characterize
systems, in that it penalizes those who would draw innovations out of the
system and contribute none.
However,
even
quality control.
1
such
In
an
a
reward
almost
system
pure
does
horizontal
not
solve
system,
the
problem
such as
of
Legitech,
"Free riders" are a problem only if the system participants object to
Sane Legitech members have complained, for instance that they are
them.
doing research for other state legislatures, thus expending valuable time
innovationA critical mass of
and resources,
for little return.
providers is essential to such a system.
-16-
only experience in the system allows users to determine what the quality of
an
innovative suggestion is likely to be, judged according to its source.
systems solve this problem by setting up
Other, less extremely horizontal
evaluative
diffused.
and
drawn
These committees are often
therefore constitute
source of centralized
a
from
the
on
reap not
They also
a
insure
participants
so that
the
in
information
quality of
quantity but
only the
be
to
population of users
the
diffusion system.
some kind of standard for innovation exchanges,
system
slated
are
review group, but they are nevertheless
peer
control
which
innovations
review
who
corrmittees
they
themselves contribute to the system.
A Hybrid Diffusion System
To
further
clarify
centralized diffusion systems,
decentralized
as
Legitech
in
1973 by a
Office
of
which
but
agricultural extension model.
now describe
we
The
far
is
a
more
horizontal
were
with
faced
to allocate the funds to local
problem
the
schools
v*io
of
"year-end
of educational
innovations were
called
of
Federal
had
been
approved
experts.
The
They decided
schools.
The local sources
About
"developer/demonstrators."
150 such developer/demonstrators were funded,
that
money,"
had developed an innovation, for
use in spreading these new ideas to other local
committee
began
quirk of bureaucratic budget handling; officials in the U.S.
Eaucation
innovation
as
the
than
Diffusion Network (NDN)
National
dollars which had to be spent by the end of the fiscal year.
an
not
vhich is
system
and
horizontal
of
characteristics
differing
the
as
modest
a
each to horizontally diffuse
"validated
Federal
practice"
funds
were
by
used
a
by
developer/demonstrators to publish brochures and other mass media messages,
to
provide
training
for
potential
innovation to other school
adopters,
teachers.
and
Few of the
to
demonstrate the local
initial
150 innovations
-17-
were "standard" innovations like those that had been promoted previously in
decade of center-periphery diffusion
a
Team
teaching,
schools
were
programmed
expected
instruction,
adopt
to
by
those
the
teacher
new
aides,
certain
ft"om
schools
the local
with
of
Eduction:
All
mainly
public
had
been
But the NDN innovations
developer/demonstrators were appropriate only to
particular
developer/demonstrators'
etc.
which
ideas,
invented by research and development laboratories.
diffusing
Office
U.S.
problems.
And
then,
even
many of
the
innovations were re-invented by other schools when
Further,
they implemented them under their local conditions.
innovations were re-labeled
with
a
local
name
by certain
some of the
adopters,
even
when an innovation's form had not really been modified very much or not at
The psychological
all.
a local
identity and to encourage pride of local ownership.
How successful
first
effect of such renaming was to give the innovation
three
is the
National
years of operation,
the
Diffusion Network?
150
popular
with
school
translated into p»litical
personnel
and
support by the
the end of its
innovations had been accepted by
several thousand adopters (Emrick and others,
very
At
the
U.S.
The NDN was generally
1977).
public.
This
popularity was
Congress, who began to give
the NDN a regular budget (with a major increase to $25 million in
precise measure of NDN's impact
different
innovations
were
was
difficult to
spontaneously
obtain
flowing out
from
demonstrators, and each of these innovations took such
The result certainly seemed to be innovation in U.S.
not
12
a
result that could be conveniently measured.
because
a
1977).
so
A
many
the developer/
variety of forms.
education, but it was
12
At least as conveniently and neatly measured as in the case of a centerpheriphery diffusion approach, where the usuaal measure of impact is the
rate of adoption of innovations promoted by a Federal agency to local
government units or to the public.
-18-
Davis, and Legitech are all characterized by a sense of local
The NDN,
control and ownership, both
of the
system
aware
been
years
seme
innivations
acceptance
of
Havelock,
1973),
Syndrome,"
they
by v^iich
for
alleviating
the
re-invention of
disseminated.
of
the
blamed
influence
for
and
others,
1974;
on
the
Havelock
and
Invented
"Not
organizational
and
innovation
which
is
considerable
Here
individual
a
means
adaptation,
Fewer
diffused.
(2)
scholars have
participation
of
so-called
syndrome by allowing
NIH
the
Innovation
Horizontal diffusion systems may provide
resistance to new ideas.
for
are
(Fairweather
often
is
of the innovations being diffused, and
conscious of the
and
which
(1)
or
built-in
assumptions about the inviolability of the original innovation are usually
found
in
the
horizontal
systems.
Whether
not
or
decentralizied
these
systems actually encourage re-invention is an empirical question.
Issues in the Study of Decentralized vs. Centralized Systems
Even
this
first,
rather
superficial
examination
of contrasting
diffusion systems raises issues which are clearly central to the question:
when
are
vertical
or
horizontal
strategies preferable?
Future
empirical
examinations of relatively centralized or decentralized systems (or of the
centralized and decentralized elements within
a
given system) will need to
consider at least the following:
relative ability of the system to bring relevant technical
expertise to bear during the implementation of an innovation by an
adopter
The
.
degree to v*iich the system is problem-centered
centered (technology pull versus technology push).
The
*
versus
solution-
The degree to which innovations undergo change ("re-invention") during
diffusion
The degree and nature of quality control
information passed through the system.
on
innovation-related
-19-
The scope of the innovations (incremental versus radical) commonly
(Perhaps horizontal systems are more likely
diffused in the system.
to diffuse incremental innovations).
presence and role of gatekeepers
facilitators or change agents?
The
expect
relatively horizontal
that
components
system
issues.
The
challenge
is
the
along
differ
will
relatively
and
dimensions
operational ize
to
systems
or
in
these
involved
(e.g.
suggested
innovations)
they
channels
vertical
concepts
the
degree of re-invention; the radicalness of the
are
information
varying
of
The
relative predominance
(interpersonal; media, etc.) in the system.
We
—
system
the
in
so
that these
issues can be developed into hypotheses to test empirically.
Summary and Implications
frcxn
In
have presented
this paper, we
In
classical
the
"center-periphery"
a
of diffusion
model
paradigm on several
this alternative method of diffusion,
which differs
key dimensions.
innovation users are not only the
Dissemina-
originators but often also the dissoninators of the innovation.
of
tion
innovation
the
occurs
through
networks,
peer
utilizing
such
mechanisms as site visits or computerized inquiry-and-response dialogues.
some possible advantages of such systems (from the
We may hypothesize
users'
support
through
of
point
for
the
the
"re-invention"
namely:
view)
innovations;
system,
of the
lower
a
sense
greater
rates
innovations.
of
variety in
of
ownership,
local
the
this
user
innovations diffused
discontinuance,
(Obviously,
hence
last
and
greater
characteristic
would not necessarily be an advantage from the originator's vantage pwint)
We
also
horizontal
expect
diffusion
to
consistently
systems:
the
find
certain
transfer
of
disadvantages
the
technical
to
.
such
know-how
-20-
necessary
implementation
to
guaranteed
not
is
there any built-in quality control.
systems,
no
innovations;
hence,
buyer
the
the
professional
or
beware.
must
system;
neither
is
extremely decentralized diffusion
In
"facilitator"
official
in
evaluator
Finally,
judges
the
extremely
in
decentralized systems such as the Legitech example in our paper, there are
few incentives for an
innovation-rich "Scrooge" to share with innovation-
needy peers besides an altruistic concern for the larger societal Good, or
the
chance to
gain
disincentives
built-in
amount
certain
a
the
to
of
rider,
free
Similarly,
fame.
innovation information from the system but
there
consistently
who
are
few
draws
out
does not pay anything
v*io
into
the system.
much
know
need
to
disadvantages
—
decentralized
system
We
the
when
is
more
former
these
about
advantages
consequently
accrue and
appropriate,
possible
and
horizontal or
a
how the latter may be overcome.
and
There are already experiments underway to counteract the Scrooge/free rider
problem by crediting an information or innovation provider and by debiting
those individuals or organizations who derive value from the system.
Several forces at work in society may serve to increase the number of
cases
in
which
appropriate.
horizontal
First,
the
diffusion
systems
for
the
potential
for
decentralized
innovations
of
information
of
the
telephone
increased
individual's verbal communication,
capacity
for
information.
increased
information
individual
Second,
problems
personalization
so
in
of
scope
or
reach
of
every
computers are multiplying the
sources
information
information
conmunications flow census showed that
the
sources
Just as the
grows as computerized conmunications diffuse through society.
invention
are
to
overload
origination
vrtiereas
the
transmit
and
will
written
lead
receipt.
to
One
production of words in
-21-
all media combined is growing at about
10 percent per annum, consimption of
words is increasing by only three percent annually.
Therefore, the average
number of audience members for each word must be decreasing (de Sola Pool,
1980).
of the
of
One response to such information overload is more careful screening
information used
—
magazines
specialized
The plethora
more sorting before consumption.
the
in
targeted audiences evidences
catering
U.S.
ever
more
carefully
increasing specialization in
form of
a
to
information flow, for example.
Therefore, it seems highly likely that our needs for pre-screening of
If past experience is any
information are going to increase, not decrease.
guide,
we
continue
will
especially peers
have
already
similar
to
fit
have
who
This
together
technological means
for
with
prefer
already
available
our own.
information,
to
interpersonal
through the
sorted
information
to
increasing need
the
ever
sources
more
their
for
of
information,
alternatives.
needs,
v*iich
are
They
very
help in selecting relevant
accessible
and
sophisticated
contacting peers across geographic distances, may
account in part for the popularity of the concept and term, "networking."
Horizontal diffusion can be viewed as the
formalized use of peer networks
to disseminate innovations.
If, as we
believe, the use of such systems is increasing, it behooves
planners and social science researchers alike to understand the dynamics of
such systems and to define those conditions under which they function best.
-22-
REFERENCES
David K. and others (1970), "Dimensions In Evaluating
Acceptability of Message Sources," Public Opinion Quarterly
Berlo,
,
the
33:
563-576.
Emrick, John
Network
.
A.
and others (1977),
Menlo Park, California:
Fairweather, G.W., D. H. Sanders and
in Mental Health Organizations
.
Evaluation of the National Diffusion
SRI International report.
Tornatzky (1974), Creating Change
Pergammon Press, Inc.
New York:
L. G.
Energy Policy for the
(1979), "Clioosing Among Conflicts:
Social
Impacts of Energy
presented at the Workshop on the
of
Man
and Science
Decentralization, June 13-16, The Institute
Rensselaerville, New York.
Luther
80s," paper
Gerlach,
Havelock (1973). Educational Innovation in the
Center for Research on
Ann Arbor, Michigan:
Utilization of Scientific Knowledge, Institute for Social Research,
University of Michigan.
Havelock,
and
R. G.
United States
M.C.
.
Information Exchange
Johnson-Lenz, Peter and Trudy (1979), "Legitech/EIES:
and Collective Knowledge Building Among State Legislative Researchers,"
in Madeline M. Henderson and Marcia J. MacNaughton (eds.) Electronic
Westview Press,
Boulder, CO:
Communication: Technology and Impacts .
AAAS Symposium Series.
Kotler,
An
Philip and Gerald Zaltman (1971), "Social Marketing:
3-12.
Social Change," Journal of Marketing
35:
to Planned
Approach
,
Dorothy (1980), "The Role of Interpersonal Communication
the Diffusion of Energy Conserving Practices and
Technologies, Proceedings of the International Conference on Consumer
Behavior and Energy Use, Banff, Alberta, Canada, September 17-20.
Leonard-Barton,
Networks
in
Perkins, Dwight and others (1977), Rural Small-Scale Industry in the
University of
California:
Berkeley,
People's Republic of China
Press.
California
.
Pool, Ithiel de Sola, (1980) "The New Structure of International
The Role of Research," paper presented to the Congress
Conmunication:
of the International Association for Mass (Communication Research,
Caracas, Venezuela, August.
Rogers, Everett M. (1977), "Re-invention of New Ideas During the Innovation
Process in Public Organizations," paper presented at the Conference on
Agenda for Research and
Technology TV-ansfer to Urban Governments:
School of Citizenship and
Maxwell
Syracuse University,
Education,
Public Affairs, May 6.
Rogers,
Everett
Innovations.
M,
with
New York:
Floyd
Shoemaker
The Free Press.
(1971)
Communication of
-23-
Stevens,
Chandler
Networking,"
H.
(1980),
World
(November-December)
von
"Many to
Future
Many
Conmunications
Society
Bulletin
,
Hippel, Eric (1976), "Ihe Dominant Hole of Users in
Instrument Innovation Process," Research Policy, July.
Through
]H,
the
Inquiry
No.
6
Scientific
u>
-V"-.
/
/:
57
033
IBASEMENT
^itB Due
fS»j
gQ-^y^^l
lAAY
V
27
M/IR
MAY 9
19!)?
1992
Tun 10 1992
AP2 7'8g
SE
NOV
AUG
1 6 '89
I4l36|»'^^v
DEC oe
3 1992
23^
li.
Lib-26-()7
HD28.IV1414 no,1214' 81
Leonard'Barton/Horizontal
diffusion
of
"11
iiiiii
3
TDflO
iiiljiiilli
002 001 Oil