-
This article was downloaded by: [Fac Psicologia/Biblioteca]On:
25 September 2012, At: 00:20Publisher: RoutledgeInforma Ltd
Registered in England and Wales Registered Number:1072954
Registered office: Mortimer House, 37-41 Mortimer Street,London W1T
3JH, UK
Human Dimensions ofWildlife: An InternationalJournalPublication
details, including instructions forauthors and subscription
information:http://www.tandfonline.com/loi/uhdw20
Development andValidation of aSpecialization Index andTesting of
SpecializationTheoryRonald J. Salz, David K. Loomis & Kelly
L.Finn
Version of record first published: 29 Oct 2010.
To cite this article: Ronald J. Salz, David K. Loomis &
Kelly L. Finn (2001):Development and Validation of a Specialization
Index and Testing ofSpecialization Theory, Human Dimensions of
Wildlife: An International Journal,6:4, 239-258
To link to this article:
http://dx.doi.org/10.1080/108712001753473939
PLEASE SCROLL DOWN FOR ARTICLE
Full terms and conditions of use:
http://www.tandfonline.com/page/terms-and-conditions
This article may be used for research, teaching, and private
studypurposes. Any substantial or systematic reproduction,
redistribution,
-
reselling, loan, sub-licensing, systematic supply, or
distribution in anyform to anyone is expressly forbidden.
The publisher does not give any warranty express or implied or
makeany representation that the contents will be complete or
accurateor up to date. The accuracy of any instructions, formulae,
and drugdoses should be independently verified with primary
sources. Thepublisher shall not be liable for any loss, actions,
claims, proceedings,demand, or costs or damages whatsoever or
howsoever caused arisingdirectly or indirectly in connection with
or arising out of the use of thismaterial.
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
-
239
Human Dimensions of Wildlife, 6:239258, 2001Copyright 2001
Taylor & Francis1087-1209 /01 $12.00 + .00
Partial funding for this project was provided by the Cooperative
State Research, Extension,Education Service, U.S. Department of
Agriculture, Massachusetts Agricultural Experiment StationProject
Number 782.
Address correspondence to David K. Loomis, Department of Natural
Resources Conservation,Human Dimensions Research Unit, Holdsworth
National Resources Center, University of Massa-chusetts, Amherst,
MA 01003-4210. E-mail: [email protected]
Development and Validation of a SpecializationIndex and Testing
of Specialization Theory
RONALD J. SALZDAVID K. LOOMISHuman Dimensions Research
UnitUniversity of Massachusetts-AmherstAmherst, Massachusetts,
USA
KELLY L. FINNCalifornia Department of TransportationSan Diego,
California, USA
Recreation specialization can be viewed as a continuum of
behavior fromthe general to the particular. Along this continuum,
participants can be lo-cated into meaningful subgroups based on
specific criteria. Previous stud-ies have defined, measured, and
segmented specialization groups in a vari-ety of ways. The research
reported here builds on the Ditton, Loomis, andChoi
reconceptualization of recreation specialization. A specialization
in-dex was developed to segment anglers into four groups based on
their ori-entation, experiences, relationships, and commitment.
Internal validationanalysis supported the use of this
specialization index as a tool for anglersegmentation. Subsequent
hypotheses tested for differences among special-ization groups in
frequency of participation, importance of activity
andnonactivity-specific elements, support for management
regulations, and side-bets. Results provide strong support for the
conceptual framework developedby Ditton et al. These findings
indicate a multidimensional index can beused to segment anglers
into discreet, meaningful specialization categories.
Keywords Recreation specialization, segmentation, specialization
index,anglers
Peer-Reviewed Articles
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
-
240 R. J. Salz et al.
Introduction
Outdoor recreation participants generally display wide variation
in their experi-ences, avidity, expertise, commitment, economic
expenditures, and social inter-actions related to a particular
activity. Connected to this variation are importantsociological and
psychological differences affecting motivations,
expectations,desired outcomes, satisfaction levels, perceptions,
and social norms. Outdoor rec-reation managers must recognize and
accommodate these differences to providesatisfactory experiences to
a widely diverse clientele. Recreation specialization isan area of
study that attempts to describe this variation through segmentation
ofparticipants into meaningful and identifiable subgroups. Bryan
(1977) was thefirst to conceptualize recreational specialization as
a continuum of behavior fromthe general to the particular,
reflected by equipment and skills used and activitysetting
preferences. The four levels of specialization he identified in a
populationof trout anglers were occasional anglers, generalists,
technique specialists, andtechnique-setting specialists. Bryan
(1977) suggested that more highly special-ized anglers are part of
a leisure social world with a shared sense of group identi-fication
derived from similar attitudes, beliefs, and experiences.
Recreation specialization studies following Bryan used a variety
of classifi-cation techniques and variables to segment participants
into specialization levels.Some studies found that a single-item
measure of specialization could be used tosegment participants. For
example, Graefe (1980) noted that frequency of par-ticipation
(i.e., avidity) was a useful surrogate for measuring angler
specializa-tion. He found that anglers who fished more frequently
(i.e., were more special-ized) had higher self-reported skill
levels, participated in more diverse fishingsettings, and had a
greater dependency on the resource. Ditton, Loomis, and Choi(1992)
also used avidity to segment recreational anglers into four
specializationlevels. Similarly, Schreyer, Lime, and Williams
(1984) used total number of riverruns as a means of classifying
river users into six groups and found differencesbetween the groups
in the type of prior river experience, motives for
participation,perceptions of conflict, and support for managerial
regulations.
Other studies took a multidimensional approach to recreation
specializationby incorporating several variables into a
specialization index. Chipman and Helfrich(1988) concluded that
investment, consumptive habits, and frequency of partici-pation
were important characteristics for determining specialization among
an-glers. Kauffman and Graefe (1984) used preferences for river
characteristics tosegment canoeists into more-specialized and
less-specialized groups. Fedler andDitton (1986) segmented anglers
into levels of consumptive orientation based onresponses to
statements regarding the importance of catching fish.
Wellman,Roggenbuck, and Smith (1982) used a specialization index
based on equipmentinvestment, past experience, and centrality to
lifestyle to segment anglers intogroups that reflect respondents
attitudes toward depreciative behavior. Virdenand Schreyer (1988)
constructed a specialization index to segment hikers basedon
equipment and economic commitment, centrality to lifestyle, general
experi-ence, and past experience variables.
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
-
241Development and Validation of a Specialization Index
Using a variety of segmentation methods, recreation
specialization studiesshowed that more specialized users differed
from less-specialized users on nu-merous attributes. These included
motives for participation (Kauffman & Graefe,1984; Schreyer et
al., 1984), importance of nonactivity-specific elements
(Fedler& Ditton, 1986), preferences for management strategies
(Chipman & Helfrich,1988; Hammitt & McDonald, 1983),
perceptions about crowding (Vaske, Donnelly,& Heberlein, 1978),
environmental preferences (Kauffman & Graefe, 1984;Schreyer et
al. 1984; Virden & Schreyer, 1988), equipment ownership and
use(Chipman & Helfrich, 1988; Wellman et al., 1982), and
centrality to lifestyle(Virden & Schreyer, 1988; Wellman et
al., 1982). In general, these studies pro-vided support for Bryans
specialization concept, and greatly advanced the gen-eral
understanding of diversity among outdoor recreation
participants.
However, the lack of any empirical testing of recreation
specialization re-mained an issue. As pointed out by Ditton et al.
(1992), any attempt to test Bryansframework for specialization was
problematic because it was tautological (circu-lar) in its
reasoning; specialization level, defined in terms of behaviors and
prefer-ences, was then used to predict specialized behaviors and
experiential preferences.As a result, recreation specialization as
a concept could never be empirically testedbecause specialization
and its subsequent propositions were both defined andmeasured in
the same terms (Ditton et al., 1992).
Ditton et al. (1992) initiated development of a testable theory
that links rec-reation specialization with elements of social
worlds as described by Unruh (1979).Unruh (1979) defined a social
world as an internally recognizable constellationof actors,
organizations, events and practices which have coalesced into a
per-ceived sphere of interest and involvement for participants.
According to this per-spective, members of the same social world
hold similar attitudes, beliefs, andmotivations that create a sense
of group identity. Unruh (1979) further suggestedthat members
within a social world could be ordered along a theoretical
dimen-sion of involvement level based on four key characteristics:
orientation, experi-ences, relationships, and commitment. For each
characteristic, Unruh (1979) de-scribes four involvement levels
that correspond to four trans-situational socialtypes: strangers,
tourists, regulars, and insiders (Table 1).
Ditton et al. (1992) reconceptualized and redefined recreation
specialization
TABLE 1 Characteristics and Types of Social World Participation
(from Unruh,1979)
Social types or subworlds
Characteristics Strangers Tourists Regulars Insiders
Orientation naivete curiosity habituation identityExperiences
disorientation orientation integration creationRelationships
superficiality transiency familiarity intimacyCommitment detachment
entertainment attachment recruitment
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
-
242 R. J. Salz et al.
as a process by which recreation social worlds and subworlds
segment and inter-sect into new recreation subworlds, and the
subsequent ordered arrangement ofthese subworlds and their members
along a continuum. Subworld types are arrangedby Ditton et al.
(1992) on a continuum from least specialized to most
specialized.
Ditton et al. (1992) developed eight recreation specialization
propositions.They tested three of these, using frequency of
participation to segment anglersinto four specialization levels.
Their results provided empirical support for spe-cialization by
showing that the four groups differed as predicted in their
resourcedependency, level of mediated interaction, and the
importance they attach to ac-tivity-specific and
nonactivity-specific elements within a recreational activity.Highly
specialized anglers were found to have a higher resource dependency
thandid less specialized anglers. The highly specialized groups
placed more impor-tance on catching big, distinctive, or trophy
fish, whereas the less specializedanglers appear to be less
interested in the rare event aspect of the fishing expe-rience.
They found that anglers who were more specialized had a greater
involve-ment in various types of mediated means of communication
than did less special-ized anglers. Finally, Ditton et al. (1992)
found that as level of specializationincreased, the importance
attached to catch-related angling motivations (e.g., catch-ing fish
of preferred size, number, or species) decreased relative to
noncatch-relatedangling motivations (e.g., to be outdoors, to
relax, to be with friends, etc.).
Although their single dimension (i.e., frequency of
participation) approachto angler segmentation proved successful,
Ditton et al. (1992) recognized thatother variables can and should
be used as a means of classifying individuals intospecialization
subgroups. A single variable (such as avidity) cannot
adequatelymeasure these distinct dimensions of specialization and
may result in highmisclassification rates. In this paper, we
suggest that the testing of recreation special-ization theory, and
its application, is advanced when using a multivariable ap-proach
to segmentation that incorporates orientation, experiences,
relationships,and commitment.
Study Objectives
The first purpose of this research was to develop and validate a
multivariablespecialization index based on a social world view of
recreation specialization.The second purpose of this research was
to use this index to test recreation special-ization theory by
re-examining one of the propositions tested by Ditton et al.(1992),
examining two other propositions that have not yet been tested, and
devel-oping and testing a new proposition. The proposition to be
retested states: as levelof specialization in a given recreation
activity increases, the importance of activity-specific elements of
the experience will decrease relative to
nonactivity-specificelements of the experience (Proposition Eight
in Ditton et al., 1992). Ditton et al.(1992) found that
more-specialized anglers placed less importance on
activity-specific elements, such as catching fish, and more
importance on the nonactivity-
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
-
243Development and Validation of a Specialization Index
specific elements of the fishing experience, such as enjoying
nature, relaxing,being with friends or family, and so forth.
The second proposition states that participants who are more
specialized wouldindicate greater support for management rules and
regulatory procedures, as wellas for social norms that identify and
often dictate acceptable behavior, than wouldless-specialized
participants (Proposition Four in Ditton et al., 1992). Temporaryor
seasonal closures due to overfishing, for example, would have a
greater impactfor more-specialized individuals than for
less-specialized individuals. Therefore,by voluntarily accepting
rules and social norms associated with the activity, par-ticipants
help to ensure its continuation (Ditton et al., 1992). The third
proposi-tion states that more-specialized anglers have higher
levels of side-bets than doless-specialized anglers. Side-bets
denote when something of value (time, money,social relations) is
invested in the activity with the condition that to discontinuethe
activity could result in a loss of the investment (Alluto,
Hrebiniak, & Alonso,1973; Becker, 1960). More-specialized
individuals are proposed to have a greaterfinancial and emotional
investment in a given activity than less-specialized indi-viduals
(Proposition Two in Ditton et al., 1992).
The new proposition that we propose here states that as level of
specializa-tion in a given recreation activity increases, frequency
of participation in thatactivity will increase. We base this
proposition on the results of previous research.Graefe (1980) found
avidity to be a surrogate measure for specialization level.Schreyer
et al. (1984) similarly used number of river runs to segment river
usersinto subgroups and found significant differences between these
subgroups .Chipman and Helfrich (1988) successfully used frequency
of participation as oneelement of determining specialization level.
Finally, Ditton et al. (1992) used avidityto segment a population
of anglers into specialization subgroups, and found sig-nificant
differences between the subgroups. We would view this as
PropositionNine as added to the eight previously stated by Ditton
et al. (1992).
HypothesesBased on the previous propositions, the following
hypotheses were generated.
Ha1(a): High-specialization anglers will attach less importance
to activity-spe-cific elements of the fishing experience than will
low-specializationanglers.
Ha1(b): High-specialization anglers will attach more importance
to nonactivity-specific elements of the fishing experience than
will low-specializationanglers.
Ha2: High-specialization anglers will have a greater support for
various man-agement tools and regulations than will
low-specialization anglers.
Ha3: High-specialization anglers will have generated a greater
value of side-bets than will low-specialization anglers.
Ha4: High-specialization anglers will have a greater frequency
of participa-tion than will low-specialization anglers.
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
-
244 R. J. Salz et al.
Methods
Specialization Index
In developing our specialization index, we chose to pursue an a
priori approachthat builds on theory, and that uses theory to
generate the index items. Our spe-cialization index items,
therefore, were derived from the four characteristics
(ori-entation, experiences, relationships, and commitment) used by
Unruh (1979) toplace participants in a particular subworld (or in
our case a particular specializa-tion level). For each
characteristic, Unruh described four subworld types of
par-ticipants: strangers, tourists, regulars, and insiders (Table
1). Based on these de-scriptions, we developed four survey
questions (i.e., corresponding to the fourcharacteristics), each
containing four possible response options (i.e., correspond-ing to
four specialization levels). Question response options, consisting
of state-ments describing a participants connection to an activity
relative to that particu-lar characteristic, were ordered from
least specialized (response option = 1) tomost specialized
(response option = 4) along a 4-point scale (Table 2). It
wasexpected that for each item, the least-specialized participants
would select re-sponse option 1, and the most-specialized
participants would select response op-tion 4.
The sum of the four responses (e.g., least specialized: 1 + 1 +
1 + 1 = 4,highly specialized: 4 + 4 + 4 + 4 = 16) was then used to
locate anglers along therecreation specialization continuum. The
actual process of developing and testingthe specialization index
used for segmentation of anglers into specialization lev-els is
described in the Results section.
Data Collection
Data were collected by way of a mail survey administered to a
random sample oflicensed Massachusetts anglers. The basic survey
design and implementation fol-lowed accepted principles based on
Salant and Dillman (1994). A personalizedadvance-notice letter was
sent to all members of the sample announcing they hadbeen selected
to participate in the survey and that they would be receiving
thequestionnaire in the mail within the following week. One week
later a set of sur-vey materials was mailed to all members of the
sample. These materials includedthe questionnaire, a cover letter
describing the intent of the survey, and a self-addressed stamped
envelope for returning the completed survey. Two weeks aftermailing
the advance notice letter, a thank you/reminder postcard was mailed
to allmembers of the sample. This follow-up served to thank those
who had alreadycompleted and returned their questionnaire, and to
request a response from thosewho had not. Five weeks after mailing
the advance notice letter, a second set ofsurvey materials was sent
to those who had not yet responded. This second surveypackage was
identical to the first, except that the personalized cover letter
wasrevised to further encourage the subject to complete and return
their survey.
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
sheila.romeroResaltado
-
245Development and Validation of a Specialization Index
TABLE 2 Recreation Specialization Index Survey Questions and
ResponseOptions
Q. Please indicate your general orientation to the sport of
fishing.1) I am an outsider. I am uncomfortable when I go fishing,
and dont really feel
like I am part of the fishing scene.2) I am an observer or
irregular participant. Sometimes it is fun, entertaining, or
rewarding to go fishing.3) I am a habitual and regular
participant in the sport of fishing.4) I am an insider to the
sport. Fishing is an important part of who I am.
Q. Please indicate how you would best describe yourself during a
fishing experi-ence.
1) I am often uncertain. I am unsure about what I can or cannot
do while fishing,or how to do it.
2) I have some understanding of fishing, but I am still in the
process of learningmore about fishing. I am becoming more familiar
and comfortable with fish-ing.
3) I have become comfortable with the sport. I have regular,
routine and predict-able experiences. I have a good understanding
of what I can do while fishing,and how to do it.
4) I am a facilitator in the sport. I encourage, teach and
enhance opportunities forothers who are interested in fishing.
Q. Please indicate how you would best describe your
relationships with otheranglers.
1) Superficial. I really dont know any other anglers.2) Very
limited. I know some other anglers by sight and sometimes talk
with
them, but I dont know their names.3) One of familiarity. I know
the names of other anglers, and often speak with
them.4) Close. I have personal and close relationships with
other anglers. These friend-
ships often revolve around fishing.
Q. Please indicate how you would best describe your commitment
to fishing.1) Almost nonexistent. I am basically indifferent about
going fishing.2) Moderate commitment. I will continue to go fishing
as long as it is entertain-
ing and provides the benefits I want.3) Fairly strong
commitment. I have a sense of being a member of the activity,
and it is likely that I will continue to fish for a long time.4)
Very strong commitment. I am totally committed to fishing. I
encourage oth-
ers to go fishing and seek to ensure the activity continues into
the future.
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
-
246 R. J. Salz et al.
Testing Specialization Theory
One-way ANOVA tests were used to test for mean differences
between special-ization groups. A significance level of .10 was
used to test the null hypotheses.This level of confidence reflects
a balance between a higher probability of com-mitting a Type I
error (rejecting a true null hypothesis) and consequently
decreas-ing the probability of committing a Type II error (failure
to reject the null when itis false). Gregorie and Driver (1979)
suggest this as being a more appropriatelevel (than 0.01 or 0.05),
so later studies would not mistakenly consider some ofthe
insignificant differences as being unimportant, when in fact they
might havebeen due to the commission of a Type II error.
Results
Response Rate
A total of 1,411 questionnaires (54.6%) were returned in usable
form (Table 3).There were 312 questionnaires returned as
undeliverable by the U.S. Postal Ser-vice, 3 were returned because
the addressee was deceased, and 29 returned byrespondents were
unusable. The remainder were nonresponses.
Index Development and Internal Validation
Frequency distributions were calculated for each of the four
index items (Figure1). On a scale of responses from 1 (least
specialized) to 4 (highly special-ized), the modal response for all
four items was 3. The proportion of responsesin the
least-specialized category (i.e., response = 1) was 2% or less for
orienta-tion, experience, and commitment. The proportion of
least-specialized re-sponses (response = 1) was considerably
greater for relationships (7.3%), al-though this was still small
compared to the proportion for the other three response
TABLE 3 Status of Sport Angler Questionnaire Response
Type of response N %
Initial sample 2,930 Mortality 344
Deceased (3)Nondeliverable (312)Not-usable upon return (29)
Effective sample 2,586 100.0Nonresponse 1,175 45.4
Usable returned surveys 1,411 54.6
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
sheila.romeroResaltado
sheila.romeroResaltado
-
247Development and Validation of a Specialization Index
options. Nearly 60% of respondents chose 3 for the question
regarding experi-ence. The other variables were more evenly
distributed across responses, exceptfor the previously noted lack
of 1 responses (Figure 1).
Bivariate relationships among the items considered for inclusion
in the in-dex were then examined to determine the degree to which
the items were related(Babbie, 1995). Correlation coefficients for
the six pair-wise comparisons rangedfrom 0.41 to 0.60 and were all
statistically significant (Table 4). This middle rangesuggests that
no two items were so similar as to warrant exclusion from the
indexto avoid redundancy. Therefore, although significant positive
relationships werefound for all pair-wise comparisons, each item
measures a somewhat differentaspect of recreation specialization.
The two lowest correlation coefficients in-volved the variable
relationships (0.41 and 0.43), whereas the highest correla-tion was
between orientation and commitment (0.60).
Another way to analyze bivariate relationships is to examine the
percent ofoccurrences when two variables differ from each other by
more than a particularamount. For each of our four variables,
possible responses ranged from 1 (leastspecialized) to 4 (highly
specialized). For all pair-wise comparisons, less than9% of all
respondents had responses for any two variables that differed by
morethan one (Table 4). This further supports the strong positive
relationships betweenall items. Most of the cases where an anglers
responses for two variables diddiffer by more than one involved the
variable relationships. Pair-wise compari-sons not involving the
variable relationships differed by more than one for onlyabout 3%
of respondents.
Index item reliability was tested using Cronbachs coefficient
alpha(Cronbach, 1951). The reliability of the final multiple-item
index was measuredwith an internal consistency coefficient
(Cronbachs alpha) of 0.78. Alpha values
FIGURE 1. Distribution of angler selections of response options
according to thefour index items.
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
-
248 R. J. Salz et al.
when a particular item was deleted were 0.68 for commitment,
0.74 for experi-ence, 0.70 for orientation, and 0.76 for
relationships. This further supported theinclusion of all four
recreation specialization social world characteristics
(i.e.,orientation, commitment, experience, and relationships) in
our index.
Based on our results from the bivariate comparisons and
Cronbachs alpha,we decided to include all four items in creating
our recreation specialization in-dex. A composite specialization
rank was calculated by summing the responses tothe four items for
each respondent (Figure 2). Composite scores ranged from 4through
16. Respondents were segmented into specialization groups based
ontheir cumulative item score as follows:
If cumulative score = 46 Index Level = 1 (least specialized)If
cumulative score = 710 Index Level = 2 (moderately specialized)If
cumulative score = 1113 Index Level = 3 (very specialized)If
cumulative score = 1416 Index Level = 4 (highly specialized)
TABLE 4 Bivariate Relationships Among Index Items
Correlation % of responses differingIndex item pair coefficient
by more than one
Relationships and Experience 0.41 8.2%Relationships and
Orientation 0.43 8.9%Relationships and Commitment 0.49
7.8%Experiences and Orientation 0.48 3.0%Experiences and Commitment
0.50 3.0%Orientation and Commitment 0.60 3.0%
FIGURE 2. Distribution of anglers according to cumulative index
score.
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
-
249Development and Validation of a Specialization Index
Again, we pursued an a priori process in developing the index.
This alsoapplied to determining which item scores should correspond
to which specializa-tion levels. We chose to make the score
brackets as equal in size as possible (lev-els 1, 3, and 4 all had
a range of 3 in their score, whereas level 2 had a range of 4).The
number of anglers classified into each specialization level is the
result of thisprocess, rather than the opposite in which some
preconceived distribution of an-glers is forced into a manipulated
set of index brackets.
This process resulted in the least specialized angler group
(Level = 1) ac-counting for only 1.2% (n = 16) of all respondents
(Figure 3). Moderately spe-cialized anglers (Level = 2) accounted
for 32.5% (n = 440), very specializedanglers (Level = 3) accounted
for 42.3% (n = 572), and highly specialized an-glers (Level = 4)
accounted for 24.0% (n = 325) of all respondents.
Internal index validation is conducted to demonstrate that an
index success-fully measures what it is intended to measure
(Babbie, 1995). A method of inter-nal validation called item
analysis was conducted to examine the extent to whichour composite
index is related to (or predicts responses to) the four items
(i.e.,relationships, commitment, experience, and orientation) that
comprise it. Itemanalyses using direct comparisons were possible
because both index scores (i.e.,index level) and item scores were
based on equivalent 4-point scales ranging fromleast to highly
specialized. The index score was identical to the item score
fororientation 72% of the time, commitment 74% of the time,
experiences66% of the time, and relationships 60% of the time. For
all items, the absolutedifference between index score and item
score exceeded one for less than 3% ofrespondents. These results
support the internal validity of our specialization index.
As a final test, correlations were computed between
specialization index leveland more traditional measures of
specialization such as avidity (freshwater daysfished in past 12
months) and total years fished. The correlation between
special-
FIGURE 3. Distribution of anglers according to specialization
level.
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
-
250 R. J. Salz et al.
ization index level and freshwater days fished in past 12 months
was 0.38, whereasthe correlation between specialization index level
and years fished was 0.18. Bothwere highly significant (p <
0.0001), indicating that our specialization index cor-relates with
these unidimensional specialization indicators. However, both
corre-lations were also fairly low, suggesting that important
differences between ourindex and these unidimensional indicators do
exist.
Testing Recreation Specialization Theory
As mentioned before, our segmentation of respondents resulted in
only 16 indi-viduals (1.2%) being classified into the
least-specialized level. Because this is aninadequate sample size
for our analyses, this group was subsequently dropped forhypothesis
testing. Therefore, hypotheses were tested using only three
specializa-tion levels: Moderately specialized (M), Very
specialized (V), and Highly spe-cialized (H); levels 2, 3, and 4,
respectively.
Hypothesis One
Seven items were used to measure the importance of
activity-specific elements ofthe fishing experience. Results show
significant differences for five of these sevenmeasures (Table 5).
However, one of these items was contrary to specializationtheory
because more-specialized anglers rated the item experience of the
catchas more important than did less-specialized anglers. For two
other items, therewas no significant difference among
specialization levels. Still, four out of thefive items with
significant differences were ordered as predicted by
specializationtheory. Based on these results, the null hypothesis
that there are no differencesaccording to level of specialization
on activity-specific measures of the fishingexperience was
rejected, and we accept hypothesis Ha1(a) as stated, but
proposethat more investigation is needed regarding the items that
did not behave as pre-dicted by specialization theory.
Ten items were used to measure the importance of
nonactivity-specific ele-ments of the fishing experience. Results
show significant differences for 9 out ofthe 10 items according to
level of specialization (Table 6). It was predicted
thatmore-specialized anglers would place greater importance on
nonactivity-specificactivities than would less-specialized anglers.
Because the results are as predicted,the null hypothesis is
rejected and Ha1(b) is accepted as stated.
Hypothesis Two
Eleven items were used to measure support or opposition to
various managementregulations. The null hypothesis, which states
that there are no differences be-tween anglers in their support and
opposition to management rules, was rejectedbecause significant
differences were found for ten of the eleven items (Table 7).
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
-
251Development and Validation of a Specialization Index
The prediction that more-specialized anglers would indicate a
greater support formanagement rules than would less-specialized
anglers was supported on 9 of the10 significant items. The mean
values for one item (restricted fishing area) weredirectly opposite
of that predicted. Because 9 of the 10 significant items
wereordered as predicted, Ha2 is accepted as stated.
Hypothesis Three
Four items relating to the cost of replacing fishing equipment
were used to mea-sure side-bets. It was predicted that
more-specialized anglers would generate agreater value in side-bets
than would less-specialized anglers. Significant differ-ences
supporting this prediction were found according to specialization
level forall four items (Table 8). Therefore, we reject the null
hypothesis. Because themean differences are as predicted, we accept
Ha3 as stated.
TABLE 5 One-way ANOVA Tests for Mean Differences in Importance
ofActivity-specific Items According to Specialization Level
Level of specialization
Items* M V H F p
For the experience of 3.500** 3.818 4.128 30.29 0.000the
catch
For the sport of fishing, 3.556 3.904 4.183 26.11 0.000not to
obtain food to eat
Im just as happy if I 4.110 4.181 4.370 7.57 0.001release the
fish I catch
I am just as happy if I dont 4.053 4.158 4.329 7.55 0.001keep
the fish I catch
A fishing trip can be 3.792 3.834 4.031 5.90 0.003successful
even if nofish are caught
When I go fishing, Im just 3.095 3.034 3.111 0.67 0.510as happy
if I dont catcha fish
To obtain fish for eating, and 1.502 1.480 1.547 0.55 0.578not
for sport
*For items 1, 2, and 7 mean scores were based on responses to
the following categories;1 = Not at all important, 2 = Slightly
important, 3 = Moderately important, 4 = Veryimportant, 5 =
Extremely important. For all other items, mean scores were based
onresponses to the following categories; 1 = Strongly disagree, 2 =
Disagree, 3 = Neutral, 4= Agree, 5 = Strongly agree.**Means
underscored by same line are not significantly different (.10)
using Tukeystest.
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
sheila.romeroResaltado
sheila.romeroResaltado
-
252 R. J. Salz et al.
Hypothesis Four
Results showed significant differences on angler frequency of
participation ac-cording to level of specialization (Table 9). The
null hypothesis is therefore re-jected as stated.
Highly-specialized anglers had significantly higher rates of
par-ticipation than did moderately-specialized anglers, who in turn
had significantlyhigher rates of participation than did
lower-specialized anglers. Because this isconsistent with what was
predicted, Ha4 is accepted as stated.
Discussion
Our results provide strong support for the theory of recreation
specialization asreconceptualized by Ditton et al. (1992), and for
use of the specialization indexdeveloped here. Results from our
hypotheses tests were as predicted for an over-whelming majority of
the items we investigated. Our study also strongly supportsthe
inclusion of all four characteristics of social worlds (commitment,
orienta-
TABLE 6 One-way ANOVA Tests for Mean Differences in Importance
ofNonactivity-specific Items According to Specialization Level
Level of specialization
Items* M V H F p
To experience adventure 3.405** 3.732 4.009 28.77 0.000and
excitement
To be close to the water 3.366 3.576 3.973 21.20 0.000For
relaxation 4.218 4.345 4.559 16.48 0.000To be with friends 3.107
3.206 3.559 13.21 0.000To experience natural 4.134 4.248 4.453
12.82 0.000
surroundingsTo experience new and 2.842 2.939 3.279 12.35
0.000
different thingsTo get away from the 3.409 3.474 3.842 10.49
0.000
demands of other peopleTo be outdoors 4.177 4.236 4.450 10.44
0.000To get away from 3.800 3.912 4.159 9.78 0.000
the regular routineFor family recreation 3.279 3.136 3.256 1.72
0.179
*Mean scores were based on responses to the following
categories; 1 = Not at all impor-tant, 2 = Slightly important, 3 =
Moderately important, 4 = Very important, 5 =
Extremelyimportant.**Means underscored by same line are not
significantly different (.10) using Tukeystest.
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
-
253Development and Validation of a Specialization Index
tions, experience, and relationships) as related and reliable
measures of recreationspecialization.
Specialization Index Development
There are several possible explanations for the fact that the
least specializedsubworld made up such a small proportion of our
sample (only 1.2%). First, weshould not rule out the possibility
that this group may, in fact, be much smaller insize than the other
groups. This would be the case if the learning curve from
leastspecialized to moderately specialized requires a relatively
short time period.Because our survey was administered to those
people who had purchased licensesduring the previous year, anglers
who were least specialized at the time of li-cense purchase had at
least a full fishing season to increase their specializationlevel
prior to receiving our survey. Another possible explanation is that
our sampledid not tap into those groups of anglers that make up the
majority of the leastspecialized group. For example, children
(under 17 years old), out-of-state an-glers, and 3-day license
holders were not part of our survey population. One mightreasonably
expect these anglers to be among the least specialized. We
consider
TABLE 7 One-way ANOVA Tests for Mean Differences in Support
andOpposition of Management Regulation Items According to
Specialization Level
Level of specialization
Items* M V H F p
Creel limit 4.109** 4.293 4.463 14.79 0.000No stocking allowed
3.559 3.673 3.935 13.70 0.000Maximum size 3.284 3.547 3.733 13.64
0.000Stock non-native fish 3.009 3.282 3.343 11.13 0.000Minimum
size limit 4.108 4.211 4.433 8.62 0.000Restricted fishing area
3.434 3.260 3.069 7.23 0.001Mandatory catch 3.122 3.162 3.439 6.67
0.001
and releaseStock native fish 4.219 4.336 4.403 6.03 0.003Slot
limit 3.117 3.190 3.388 5.86 0.003Voluntary catch and 3.874a 4.028b
4.022a,b 2.83 0.059
releaseProhibit use of certain 3.612 3.545 3.581 0.43 0.653
gear
*Mean scores were based on responses to the following
categories; 1 = Strongly oppose,2 = Oppose, 3 = Neutral, 4 =
Support, 5 = Strongly support.**Means underscored by same line or
same superscript are not significantly different(.10) using Tukeys
test.
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
-
254 R. J. Salz et al.
these explanations to be the most likely reasons for the small
size of the leastspecialized group.
Nonresponse bias could also be a possible explanation if the
probability ofan angler returning our survey was positively
correlated to the anglers specializa-tion level. However, in a
study of nonresponse bias on angler surveys, Fisher(1996) found
that species preferences and scores from summated Likert scaleswere
independent of response probabilities. Finally, the choice of words
we usedfor the least specialized response options could explain the
low percent of re-spondents selecting those options. Anglers may
have felt embarrassed to identifythemselves with words such as
outsider, uncomfortable, unsure or uncer-tain, all of which may
have strong negative connotations. Our results suggestthat least
specialized subworlds may be more difficult to sample for a variety
ofreasons. A special sample design may be needed in certain
situations to adequatelyaddress this group.
Our results showed that although all four social world
characteristics (rela-tionships, orientation, experience, and
commitment) should be included in theindex, the relationships
dimension behaved somewhat differently from the otherthree.
Specifically, some anglers scored least specialized for
relationships butwere in the middle-to-high range of specialization
for the other three dimensions.This suggests that for the activity
of freshwater fishing, having personal relation-
TABLE 8 One-way ANOVA Tests for Mean Differences in the Cost of
ReplacingFishing Equipment with Similar Equipment Between
Specialization Level
Level of specialization
Items* M V H F p
Replace reels $119.33* $229.49 $455.80 90.00 0.000Replace tackle
114.80 282.28 579.84 78.65 0.000Replace rods 138.31 284.52 555.28
38.18 0.000Replace electronic 262.00 436.65 580.42 6.95 0.001
equipment
*Means underscored by same line are not significantly different
(.10) using Tukeys test.
TABLE 9 One-way ANOVA Tests for Mean Differences in Frequency
ofParticipation According to Specialization Level
Level of specialization
Items* M V H F p
Mean total days 15.566* 36.656 56.609 105.54 0.000fished
*Means underscored by same line are not significantly different
(.10) using Tukeys test.
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
-
255Development and Validation of a Specialization Index
ships with other anglers may not be as important of a component
when advancingto higher specialization levels as previously
thought. Although interaction andcommunication relate to social
world boundaries (Unruh, 1980), in todays worldthese can be readily
achieved through mediated channels instead of personal con-tact.
Some highly specialized anglers may rely on journals, magazines,
cable tele-vision, and the Internet to acquire and exchange
information about fishing. If so,our question measuring
relationships, which focuses only on personal contacts,may have to
be expanded to include a wider range of interactive and
communica-tive possibilities.
The characteristics included in our index were derived directly
from the so-cial worlds literature. Still, the question of which
specific measures should beused to define specific characteristics
of a specialization index is open to interpre-tation (Kuentzel
& McDonald, 1992). For example, commitment to an activityhas
been measured as the number of related magazines one subscribes to
(Bloch,Black, & Lichtenstein, 1989), the level of activity
involvement (Williams &Huffman, 1986), the centrality of the
activity to ones lifestyle (Chipman &Helfrich, 1988), the
number of side-bets invested in, and an affective attach-ment to
the activity (Buchanan, 1985). Similarly, one could come up with
mul-tiple ways to define and measure orientation, experience, and
relationshipsrelated to a particular activity.
Specialization dimensions can also be measured using either
behavioral orcognitive measures. One of the main features of social
world involvement is vol-untary identification, meaning one chooses
to become a member of a social worldrather than it being a
requirement (Unruh, 1980). The necessity of voluntary
iden-tification suggests a strong cognitive component to entry into
a social world andmovement between subworlds within that social
world. This cognitive compo-nent is reflected in the questions we
used in this study to measure specializationdimensions. For
example, rather than measure commitment through other vari-ables as
described above, anglers were asked directly to choose the
statementsthat best describe their involvement in the sport.
Approaching specialization from a social worlds perspective may
add sub-jectivity to the index because words like commitment,
insider, and orienta-tion can mean different things to different
people. However, this subjectivitydoes not necessarily bias the
segmentation process, but rather, it may redefinespecialization in
a new way. The assumption that a specialization index derivedfrom
objective measures (i.e., gear used, days fished, magazines
purchased) ispreferable to one that uses more subjective, cognitive
measures should not auto-matically be made. The decision of which
index to use should, perhaps, be basedon the goals of the
particular study and the research purpose or management
ap-plication it is intended for. The general lack of consistency in
measuring special-ization in the outdoor recreation literature
supports this contention. For futurestudy, it would be interesting
to compare participant segmentation using our in-dex with previous
specialization indices using the same survey population.
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
-
256 R. J. Salz et al.
Testing Recreation Specialization Theory
Results indicated that more-specialized anglers were more
interested in a qualita-tive experience, whereas less-specialized
anglers had a more simplistic view offishing that did not consider
other intrinsic elements of the experience to be quiteas important.
This supports specialization theory and further reconfirms the
re-sults of Ditton et al. (1992).
Frequency of participation was also shown to increase as
specialization lev-els increase. Our results are consistent with
previous work and justify the additionof our proposed Proposition
Nine. Individuals are likely to increase their fre-quency of
participation when they feel some sort of attachment to an
activity. Asspecialization level increases, alternative activities
will be rejected as the commit-ment to participating in the primary
activity increases (Buchanan, 1985; Unruh,1979).
It appears that more-specialized anglers are more receptive to
managementregulations than are less-specialized anglers. The
support for management regu-lations was shown to increase as
specialization increases. The former group ismore likely to be
impacted than the latter group if fishing activities were
discon-tinued; therefore, as predicted from specialization theory,
the former would bemore supportive of rules and regulations issued
from fisheries managementagencies.
Finally, as predicted, side-bets anglers appropriated for
fishing equipmentwere shown to increase as level of specialization
increased. Because of a greaterinvolvement within the activity,
more-specialized anglers will commit greater fi-nancial costs
towards fishing than will less-specialized anglers.
Management Implications
There is potential here for fisheries managers to gain an
understanding of groupdifferences on a variety of issues to
efficiently improve services already provided.By developing and
promoting services based on some aggregation of anglers,
theinterests of many anglers are ignored. Managers may then be
confronted with afairness issue, where some anglers perceive that
resources are allocated unfairly.Segmentation by specialization
recognizes that different groups have differentattributes that
require different marketing schemes. Through a better
understand-ing of the angling constituency, managers can avoid
making resource allocationdecisions that may result in the loss of
credibility for the fisheries agency (Ditton,1996; Loomis &
Ditton, 1993). The results of this study provide strong supportfor
the use of a multidimensional index as a means of classifying
participants intohomogeneous groups, based on the recreation
specialization theory developed byDitton et al. (1992). Such
insight to anglers can also be used to effectively evalu-ate
current management objectives and services.
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
sheila.romeroResaltado
-
257Development and Validation of a Specialization Index
References
Alluto, J. A., Hrebiniak, L. G., & Alonso, R. C. (1973). On
operationalizing the conceptof commitment. Social Forces, 51,
448454.
Babbie, E. (1995). The practice of social research. Belmont, CA:
Wadsworth PublishingCompany.
Becker, H. S. (1960). Notes on the concept of commitment.
American Journal of Sociol-ogy, 66, 3240.
Bloch, P. H., Black, W. C., & Lichtenstein, D. (1989).
Involvement with the equipmentcomponent of sport: Links to
recreational commitment. Leisure Sciences, 11, 187200.
Bryan, H. (1977). Leisure value systems and recreational
specialization: The case of troutfishermen. Journal of Leisure
Research, 9(3), 174187.
Buchanan, T. (1985). Commitment and leisure behavior: A
theoretical perspective. Lei-sure Sciences, 7, 401420.
Chipman, B. D., & Helfrich, L. A. (1988). Recreational
specialization and motivations ofVirginia river anglers. North
American Journal of Fisheries Management, 8, 390398.
Cronbach, L. J. (1951). Coefficient alpha and the internal
structure of tests. Psychometrika,16, 297334.
Ditton, R. B. (1996). Understanding the diversity among
largemouth bass anglers. Ameri-can Fisheries Society Symposium, 16,
135144.
Ditton, R. B., Loomis, D. K., & Choi, S. (1992). Recreation
specialization:Reconceptualization from a social worlds
perspective. Journal of Leisure Research,24(1), 3551.
Fedler, A. J., & Ditton, R. B. (1986). A framework for
understanding the consumptiveorientation of recreational fishermen.
Environmental Management, 10(2), 221227.
Fisher, M. R. (1996). Estimating the effect of nonresponse bias
on angler surveys. Trans-action of the American Fisheries Society,
125, 118126.
Graefe, A. R. (1980). The relationship between level of
participation and selected as-pects of specialization in
recreational fishing. Doctoral dissertation. Texas
A&MUniversity, College Station.
Gregorie, T. G., & Driver, B. L. (1987). Type II errors in
leisure research. Journal ofLeisure Research, 19(4), 261272.
Hammitt, W. E., & McDonald, C. D. (1983). Past on-site
experience and its relationshipto managing river recreation
resources. Forest Science, 29, 262266.
Kauffman, R. B., & Graefe, A. R. (1984). Canoeing
specialization, expected rewards, andresource related attitudes. In
J. S. Popodic, D. I. Butterfield, D. H. Anderson, & M.R.
Popodic (Eds.), National River Recreation Symposium Proceedings
(pp. 629641). Baton Rouge, LA: Louisiana State University.
Kuentzel, W. F., & McDonald, C. D. (1992). Differential
effects of past experience, com-mitment, and lifestyle dimensions
of river use specialization. Journal of LeisureResearch, 24,
269287.
Loomis, D. K., & Ditton, R. B. (1993). Distributive justice
in fisheries management.Fisheries, 18, 1418.
Salant, P., & Dillman, D. A. (1994). How to conduct your own
survey. New York: Wiley.
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012
-
258 R. J. Salz et al.
Schreyer, R. M., Lime, D. W., & Williams, D. R. (1984).
Characterizing the influence ofpast experience on recreation
behavior. Journal of Leisure Research, 16, 3450.
Unruh, D. R. (1979). Characteristics and types of participation
in social worlds. SymbolicInteraction, 2, 115130.
Unruh, D. R. (1980). The nature of social worlds. Pacific
Sociological Review, 23(3),271296.
Vaske, J. J., Donnelly, M. P., & Heberlein, T. A. (1978).
Perceptions of crowding andresource quality by early and more
recent visitors. Leisure Sciences, 3, 367381.
Virden, R. J., & Schreyer, R. M. (1988). Recreation
specialization as an indicator ofenvironmental preference.
Environment and Behavior, 20, 721739.
Wellman, J. D., Roggenbuck, J. W., & Smith, A. C. (1982).
Recreation specialization andnorms of depreciative behavior among
canoeists. Journal of Leisure Research, 14,323340.
Williams, D. R., & Huffman, M. G. (1986). Recreation
specialization as a factor inbackcountry trail choice. In R. Lucas
(Ed.) Proceeding of the National WildernessResearch Conference
Current Research, (General Technical Report INT-211, pp.339344),
Ogden, UT: USDA Forest Service Intermountain Research Station.
Dow
nloa
ded
by [F
ac Ps
icolog
ia/Bi
bliote
ca] a
t 00:2
0 25 S
eptem
ber 2
012