‘13". 7‘14
"‘32:?
..

, i 7 ...; ~«2.3.. ‘.’. 5 3L U1“ .L
y -..?2. _ 2 1 ,. 3- 1 .- ' ' ,"V . ’ " ” ~ 1 -‘ "‘ '1‘ '2‘?“ 71: W
:.m ,. 2 . < r: ' v - , - , ‘v . 0::- '“' ‘.‘ y .Cvu “‘1 "‘ M
V . _.,._. .‘ , , ': -. - , - ~~ .. -. ~‘r. -. ....» ~ ‘ ’ V “3&7”:
' a»- It - >~ ' , - ' " g, ' “ ‘ V. V “‘3’ ‘:“.*1a.1:‘::«' - ~'-‘ ' :7 ‘53,. §m"‘?fim“‘m»‘:‘“% .3:
H. v L . _. 7. . 4 . ' .. .‘ ' . - ‘ _ .. . a A . m ‘
: r _ ", V ‘ . .‘ L , --‘ 2. '5 , . k t . *r‘“. ....2 . 33 ‘ . m 2». w “ 31:3... ...;_«--
. .. L .. . ,5 a . - ‘-- L , : ~— ~ ‘14 SW: * ~- . ~.
7 > . 2 7-.....“

 

u. w \- uhl--.‘r£~|.‘~.-ko

w
. ‘ .. ., -..... 1:222.
1:413? 3&égfifi2‘mw _ ”.sz

xiluvrui u
... . r- .,
3.‘ s“,

.. «I Wt.” x. ..
, A ’ ‘ :1 ‘ ~ 1 . .g:z.~~:r.=1~3:‘33§ 1%
_ . . v , aluuun .1- _~, 1‘ "
’ ‘ ,- ' . '. ‘ . 4 ,, -'. ”q' "V333"; t'?‘ .5‘ ”3., , : 4.13555}: \m‘; shtmfi"; .- “*‘am
- A ~' , u - . .-
. . L ).. _ _ 7 - 7-5 . , .. M $3...-.’.;.‘::.‘.““ ‘12... "r “flaw-“
.

. . ‘4‘. hug-Lu
wan-.2531:
..

I4

*h“.
ex "3"“...7312...

 

r.

.- h .

.- . ‘ A. ....V W
x V .L . ,, f. w“ '7 - - * '.. '

J: :x .32.“, - .- , . ‘ ”If": Jan:-
.

,.:;::I:"“‘“

. § ...:

l-v ,.

u. u” 5'. “l""- =

I . , ‘2‘ ~. ~. ¢ JN~ .

~ , " ~ . ‘ ‘ .. “'r. L ~ ‘ ' L T1."‘L2..'"“~.. ““1: grrnlt‘h --

‘ 4' «r : . ‘ ,. Efi.:..'.;. ”1.2. -. has

... .. .....- U ~ su.~u~—~~\
at? L Ina-5‘
*1“ . ; b «:2- n: “by“

5. 2:.
’ ~“_3";IE“~3~:~-"‘

 

   

n

3
.... - * re
. LurBa ”Labia.
. .LHLu
2 ...- ,

i
1

Ana

 

 

2:7
J.
‘5
r:
.
g . u
, 4.7 ”J...” 41):.31 2 .,
‘1. x.:..wl-‘I.Uuuw fl.
1“— fluJM-fl‘ NJ ‘3
mm n’. .t m
' k 9L 2 1 {5:53
" )— ' 1 ' . ~ 3 I" , r _ .. a ‘ ‘
.1. , > A H .« E ._ ...En- " , J'LKA. :wia ‘3:
.L ‘ . « A r ' . ' .> L. an « ... «L4. - ~ - - . 2 "9-24: §<kuwfi u
._;> v. . u .7 . .V - v . ‘ w _ 5: .v '4“ A .q {‘ta-;#(:u’}m.f film-y
, , . .. »- L, , ‘- - 2 a; 2 xvi-hub. -:.; 13“ ”vi“— L "4" “"3.”
. . 2‘: - L. 4-4. N 'Lr, e . ' an ”El-f, «M
- _ ' J~“1"' v~~’ "- '| M ’ ' ‘WflW Av ' ' “ "w“ {’8‘
. . ~ . . .‘ "I""" ,1 u.
. ,
f
“u
.
K
5
.
2'
.
l
§
5 '.

 

Au..-

 

 

 

 

 

 

 

   

It. a'
:

x
4 n
I.
l
‘
' 1
f .
_ .
. .7 _
f ’13: -7 Iw
j-vw v "41'; vmp 1"“un r.- ~ '-
.
J ’: ' . .' [.1 ‘ "I "’l 'n {33"
’ Kw 4" V: ." l "‘V I I" 'U
‘ ..u- ....A; . ... ..If...’n.. ,.
a ' r y ..,f, ‘. "” ‘ if, ‘ ' ‘ ”2‘3‘;'“""':!"“’"'
". . ‘y
t u 'p ' V “Afr
‘ L . ..
h v - ..v. . J_
... nab. ‘ ~ ' 7 ”“qu;

.1 . . . . . . ..a ‘ - ’ ‘- .. . ' «2t: [NV-r1:
. 7 . :- . L ., _ Z.

. z ‘ . . r ' < r. .(

I l"l:‘“ ':'

 

tux-7* rm

 

“:41.

 

. .

“flu/.5“ .5. J2. (... L‘

. 1' n: ”I: ' J «7 "a uFYKt’JI'F24”. .
L' A ‘. . ‘ ' 9- -‘1w- rm

     

 

This is to certify that the

thesis entitled

Regional Analysis of 'Iburism Potential:
An Application to Michigan Counties

presented by

Petra Goebel

has been aecepted towards fulfillment
of the requirements for

Master of—Science__ degree in .P__ark and Recreation
‘ Resources

/ , é?

Major ofessor

Date 9-26—1991

0-7639 . MS U is an Affirmative Action/Equal Opportunity Institution

 

UBRAR‘!
Michigan 51.
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE I

       

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MAY 05 2003
(l 5 0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

—T l

MSU Is An Affirmative Action/Equal Opportunity Institution
cmmut

 

 

 

 

REGIONAL ANALYSIS OF TOURISM POTENTIAL:

AN APPLICATION TO MICHIGAN COUNTIES

BY
Petra Gébel

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Park and Recreation Resources

1991

ABSTRACT

REGIONAL ANALYSIS OF TOURISM POTENTIAL:'
AN APPLICATION TO MICHIGAN COUNTIES

BY
Petra Gébel

The analysis of tourism regions identifies important
variations in the ability of counties to attract and service
tourists. Therefore, regional analysis is a useful tool in
tourism development, planning, and evaluation.

The purpose of the present study was (1) to develop a
model that predicts tourism potential of Michigan counties,
(2) to explore relationships between measures of tourism
potential and tourism components at the county-level, and
(3) to describe spatial variations in tourism potential and
tourism components across Michigan counties.

A regression analysis along with a spatial analysis
were performed using secondary data. Per capita lodging
supply and per capita hotel/motel use tax were used as
measures of tourism potential and served as two distinct
dependent variables. Results suggest that tourism in
Michigan is concentrated in the Upper Peninsula and the
northern Lower Peninsula. Natural resources, especially
miles of Great Lakes shoreline accounted for most tourism

potential.

TABLE OF CONTENTS

LIST OF TABI‘ES O O O O O O O O O O O O O 0

LIST OF FIGURES O O O O O O O O O O O O 0

Chapter

I.

II.

III.

INTRODUCTION . . . . . . . . . .

Significance of Tourism . . . . .
Tourism in Michigan . . . . . . .
Problem Statement . . . . . . . .

Study Applications . . . . .
Study Objectives . . . . . .
Definitions . . . . . . . .

LITERATURE REVIEW . . . . . . . .

Tourist Decision Making Process .
Tourism Components . . . . . .

Characteristics of Tourism in Michigan

Accommodations Use as a Tourism Performance

Measure . .

Population and the Relative Importance

of Tourism . . . . . . . . .
Regional Analysis . . . . . . . .
Destination Images . . . . . . .
Importance of Tourism Components

Conclusions from the Literature Review

Hypotheses . . . . . . . . .
ANALYSIS TECHNIQUES AND PROCEDURES

Research Variables . . . . . . .
Regression Analysis . . . . . . .

iii

Page
vi

vii

13

15
16
20
21
23

24
26

26
27

Chapter

IV.

VI.

VARIABLE MEASUREMENT . . . . . . . . . .
Dependent Variables . . . . . . . . . . .
Lodging Supply . . . . . . . . . . .

Supply of Guest Rooms . . . . .
Supply of Campsites . . . . . .

Hotel/Motel Use Tax . . . . . . . .
County Population . . . . . . . . .

Spatial Distribution of Per Capita

Figures . . . . . . . . . . . .
Independent Variables . . . . . . . . . .

Public Lands Ratio . . .
National and State Parks
Water Resources . . . .
Attractions . .
Golf Courses . .
Highways . . . .
Ski Areas . . .

Limitations of the Study . . . . . . . .
FINDINGS AND DISCUSSION . . . . . . . . .

Correlation Analysis . . . .
Attraction Index Development
Division of the Data Set . .
First Regression Analysis . .
Analysis of Influential Cases
Second Regression Analysis .
Analysis of Residuals . . . .
Test of Regression Results . . . .
Comparison of Tourism Potential Measur
with Images of Michigan Tourism
Destinations . . . . . . . . . . . .
Discussion of Results . . . . . . . . . .

e

CONCI’US ION O O O O O O O O O O O O O O 0

Summary of the Study . . . . . . . . . .
Conclusion . . . . . . . . . . . . . . .
Implications for Tourism Planning and
Management . . . . . . . . . . . . .
Future Research . . . . . . . . . . . . .

iv

Page
32
34
34

35
38

4O
43

45
48

48
51
52
55
58
59
60

61
63

63
65
67
67
69
69
73
76

79
82
85

85
87

88
9O

BIBLIOGRAPHY . . . . . .
APPENDICES . . . . . . .
Appendix

A. Data . . . .

B. Maps on Spatia

Variable .

Distr

ibution 0

Each

Page
92

97

97

101

LIST OF TABLES

Tourism Components

Operators of Campgrounds and Lodging

Establishments .

Trend in Unadjusted Hotel/Motel Use Tax
Receipts in Michigan (1985 - 1989)

Variable Description . . . . . .

Correlation Matrix of Study Variables .

Correlation Matrix Including Attraction

Index . . . . . .

First Regression:

Capita and Tourism Components . .

First Regression:

Capita and Tourism Components . .

Second Regression:
Capita and Tourism Components . .

Second Regression:
Capita and Tourism Components . .

Test of Regression Results for Per Capita

Hotel/Motel Use Tax Per

Lodging Supply Per

Hotel/Motel Use Tax . . . . . . .

Test of Regression Results for Per Capita

Lodging Supply .

Dependent Variables . . . . . . .

Independent Variables . . . . . .

vi

Lodging Supply Per

Hotel/Motel Use Tax Per

page

11

37

41
47

64

67

68

68

70

7O

78

78
97

99

8.10

3.11

B.12

LIST OF FIGURES

Counties and Basic Regions in Michigan .

Residuals for Per Capita Hotel/Motel
use Tax 0 O O O O O O O O O O O O I O O O

Residuals for Per Capita Lodging Supply .
Image Scores of Michigan Counties . . . .

Spatial Variations in the Location of
Guest Rooms . . . . . . . . . . . . . . .

Spatial Variations in the Location of
Campsites . . . . . .

Spatial Variation in Hotel/Motel Use Tax
Spatial Variation in County Population .

Spatial Variation in Per Capita Lodging
Supply . . . . . . . . . . . . . . . . .

Spatial Variation in Per Capita Hotel/Motel

use Tax 0 O O O I O O O O I O O O O O O 0
Spatial Variation in Public Lands Ratio .

Spatial Variation in the Location of
National and State Parks . . . . . . . .

Spatial Variation in Miles of Stream . .
Spatial Variation in Miles of Shoreline .
Spatial Variation in Acres of Lakes . . .

Spatial Variation in the Number of Public
Water Access Points . . . . . . . . . . .

Spatial Variation in the Attraction Index
score 0 O O O O O O O O O O O O O O O O I

vii

Page

33

74
75

81

101

102
103

104

105

106

107

108
109
110

111

112

113

Page
Spatial Variation in the Number of Golf
courses I O O O O O O C O O O O O O O O O 114

Spatial Variation in the Existence of
Highways O O O O O O O O O O O O O O O O 1 1 5

Spatial Variation in the Number of Ski
Areas 0 O O I O O O O O O O O I O O O O O 116

viii

CHAPTER I

INTRODUCTION

Significance of Tourism

Tourism is a unique and special kind of service
product. The concept of tourism according to Burkhart and
Medlin (1981) involves five main characteristics. First,
tourism arises from the movement of people and their main
stay in various destinations. Second, there are three
distinguishable elements involved in tourism: travel to the
destination, the stay, and activities at the destination.
Third, the journey and the stay take place outside the
normal place of residence and work, so that tourism gives
rise to activities which are distinct from those of the
resident and working populations. Fourth, the movement to
destinations is of a temporary, short-term character, with
the tourist's intention being to return within a few days,
weeks, or months. Lastly, destinations are visited for
purposes other than taking up permanent residence or
employment. This definition does not include business

travelers.

2

These characteristics lead to the following
implications: Tourism is an invisible export industry; it
is an intangible good that cannot be stored and involves the
participation of the customer or tourist. In other words,
the cyclical patterns of demand for tourism goods and
services have obvious implications for employment and
investment. Moreover, tourism is a fragmented industry. It
is subject to seasonal variations, unpredictable influences
from external forces, the heterogeneous nature of tourism
motivations, expectations and images, and it is highly
elastic with respect to both price and income. These
factors promote a low level of customer loyalty with respect
to destinations, modes of travel and accommodations.
Therefore, tourism is a powerful tool for economic, social,
and physical change within a community (Mathieson and Wall,
1982). Types of tourism facilities include parks, hotels,
campgrounds, resorts, ski areas, restaurants, and
entertainment centers. In Michigan, tourism includes an
abundance of activities such as boating, fishing, swimming,
skiing, sightseeing, and shopping. In addition, tourism
includes many different providers, such as facility owners,
travel agencies, transportation companies, hospitality
services, and tour operators. Thus, tourism is a complex,
comprehensive, and dynamic phenomenon (Mathiesen and Wall,

1982).

3
Tourism in Michigan
Tourism is an important industry in Michigan,
significantly contributing to the state's economy. With a
decreasing automobile industry, recreation, tourism and
related businesses are a major source of sustained economic
growth and job creation. Michigan has many advantages in
the competition for tourists compared to other midwestern
states. It has a large population base, cultural centers,
attractions and scenic areas, particularly along its
thousands of miles of lake shoreline that are unmatched in
the Midwest (Hudson Institute, 1985). Michigan has seasonal
and year-round recreational sites in all parts of the state.
Unfortunately, many of these sites are minimally developed
and underused (Public Sector Consultants Inc., 1987). These
areas being within a six-hour drive for more than 25 million
people, there is the potential for substantial growth of a

broad range of facilities to serve the tourist.

Problem Statement
As O'Halloran (1988) indicates, the majority of
professionals rely on the state for tourism data; therefore
they use data in an aggregated form which is not specific to
their respective counties. This reliance on statewide
rather than county-specific data can lead to misleading
recommendations. Since counties are competing for tourism

dollars, a better understanding of regional tourism is

4
necessary. To understand tourism regions made up of
counties, the question needs to be asked: What are the
characteristics of a successful tourism county?

An analysis of tourism regions identifies important
regional variations not only in the ability of places to
attract and service tourists but also in the ability to
benefit from tourists. Therefore, regional analysis is a
useful tool to assist with product development, tourism
planning, and policy evaluation.

The nature of this study is descriptive and
exploratory. The intent is to document relationships that
may exist between county-level development variables and
tourism potential. Two different measures of tourism
potential will be examined and their usefulness in
explaining spatial differences in tourism potential will be
discussed. Major county-level variables include:
hotel/motel use tax, number of guest rooms and campgrounds,
population, miles of Great Lakes shoreline and of streams,
acres of lakes, public access to water resources, percentage
of public land, acres of National and State Parks, presence
of freeways, number of ski areas and golf courses, and,
finally, the number of entertainment facilities.

A regression model will be developed which interrelates
these variables. This analysis will be accompanied by a
spatial analysis comparing the results of the regression

model with a visual representation of the spatial patterns

5
of key variables. This comparison will assist in
interpreting the regression analysis findings.

The study takes advantage of secondary data already
available. The data has been provided by the Travel Tourism
and Recreation Research Center and the Center for
Redevelopment of Industrialized States at Michigan State
University, the Michigan Travel Bureau, and the Michigan
Department of Transportation.

Specifically, the purpose of this research project is
to answer the following questions: "Is tourism potential in
Michigan counties related to specific county level tourism
attributes?", "Can a model be developed which is useful in
explaining spatial differences between counties?" and,
"Which attributes are the most important in explaining
tourism potential?", meaning, "What attributes make a county

a successful tourism destination?"

W

This analysis will aid in recognizing tourism
potentials and development strategies. For example, the
Michigan Travel Bureau could use the results to better focus
its efforts in the "Say Yes to Michigan" campaign. Regional
tourism associations might use the findings to further
develop tourism promotion in their areas, and Chambers of
Commerce might find applications for this research in their

discussion of development issues in their own communities.

6
In general, the findings could assist any agency determining
potential tourist markets, planning long term touristic
developments in a region, or developing guidelines to
improve or strengthen the success of a tourism destination

(Hunt, 1975).

Study ijegtiyeg

1. Develop a model that predicts a county's potential for
tourism:

2. Explore relationships between measures of tourism
potential and identified tourism components: and

3. Describe spatial variations of tourism potential and

tourism components across Michigan counties.

Definitions
Tourism Components: Destination attributes and
characteristics that play a vital role in the character
of a tourism region. They may also be called
destination features. Examples are lakes, mountains,
beaches, marinas, museums, or restaurants.
Tourism Potential: The ability of a county or region to

attract tourists.

CHAPTER II

LITERATURE REVIEW

In this chapter selected literature topics related to
this study are reviewed. These topics include: (1) tourist
decision-making process, (2) tourism components, (3)
characteristics of tourism in Michigan, (4) accommodations
use and tourism potential, (5) population and the relative
importance of tourism, (6) regional analysis, (7)
destination images, and (8) dominance of tourism components.
The first section of this chapter reviews the tourist
decision-making process and provides an explanation of how
this study fits into the overall concept of tourism. In the
next section, tourism components are identified from which
the independent variables in this study are derived. The
third section describes the characteristics of tourism in
Michigan. The fourth and fifth section are related to the
choice of the dependent variables. The sixth section covers
important studies in the field of regional analysis followed
by a section on tourism destination images as they are

related to the study. Finally, the findings of related

8
studies are reviewed to suggest the dominance of alternative

tourism components in explaining the model.

Tourist Decision Making Process

To better understand the implications and limitations
of this study, it is important to understand the tourist
decision-making process and to realize how the variables
under discussion fit into this concept. Mathieson and Wall
(1982) define five principle decision-making phases. In
phase one, there is a felt need or travel desire; reasons
for and against meeting that desire are weighted. The
second phase involves information collection and evaluation
through advertisement and word-of—mouth recommendations.
The third phase includes the actual travel decision where
choices are made regarding destination, mode of travel, type
of accommodation and activities. Phase four involves travel
preparations and experiences. In the final phase, travel
satisfactions and evaluations are established which will
influence future travel decisions of individuals. Factors
influencing the decision-making process include attributes
and characteristics of destinations as well as past
experiences. Each of these factors is highly interrelated,
which makes the process even more complicated (Mathieson and
Wall, 1982).

The outcome of these five decision-making phases is

greatly influenced and modified by tourist profiles, travel

awareness, and trip features. The tourist profile describes
socio-economic and behavioral characteristics of tourists.
Motivations, attitudes, needs and values of tourists are of
crucial importance in contributing to and influencing the
decision—making process (Mathieson and Wall, 1982).

Potential tourists may be motivated to travel; however,
unless they are informed of available opportunities, they
might not travel at all. Travel awareness depends on the
availability of information and the credibility of these
information sources. Generally, there are formal and
informal sources of information, formal sources including
all means of promotion and informal sources referring to
comments and recommendations by relatives, friends and other
travelers.

Trip features include such factors as distance,
duration of stay, time constraints, trip cost and value for
price, party size, safety at the destination and confidence
in travel arrangements and travel intermediaries.

Tourism7components play a vital role in the assessment
of alternatives and in the final choice of the destination.
These components will be the focus of this study. To what
degree and in which combination are tourism components
related to the tourism potential of a destination? However,
as this discussion shows, tourism components are only one of
many factors influencing the individual tourist's decision

regarding when, how and where to travel.

10
Tourism Components
Many authors take different tourism components into
account as illustrated in the following section. McIntosh
and Goeldner (1984) identify natural resources,
infrastructure, superstructure, transportation and
hospitality resources. Burkhart and Medlik (1981) discuss
attractions, accessibility, and amenities. Robinson (1976)
identifies good weather, scenery, amenities, historical and
cultural factors, accessibilities and accommodations. Each
of these authors mention more or less the same components,

they simply label them differently (see Table 2.1).

Characteristics of Tourism in Michigan

Holecek (1991), summarizes research on travel and
tourism in Michigan and characterizes pleasure travel to and
through Michigan. Based upon his review of available data,
he draws the conclusion that Michigan's prime travel market
-- drawing 80 percent of its overnight trips -- includes
Illinois, Indiana, Michigan itself, Minnesota, Ohio, and
Wisconsin. That means there are fifty million potential
consumers. About half of the overnight trips occur on
weekends, primarily in July, August, and September (40
percent of all overnight trips). The large majority (80
percent) travels by car, truck, or recreational vehicle for
an average length of stay of approximately four nights.

Forty-five percent of pleasure trips were taken to visit

Table 2.1:

11

Tourism Components

McIntosh and
Goeldner

Burkhart and
Medlik

Robinson

Natural Resource

Infrastructure

Superstructure

Transportation
Hospitality
Resources

Attractions

Accessibility

Amenities

Good Weather
Scenery
Amenities

Historical and

Cultural Factor

Accessibility

Accommodations

landform, fauna,
beaches, climate
communication, roads,
parking, marinas,

bus stations,...
resorts, hotels,
restaurants, malls,
entertainment places,
transportation equipment
attitude of residents,
festivals, history,
dancing, museums..

flora,

climate, scenic beauty,
history, events,
exhibitions,sports,...
distance from major
cities, external
transportation and
communication systems
accommodation, catering,
entertainment, internal
transportation and
communication

lakes, mountains, forests
all facilities for
recreation and amusement
ruins, castle, festivals,
cathedrals, art galleries
malls...

closeness to major cities
and easy accessibility
food and lodging

Source:

McIntosh and Goeldner (1984), Burkhart and Medlik

(1981) and Robinson (1976).

12

friends and relatives, 29 percent had the primary purpose of
outdoor recreation, and 20 percent were mostly for
sightseeing/tours and special attractions. Forty-three
percent of the overnight trips involved staying over at
friends and relatives, 30 percent involved lodging in hotels
or motels, 11 percent involved camping, and 9 percent
involved the use of a vacation home. The recreation
activities most likely to be considered when visiting
Michigan on a pleasure trip were fishing (23%), swimming
(22%), boating/ canoeing (18%), hiking (18%), and camping
(6%). Attendance of attractions and events most frequently
reported when on a pleasure trip were dining and evening
events (23%), natural attractions (18%), landmarks and
historic sites (17%), developed attractions such as museums
(13%), and festivals, fairs and cultural events (11%).

Michigan consists of many diverse and attractive
geographical regions, ranging from the natural beauty of the
Upper Peninsula to the metropolitan area around Detroit.
Spotts (1991) suggests that, as opposed to visitors of the
Upper Peninsula, visitors travelling to the southeastern
portion of the state are less likely to enjoy outdoor
recreation or to visit natural attractions on their trip.
In contrast, they are most likely to report that they are
visiting friends or relatives or specific attractions as a
trip purpose. The utilization of friends' and relatives'

homes for overnight stays is highest in southeast Michigan,

13
medium in eastern and western Michigan, and lowest in the
Upper Peninsula, reflecting the spatial distribution of
Michigan's population.

Fridgen (1987) identifies major tourism destinations as
perceived by tourists who are familiar with Michigan as well
as by those less familiar with Michigan. For travellers
familiar with Michigan, designated tourism and recreation
centers were Grand Traverse, Mackinac, Keweenaw, Charlevoix,
Marquette, Antrim, Wayne, and Roscommon counties. Less
familiar travelers gave slightly different responses. They
failed to identify small regions along the northwestern
coast of the Lower Peninsula as well as the central lakes
area in the center of the state. Instead they marked the
Thumb Region and the region along Lake St. Clair. The
majority of the identified counties were coastal counties.
However, it was difficult to determine whether neighboring
counties of tourism centers identified by respondents were

also perceived as part of the tourist region.

Accommodations Use and Tourism Performance
Accommodations use as a measure of tourism potential
has been widely discussed. Choy (1985) suggested that the
performance of the hotel industry can be used as an
indicator of general trends in travel and tourism-related

activities for a particular area.

l4

Pearce (1987) outlines major reasons for the widespread
application of accommodation use as an indicator for tourism
activity. In his view, accommodations are one of the more
visible and tangible elements of tourism. A list of
accommodations is generally readily available, with
statistics on commercial accommodation being compiled for
fiscal and other reasons. The use of accommodation is also
logical, since a stay away from home is one of the defining
characteristics of tourism. Except where day trips are
dominant, the distribution of accommodations provides a good
measure of the distribution of tourism activity across a
region, state, or county. This distribution is also a
reasonable measure of economic impact, since accommodation
expenditures usually account for one-third to one-half of a
tourist's travel budget (Pears, 1987).

Brown (1981) justifies the use of lodging receipts as
an indicator of total tourism-related expenditures. He
argues that tourists, as well as local residents, trade at
the same restaurants, recreation establishments, and other
sectors of the retail economy. Also, because there are many
entrances and exits into and out of counties, few other
measures are as good an indicator of tourism volume.

Bishop and Spotts (1990) also point out the usefulness
of hotel/motel use tax collections as an important indicator
of the performance of the travel industry as a whole.

However, the use tax is not a comprehensive indicator of

15

travel activities since much travel activity takes place on
day trips, and portions of overnight travel involve lodging
in second homes, with friends or relatives, or at
campgrounds. Furthermore, the tax data reflects the price
charged as well as the amount of traveler spending in hotels
and other establishments. Based upon the arguments and work
of these and other authors it appears that hotel/motel use
tax receipts are one of the better indicators available for

the researcher to use in tourism investigations.

Population and the Relative Importance of Tourism

Although data on the distribution of accommodations
give a useful indication of where tourism is important in
terms of a county or region, absolute figures do not
necessarily reflect the significance of tourism within a
particular county. Clawson and Knetsch (1969) recognized
the impact of a population base on tourism performance quite
early in their work. They suggested that the influence of
the population distribution be removed within the potential
zone by putting the variable on a per capita basis. Looking
at lodging data on a per capita base makes it possible to
determine the importance of tourism for a region and to
eliminate the effect population centers have in terms of
attracting visitors in and of themselves.

Keogh (1984) suggests that researchers apply Defert's

tourist function index. This index is a measure of tourist

16
activity or intensity as reflected by two populations, the
visitors and the visited. It is derived by comparing the
number of beds available to tourists in a county (N) with
the resident population of that county (P) according to the

formula : TF = ( N * 100 ) / P.

Regional Analysis

Through regional analysis important regional variations
can be identified, including a region's ability to attract
and service tourists as well as to benefit from
expenditures and contact with tourists. In understanding
tourism at the regional level it is essential to know how
one destination relates to another, whether or not a region
is a primary destination itself and what is the nature and
composition of regional tourism components. Therefore,
regional analysis is useful for the purpose of product
development, tourism planning, and policy evaluation (Smith,
1987).

In North America, Gunn (1982) was one of the first to
do research in the field of tourism regional analysis. He
proposed a methodology for identifying regions with
potential for tourism development in Canada based on the
fact that tourism implies travel from origins to
destinations and that destinations imply a sense of space.
He declares that a fundamental requirement of tourism

planning is to understand, first, the location of potential

17
tourism places with respect to other features in other
locations and, second, the natural and cultural resources
and the popular images of those places.

In his study "An Approach to Regional Assessment of
Tourism Potential" Gunn (1980) describes a process that
determines geographical zones. Because of the strength of
certain locational and organizational factors, some regions
have a greater potential for future tourism than others.
His assumption is that the tourism industry depends on the
flow of tourists who seek things to see and do, like parks
and recreation resources, or commercial attractions and
events. Attractions are primarily physical land
developments. Gunn developed an inventory of regional
resources for 20 counties in south central Texas. He used
mapping procedures to describe zones with greatest tourism
potential in order to help with future policy decisions. In
the study by Gunn nine factors were used that are closely
related to the variables used in this study, transportation
and access to the region being the most crucial ones as
identified by Gunn.

In yet another study, Brown (1981) successfully uses
lodging receipts to measure the relative and absolute
importance of tourism to the economies of New York counties.
He employs a spatial analysis and develops two maps. One
map shows net changes in lodging receipts for New York

counties. The other map illustrates lodging receipts of a

18
county in percentage of the New York states total lodging
receipts.

Smith (1987) describes a procedure for defining tourism
regions on the basis of county-level resource patterns. He
identified 16 variables most of which are similar to the
variables used in this study. He counted facilities and
resources while neglecting the capacity or quality of each
individual facility. Next, a principal component factor
analysis was performed on these variables, attaching a score
to the factor loadings for each county. These scores were
then mapped to illustrate county variations. Thirdly,
tourism regions were identified through a cluster analysis
by grouping counties with similar resource patterns.
Finally, with a regression model, resource patterns were
correlated with variations in the economic importance of
tourism in each county to help identify the types of tourism
resources important for a strong economy. Smith employed
two measures of economic importance. One was the percentage
of local retail receipts attributable to tourist
expenditures as an indication of the significance of tourism
in the area. The other measure was the percentage
contribution of tourism expenditures in a county to total
provincial receipts to measure the magnitude of each
county's tourism development in terms of its contribution to

the total provincial tourism economy.

19

His findings suggest that there are strong regional
variations across the tourism resource base. Counties are
not created equal when it comes to tourism components. Some
counties are poor in one type of tourism resource, while
rich in another. In terms of economic impact, it was
observed that urban tourism and urban fringe tourism had the
highest economic impact, far higher than wilderness and
outdoor recreation resources (Smith, 1987). Smith concludes
that tourism is an industry with important geographical and
regional aspects. These aspects must be recognized and
understood if weaknesses in a county or state tourism
industry are to be corrected and if the strengths are to be
fully utilized.

Livingstone and Mitchell (1989) used the same
methodology as Smith in their study of South Carolina
counties. They also found similar results. First,
considerable regional differences were observed. Second,
the regression analysis showed that urban tourism had the
highest positive correlation with the dependent variable.
They used number of visitors, visitor expenditure, and state
and local taxes generated by visitors as dependent variables
in four separate regression analyses. Counties that fall
into the urban tourism category have large numbers of
accommodations, activities, facilities, and programs which
attract the largest numbers of in- and out-of-state

visitors.

20
Destination Images

A tourism image is defined as the impressions that a
person holds about a region in which he or she does not
reside (Hunt, 1974). Although tourism resources are the
basis for tourism travel, images play an important role in
the tourist decision-making process. Hunt (1974) designed a
study to determine whether some recreation areas command
higher use than others simply because of a perceived status.
Data indicated that there were differences in preference for
areas. In a study of nine regions, Goodrich (1978) explored
the relationship between perceptions of an area and
preferences for an area as a vacation destination. Results
indicate that favorable impressions of a tourist area
increase the probability of choice or preference for that
area as a vacation destination.

Raitz and Dakhil (1988) administered a questionnaire to
college students throughout the United States to find out
their preferences for specific physical environment types
for high quality recreational experiences. Seashores and
lake environments were the most preferred. Assuming that
these findings can be generalized for all tourists, Michigan
has the potential to build upon these preferences in the
competition for tourists.

Specifically in Michigan, Deale (1983) conducted a
study of auto travelers' perceptions using a cognitive

mapping technique. Her major finding was that counties

21
bordering the Great Lakes were identified more often as the
centers of recreation and tourism regions than inland
counties. She reports that respondents associate these
shoreline regions much more often with recreation and
tourism then regions located inland. Mackinac Island and
Grand Traverse County received special attention. Also,
western Lake Michigan counties had a more positive image
than eastern Lake Huron counties. The northwestern counties
of lower Michigan were believed to be more scenic, and more
popular destinations, having better accommodations and
friendlier people, as well as better boating, camping,
sailing, beaches, festivals and events than counties located

on the Lake Huron side of Michigan.

Importance of Tourism Components

The tourism system is activated by attractions. Only
in rare cases do people travel without being stimulated by
attractions. The attraction is the primary trip purpose
(Crompton, 1990). There are many forms of attractions;
researchers are still arguing which types of attractions are
most important in generating tourism visitation.

Richie and Zins (1978) administered a survey to
managers from various sectors of the tourism industry and
found that natural beauty was clearly the single most
important determinant of the attractiveness of a given

region. They found that cultural and social characteristics

22
were the second most attractive feature of a tourism
destination.

Var (1974, 1977) established an index of touristic
attractiveness on the basis of informed judgments for Turkey
and for the Canadian Province of British Columbia. Natural
factors and food and shelter were ranked highest in both
countries. Historical prominence, which was quite important
for Turkey, was also rated an important factor in British
Columbia. Natural beauty, however, was ranked first in both
cases. On the other hand Smith's (1987) findings
illustrate, that urban tourism based on theater, shows,
festivals or historic sites contributes most to the tourism
performance of a county, far more than outdoor recreation.
But then he did not consider the role of population when
developing the dependent variable.

Makens (1987) discusses the importance of historic
sites as visitor attractions in his article. He found that
historic sites were primary visitor attractions and act as
important vehicles for preserving and transmitting the
values of our cultural heritage to each emerging generation.

Curtis (1990) acknowledges the major success of
commercial theme parks as visitor attractions. He describes
them as challenges to the public recreation sector. Without
integrating such attractions, the public sector will fall

behind its commercial competitors.

23

A survey on Michigan tourism administered by Ross Roy,
Inc. in the mid-808 (Holecek, 1991) estimates that 48
percent of all travelers taking pleasure trips in Michigan
were visiting a natural attraction, developed attraction,
historic site, landmark, or museum. Recreation activities
most likely to be considered when on pleasure trips in
Michigan were fishing, swimming, boating, canoeing, hiking,
and camping. This emphasized the natural resource-based
recreation in Michigan. Attractions of interest when
traveling in Michigan were dining and evening entertainment
(32%), natural attractions (18%), landmarks and historic
sites (17%), developed attractions and museums (13%), and
fairs, festivals, and cultural events (11%) (Holecek, 1991).

Apparently the importance of a single tourism component
varies with the region under consideration. However,
natural resources as well as developed attractions play an
important role with respect to the tourism potential of a

region.

Conclusions from the Literature Review
Regional analysis is an important research technique
which can be used to understand the location of potential
tourism destinations with respect to other features. It also
helps the researcher understand the natural and cultural
resources as well as the popular images of those places

(Gunn, 1982). The literature review suggests that the

24
tourism performance in Michigan counties or elsewhere could
be partially explained by the existence of physical land
developments. It has to be recognized, however, that there
are also other reasons for choosing a specific tourist
destination (Mathieson and Wall, 1982). The major tourism
components that will have to be incorporated into the model
are accommodation features, developed resources or
attractions, natural resources, and the accessibility of
tourism destinations (Robinson, 1976). Natural resources
are likely to be the dominant tourism component in Michigan
(Ritchie and Zins, 1987, Var, 1974, 1978), especially the
Great Lakes shoreline (Deale, 1983, Raitz and Dakhil, 1988).
Several measures of tourism potential have been suggested in
literature, namely hotel/motel use tax, which is a
performance measure (Bishop and Spotts, 1990), and volume of
accommodations, which is a supply measure (Pearce, 1987),
each of which can be expressed as a per capita figure
(Clawson and Knetsch, 1969). Relationships between these
variables and regional differences in tourism potential
across Michigan counties are the subject of the hypotheses

for this study.

t s s
Based upon previous works which have explained the
relationships between attractions, natural resources and

tourism growth, two general hypotheses can be proposed. The

 

25
consistent findings by others that tourists are attracted to
natural resources, scenic beauty, and attractive landscape
elements, provide the background for hypothesis number one,

which is stated as follows:

Hypothesis 1:
Natural resource components as opposed to developed
attractions will be the strongest explanatory contributor to

county-based tourism potential.

Findings also suggest that visual as well as physical
access to the Great Lakes are a very important factor in
attracting tourists. The powerful presence and beauty
associated with the Great Lakes provides the basis for the

next hypothesis which is as follows:

Hypothesis 2:
Great Lakes shoreline will be the single most important
tourism component in explaining tourism potential in

Michigan counties.

CHAPTER III

ANALYSIS TECHNIQUES AND PROCEDURES

In this Chapter research variables and analysis
techniques, particularly regression analysis are outlined
and discussed. The first section of this chapter reviews
variable measurement and discusses the spatial distribution
of variables in Michigan. The second part of the chapter
outlines the model building process and describes steps used

in the regression analysis.

Research Variables

The selection of specific variables from among the many
suggested in the tourism literature is very crucial to the
success of the study. The choice of dependent as well as
independent variables will be described in detail. Each
variable will be first defined and then operationalized. To
better understand spatial differences between counties and
to better interpret the results of the regression analysis a
spatial analysis of each variable will follow.

Spatial analysis is a method used to reveal general

geographical variations of variables. By mapping tourism

26

27
components on a county bases a broader picture of the
spatial structure of tourism components in Michigan can be
illustrated. Pearce (1987) provides many examples of
spatial analysis applications in tourism; Herbin (cited in
Pearce, 1987) has produced maps showing the distribution of
resorts, ski lifts, and occupancy rates in the Alps.
Ashworth's maps (cited in Pearce, 1987) show the
distribution of accommodations and attractions in the
Netherlands, and Ishii (cited in Pearce, 1987) deals with a
range of recreational resources and facilities in Japan.
Also Smith (1987) as well as Var (1974, 1977) mapped spatial
variations in their tourism components. Spatial analysis is
a technique, widely used by tourism researchers, that

presents variables in their spatial context.

Regression Analysis

A multiple linear regression model will be the primary
analysis instrument in this study. It relates multiple
independent variables to one dependent variable. The model
can be expressed as:
Yi = so + nix1i + 32X” + + ppxpi + ei
The notation Xpi indicates the value of the pth independent
variable for case i. The 8 terms are unknown parameters and
the ei terms are independent random error terms which are

2

normally distributed with mean 0 and constant variance 0 .

The model assumes that there is a normal distribution of the

28
dependent variable for every combination of the values of
the independent variables in the model. The number of
observations has to exceed the number of coefficients
estimated to come to reliable results (Chou, 1989).

One of the first steps in calculating an equation with
several independent variables is to calculate a correlation
matrix for all variables. Such a matrix shows the
correlation between the dependent variable and each
independent variable, as well as intercorrelations among
independent variables. Large intercorrelations between the
independent variables of 0.5 and more are a threat to the
results of the multiple regression analysis. Without adding
much to the overall fit of the model, they inflate the
variances of the estimates, making individual coefficients
quite unreliable. Therefore, highly correlated independent
variables will have to be modified (Kachigan, 1982).

Partial regression coefficients or B-coefficients will
be estimated with the method of least squares. To overcome
the difficulty of direct comparison of the relative
importance of explanatory variables when different units of
measurement are used, a transformation into units of
standard deviation is necessary. The Beta coefficients can
be calculated directly from the regression coefficients
using

Betak = B1K * (Sh / Sy)

29
where Sk is the standard deviation of the kth independent
variable (Ott, 1977).

Having obtained the sample regression equation and
established the varying degrees of significance of the
regression coefficients, the closeness of fit of the
regression model needs to be determined by calculating the
coefficient of multiple determination, R square. R square
is defined as the ratio of the explained variation due to
regression to the total variation in the dependent variable.
Thus, R square can be interpreted as the portion of the
total variation in the dependent variable that is associated
with or explained by the regression of the dependent
variable on all independent variables. Therefore, the
closer the value of R square is to one, the smaller the
scatter of points about the fitted regression and the better
the fit is. The sample R square tends to be an optimistic
estimate of how well the model fits the population. The
model usually does not fit the population as well as it fits
the sample from which it is derived. The adjusted R square
corrects R square to more closely reflect the goodness of
fit of the model in the population (Kachigan, 1982).

The coefficients in the sample multiple regression
equation are subject to sampling error. Before it can be
used as a predictive device, it needs to be determined
whether the sample regression coefficients are statistically

significant. This will be accomplished by testing the joint

30
effects of the independent variables on the dependent
variable. Thus, a F-test of
1% : Bl== £3== ... = fip== 0
would be a test of the null hypothesis that the dependent
variable is not linearly related to the independent
variables. Large values of F will indicate rejection of the
null hypothesis (Chou, 1989).

In order to look for evidence that the necessary
assumptions are not violated, it is important to plot
residuals. A residual is the difference between an observed
value and the value predicted by the model. Residuals
should be plotted against predicted values. The
distribution of residuals should be examined for normality.
The relative magnitude of residuals is easier to judge when
they are divided by estimates of their standard deviation.
The resulting standardized residuals are expressed in
standard deviation units above or below the mean (Ott,
1977). Standardized residuals will not only be plotted by
case but also mapped. The mapping criteria is whether the
standardized residual for a county is above or below the
mean. Mapping makes it possible to detect spatial patterns
of over- or underestimation of the dependent variable for
certain counties.

When building a regression model it is also important
to identify cases that are influential, or that have a

disproportionately large effect on the estimated regression

31

model. One way of doing so is to compute the leverage of a
case. The leverage for a case describes the impact of the
observed value of the dependent variable on the prediction
of the fitted value. Ideally, each observation should have
an equal influence. Therefore, all leverages should be near
i / p, i being the number of cases and p being the number of
independent variables (Kachigan, 1982).

Finally,it is useful to split the sample randomly into
two parts. One part is then used to estimate the model,
while the remaining cases would be reversed for testing the
goodness of fit. This is important since a model usually
fits the sample from which it is derived better than it fits
the population (Ott, 1977). In this study splitting the
data set into two parts would result in a very small sample
size for the model building process, and therefore, only
five counties will be reserved to test the model. Of course,
these five counties will be excluded from the model building
process. These five counties will be chosen randomly from
counties that present the middle range of population sizes

among Michigan counties.

CHAPTER IV

VARIABLE MEASUREMENT

This chapter which focuses upon variable development
and measurement,is divided into three sections: (1)
dependent variables (2) independent variables, and (3)
limitations of the study. In the first two sections
variables will be derived from major tourism components that
have been identified in the literature review. Those
considered for the study include: accommodation features,
developed attractions, natural resources, accessibility, and
county population. Variables will be defined and a
rationale for including the variable in the model will be
provided. Then variables will be operationalized. The
spatial distribution will be shown by mapping each variable
in its geographical context, and finally, a general
discussion of the variable will follow. The last section of
this chapter discusses limitations of the study. Michigan
counties and regions are illustrated in Figure 4.1. Similar
base maps will be used in the presentation of the spatial

analysis.

32

 

 

  

 

 

 

Lower m
Peninsula
{you I.“ (m! m W
W

Km"

 

 

 

 

 

 

 

 

 

 

lama.-

 

 

 

all!“
Southern
Lower

Peninsula {W II]: M

 

 

 

 

 

 

 

 

 

mm

 

 

 

 

 

 

Figure 4.1: Counties and Basic Regions in Michigan

33

34
Dependent Variables

For the purpose of this study two different dependent
variables have been chosen, per capita hotel/motel use tax
and per capita lodging supply. It is assumed here that both
are measures of tourism potential; the first, however, is a
performance measure, dividing the amount of hotel/motel use
tax receipts by the population base of a county. The second
is a supply measure, dividing the number of guest rooms and
campsites in a county by its population base. While per
capita use tax takes only travelers who stay in hotels or
motels into account, per capita lodging supply recognizes
the importance of camping as a lodging alternative.
Unfortunately, though, data on campground visitation is not
available. However, since tourism is a growing industry in
Michigan, per capita lodging supply should provide an
appropriate indication of tourism potential for counties and
regions in the state. In the following, each part of the
two measures will be discussed separately: (1) lodging
supply, (2) hotel/motel use tax, and (3) relationship to
county population. The data on the dependent variables is

displayed in Table A.1.

W291):
Camping in Michigan is not only an important recreation
activity and base for other forms of recreation, it is also

a type of lodging. According to a survey by Ross Roy, Inc.,

35
14 percent of all pleasure trips to Michigan between 1983
and 1986 included an overnight stay in a campground. In
1985, 88,053 campsites were available to the traveler
compared to 96,503 guest rooms in lodging facilities in 1986
(Spotts, 1991). Therefore, it is important to include the
number of campsites as well as the number of guest rooms in

any analysis of tourism lodging supply.

Su G st oms

It is estimated that 96,503 guest rooms are available
to the tourist visiting and traveling in Michigan.
According to Spotts (1991), 59 percent of the establishments
were hotels/motels, 36 percent were cabin/cottage
establishments, 5 percent were bed & breakfasts/historic
inns, and less than 1 percent were condominiums. Overall,
87 percent of guest rooms were provided by hotels and motels
as opposed to cabins/cottages and condominiums, which
emphasizes the more urban orientation of guest rooms.
Operators of lodging establishments are listed in Table 4.1.

It is also important to note that some lodging
establishments are not open year round. For instance, in
rural Mackinac County only 22 percent of the lodging
establishments operated year round as compared to 100
percent in the more densely populated counties Clinton,
Isabella, Lapeer, Midland, Saginaw, Lenawee, and Washtenaw.

These percentages demonstrate the seasonal character of

36
resource-based tourism as compared to year-round tourism and
business traffic in urban counties.

The data set was obtained from the Travel, Tourism, and
Recreation Resource Center at Michigan State University.

The original data sources were the Yellow Pages of phone
directories, the American Hotel and Motel Association's
annual directory, a publication by the Michigan Travel
Bureau entitled "Michigan's Bed & Breakfast and Historic
Inns," the Michigan Lodging Association's membership roster,
and a printout of taxpayers supplied by the Department of
Treasury. Although the guest room inventory dates back to
1986, it can still be used as a good approximation of 1989's
guest room distribution over all Michigan counties.

The total supply of guest rooms is highly concentrated
in urban areas, reflecting the high volume of business
travel, as can be seen in Figure 8.1. The highly traveled
interstate highways may also account for the relatively
large number of guest rooms in these counties. On the other
hand, rural counties with a high supply of guest rooms like
Mackinac, northwest Lower Michigan, Gogebic, Chippewa,
Roscommon, and Iosco, all contain important tourist sites
and attractions. The Michigan Straits region includes major
historic attractions; Sleeping Bear Dunes National Lakeshore
and many popular resorts are located in northwest Lower
Michigan; most ski areas and the only mountains in Michigan

are found in Gogebic County; Chippewa County includes the

37
heavily visited 800 Locks and Tahquamenon Falls State Park;
Roscommon County provides two large inland lakes: and Iosco
County offers plentiful Great Lakes fishing opportunities
(Spotts, 1990). Counties like Berrien, Grand Traverse,
Marquette, Muskegon, and Saginaw all contain sizable cities
as a population base as well as important tourist
attractions. Relatively few guest rooms are located in the
central Upper Peninsula, northeast Lower Michigan, west
central Michigan, the Thumb region, or most of the extreme
southern part of the Lower Peninsula. Each of these areas
lack a major city, major tourist attractions, or an

interstate highway.

Table 4.1: Operators of Campgrounds and Lodging

Establishments

Campground % of total Lodging % of total
Provider Sites Provider Establishments
Commercial 58.9 Hotel/Motel 59
State Park 15.5 Cabin/Cottage - 36
Local Public 14.7 B & B/Inn 5
Non-Profit " 5.1 Condominium < 1
State Forest 3.4
Nat. Forest 2.0
Nat. Park 0.3

Total 100.0 Total 100

Source: Spotts, D. (1991). Travel and Tourism in Michigan -
A Statistical Profile, Second Edition. East Lansing,
Michigan: Research Monograph # 2, Travel, Tourism and
Recreation Research Center, Michigan State University.

38
W

Campsites are the second component of lodging supply.
They are provided by the private sector, local communities,
and by State and National Parks, as well as by State and
National forests. Operators of campgrounds are described in
Table 4.1 (Spotts, 1991). Camping is typically a more
resource-based type of lodging compared to the location of
guest rooms. Furthermore, travelers who camp are less
likely to include business travelers or visit friends or
relatives compared to those staying at commercial lodging
facilities. Like some of the more rural lodging facilities,
campground use is very seasonal and restricted by changes in
weather. When comparing the number of campsites and the
number of guest rooms one must bear in mind that campgrounds
are open approximately six months out of the year and
sometimes less.

The data for the campsite variable was obtained from
various sources. The number of commercial and local public
campsites was provided by the Michigan Department of Public
Health, where all commercial campgrounds must be licensed
and registered. The data dates back to 1985, but it should
provide a reasonable estimate of current campsite
distribution. It should be noted that the important issue
are differences in the spatial distribution of campsites
across Michigan counties, not the exact number of campsites.

Campsites which are offered by the non-profit sector, like

39
fraternal organizations, recreation or religious
organizations, are excluded from the count, because they are
not open to the general traveling tourist.

An estimate of the number of State and National Forest
campsites was provided in the brochure "Michigan Campground
Directory" published by the Michigan Travel Bureau in
connection with the Michigan Association of Private
Campground Owners. The information is dated 1991.

State Park campsites were estimated by content
analyzing the brochure "Michigan State Parks" (1990)
published by the Michigan Department of Natural Resources
and distributed by the Michigan Travel Bureau. In summary,
the calculated number should provide a reasonable estimate
of all campsites available to the traveling public in each
of Michigan's counties.

The spatial distribution of campsites is illustrated in
Figure 8.2. Campsites are concentrated in the eastern Upper
Peninsula, primarily in Chippewa and Mackinac counties,
which both have very attractive tourist features. Hardly
any campsites are located in the inland counties in the
middle part of Michigan -- Midland, Gladwin, Gratiot,
Clinton, or Eaton. These are primarily agricultural
counties with few lakes and streams to foster campgrounds.
Most campsites are located along the west coast in Mason,
Oceana, Muskegon, Ottawa, or Allegan and at the beaches of

the northern Lower Peninsula in counties like Grand

40
Traverse, and Cheboygan. Campsites are also located within
easy reach of population centers. There is also a large
number of campsites in the southernmost counties of Michigan
which are likely to serve Ohio and Indiana residents as well
as Michiganders from major cities. These counties have also
a good number of miles of streams and acres of lakes as well

as several State Parks.

ote t s

The second key dependent variable in this study is
hotel/motel use tax revenues. The room tax is a 4 percent
tax levied on "rooms or lodging furnished by hotelkeepers,
motel operators and other persons furnishing accommodations
that are available to the public on the basis of a
commercial business enterprise, irrespective of whether or
not membership is required for use of the accommodations,
except rooms and lodging rented for a continuous period of
more than one month" (Michigan Department of Treasury,
1984). The types of establishments required to pay the room
tax include inns, motels, tourist houses, nudist camps,
apartment hotels, resort lodges, and cabins as well as
certain camps operated by other than non-profit
organizations. The data are collected and recorded by the
Department of Treasury; it is also maintained by and

available through the Travel, Tourism and Recreation

41
Research Center at Michigan State University for all
Michigan counties.

The trend in use tax collections shows a steady
increase between 1985 and 1989. However, in 1990 an
economic downturn was felt due to the beginning recession
and the Gulf War (Spotts, 1991). The trend is portrayed in
Table 4.2. It has to be kept in mind that changes in the
data reflect changes in the prices charged at lodging
establishments as well as changes in the amount of spending

therein.

Table 4.2: Trend in Unadjusted Hotel/Motel Use Tax Receipts
in Michigan (1985 - 1989)

Year Hotel/Motel Use Tax in $
1985 17,127,335
1986 18,737,261
1987 18,737,261
1988 19,978,770
1989 21,812,110

Source: Spotts, D. (1991). Travel and Tourism in Michigan -
A Statistical Profile, Second Edition. East Lansing,
Michigan: Research Monograph # 2, Travel, Tourism and
Recreation Research Center, Michigan State University.

The following limitations in the data need to be
recognized (Spotts, 1986). First, there is the problem of
taxes paid by hotel/motel chains. In the case of a
franchise, use taxes are paid in one sum by the corporate
headquarters rather than by each of their affiliated

properties. Consequently, these payments are recorded for

42
the county or state containing the corporate headquarters.
If franchise businesses do a significant share of the
business in a given county, the hotel/motel use tax for a
county will underestimate the volume of business activity.

The second limitation is the classification problem.
When businesses apply for a use tax registration they
complete an application that requires them to state which of
the Department of Treasury's Business Classification Codes
best describe their principal business activity. These
responses are the basis for the Department of Treasury's
categorization of use tax collections under specific
business types. Therefore, in most instances, someone
operating a hotel would select BCC 701 - "Hotels, Tourist
Courts, Motels" - as the best description of his or her
business. However, if this hotel contains a restaurant
which the applicant considers as the principal business
activity, he or she could choose another Business
Classification Code.

Finally, use tax collections might not fully reflect
business transactions because some business owners may not
fully report all of their taxes. The exact magnitude of
this problem is unknown, but it is assumed to exist at some
level.

These difficulties limit the quality of the data as an
indicator of tourism volume and, therefore, hotel/motel use

tax receipts should be considered only as an approximate

43
indicator of tourism volume, particularly in counties that
have many franchise businesses. Nevertheless, if it is
assumed that changes in tax volume are not due to one of
these three problems, then changes in taxes collected above
price changes can be attributed to shifts in business volume
and can be viewed as a tourism performance measure in that
county (Spotts, 1986).

The use tax has a similar spatial distribution pattern
as Michigan guest rooms, as can be seen in Figure 8.3. The
use tax, however, emphasizes the importance of year-round
tourist attractions and business travel in the urban areas,
which are independent from vacation time or weather changes.
Therefore, summer vacation destinations in the Upper
Peninsula that have a high number of guest rooms like
Chippewa, Mackinaw, and Marquette, as well as Gogebic, which
has many winter sports resorts, do not have extremely high
tax revenues although they have many rooms. The same is
true for the counties in the northern part of the Lower
Peninsula. In addition, it is possible that room charges
are not as high in these counties when compared to more

urban counties.

gonnny Popnlntinn

The previous literature review suggests that population
size is an important variable to be considered when studying

tourism destination regions.

44

By including the population base of each county as part
of the dependent variable, population-based differences of
counties can be controlled. By controlling for population,
attention is paid to the fact that cities are an attraction
in and of themselves. Recreation facilities, events, and
other attractions within population centers serve primarily
their own population and only secondarily tourists.

The population data for each Michigan county was
obtained through the Center for Redevelopment of
Industrialized States located at Michigan State University.
It represents 1990 census data. Compared to the 1989
population estimate derived from the 1980 census which is
also available, the 1990 census data seems to provide a
better estimate of the actual 1989 population in each county
(Menchik, 1990). Based upon the 1990 estimate, Michigan has
a population of 9.1 million people. Counties range in size
from 1,707 (Keweenaw) to 2.11 million people. The largest
counties by far are Wayne (2.11 million) and Oakland (1.08
million).

Michigan's population is concentrated in the southern
portion of the Lower Peninsula, specifically in the three
counties that constitute the Detroit metropolitan area and
in Kent county. The Thumb region and the southernmost
counties of the lower Peninsula, Cass, Branch, and
Hillsdale, are exceptions to the overall density of the

populated southern Lower Peninsula. The Upper Peninsula and

45
the northern portion of the Lower Peninsula are less densely
populated. Exceptions include Marquette and Grand Traverse
counties, which are popular tourism destinations; in
addition, Bay County, Midland, and Isabella counties have
moderate population densities. In the Upper Peninsula
several counties have minimal populations, particularly
Alger, Baraga, Keweenaw, Luce, and Schoolcraft. The
population distribution in Michigan can be seen in Figure

B.4.

Spatial Distribution of Per Qaniga Eignres

A completely different picture emerges when using per
capita figures. In terms of per capita lodging supply the
northern counties dominate (see Figure 8.5). Tourists drawn
to the natural beauty, large public land holdings, abundance
of lakes, beautiful shoreline and extensive forests in these
regions may help to support a higher per capita lodging
supply. High per capita lodging supply is observed in most
of the eastern counties in the Upper Peninsula as well as
Keweenaw and Gogebic counties. The northern portion of the
Lower Peninsula, particularly Charlevoix, the Straits
region, western coastal counties and selected inland
counties with extensive state forests and plenty of inland
lakes all have higher per capita lodging supply. Per capita
lodging supply in the southern Lower Peninsula with the

exception of Saginaw County with Frankenmuth, its major

46
tourist attraction, is quite limited. As illustrated in
Figure 8.5, limited per capita lodging supply can be
observed in the Thumb counties of Huron and Salinac and in
the southernmost counties of Michigan which serve as nearby
recreation areas to the Detroit metropolitan area.

A similar picture emerges for per capita hotel/motel
use tax (see Figure 8.6). The Upper Peninsula, especially
the eastern portion has a high per capita hotel/motel use
tax. Leelanaw, Grand Traverse, Antrim, Charlevoix, Emmet,
Crawford, and Otsego counties dominate per capita use tax
receipts in the northern Lower Peninsula. Generally
speaking, inland counties are less successful than coastal
counties, except for Crawford and Otsego County where there
are extensive State Forests and plenty of inland lakes. The
southern Lower Peninsula, relatively speaking, has neither
high lodging nor use tax per capita. Saginaw county appears
to be the exception in the lower portion of the Lower
Peninsula, due to an active business environment as well as
a major attraction. Counties in the Detroit region reveal a
modest per capita use tax distribution due to the general
business activity along with tourists. Washtenaw and Ingham
counties probably have more tax receipts than the
surrounding counties due to their universities and

associated activities.

47

Table 4.3: Variable Description

d V ' b e
Use Tax p.c.

Lodging Supply p.c.

Ingenengent Vazianle
Public Land

National and State Parks

Streams

Great Lakes Shoreline
Lakes

Access Points

Attractions
Cultural Attractions
Entertainment Places
Eating / Drinking Places

Golf Courses

Highways

Ski Areas

hotel/motel use tax divided by
county population

number of guest rooms plus
campsites divided by county
population

public lands in percentage of
county area

acres of National and State
Parks

miles of streams

miles of Great Lakes shoreline
acres of lakes

number of public water access
points

(see Attraction Index
Development in chapter 5)

number of golf courses
existence of interstate
highways or freeways

number of downhill ski areas

48
Independent Variables
Out of the abundance of variables mentioned in
literature and other studies ten independent variables have
been chosen for the purpose of this study. The variables
represent an inventory list of resources which will be used
in explaining a county's potential for tourism. Table 4.3
provides a summary of the variables and Table A.2 contains

the actual county-level data.

b ' s '0

Michigan contains large acreage of public lands, these
lands provide a great variety of recreational opportunities
and attract tourists regionally and nationally. They also
contribute to Michigan's image as an interesting tourist and
recreation destination. For instance, Michigan has more
public lands than Illinois or Wisconsin, its principal
competitors for tourists in the Great Lakes region (Wells
and Eidelson, 1991).

The variable public lands as percentage of the total
county area was chosen for the following reasons. First,
public lands are tourist attractions. Second, public lands
provide for an abundance of nature-based recreation
activities, like fishing and hunting, skiing, snowmobiling,
hiking, swimming and boating. Third, by using the
percentage of public lands rather than the acreage of public

lands alone, it is possible to identify tourism development

49
issues, because generally the more land under public
management, the less space is available for commercial
recreation and tourism developments, such as resorts,
lodging and eating establishments, theme parks or other
attractions. Use of the ratio also controls for the county
size.

Acreage estimates were acquired through the Travel,
Tourism and Recreation Research Center. According to Wells
and Eidelson (1991), public lands contain all lands managed
by federal or state agencies that are open to the public for
recreation purposes. However, that does not mean that
recreation is the sole use of these lands. Under the
multiple use philosophy, national and state forests are
managed to provide timber, wildlife, minerals, watershed
protection, and other benefits including recreation
opportunities. Not included in the definition of public
lands, however, are federal lands that are closed to the
public (e.g. military bases, Indian reservations).

Public recreation lands include state forests, state
game and wildlife areas, National Park Service areas,
National Forests, national wildlife refuges, the Soo Locks,
State Parks, state recreation areas, and state boating and
fishing sites. Also included are specially designated
areas, national wilderness areas, national landmarks,
national wild and scenic rivers, state wilderness, state

natural rivers, and Great Lakes bottomland preserves.

50
However, in Michigan almost 90 percent of public land
acreage is contained within National and State forests.

The ratio of public lands as a variable is purely
quantitative. It is not a comprehensive measure of
recreation opportunities. Qualitative aspects, such as
recreation facilities and services provided therein, the
abundance of fish and wildlife resources they contain, and
private sector support facilities are not represented by the
variable although these are the qualitative factors that do
contribute to the attractiveness of public lands.
Transportation facilities that provide access to these areas
and information provided to the public about all of the
above are also not included in the variable, which may limit
its usefulness as a predictor of tourism potential within a
county.

The percentage of public lands is logically highest in
counties that are not very densely populated. Therefore,
counties in the Upper Peninsula have the highest percentage
of public lands, with the exception of Menominee and
counties in the northern part of the northern Lower
Peninsula. In Schoolcraft, Mackinac, Oscoda, Crawford, and
Roscommon counties, over 50 percent of the counties area are
held in public lands because of extensive state and national

forests (see Figure 8.7).

51
'0 nd S ta 8

Michigan's State and National Parks contribute
significantly to the state's attractiveness as a tourist
destination (Spotts, 1991). They provide a wide variety of
outdoor recreation opportunities and also an abundance of
scenic and historic attractions. Some of Michigan's most
outstanding natural attractions such as Tahquamenon Falls,
the Porcupine Mountains, Pictured Rocks, or the Sleeping
Bear Dunes are located within the Michigan State Park system
or the National Park system.

Parks provide for a diverse array of recreation
activities such as hiking, fishing, swimming, sightseeing,
picnicking, and boating. They also serve as attractions
themselves (Spotts, 1991).

The variable acres of National and State Parks has been
chosen in addition to the percentage of public land to add a
qualitative component. The existence of park units serve as
indicators of natural beauty and historic values.
Furthermore, parks are usually more developed for tourism
purposes than national or state forest facilities, and are
more accessible.

There are sixty-seven State Parks, one National Park
and two National Lakeshores in Michigan. This data was made
available through the Travel, Tourism, and Recreation

Resource Center at Michigan State University.

52

National and State parks are established because of two
reasons; first, to preserve outstanding natural beauty and
historic values, and second, to provide recreational
opportunities in natural areas for the public. Therefore,
these parks can be found primarily in the very sparsely or
densely populated regions indicated in Figure 8.8. Rural
counties in the Upper Peninsula and northern Lower Peninsula
such as Keweenaw, Ontonagon, Alger, Luce, Chippewa, Benzie
and Leelanaw have extensive park lands. On the other hand,
large amounts of park land are also located within easy
reach of major metropolitan areas in Oakland, Livingston,
Washtenaw, and Jackson counties. Extensive park land can
also be found in the northernmost portion of the Lower

Peninsula and in Crawford county.

Wate; Besonrges

As mentioned previously, surveys indicate that water—
based recreation such as boating, fishing, swimming, and
canoeing are very popular recreation activities in Michigan
(Spotts, 1990). Boating, for instance, is one of the
largest industries in Michigan, with 709,130 pleasure crafts
registered in 1989 (Talhelm, 1990).

Detailed statistics are available on fishing licence
sales and on boating activities. However, these indicate a
performance and, therefore, are of no use to this study,

since a performance figure such as use tax receipts cannot

53
be explained with another performance figure. Instead the
study takes advantage of the statistics available on
different water resources such as the miles of Great Lakes
shoreline, which are a major attraction for boaters,
fishermen, and swimmers, miles of stream in each Michigan
county, which is important for angling, swimming and
canoeing, and the acreage of inland lakes. To add a
qualitative element, the number of public access areas will
be included as a variable as well -- hundreds of miles of
shoreline are less valuable to tourists if there is no
access. In total, four water resource variables will be
included in the model building process: miles of Great
Lakes shoreline and of streams, acres of lakes, and number
of public access sites.

Water resource data was obtained through the Travel,
Tourism, and Recreation Resource Center; however, each
variable originally had a different origin. The acreage of
inland lakes was provided by the Michigan Department of
Natural Resources. The number of public access sites was
taken from the Michigan Department of Natural Resources'
"Updated Public Access Site Master List" (1990) which gives
an estimate of access sites available in 1989. Stream
mileage was provided by Brown (1944) in his "Michigan
Streams -- Their Lengths, Distribution and Drainage Areas".

Humphreys (n.d.) provided information on miles of Great

54
Lakes shoreline. However, the shoreline of Isle Royal is
not included in this document.

Figures 8.9 through 8.12 show the spatial distribution
of water resources in Michigan. Counties in the Upper
Peninsula have the most water resources of all kinds
available. Most streams and lakes are concentrated in the
western part of the Upper Peninsula due to its larger land
mass. Eastern counties of the Upper Peninsula have
comparably more shoreline, but they also have a considerable
number of lakes and miles of streams compared to other parts
of the state.

The coastal counties in the southern Lower Peninsula
are also rather well endowed with water resources,
especially with long shorelines and an abundance of inland
lakes. In this respect, Roscommon, however, dominates with
its two huge inland lakes, followed by Cass County with its
many small lakes. The Thumb region also has plentiful water
resources, Huron and St. Clair counties both having an
extensive shoreline and many miles of inland streams.

Compared to the water resources, the number of public
access areas is unevenly distributed, apparently centering
not only around the water resources themselves but also
around population centers. Oakland, Washtenaw, and
Livingstone, for instance, have quantities of access points
that are out of proportion to their water resources. The

western half of the Lower Peninsula has far more access

55
areas than the eastern half. The Upper Peninsula has
numerous access points, but in proportion to its water
resources there are fewer access points than might be

expected.

Attractions

Visiting a tourist attraction is one of the principal
motivations for taking pleasure trips in Michigan.
Developed attractions play a vital role in the
attractiveness of a region as a tourism destination (Gunn,
1974)

Unfortunately, there is no complete and available
listing of Michigan developed attractions. To obtain an
estimate, the 1989 "Michigan Travel Planner" published and
distributed by the Michigan Travel Bureau as part of the
"Yes Mlchlgan" campaign was content analyzed. The planner
provides information about Michigan's major attractions and
tourist destinations, including an overview of communities
and a list of contacts for more information. Ferrario
(1980) recommends the method of content analysis to estimate
the number of attractions in a region. Since this planner
is distributed free of cost by the Michigan Travel Bureau,
the primary provider of tourist information, and is easily
available to all tourists, tourists are very likely to plan
their vacation with this published planning guide. It

provides an overview of main attractions. It is assumed

56
that those attractions listed are the most likely to attract
tourists to individual counties. Smaller attractions that
might not be listed function as complementary entertainment
rather than primary tourist drawing facilities. In any
case, a complete guide is not necessary for the analysis;
more important than an exact number is the spatial
distribution of attractions, even if they are estimates.
The different attractions are grouped into cultural
attractions, which are art galleries, historic sites,
museums, performing arts, bridges and tunnels; and other
entertainment facilities, which are amusement parks, tours,
general entertainment places, professional sports,
festivals, zoos, agricultural exhibitions, arboretums, and
botanical gardens. This grouping of attractions was the
same structure used in the planner itself.

A threat to the validity of the data is that only
members of the Chamber of Commerce who apply get listed in
the tourism planner (Michigan Travel Bureau, 1990). Many
small enterprises, especially, consider the fee which
accompanies the listing too high and refrain from applying.
On the other hand, these smaller businesses may not be
primary tourism attractions within the individual counties.

Eating and drinking establishments represent a
different type of attractions. Restaurants obviously
provide an essential service to travelers; dining in such

establishments often constitutes a recreational experience

57
which helps to maintain high levels of satisfaction among
travelers and encourages repeat visits (Spotts, 1991).

A full listing of eating and drinking places in
Michigan counties is available from the Michigan Department
of Public Health, since all food service establishments have
to register by county and type of establishment. Forty-one
percent of all establishments can be considered table
service restaurants, and 34 percent are fast food businesses
(Spotts, 1991).

Again, all of these numbers are quantitative in nature.
The data fails to indicate qualitative differences between
highly visited attractions like Bronner's Christmas Market
or Zender's in Frankenmuth and low visitation attractions
such as smaller regional festivals. However, most often a
highly visited attraction is accompanied by many
complementary attractions which are also listed.

Most attractions are located in Michigan's population
centers (see Figure 8.13). Nearly all are located in the
southern Lower Peninsula in cities along major highways.
Several major attractions can also be found in the Upper
Peninsula, primarily in Keweenaw, Houghton, Chippewa, and
Marquette counties. Grand Traverse county also features a
large proportion of attractions followed by Alcona, Otsego,
Antrim, and Leelanaw counties in the northern Lower
Peninsula, all are popular tourist destinations. However,

when looking at the spatial distribution of the attractions,

58
one has to keep the limitations of the data in mind; only
members of the Chamber of Commerce are included in the
planning guide. Counties that do not have a Chamber of
Commerce are therefore underrepresented. Second, numbers of
attractions may have nothing to do with the importance of an

attraction or its visitation.

9.911.923.2152:

Golf is an increasingly popular sport. Recently,
Michigan has been heavily promoted as a golfing vacation
destination, particularly through the Michigan Travel
Bureau's advertisement campaigns. The allocation of
promotion dollars amounted to 30 percent of the Bureau's
advertising budget (Spotts, 1991). The number of golf
courses in 1989 was based upon data gathered by Rasmussen,
Roy and Rasmussen in their "1990 Michigan Golfers Map &
Guide": it provides a complete listing of all golf courses
in Michigan.

Although Michigan is marketed as the golfer's paradise,
golf courses are concentrated exclusively in the population
centers of the southern Lower Peninsula, primarily in the
Detroit region and its neighboring counties, as well as in
Kent, Ingham and Jackson Counties. However, a handful of
golf courses are located in the tourist counties Grand
Traverse, Antrim, Charlevoix, and Otsego, as Figure 8.14

illustrates.

59

mm

The road system plays a vital role in providing fast
and convenient access to tourism regions and destinations.
Tourists cannot reach a destination without access roads.
The existence of highways in a county is therefore a good
indicator of its accessibility. A highway provides access
to a region; unfortunately it does not indicate whether
tourists actually stay in a county or whether they only pass
through.

The highway variable is a dummy variable which
indicates whether or not a county is connected to the
interstate highway of freeway system. It was obtained with
the help of a map study.

The spatial distribution of interstate highways and
freeways is documented in Figure 8.15. With the exception
of Mackinac and Chippewa Counties the Upper Peninsula is not
connected to the highway system as defined by this study.
The Lower Peninsula, in contrast, is very accessible by
highway. With the exception of the Thumb region, the
northeastern Lower Peninsula, and some counties in the
northwestern part nearly all counties are connected to the
highway system, with I-75 and I-94 being the most important

freeway systems.

60
rea

Winter sports are an important element of Michigan
tourism. They are important for the tourism industry
because they help to minimize the effect of seasonality.

The number of downhill ski areas was chosen as an
independent variable. Cross country skiing is an important
winter sport as well, but it is not limited to special
facilities. Cross country skiing can take place in any open
area, and therefore, it is difficult to derive a
comprehensive measure to use in a study such as this.

The names of existing ski areas were obtained from the
Michigan Travel Bureau which were then located on a map in
order to determine the county these ski areas are in.

Ski areas can be found in the Detroit area, in the
southwest corner of the lower Peninsula, which serves the
Chicago population, and close to cities in the center of the
southern Lower Peninsula (e.g. Lansing, Kalamazoo and Battle
Creek). The northwestern part of the Lower Peninsula has
many ski areas as well. This is where Michigan's hills are
located. The highest concentration of ski areas is located
in the western counties of the Upper Peninsula, where
Michigan's only mountains are to be found. These areas are
within easy reach for people from other Midwestern states,
particularly Wisconsin and Minnesota tourists (see Figure

8.16).

61
Limitations of the Study

This study is an attempt to relate various tourism
components to the success of a county in terms of tourism
potential. Despite the care with which the identification
of tourism components was undertaken, the ultimate choice of
variables remained to a large extent a question of arbitrary
judgment. As a result, it is possible that certain items
are missing from a list of those that could be studied.
Also, most of the variables utilized in this study are
quantitative in nature, representing tourism components but
not including information about the degree of usefulness of
the resource as a tourism component.

Second, this study takes advantage of secondary data
which was not collected for the purpose of this study.
Therefore, accuracy and appropriateness of selected
variables could be problematic. This may be a problem of
accuracy and validity related to the attraction data derived
from the "Travel Planner" which is distributed by the
Michigan Travel Bureau.

Third, this is a study of Michigan counties which can
aid many counties in their tourism planning process.
Unfortunately, tourism regions do not necessarily follow
political county boundaries. For instance, all lodging
facilities of a county might be concentrated in one part of

that county and not be equally distributed.

62
Lastly, the sample size is small, since there are only
83 counties in Michigan. Only five counties could be
reserved to test the regression model. This provides for

only a crude estimate.

CHAPTER V

FINDINGS AND DISCUSSION

In this chapter general findings related to tourism
components and their relationships to tourism potential of
individual counties are provided. This chapter is divided
into nine sections according to the steps of the regression
analysis: (1) correlation analysis, (2) attraction index
development, (3) division of the data set, (4) first
regression analysis, (5) analysis of influential cases, (6)
second regression analysis, (7) analysis of residuals, (8)
test of regression results, and (9) discussion of measures

of tourism potential.

Correlation Analysis
Twelvé‘independent variables were identified for the
study. The correlation matrix for the variables is
presented in Table 5.1.
As can be seen in Table 5.1 major significant
intercorrelations that could be threatening to the results
of the regression model exist between the number of golf

courses, eating and drinking establishments, and cultural as

63

64

Table 5.1: Correlation Matrix of Study Variables

Public Parks Shore Stream Lakes High
Land Line ways

Public L. 1.0000 .2005 .3497 .1304 .3469 -.4086

Parks .2005 1.0000 .2790 .0560 .0599 -.1565
Shoreline .3497 .2790 1.0000 .2188 .1561 -.1118
Streams .1304 .0560 .2188 1.0000 .1904 -.2391
Lakes .3469 .0599 .1561 .1904 1.0000 -.1936
Highways -.4086 -.1565 -.1118 -.2391 -.1936 1.0000
Ski Areas .1789 .0880 -.0630 .2307 .2360 -.0238
Golf -.4652 -.0311 -.1232 -.0412 -.0975 .4195
Entertain. -.3627 -.0782 -.0082 -.0322 -.1879 .3610
Culture -.2074 .0390 .1871 .0345 -.0239 .2804
Eat/Drink -.3021 -.0112 -.0291 .0027 -.0874 .2899
Access .2557 -.0409 .1172 .3255 .3942 -.2166
Ski Golf Enter- Culture Eating/ Water
tainm. Drinking Access

Public L. .1789 -.4652 -.3627 -.2074 -.3021 .2557

Parks .0880 -.0311 -.0782 .0390 -.0112 -.0409
Shoreline -.0630 -.1232 -.0082 .1871 -.0291 .1172
Streams .2307 -.0412 -.0322 .0345 .0027 .3255
Lakes .2360 -.0975 -.1879 -.0239 -.0874 .3943
Highways -.0238 .4195 .3610 .2804 .2899 -.2166
Ski Areas 1.0000 .1559 -.0038 -.0368 .0815 .3037
Golf .1559 1.0000 .5391 .6250 .8389 -.0569
Entertain. -.0038 .5391 1.0000 .7331 .6445 -.1652
Culture -.0368 .6250 .7331 1.0000 .6777 .0122

Eat/Drink .0815 .8389 .6445 .6777 1.0000 -.0441
Access .3037 -.0569 -.1652 .0122 -.0441 1.0000

65
well as entertainment facilities. Therefore, the number of
eating and drinking establishments, cultural and
entertainment facilities will be combined into one
attraction index, since they all represent developed
attractions/entertainment. Although the number of golf
courses is highly related to these attractions, it was not
integrated into the index. This was done purposely in order
to better determine the importance of golf as a tourism
component.

Unlike the entertainment variables, the four variables
that represent water resources cannot be combined into an
index without losing important information because they are
not strongly correlated and have a different distribution

over Michigan counties.

Attraction Index Development

An index is a composite measure of a variable and a
data reduction device. It combines several empirical
indicators of a variable into one single measure while
maintaining the specific details of all individual
indicators (Babbie, 1983).

The assumptions for creating an index are that all
items have to be highly empirically related to each other
and that each component has to add some new meaning to the
evaluation. Items of the attraction index in this study are

the variables eating and drinking places and culture as well

66

as entertainment facilities. As the correlation matrix
shows, the variables are highly interrelated, and since all
variables represent different types of attractions preferred
by different types of tourists, new meaning is added by each
component as well. Therefore, the assumptions for creating
an index are fulfilled.

The index was built according to the following rules:
In order to level the scales of eating and drinking
establishments with cultural and other attractions, a
percentage figure was calculated by dividing the number of
facilities available throughout Michigan into the number of
facilities contained by individual counties. The resulting
figures were then multiplied by one hundred for each county.
Finally, a weighted average for all three components was
calculated, devaluating the eating and drinking places
variable by 50 percent. Fifty percent was chosen for two
reasons: First, restaurants are essential, since all
tourists have to eat and drink, but they are not the main
attraction that attracts tourists to an area. It is rather
a factor that contributes to their comfort and satisfaction,
stimulating them to repeat visits (Ferrario, 1980). Second,
restaurants are likely to be more heavily used by residents
than other attractions.

The corrected correlation matrix is presented in Table
5.2: it reveals moderate relationships between the other

nine variables and the newly created attraction index.

67

Table 5.2: Correlation Matrix Including Attraction Index

Attraction

Index
Public L. -.3251
Parks -.0187
Shoreline .0619
Streams .0021
Lakes -.1121
Highways .3496
Ski Areas .0109
Golf .7394
Attractions 1.0000
Access -.0744

Division of the Data Set
According to the rules established in the procedure
section, Berrien, Crawford, Hillsdale, Kalkaska, and Ottawa
counties have been randomly chosen for exclusion from the
model building process in order to test the partial

regression coefficients later in this chapter.

First Regression Analysis

Two regression runs were performed, one using per
capita hotel/motel use tax as the dependent variable, the
other employing per capita lodging supply.

The results of the multiple regression analysis are
presented in Tables 5.3 and 5.4. The multiple regression
coefficients of .69 for the per capita use tax and .84 for
the per capita lodging supply are significantly high,

implying that tourism components in this study predict the

Table 5.3:

68

First Regression: Hotel/Motel Use Tax Per
Capita and Tourism Components

8 Beta T Sig T
Access -4.9239 -.08 -.71 .46
Golf -25.4507 -.27 -1.82 .73
Parks .0025 .07 .71 .48
Shore 3.9323 .43 4.05 .00
Ski Areas 238.2818 .32 3.17 .00
Streams -.6238 -.29 -2.91 .00
Lakes .0118 .16 1.51 .14
Highways 19.3041 .01 .14 .89
Public Lands 6.1051 .16 1.33 .19
Attractions 109.4750 .28 2.05 .05
(Constant) 297.3129 1.56 .12

Multiple R = .69039; Adjusted R Square = .39853

Note:

Table 5.4:

8 stands for partial regression coefficient;

Beta refers to standardized partial regression
coefficient:

R Square represents coefficient of determination:
Sig T refers to the observed significance level

First Regression: Lodging Supply Per Capita
and Tourism Components

8 Beta T Sig T
Access -.0107 .02 .20 .84
Golf -.2433 -.24 -2.15 .04
Parks .0002 .42 5.95 .00
Shore .0113 .11 1.43 .16
Ski Areas -.5192 -.06 -.85 .40
Streams -.0039 -.17 -2.25 .03
Lakes .0002 .24 3.13 .00
Highways 2.4773 .18 2.17 .03
Public Lands .1898 .47 5.10 .00
Attractions .6347 .15 1.47 .15
(Constant) .9597 .62 .54
Multiple R = .84008; Adjusted R Square = .66182
Note: 8 stands for partial regression coefficient;

Beta refers to standardized partial regression

coeff

icient;

R Square represents coefficient of determination:
Sig T refers to the observed significance level

69
level of tourism potential quite well. When looking at the
adjusted R squares, however, 66 percent of the total
variance in tourism potential can be explained using per
capita lodging supply as a measure: only 40 percent of the
total variance can be explained when using per capita use

tax as a measure for tourism potential.

Analysis of Influential Cases

Both models were checked for influential cases by
calculating the leverage for each county as defined in
Chapter III. Only the three most influential cases have
been excluded from the analysis since the sample size is
already small. Keweenaw, Chippewa, and Wayne counties have
been extremely influential cases in each model. It appears
their influence may be due to the fact that Keweenaw and
Chippewa are very successful tourism counties with very low
population bases. Wayne county is also a successful tourism
county due simply to its large population base and many

attractions.

Second Regression Analysis
After excluding these three counties from the model
another two regression were run which improved the results
markedly (Table 5.5 and Table 5.6). The multiple regression
coefficients were significantly higher in this later run

compared to the first. The adjusted R square for the per

70

Table 5.5: Second Regression: Hotel/Motel Use Tax Per
Capita and Tourism Components

8 Beta T S1g T
Access -2.5372 -.04 -.48 .63
Golf -28.8343 -.27 -2.50 .02
Parks .0029 .04 .51 .61
Shore 7.7808 .61 7.40 .00
Ski Areas 212.4990 .29 3.48 .00
Streams -.6131 -.29 -3.58 .00
Lakes .0076 .10 1.21 .23
Highways 109.5359 .08 .94 .35
Public Lands 6.1174 .16 1.68 .10
Attractions 109.4750 .37 3.44 .00
(Constant) 211.5491 .47 .64

Multiple R = .82455; Adjusted R Square = .62986; F = 13.59

Note: 8 stands for partial regression coefficient:
Beta refers to standardized partial regression
coefficient;
R Square represents coefficient of determination:
Sig T refers to the observed significance level

Table 5.6: Second Regression: Lodging Supply Per Capita
and Tourism Components

8 Beta T Sig T
Access -.0036 .00 -.08 .94
Golf -.1746 -.19 -1.74 .08
Parks .0000 .02 .27 .78
Shore .0390 .34 4.26 .00
Ski Areas -.2202 -.03 -.42 .68
Streams -.0030 -.15 -2.01 .05
Lakes .0002 .29 3.66 .00
Highways 2.9424 .25 2.91 .00
Public Lands .1848 .53 5.82 .00
Attractions .6812 .13 1.27 .21
(Constant) -.1740 -.13 .90

Multiple R = .84294: Adjusted R Square = .66531: F = 15.71

Note: 8 stands for partial regression coefficient:
Beta refers to standardized partial regression
coefficient;
R Square represents coefficient of determination;
Sig T refers to the observed significance level

71

capita use tax increased from 0.39 to 0.63. For per capita
lodging supply the R square remained the same (0.66). This
implies that 63 percent and 66 percent of the total
variance, respectively, can be explained by the tourism
components in each of the counties. In both cases the high
F-coefficients (F = 13.59 and F = 15.71, respectively)
suggest that the null hypothesis can be rejected -- the
dependent variables are not linearly related to the
independent variables.

In selecting an individual tourism component, the
variable, miles of shoreline had the most explanatory power
(Beta = .61) for the dependent variable of per capita use

tax, followed by attractions (Beta = .37). Ski areas (Beta

= .29), streams (Beta = -0.29) and golf (Beta -0.27) have
the same explanatory power, but it is only half of the
shoreline variable. The other variables have no
statistically significant impact. This does not necessarily
mean that these components do not have any explanatory power
related to tourism potential: rather, the statistically
insignificant results could be explained by the small sample
size or measurement problems related to the variable itself.

When using lodging supply per capita as a measure of

tourism potential, the total variance is explained

differently. Here, public lands (Beta .53) followed by

shoreline (Beta = .34) and lakes (Beta .29) have the most

explanatory power. In this model, attractions have less

72
explanatory power when measuring tourism potential in form
of lodging supply per capita (Beta = .13) and ski areas are
of practically no importance (Beta = -.03). Instead the
highway variable gains importance (Beta = .25). However, in
both models natural resources dominate over developed
attractions.

Both models show negative Beta scores for golf courses
and streams and very low Betas for parks. These facts need
further explanation. Golf courses are primarily located in
the populated southern Lower Peninsula and, therefore, have
a negative sign since the dependent variables are per capita
figures. The miles of stream are a very quantitative
measure and give no indication of the natural beauty or
attractiveness associated with a mile of stream. The stream
variable includes all streams. Many streams might be too
small or inaccessible for recreation purposes. Others might
flow through unattractive cities. These facts might explain
the negative Beta scores for streams. The lack of
importance of parks as a variable in the models could be
explained by the fact that Michigan Parks are located in or
near rural as well as urban areas. This would nullify the
relationship with the dependent variables. Therefore, the
Beta scores are diminished. By excluding Keweenaw and
Chippewa in the model building process, two counties have
been excluded that have extensive park lands which may

further lower the explanatory power of parks as a variable

73
in the model. Furthermore, acres of parks might not be the
ideal measure of touristic attractiveness, since it fails to

convey a qualitative characterization.

Analysis of Residuals

In order to test the regression assumptions a residual
analysis was performed. A normal probability function which
plots standardized residuals against predicted values and a
histogram of standardized residuals were calculated to
insure linearity and normality for the developed model. The
results are satisfactory and consistent with expectations
and assumptions for the model.

Furthermore, standardized residuals were calculated for
each county. These standardized residuals were mapped to
test whether any systematic error occurred in developing the
model. A standardized residual describes the difference
between an observed value and the value predicted by the
model. These differences are expressed in standard
deviation units above or below the mean. Therefore, a
negative sign indicates an underestimation of the predicted
value and a positive sign stands for an overestimation of
the predicted value by the model. Ideally, standardized
residuals have random variations, indicating that there is
no systematic error. Figures 5.1 and 5.2 show the spatial
distribution of standardized residuals which makes is easier

to detect regional patterns of residuals relating to the

 

 

 

 

 

 

 

 

 

  
 

Northern
Lower
Peninsula 2 + +

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 5.1: Residuals for Per Capita Hotel/Motel Use Tax

74

 

 

 

 

 

 

 

 

 

Northern

Lower __ _ _ _
Peninsula / + + r}

 

 

 

 

 

 

 

i++++

> 0 + + + —
l— < 0 ......
- .Test County + "

II Influential Case Southern T + l- -
Lower

Peninsula/ _ I .4. l _ l

  
    
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 5.2: Residuals for Per Capita Lodging Supply

75

76
variables in this study. As can be seen the spatial
distribution of residuals does not correspond to spatial
variations of any independent variables in this study.
Therefore, it can be assumed that no systematic error

occurred in the model development process.

Test of Regression Results

To confirm the regression results five counties were
excluded from the model development. The variables of these
counties were entered into the regression equation to
determine how well the model predicts the dependent
variables. The predicted values were then compared to the
actual observed value for each county. Since the regression
results were only tested on five counties, the results are
not statistically significant, but they do provide an
indication of how well the model predicts the dependent
variables. The equations used for the test were derived
from the regression analysis. For the per capita use tax

the equation was:

Y = 72.35 - 2.54 * xAC + 211.55 * x“. + 0.0029 * xPA + 7.78
* x33 - 0.13 * X81‘ + 212.5 * x9’" + 0.0076 * xL + 109.54
* x3 + 6.12 * xpL - 28.83 ... xG

For the per capita lodging supply the equation was:

77
Y = - 0.17 - 3.6 * XAC + 0.68 * x“. + 0.00001 * pr + 0.39 *
x33 - 0.003 * Xs-r - 0.22 * x3K + 0.0002 * xL + 2.94 * xH

+ 0.19 * XPL " 0.18 * XG

X103 number of public water access points

er‘ attraction index score XL : acres of lakes

XRA: acres of national and XE : existence of highways
state parks XPL: public lands ratio

)%H‘ miles of Great Lakes )gK: number of ski areas
shoreline XG : number of golf courses

XST: miles of streams

The test results are presented in Tables 5.7 and 5.8.
Since the predicted values are calculated from an equation
that consists of positive as well as negative sums it is
possible that the final value is negative. Hillsdale, for
instance, has negative values because of its large number of
golf courses. Due to the relatively small sample size of
five there are only three ways of judging the results.
First, the observed and predicted value can be compared in
absolute terms. Second, the standardized residuals can be
compared; the higher they are, the more inaccurate the
estimates. Third, the ranks of the five counties can be
compared to determine whether the counties are ranked the
same when using the observed or the predicted value as a

sorting criterium. Although the residuals are somewhat off

78
in both cases, when comparing standardized residuals the
observed values are reasonably close to the regression line.
Using the relative rank as a criterium, it becomes clear
that the regression results should not be used to predict
the absolute value of either variable but rather to indicate
a relative rank or trend. According to the five test
counties, both measures seem to predict the tourism

potential of a county equally well.

Table 5.7: Test of Regression Results for Per Capita

Use Tax

County Observed Predicted Residual Standardized

Value Value Residual
Berrien 426 693 - 267 - 0.72
Crawford 1288 773 515 1.72
Hillsdale 35 - 162 197 0.53
Kalkaska 118 176 - 58 - 0.16
Ottawa 157 726 - 559 - 1.53

Note: Sample Standard Deviation of Residuals = 372

Table 5.8: Test of Regression Results for Per Capita
Lodging Supply

County Observed Predicted Residual Standardized
Value Value Residual
Berrien 1.97 2.25 - 0.28 - 0.09
Crawford 16.75 13.21 3.54 1.09
Hillsdale 1.60 - 0.93 2.53 0.78
Kalkaska 2.85 7.56 - 4.71 - 1.45
Ottawa 1.74 3.81 - 2.07 - 0.64

Note: Sample Standard Deviation of Residuals = 3.24

79
Comparison of Tourism Potential Measures with
Images of Michigan Tourism Destinations

Comparing mental images of Michigan tourism
destinations with the dependent variables in this study
sheds light on whether the variables used in this study are
adequate measures of tourism and whether these mental images
reflect reality. However, an image is a mental
representation of an object, person, place or event which is
not physically before the observer (Fridgen, 1987).
Therefore, an image is a perceived reality biased by
motivations, attitudes, or demographics of an individual and
based upon impressions and past experiences.

Fridgen (1987) surveyed automobile travelers stopping
at Travel Information Centers in Michigan during the summers
of 1982 and 1983. In a cognitive map task, respondents
indicated which parts of Michigan they perceived to be
recreation and tourism regions by first, circling counties
that together form distinct tourism regions and by second,
placing the letter X in the county that constitutes the
center of that perceived tourism region. The data generated
from the survey was combined to form a Tourism Location
Score which made it possible to compare travelers
perceptions of where tourism regions are located within
which counties.

The Tourism Location Score in the Fridgen study was

calculated according to the following equation:

80
TLS = ( 0.4a + 0.48 + 0.2c ) * 1000,
A being percentage of X's received by each county, 8 being
the percentage of times a county was completely circled, and
C being the percentage of times a county was partially
circled.

The survey was conducted in 1982 and 1983. This study,
however, makes use of 1989 data. Over the course of the
years images as well as facts related to tourism development
are likely to have changed somewhat in some counties.
Nevertheless, by comparing the image scores of the Fridgen
study with the dependent variables in this study it is
possible to get a general impression of how well perceptions
correspond to use tax distributions and lodging supply.

As can be seen in Figure 5.3 the range of Tourism
Location Scores is highest in coastal counties, primarily
along Lake Michigan, at the connecting waters of Lake
Michigan and Lake Huron, or on Lake Superior. The highest
inland county is Roscommon County; it is located in the
center of the state and contains two of the larger inland
lakes within Michigan.

The image score clearly underestimates the importance
of inland counties as can be seen when looking at Figures
8.5 and 8.6. Besides coastal counties the tourism potential
measures in this study also identify inland counties in the
central northern Lower Peninsula, but they neglect urban

counties in the Detroit region as well as the Thumb

 

 

 

 

 

30 < Southern
wer

 

 

 

 

 

Peninsula

 

 

 

 

 

 

 

 

Figure 5.3: Image Scores of Michigan Counties

81

82
counties. Generally speaking though, there are
corresponding results, namely the importance of the Upper
Peninsula, the northern Lower Peninsula, and other coastal

counties, especially in western Michigan.

Discussion of Results

Two different measures of tourism potential have been
employed in this study. As can be seen from the regression
results, both measures have significantly high R squares.
Per capita lodging supply has a slightly higher R square
than the per capita hotel/motel use tax. When comparing the
test results and the results of the comparison with the
image scores, both measures do well.

However, there are distinct methodological and
measurement differences involved which are reflected by the
Beta-coefficients of the two measures. The per capita
hotel/motel use tax is not a comprehensive measure of
tourism. Although hardly any residents stay at hotels or
motels, a report (van Doren and Gutske, 1982) indicates that
the major market for hotels and motels in the United States
in 1978 consisted of only 32 percent tourists. Forty-three
percent are business travellers and 17 percent are
conference participants. Therefore, urban tourism is
overemphasized, which is reflected by very high Beta-
coefficients for attractions. On the other hand, other

important tourism lodging alternatives like camping are

83
neglected. The per capita lodging supply variable includes
hotel/motels as well as camping. Unfortunately, it is a
supply figure rather than a performance figure, and,
therefore, is not a complete measure either. Especially
when there are high variations in occupancy rates, supply
figures have to be interpreted with care (Koegh, 1984).
With the very seasonal usage of campsites, the Beta-
coefficient for public lands might be exaggerated, and the
importance of attractions and ski areas might be
underestimated.

However, the results of both models shed light on
Michigan tourism. First, natural resources are very
important tourism components in Michigan. Michigan has all
the right to be called the "water state." Both models reveal
the major importance of the Great Lakes shoreline. Streams
and public water access points, though, did not contribute
significantly to a county's tourism potential in this study.
Public lands are another valuable natural resource.

However, parks were not a significant tourism component in
this model compared to public lands.

The importance of natural resources does not imply that
attraction are unimportant. Considering the fact that there
is only one attraction variable compared to six natural
resource variables and that the attraction index is a rough

estimate of all the different kinds of attractions in the

84

state of Michigan, the influence of attractions on the
tourism potential of Michigan counties is considerable.

Since both measures convey slightly different
interpretations of the variables, there probably are
different users for each measure. Hotel managers who are
interested in predicting the success of the hotel industry
would appear to benefit from using the per capita use tax
model. The Campground Association, on the other hand, would
benefit more from the use of the per capita lodging supply
model to study their respective contribution to tourism in
the counties of Michigan. Regional planners should pay
attention to the composition of their county's lodging
supply. If counties have primarily hotels and motels and
few campgrounds, tourism planners in these counties should

use per capita use tax and vice versa.

CHAPTER VI

CONCLUSION

In this concluding chapter, a summary of the study is
provided along with a brief review of the findings and
conclusions. Implications of these findings are discussed

along with recommendations for further research.

Summary of the Study

The present study was based on secondary data provided
by various Michigan agencies. It deals with the analysis of
tourism potential of Michigan counties. Ten tourism
components were identified which were then related to
measures of tourism potential. Measures of tourism
potential in this study were per capita hotel/motel use tax,
a performance figure, and per capita lodging supply, a
supply figure which takes hotel/motel lodging as well as
camping into account.

This study was conducted in two phases. In the first
phase variables were identified and discussed in their
geographical context. A spatial analysis was conducted to

better understand regional differences of the variables

85

86
employed in the study. In the second phase the two measures
of tourism potential were regressed on the ten tourism
components identified and key variables which best explain
levels of tourism potential were identified.

There were three objectives stated in Chapter I. The
first specific study objective was to develop a model that
predicts a county's tourism potential. Two models were
developed which made use of the two different tourism
potential measures. The ten independent variables were the
public lands ratio, acres of National and State Parks, Miles
of streams and Great Lakes shoreline, acres of lakes, number
of public water access points, ski areas and golf courses,
attraction index score, and the existence of highways.

These variables predict both tourism potential measures
equally well (R square = 0.82 and R square = 0.84,
respectively).

The second specific objective was to explore
relationships between measures of tourism potential and
tourism components.' Due to the different nature of the
tourism potential measures the Beta-coefficients derived
from the regression analysis vary for both measures.
However, both models reflect the importance of natural
resources in the state of Michigan. The shoreline and
public lands are of particular importance. Attractions have
also a considerable influence on tourism potential.

National and State Parks as well as streams, did not have

87
significant explanatory power in this model. Neither do
developed facilities such as water access points and golf
courses (see Tables 5.5 and 5.6).

The third specific objective was to describe spatial
variations of tourism in Michigan. The spatial analysis
clearly shows the power of the population variable. The
population is concentrated in southern Lower Michigan as
well as most attractions and golf courses, whereas most
natural resources are located in the northern Lower
Peninsula and in the Upper Peninsula. The Upper Peninsula
and the Straits region clearly have the highest potential
for tourism. Middle Michigan, on the other hand is
primarily an agricultural area with very little to attract
tourists, although it is conveniently close to major

population centers.

Conclusion

The existence of tourism components as identified by
Robinson (1976), McIntosh and Goeldner (1984) and others
explains a county's potential for tourism very well.
Population is a very powerful factor in Michigan tourism;
therefore, it is important to control for differences in
county population by introducing per capita figures (Clawson
and Knetsch, 1969). The two tourism potential measures,
hotel/motel use tax receipts suggested by Bishop and Spotts

(1990) and accommodation numbers (Pearce, 1987), do equally

88
well. However, both have different uses due to the
different nature of the two measures.

As Gunn (1982) writes, regional analysis helps to
understand the location of potential tourism destinations
with respect to other features and to understand the natural
and cultural resources as well as the popular images of
those places. The regression results confirm the findings
of Var (1977) and Richie and Zins (1987). Natural resources
are of major importance in the State of Michigan. The
findings also comply with the results of Deale (1983) as
well as Raitz and Dakhil (1988), who emphasize the
attractiveness of seashores and beaches. However, the
findings of this study are specific to Michigan tourism. In
order to generalize the conclusion to tourism in general,
similar studies must be conducted to examine the validity of

the model for different destinations, counties or states.

Implications for Tourism Planning and Management
Proper'management and planning are important in order
to maximize benefits from tourism. Attempts to manage and
plan the growth of tourism will be enhanced if attention is
paid to patterns and spatial interactions of tourism. The
geographical study of Michigan tourism may lead to a better
understanding of the role of the tourism industry in a
county. As such, it may signal the success or failure of

regional development strategies or indicate the need for

I). \l t) .i-r.-l’i.’

\f

|,.\’ Ill .I ilk’ll'lll l.vui . ti‘ 'Ilrt\l|i'.lli.l\ iall’l‘ ll

 

 

89
these, alert authorities and researchers to likely problem
areas and suggest solutions that might be adopted.

To promote a destination effectively it is essential to
know the "prime magnets" around which a marketing strategy
might be developed. On the national level the results of
this study might aid the Michigan Travel Bureau in better
promoting Michigan. On the county level it could aid
counties in better promoting their assets, attractions and
potentials. Counties in the northern lower Peninsula, for
instance, could more heavily promote the beauty of their
public lands and the recreation activities that they
facilitate. In so doing, these counties could increase
tourism revenues and better compete with neighboring
counties in the Upper Peninsula.

The model might also aid hotel/motel managers or
campground owners in selecting potential locations for their
enterprises. A campground owner, for example, could examine
the data and methods included in this study to determine the
most opportune locations for new lodging ventures. Also, by
inserting their own tourism component data into the model
equation, revenue trends in the lodging industry could be
predicted by county planners. These trends could then lead
to alternative, perhaps more effective, development
strategies for individual counties, allowing area tourism to

develop to its fullest potential.

90

The major importance of public lands as a tourist
attraction may also help the Forest Service and other public
agencies to focus their multiple use philosophy more on
recreation and tourism.

Realizing the complexity of tourism in the state of
Michigan, this study might help to establish tourism
regions. Developing regions with similar tourism structures
might help counties to pool their efforts in tourism

development, promotion, and planning.

Future Research
Finally, research approaches to overcome and improve
study limitations as mentioned in Chapter III are
recommended in the following:

(1) Attractions are an important tourism component. For
the sake of future research it is necessary to work
towards a comprehensive inventory of developed
attractions in the state of Michigan and elsewhere. It
is also important to distinguish between different
types of attractions. Then the impact of attractions
on the tourism potential of a county should be
reconsidered.

(2) Future studies should work towards a qualitative
assessment of the research variables and a refinement
of measurements. Instead of using acres of park land,

for instance, the number of parks could be used

,’I\l.l|l.2\- 1’.) .\’i'

.(ll) lullltl Illillltllyi

 

 

ll ll. . i]\tl.. ll"! rill

(3)

(4)

(5)

91
assigning different weights to State and National
Parks. It would be also useful to develop a variable
that qualifies the miles of streams since not all
sections of streams are equally attractive.
Both measures of tourism potential employed in this
study are not "perfect." In order to combine the
advantages of both measures, data on overall lodging
performance should be collected. Visitation of
hotel/motels as well as of campgrounds should be
measured.
Future studies should deal not only with county-level
tourism. Instead or additionally, tourism regions
within counties should be identified since tourism
regions are not defined in terms of county borders.
In order to validate the results of the study, it
should be repeated for other states. The Beta-
coefficients should be tested on other Great Lakes
states with tourism industries similar to Michigan's

like IIlinois and Wisconsin.

BIBLIOGRAPHY

BIBLIOGRAPHY

Babbie, E. (1980). Inn Eznggige Q: Sgcini Rgseancn. Belmont:
Wadsworth Publishing Company.

Bishop,G. and Spotts, D. (1990). Micnigan Travei Activity;
i989 -1990 Winte; Senggn Began; - Seasonal Inavgi
ngnitoring Bgnong i 12. East Lansing: Travel, Tourism

and Recreation Resource Center, Michigan State
University.

Brown, C. J. D. (1944). Michigan Streams. Their Lengths,
Distribution and Drainage Areas. ' se 'on,
13 (5): 9 - 10.

Brown, T. (1981). Assessing Changes in Tourism in New York
State. Se :A 'cu tu e, (13): 1 - 9. Ithaca:
Cornell University Agricultural Experiment Station.

Burkhart, A.J. and Medlik, S. (1981). Ignzi§n_;_£n§;y
Engsenn_nng_£ninzg, Second Edition. London: Heinemann.

Chou, Ya-lun (1989). at's c s's s ss and
Eggnnnig_. New York: Elsevier.

Choy, D.J.L. (1985). ec s 0 us a ce.
Tourism Management.
Clawson, M. and Knetsch, J. (1969). Econonics Q: Qntdoor

Becneatinn. Baltimore: John Hopkins Press.

Crompton, J. (1990). Claiming our Share of the Tourism

Dollar. Banks ann Regneagign, March 1990: p. 42.

Curtis, J. (1990). Mo Rec - The Challenge of the 905. Enyn
nng Recnention, January 1990: p. 52.

Deale. C S- (1983): W

W East Lansing:
Department of Parks and Recreation, Michigan State

University, Thesis.

Enzel. R-G- (1989). W
annnn. Washington, D.C.: Inter-Ski Services, Inc.

92

 

 

 

 

93

Ferrario, F. (1980). Tourist Potential and Resource
Assessment. In D. Hawkins, E. Shafer and J. Rovelstad,

Tourism Planning nnd Developmenn issnes, Washington
D.C.: George Washington University.

Fridgen, J. (1987). Use of Cognitive Maps to Determine
Perceived Tourism Regions. Lgignze Science. 9 (2): 101
- 107.

Goodrich, J. (1978). A New Approach to Image Analysis
Through Multidimensional Scaling. Jpnrnnl_pf;1:nyel
Resenggn, 16 (3):3 - 5.

Gunn, C.A. (1979). Tpnrisn_21nnning. New York: Crane, Russak
& Co.

Gunn, C. A. (1980). An Approach to Regional Assessment Of
Tourism Development Potential. In D. Hawkins, E. Shafer

and J- Rovelstad T2uri_n_Plannins_and_2s_slsnnsnt
issues, Washington D. C.: George Washington University.

Gunn, C.A. (1982). A P o as d M t o c or Id n
Areas of Tourism Development Eotennini in Qanana.

Ottawa: Government Office of Tourism.

Holecek,D. (1991). Characteristics of Michigan's Travel
Market. In Spotts, D. v Tour s ' ‘chi n -

A Stntistignl Ppofiile, Second Edition. East Lansing:

Research Monograph # 2, Travel, Tourism and Recreation
Resource Center, Michigan State University.

Hudson Institute (1985). Micnignn Beypng 2909; Einnl
ngpng. Indianapolis: Hudson Institute.

Humphreys, C.R., Ali Razaque, J.O. Vaeteh, and Rex Yu. n.d.

elimin e ort: Stat ca a a ela e
Michigan angp Resonnges nng Wnie: use . Lansing:
Department of Resources Development, Michigan State
University.

Hunt, J. (1974). Ima e: C ° ' flue c o o s

Qeveiopmenn. Logan: Institute for Outdoor Recreation
and Tourism, Utha State University.

Hunt, J. (1984). An Analysis of State Image Change over a

Twelve Year Period (1971 - 1983). 0 ve
3g§gnppn, Fall 1987: 15 - 19.
Kachigan. S K- (1982)-uultixar1ats_Statistical_Analxsisi__A

Congepguai Ingponuggi n. New York: Radius Press.

94

Koegh, B. (1984). The Measurement of Spatial Variations in

Touristic Activity- Annels_2f_Tsurism_Bessarsh. (11):
267 -282.

Livingstone, P. and Mitchell, L. (1989). A Regional Analysis
of South Carolina Tourism. 0 '5 se h,
10: 301 - 317.

Makens, J. (1987). The Importance of US Historic Sites as

Visitor Attractions. lournal_sf_Traxsl_8sssarsh. Winter
1987: p. 8.

Mathieson, A. and Wall, G. (1982). Ingrisn;__figgngnigy
Phxsisal_and_§osial_lnnacts New York: Longman InC-

Menchik, B. (1991). Director of the Qen;gryj_y_geneygippnen_
of innusppiaiized States, Michigan State University.

Personal Interview.

Michigan Association of Private Campgrounds, Inc. (1991).
. Williamsburg, Michigan:
Michigan Association of Private Campgrounds.

Michigan Department of Natural Resources. n.d. MignignnL§
Egnegt Bgsonpces i222; An Assnssngnp. Lansing:

Michigan Department of Natural Resources.

Michigan Department of Natural Resources. n.d. Mignignn
Sinte Banks. Lansing, Michigan: Michigan Department of
Natural Resources, Parks Division.

Michigan Department of Natural Resources (1974). nignignn
Lnke Inventppy. Lansing: Michigan Department of Natural
Resources, Office of Planning Services.

Michigan Department of Natural Resources (1990). upgnpgg
Public Access Sign Master Linn, Serial Number 624-90.

Lansing: Michigan Department of Natural Resources,
Recreation Division. Lansing, Michigan: Michigan
Department of Transportation.

Michigan Department of Transportation (1991). ' l n:

Michigan Department of Treasury (1985). finsing§§_§nine_np

n erst di M' ' a '8 es Us .
Michigan Travel Bureau (1989). t de
and.§alsndar_of_zxsnts. East Lansing. Michigan:

Michigan Travel Bureau.

 

 

 

95

Michigan Travel Bureau (1990). Mignignn_1;nyel_£1nnngp. East

Lansing, Michigan: Michigan Travel Bureau.

O'Halloran, R.M. (1988). ' ° ' o

An_Analxs1s_of_uishisan_9_untx
Sgcongnzy Dnta with Recommendntions for Utilization by
Bnblic and Pnivate Tourisn Planners and unnageng. East

Lansing: Department of Parks and Recreation Resources,
Michigan State University, Dissertation.

ott. L. (1977). An_Introdusiion_to_Statistisal_nethods_and

Qnta Anniysis. Belmont: Duxbury Press.
Pearce. D- (1987)- T2urisn_I2sa21__A_§sosranhisal_Analxsis.

New York: John Wiley & Sons.

Public Sector Consultants (1987). Qgpnping_nipnignn;§
Euture: Towards the Xea; 2000. Lansing:Public Sector

Consultants.

Rasmussen, R. and B. Rasmussen (1990). i229_ni§nignn_§plfig;§
nap g guide. Detroit: RSG Publishing, Inc.

Raitz, K. and Dakhil, M. (1988). Recreational Choices and

Environmental Preferences. Annels_of_Touri§m_Bsssarch.
(15): 357 - 370.

Ritchie, J. and Zins, M. (1978). Culture as a Determinant of
the Attractiveness of a Tourism Region. Annn1§_pfi

Tourisn_8sssarsh. 5 (2): 252 - 267-

Robinson, H. (1976). A Geograpny of Tourisn. Plymouth:
MacDonald and Evans Ltd.

Smith, C. (1982). Regional Analysis of Tourism Resources.

Annals pf Tounisn Besgnzgn, 254 - 273.

Spotts, D. (1986). Tnavei ann Tounisn in nicnigan - A
snngisnigni Ppofile, First Edition. East Lansing:
Research Monograph # 1, Travel, Tourism and Recreation
Resource Center, Michigan State University.

Spotts, D. (1991). ve o 's ' ' -
singisgicnl Ppofile, Second Edition. East Lansing:
Research Monograph # 2, Travel, Tourism and Recreation
Resource Center, Michigan State University.

Talhelm, D. (1991). Recreational Boating in Michigan. In
Spotts, D. Travel ann Ioupisn in nicnignn - A
spatisiigni Profile, Second Edition. East Lansing:
Research Monograph # 2, Travel, Tourism and Recreation
Resource Center, Michigan State University.

96

Tull, D.S. and Hawkins, D.I. (1990). Mn:3e§ing_3g§gnzgn

ngngnzgngn§_nnn_n§nnpn§. New York: Macmillan Publishing
Company.

van Doren, C.S. and Gutske, L.D. (1982). Spatial Analysis of
the 0.8. Lodging Industry, 1963 - 1977. Annn1§_pfi
Tourism_8essarsh. 9 (4): 543 -563-

Var, T. (1974). Establishing a Measure of Touristic

Attractiveness. J2urnal_2f_Traxsl_Besearsh. 12 (4): 1 -
8.

Var, T. (1977). Determination of Touristic Attractiveness of
the Touristic Areas in British Columbia. Jpnrnnl_pf

Traxel_8esearsh. 15 (3): 23 - 29-

Wells, P. and Eidelson, M. (1991). Federal and State
Recreation Lands in Michigan In Spotts, D. ave nd
Tourism in Michigan - A Statispicai Enptiie, Second
Edition. East Lansing: Research Monograph # 2, Travel,
Tourism and Recreation Resource Center, Michigan State
University.

APPENDIX A

DATA

Table A.1: Dependent Variables

97

Mulber- of Guestrooms All Lodging
plus divided by
Campsites Population

County Use Tax Use Tax
(in Dollars) divided by

Population

Alcone 11,505 113
Alger 124,660 1,389
Allegan 107,703 119
Alpene 68,428 224
Antrin 289,960 1,595
Arenac 14,270 96
Baraga 29,456 370
Barry 7,599 15
Bay 61,988 55
Benzie 52,001 426
Berrien 391,539 243
Branch 61,274 148
Calhoun 102,517 75
C888 7,696 16
Charlevoix 308,624 1,438
Cheboygan 67,597 316
Chippewa 131,607 380
Clare 29,130 117
Clinton 10,240 18
Crawford 157,937 1,288
Delta 122,610 325
Dickinson 90,095 336
Eaton 164,721 177
Emmet 548,239 2,189
Genesee 337,398 78
Gleduin 11,091 51
Gogebic 97,865 542
Grand Traverse 1,252,970 1,949
Gratiot 43,880 113
Hillsdale 15,017 35
Houghton 85,389 241
Huron 45,227 129
lnghem 124,506 44
Ionic 15,528 27
Iosco 157,006 520
Iron 28,786 218
lsabelle 195,290 358
Jackson 163,020 109
Kalamazoo 71,159 32
Kalkeska 15,914 118
Kent 1,063,379 212
Keweeneu 21,079 1,235
Lake 22,893 267
Lapeer 8,003 11
Leelanau 257,898 1,560
Lenawee 34,341 38
Livingston 68,296 59
Luce 20,257 352
Mackinac 443,168 4,152
Macomb 517,625 72
Menistee 110,350 519
Marquette 148,237 209
Mason 90,103 353
Mecosta 50,197 135
Menominee 27,335 110
Midland 212,863 281
Miseaukee 14,807 122

Campsites

210

3,419
2,114
4,781

343

917
1,159

4,457

douse-onus

d

-I am
O VudM-D
C O O I Q

d
. . . .

d-OOOOEI‘UOOUI

d
.

838233283238833338383833333888i:3&381'183013888833838358888

.b
Og-IN-IN-DONlO-DUN-DOV

..Sfieree

d
5-IN5-INQO
.

98

Table A.1 (cont'd.): Dependent Variables

County Use Tex Use Tax Number of Number of Guestrooms All Lodging Population
(in Dollars) divided by Guestrooms Campsites plus divided by
Population Campsites Population
Monroe 111,233 83 655 1,207 1,862 1.39 133,600
Montcalm 5,044 10 205 1,226 1,431 2.70 53,059
Montmorency 28,737 322 289 477 766 8.57 8,936
Muskegon 79,892 50 1,444 1,514 2,958 1.86 158,983
Meweygo 11,248 29 219 1,696 1,915 5.01 38,202
Oakland 1,697,966 157 9,714 3,530 13,244 1.22 1,083,592
Oceene 30,433 136 396 2,166 2,562 11.41 22,454
Ogemew 7,778 42 291 1,003 1,294 6.93 18,681
Ontonagon 70,866 800 425 394 819 9.25 8,854
Osceola 3,970 20 83 593 676 3.36 20,146
Decode 8,797 112 192 497 689 8.79 7,842
Otsego 217,976 1,214 1,177 642 1,819 10.13 17,957
Ottewe 295,266 157 843 2,426 3,269 1.74 187,768
Presque isle 19,033 138 183 486 669 4.87 13,743
Roscommon 94,702 479 1,298 1,236 2,534 12.81 19,776
Saginaw 332,284 1,514 1,795 261 2,056 9.37 21,946
St. Clair 246,074 169 1,045 1,395 2,440 1.68 145,607
St. Joseph 42,881 73 619 1,256 1,875 3.18 58,913
Senilec 18,092 45 318 1,080 1,398 3.50 39,928
Schoolcraft 98,945 1,192 591 602 1,193 14.37 8,302
Shiewessee 37,267 53 185 193 378 0.54 69,770
Tuscole 10,232 18 110 184 294 0.53 55,498
Van Buren 83,897 120 582 736 1,318 1.88 70,060
Washtenaw 976,272 345 3,216 438 3,654 1.29 282,937
weyne 2,508,767 119 15,795 810 16,605 0.79 2,111,687
Hexford 150,041 569 669 1 205 1,874 7.11 26,360

Total 15,919,966 96,503 4,689,240 3,769,741 9,105,303

 

99

Table A.2: Independent Variables

County Public National Utter Attrac-
Land 8 State Shore- Access Highway Ski tion Golf
Ratio Parks line Streams Lakes Areas Existence Areas index Courses

Alcona 27.2 549 27.4 362 11,842 8 0 0 0.03 2
Alger 40.7 36,740 120.4 709 8,679 9 0 0 0.22 1
Allegan 8.7 909 24.0 517 7,031 28 1 1 0.99 12
Alpena 11.7 1,270 61.0 301 12,747 9 0 0 1.25 1
Antrim 13.2 0 26.5 264 17,879 32 0 2 0.80 8
Arenac 12.1 0 47.3 156 0 9 1 0 0.04 2
Baraga 20.7 7,593 69.9 696 6,651 22 0 0 0.44 1
Barry 6.4 5,000 0.0 272 8,565 33 0 1 1.08 6
Day 1.6 143 36.2 50 112 8 1 0 1.59 6
Benzie 32.3 10,991 24.8 104 17,016 23 0 1 0.45 4
Berrien 0.6 2,066 13.7 500 2,271 13 1 1 2.92 11
Branch 0.1 413 0.0 325 7,468 17 1 0 0.86 3
Calhoun 0.0 0 0.0 540 3,688 17 1 0 2.91 9
Cass 1.1 0 0.0 229 8,250 24 0 1 0.10 6
Charlevoix 19.8 3,188 102.4 215 22,266 20 0 1 0.26 6
Cheboygan 35.8 1,430 34.5 420 48,523 23 1 0 0.28 3
Chippewa 40.3 26,228 456.0 800 9,015 26 1 0 1.72 5
Clare 14.0 36 0.0 331 4,041 16 1 2 0.63 3
Clinton 2.5 2,678 0.0 319 484 6 1 0 0.28 5
Crawford 60.9 9,724 0.0 204 2,730 31 1 2 0.03 2
Delta 40.3 712 198.7 514 3,092 24 0 1 0.72 4
Dickinson 44.8 0 0.0 645 4,992 26 0 1 0.92 3
Eaton 0.1 0 0.0 207 407 5 1 0 1.35 7
Emmet 24.5 7,871 75.0 98 8,288 7 1 3 0.30 3
Genesee 0.0 0 0.0 355 3,413 5 1 0 2.87 16
Gladwin 26.2 365 0.0 473 6,349 12 0 0 0.23 2
Gogebic 43.8 9,767 30.1 1,204 27,253 15 0 4 0.07 2
Grand Traverse 21.2 746 55.5 168 13,464 26 0 0 2.13 6
Gratiot 3.7 0 0.0 241 1,165 3 1 0 0.64 4
Hillsdale 0.7 0 0.0 298 2,776 9 0 0 0.29 4
Moughton 30.2 516 50.6 923 21,391 29 0 1 2.11 3
Huron 2.3 1,278 92.5 942 76 13 0 0 0.92 6
ingham 1.3 34 0.0 234 533 4 1 0 3.96 19
lonia 2.8 4,018 0.0 464 1,727 15 1 0 0.67 4
losco 37.9 70 36.2 259 10,102 11 0 0 0.87 3
Iron 33.0 316 0.0 902 21,376 31 0 1 0.44 2
isabella 0.5 0 0.0 330 1,511 3 1 0 0.49 5
Jackson 3.3 12,963 0.0 324 8,523 9 1 0 2.51 17
Kalamazoo 1.6 3,086 0.0 365 8,477 19 1 0 2.08 13
Kalkaska 42.4 0 0.0 284 3,733 26 0 0 0.40 1
Kent 1.5 0 0.0 772 5,786 24 1 2 2.90 23
Keweenaw 36.5 133,964 100.5 271 4,957 11 0 0 1.01 1
Lake 46.7 0 0.0 250 2,639 36 0 0 0.23 1
Lapeer 2.9 4,061 0.0 594 3,448 8 1 0 1.10 7
Leelanau 22.1 45,589 151.4 58 17,003 17 0 1 0.67 2
Lenawee 0.9 3,498 0.0 622 3,934 8 0 0 0.79 8
Livingston 4.7 12,059 0.0 469 7,725 11 1 1 0.93 10
Luce 50.0 10,688 31.0 658 12,478 29 0 0 0.42 1
Mackinac 51.2 1,951 298.4 347 21,467 22 1 0 2.69 4
Macomb 0.7 1,984 26.5 296 895 6 1 0 1.89 25
Manistee 30.9 201 25.4 276 7,041 23 0 0 0.90 4
Marquette 23.3 1,124 79.4 1,906 22,401 62 0 2 0.97 2
Mason 19.9 4,116 27.5 238 8,180 26 1 0 0.91 2
Mecosta 4.2 0 0.0 293 8,497 25 1 0 0.08 2
Menominee 13.8 650 41.3 815 3,803 17 0 0 0.26 0
Midland 12.4 0 0.0 309 2,409 2 1 0 0.71 2

100

Table A.2 (cont'd.): independent Variables

Highway

Ski

Existence Areas

d
. . . C Q C . . . . U .

Golf

County Public Mational Hater
Land 8 State Shore- Access

Ratio Parks line Streams Lakes Areas

Missaukee 27.5 0 0.0 209 4,287 7
Monroe 2.0 1,035 56.6 459 265 10
Montcalm 4.8 0 0.0 477 5,099 29
Montmorency 37.4 290 0.0 306 8,846 20
Muskegon 7.0 2,541 26.9 394 9,966 12
Hewaygo 20.7 0 0.0 484 7,816 20
Oakland 4.9 28,062 0.0 469 17,792 36
Oceans 16.6 2,871 27.4 224 2,938 5
Ogemaw 26.3 4,329 0.0 381 4,122 28
Ontonagon 38.1 49,610 56.2 1,282 10,391 8
Osceola 5.1 0 0.0 301 1,611 19
Oscoda 56.4 0 0.0 219 2,380 10
Otsego 29.6 62 0.0 198 4,905 19
Ottawa 0.5 376 25.0 307 4,709 15
Presque isle 19.7 5,822 68.7 301 13,594 20
Roscommon 54.8 960 0.0 204 37,536 9
Saginaw 4.6 0 0.0 593 1,400 8
St. Clair 2.5 1,862 164.2 1,007 0 4
St. Joseph 0.7 0 0.0 734 23,121 27
Sanilac 1.4 238 40.5 307 631 0
Schoolcraft 64.8 956 64.1 959 385 12
Shiawassee 0.3 0 0.0 292 8,136 18
Tuscola 5.6 0 20.1 184 908 7
Van Buren 0.4 661 13.4 324 4,737 28
washtenaw 3.3 14,134 0.0 372 6,810 18
Mayne 0.4 944 75.4 391 1,886 10
Hexford 38.9 248 0.0 254 6,297 14
Total 20.2 485,586 3,003 6,350 671,137 1,406

c>ca-a--cacac>cac:-UIc>c>na-c:cac>c>c:a~c>nic>ca-c:
83333383383'8’1’1288388121383838338

..a
O
C

APPENDIX B

MAPS ON SPATIAL DISTRIBUTION OF EACH VARIABLE

 

     

Upper Peninsula

 

 

 

 

Figure B.l: Spatial Variation in the Location of Guest Rooms

101

 

 

 

Northern
wer
Peninsula

 
 
 

250 <
500 <
1,000 <
1,500 <

 

Figure B.2: Spatial Variation in the Location of Campsites

102

 

 

     

E] s 50,000

 

50,000 4:] 3 100,000
100,000 < s 200,000
200,000 < 3 500,000

 

 

500,000 <-

 

 

Figure B.3: Spatial Variation in Hotel/Motel Use Tax

103

 

     
      
  

Upper Peninsula

 

 

 

 

Peninsula

 

 

 

 

[:l 3 10,000

 

 

 

 

10,000 4: 3 50,000
50,000 < s 100,000
100,000 <. 3 500,000
500,000 s m
<- 33,2?
Peninsula

Figure B.4: Spatial Variation in County Population

104

 

    
      
    
  
 

Upper Peninsula

 
 

.~-.

9 0“
i;

.o

  

Peninsula

 

 

 

 

 

 

 

 

 

Peninsula

 

 

 

 

 

 

 

Figure B.5: Spatial Variation in Per Capita Lodging Supply

105

 

    
 

Upper Peninsula

I
15¢»?ch

 

Peninsula

 

 

 

 

Southern

 

 

 

 

 

 

Lower
Peninsula

 

 

 

 

 

 

 

 

 

 

 

Figure B.6: Spatial Variation in Per Capita Hotel/Motel Use Tax

106

 

    
     
   
 

Upper Peninsula

o'éé- In.

Northern
Lower
Peninsula

as,

 

' by: «$93,.

 

 

 

 

 

Peninsula

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure B.7: Spatial Variation in Public Lands Ratio

107

 

 

      

«as,

        
 

Peninsula

  

 

 

 

 

 

 

 

 

= 0
0 < 5 1,000
1,000 < S 5,000
5,000 < 5 10,000
10,000 <

 

Figure B.8: Spatial Variation in the Location of National and State Parks

108

 

     
    
 

Upper Peninsula

 

 

Peninsula

 

 

 

 

 

 

 

Southern
Lower
Peninsula

 

 

Figure B.9: Spatial Variation in Miles of Stream

109

 

     
   
 

Upper Peninsula

 

Peninsula

 

 

 

 

 

 

 

 

 

 

 

 

 

Peninsula

 

 

 

 

 

 

 

 

Figure B.10: Spatial Variation in Miles of Shoreline

110

 

     
      
 

Upper Peninsula

[:l ,

 

1,000 4:] g 5,000
5,000 <
10,000 <
20,000 < So them
I nger

Peninsula

Figure B.11: Spatial Variation in Acres of Lakes

111

 

    
      
  

Upper Peninsula

Peninsula

 

30 < Southern
wer
Peninsula

Figure B.12. Spatial Variation in the Number of Public Water Access Points

112

 

     
      
  

Upper Peninsula

Peninsula

 

 

 

Southern
Lower
Peninsula

Figure B.13: Spatial Variation in the Attraction Index Score

113

 

    
     
   
 

Upper Peninsula

 

 

Peninsula

i

 

 

 

 

 

 

 

 

 

Southern
Lower
Peninsula

 

Figure B.14: Spatial Variation in the Number of Golf Courses

114

 

 

 

 

N orthem

Lower
Peninsula
[:1 No Highways
I Highways
Southern
Lower
Peninsula

Figure B.15: Spatial Variation in the Existence of Highways

115

 

 

 

 

Northern ,
Lower
Peninsula

 

 

 

 

Southern
Lower
- Peninsula

 

 

 

 

 

Figure B.16: Spatial Variation in the Number of Ski Areas

116

 

"Illlllllllllllllllll