Running Head: INNOVATION DIFFUSION: A PROCESS OF DECISION-MAKING 1 Innovation Diffusion: A Process of Decision-Making The Case of NAQC Jonathan E. Beagles, M.S. Ph.D. Candidate 520-975-1224; [email protected]School of Government and Public Policy University of Arizona Keith G. Provan, Ph.D. McClelland Professor of Management & Organizations Eller College of Management and School of Government and Public Policy University of Arizona Scott F. Leischow, Ph.D. Professor, Family and Community Medicine Arizona Cancer Center University of Arizona Work on this paper was funded by a grant from the National Cancer Institute (R01CA128638- 01A11) and an Arizona Cancer Center Support Grant (CCSG - CA 023074)
42
Embed
Running Head: INNOVATION DIFFUSION: A … Head: INNOVATION DIFFUSION: A PROCESS OF DECISION-MAKING 1 Innovation Diffusion: A Process of Decision-Making …
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Running Head: INNOVATION DIFFUSION: A PROCESS OF DECISION-MAKING 1
Innovation Diffusion: A Process of Decision-Making
If this is indeed the case, we could expect ties to the NAO and to national policy and
funding organizations to serve more than just an information sharing function. In addition to
information sharing, we would suspect ties with these powerful organizations to influence a
INNOVATION DIFFUSION: A PROCESS OF DECISION-MAKING 12
decision-makers valuation criteria. Specifically, in the context of our study, we would suspect
ties to the NAO and these national organizations, more than ties to other organizations, to
increase the likelihood of an organization adopting evidence based practices while controlling for
its level of awareness. Stated in the form of hypotheses:
Hypothesis 3a: Organizations connected to the network administrative organization will
be more likely it will be to adopt innovative practices.
Hypothesis 3b: The greater the number of connections an organization has to National
Organizations, the more likely it will be to adopt innovative practices.
Internal decision-making processes such as values and goals and other evaluative criteria
are likely to have their biggest impact at the decision stage of the innovation diffusion process. It
is at this stage where diffusion researchers suggest we will see the culmination of an
organization‟s process of evaluating an innovation based on the knowledge it has gleaned.
However, to understand this evaluation process, it is important to be familiar with the evaluative
criteria organizations are likely to use. Three criteria are prevalent in the literature: efficiency,
effectiveness and prestige. Underlying the rational decision-making perspective is the idea
decision-makers will choose the alternative they perceive to be in their best interest. This concept
of best interests is commonly understood to be the most efficient (greatest benefit for least cost)
decision. Another criterion used to evaluate alternatives is by their perceived effectiveness. This
criterion differs from efficiency in that it pays less attention to the costs of an alternative. In
practice, effectiveness is often evaluated based on perceptions of consistency with an
organization‟s mission. Finally, in the innovation diffusion literature, research suggests
prestigious organizations are more likely to adopt new innovations; especially when they are
perceived as being consistent with the norms of the community (Rogers, 2003). Depending on
INNOVATION DIFFUSION: A PROCESS OF DECISION-MAKING 13
which criterion an organization utilizes, we can expect differences in its adoption/rejection
decisions. Specifically, organizations placing a greater emphasis on efficiency are likely to
require an innovation to meet more rigorous requirements than those placing an emphasis on
mission fit as an evaluative criterion in deciding whether or not to adopt a particular evidence
based practice. While both may see benefits in implementing a practice those focusing on
efficiency also place a great deal of concern on the cost side of the equation. Prestigious
organizations may also have less rigorous requirements for adopting new practices because of
the additional perceived benefit of maintaining their status within the network. Stated in the form
of hypotheses:
Hypothesis 4a: The greater the importance an organization places on rational
factors(efficiency), the less likely an organization will adopt new evidence based
practices.
Hypothesis 4b: The greater the importance an organization places on mission fit, the
more likely an organization will adopt new evidence based practices.
Hypothesis 4c: The greater an organization’s reputation within the network, the more
likely an organization will adopt new evidence based practices.
Capacity and Implementation
The final stage of the innovation-decision process with which we are concerned has to do
with implementation. At this stage information about the practice has been gathered and it has
been evaluated in light of the evaluative criteria of the organization. Here we suspect the capacity
of an organization will play a crucial role in determining whether or not an organization is able
to implement a practice it has decided to adopt. Along with internal capacities such as technical
expertise and finances, network and diffusion researchers have pointed to the importance of
INNOVATION DIFFUSION: A PROCESS OF DECISION-MAKING 14
network relations at the implementation stage. Specifically, these findings suggest the ability of
an organization to implement a new innovation effectively is enhanced when it can communicate
with others who have gone through or are going through the same process (Ducharme, et al.,
2007). A second way in which implementing organizations can gain the information they need is
through connections to the NAO; since this organization often plays a central role in the network
and is charged with network coordination and the dissemination of information. In the case of
quitlines, these connections are likely to be most important for provider organizations because of
their direct involvement in the implementation and reinvention process. Also, because
reinvention is an important part of successful implementation, the involvement of implementing
organizations in the decision-making process should enhance the effectiveness of reinvention
decisions and thus increase the likelihood of successful implementation. Stated in the form of
hypotheses:
Hypothesis 5a: The greater the number of connections a quitline’s provider organization
has with other providers the greater the number of innovative practices successfully
implemented.
Hypothesis 5b: Quitlines with provider organizations connected to the network
administrative organization will successfully implement a greater number of innovative
practices.
Hypothesis 5c: The more a quitline’s provider organization is made part of the decision-
making process, the greater the number of innovative practices it will successfully
implement.
Figure 3 provides a visualization of the hypotheses.
INNOVATION DIFFUSION: A PROCESS OF DECISION-MAKING 15
---------------------------
Figure 3
---------------------------
Data
The data for this analysis was collected during the summer of 2009. It is the first of three
rounds of data collection, which will ultimately allow for longitudinal analysis and a better
understanding of the diffusion process. The network consists of numerous individuals and
organizations filling a variety of roles. However, our focus on the adoption and implementation
of innovations guided our decision to limit the collection of data to only the organizations
directly involved in this particular decision-making process along with the network
administrative organization (NAO).
The organizations surveyed (n=95) consisted of 73 funder organizations (some quitlines
had multiple funders), 20 service providers and one organization serving in both capacities as
well as the NAQC NAO. Depending on organization size, data were collected from 1 to 6
respondents (identified beforehand as the top decision-makers regarding quitline issues) at each
organization. Primary data were collected using a web-based survey developed expressly for this
project but based on methods and measures utilized previously by Provan and colleagues
(Provan and Milward, 1995; Provan, et al., 2009). In addition, questions and methods were pre-
tested on a “working group” of key quitline members who agreed to provide initial feedback.
After extensive follow-up efforts using email and telephone, our final results included completed
surveys from 186 of 277 individual respondents (67.1% response rate), representing 85 of 94
quitline component organizations (90.4%) plus the NAQC NAO, and at least partial data (at least
one component organization) from 62 of the 63 quitlines (98.4%).
INNOVATION DIFFUSION: A PROCESS OF DECISION-MAKING 16
Our unit of analysis is the quitline, represented by the funder organization. We focused
on the funder for a number of reasons. First, we had complete data on network ties as well as
responses about awareness, adoption, rejection and implantation of evidence-based practices
from 60 of the 63 quitlines, but only partial data from a number of the larger, multi-quitline
provider organizations. In particular, one of these large providers did not complete the practice
questions since its management felt strongly that because the funder organization initiates the
contract and pays the bills, it is the funder who decides what practices to use. We used this logic
as well in our decision to focus on the funder. Second, many of the providers served multiple
states and provinces, making it difficult to disentangle the effects of the role of these providers
relative to one of its quitlines versus another. Each U.S. state (and territory) and each Canadian
province is represented by a quitline funder organization, each with its own separate budget and
network connections, making it possible to compare meaningfully across quitlines and thus, test
our hypotheses. Finally, while providers represent public, nonprofit, and for-profit entities, all
quitlines are predominantly funded by a public entity, allowing us to examine the impact of
public contracting on service awareness. Hence, our analytical focus is the funder organization
as the representative of each quitline.
Measures
Innovation decision stages. To gather information at each stage of implementation, we
asked respondents where they believed their quitline was in the implementation process
regarding 23 practices identified by the network NAO and „project working group‟. These
practices ranged from the provision of proactive counseling to the use of text messaging and the
referral of callers to health plans. However, for this study we excluded six practices from the
analysis: two because they pertained to US quitlines only; two because they were pharmacology
INNOVATION DIFFUSION: A PROCESS OF DECISION-MAKING 17
related practices; and two due to lack of evidence supporting their effectiveness. The remaining
17 practices related to behavioral therapy and related management practices consistent with the
core mission of quitlines (See Appendix A for a complete list of practices).
In completing this section of the survey, respondents were first asked to indicate „Yes‟ or
„No‟ regarding whether or not they were aware of a practice. If respondents indicated „Yes‟ they
were aware of the practice, they were then presented with a follow-up question asking them to
indicate at what level of the decision-making process their quitline was at. To answer this
question, they were provided four response options: „Have not yet discussed‟, „In discussion‟,
„Decided not to Implement‟ or „Decided to Implement.‟ If the respondent indicated a decision
had been made to implement a particular practice, they were next presented with a 5-point scale
1=No progress has been made yet to 5=Fully implemented (the practice has become part of the
quitline’s policy or standard operating procedures for all eligible callers) and asked to indicate
what level of implementation they felt their quitline had achieved regarding the practice. From
this information we created four binary variables for each practice for each quitline1. A quitline
was considered AWARE2 of a practice and received a score of 1 if at least one respondent from
the quitline marked „Yes‟ to the first question. A quitline was considered to REJECT a practice
and received a score of 1 for the practice if the majority of respondents within the quitline
indicated „Decided not to Implement‟ in the second question. Likewise, a quitline was
considered to ADOPT a practice and received a score of 1 for the practice if a majority of
respondents indicated „Decided to Implement‟ in the second question. Finally, for
1 While providers were asked to respond to these questions separately for each quitline they served, due to the
abstention from these questions by one of the large providers, we chose to analyze the funder‟s responses as the
quitline‟s response except where noted otherwise. 2 While the measure described is a count of the number of practices a quitline is aware of, analysis was done on
UNAWARE (the number of practices a quitline was unaware of) to better suit negative binomial modeling 2 While the measure described is a count of the number of practices a quitline is aware of, analysis was done on
UNAWARE (the number of practices a quitline was unaware of) to better suit negative binomial modeling
INNOVATION DIFFUSION: A PROCESS OF DECISION-MAKING 18
IMPLEMENT, quitlines received a 1 if the majority of respondents indicated a 4 or higher on the
5-point scale in the last question. Once these variables were constructed, a count variable was
calculated to indicate the number of practices in which a quitline received a score of 1 at each
stage. Scores could range from 0 to 17 (See Table 1 for a summary)
--------------------------
Table 1
--------------------------
Information sharing. Data on network relationships were collected based on receipt of
information in four areas: financial, general management, service delivery, and
promotion/outreach. Respondents were presented with a list of all quitline funders, then provider
organizations, and then other national non-quitline member organizations having a major tobacco
control focus and involvement. For each organization listed, respondents were asked to indicate
whether they received information from that organization, which of the four types of information
they received, and the level of intensity of the relationship in terms of frequency and importance
(scored on a 1 to 3 scale). Only responses scored at a high level of intensity (3) were utilized in
the final analysis.
Because some quitlines consist of multiple funders or multiple providers, we found it
necessary to aggregate these multiple responses to obtain a single funder or single provider
response for each quitline. Of the 62 quitlines from which we received at least partial data, we
received multiple funder responses from six and multiple provider responses from one. To obtain
a single funder and single provider response from each quitline, we aggregated individual
responses from the multiple organizations as if the respondents came from the same
organization. These aggregations left us with 60 funder and 17 provider responses.
Because responses were provided by individuals and the analysis for this paper is
INNOVATION DIFFUSION: A PROCESS OF DECISION-MAKING 19
presented at the organization level, a tie was considered to exist at the organization level if at
least one respondent from that organization reported receiving information from that
organization. We apply this rule based on the presumption that the segregation of duties within
an organization often necessitates a single individual be the primary person responsible for
maintaining a relationship with a particular organization (Broshack, 2004; Maurer & Ebers,
2006).
Using these data, a series of network variables were constructed for both the funders and
providers of each quitline. Consistent with the survey data and the hypotheses, the following
five distinct types of network variables were constructed for each funder, all based on indegree
(information received) centrality and all based on the highest level of intensity of involvement:
funder ties to the NAO (fnNAO: coded 0 or 1); the number of funder ties to other funders
(fnFUNDERS); number of funder ties to other providers (fnPROVIDERS); number of funder ties
to the 12 national organizations that were NAQC members, but which were not part of a specific
quitlines, like the RWJ foundation, CDC, American Legacy Foundation, and Health Canada
(fnNATIONAL); and the number of funder ties to the 10 most highly connected tobacco control
researchers (fnRESEARCH) (from a drop-down list of 42 tobacco control researchers previously
identified). For this last measure, each quitline respondent was allowed to list up to five
researchers but responses were weighted so no quitline organization could score more than a
single point for any one researcher and no more than five points total.
Because our hypotheses regarding the effect of providers‟ connections is based on their
ability to observe and discuss implementation related information we constructed the following
two variables: provider ties to the NAO (prNAO: coded 0, 1); provider ties to other providers
INNOVATION DIFFUSION: A PROCESS OF DECISION-MAKING 20
(prPROVIDERS). Ties to these two knowledge pools are suspected to be most important for
implementation and reinvention decisions.
Search. To capture an organization‟s involvement in quitline decision-making, we asked
each respondent the following question: “When deciding whether or not to implement a new
treatment practice, please indicate how decisions are usually made between your organization
and your quitline partner organization(s).” Responses were provided using a likert-type 5-point
scale with 1 = ‘Funder Decides’, 5 = ‘Service Provider Decides’, and 3 = ‘Decision is Shared
Equally’. After taking the average individual score within the organization as the organization‟s
response, we created a dummy variable, WHO, with organizations scoring a three or higher
receiving a 0 indicating the provider is heavily involved in decision-making and organizations
scoring less than 3 receiving a 1 indicating the funder dominates decision-making. Twenty-six of
the 60 funders reported the provider was heavily involved in decision-making.
Valuation criteria. In addition to the information sharing data, we asked 12 questions
regarding a quitlines‟ decision-making process (see Appendix B for a copy of the questions). The
12 items (4 items each) were designed to capture the three components of the Theory of Planned