A VESITATION MODEL
FOR SELECTED CAMPGROUNDS N
THE PROVINCE OF QUEBEC

Dissertation for the Degree of Ph. D.
MICHIGAN STATE UNIVERSITY
JACQUES ARTHUR AUGER
1974

Eflmfiui 6.3a;- “rifle-a "‘ '

s2..~z

w HER-AR; y ;._

    
 

 

 
 
   
      

 
 

y magma av g”
"DAB & SUNS'
303K BINDER?" INC.
1 33:23.1},8135525.

  

 

 

 

 

 

 

.
,II

 

 

 

ABSTRACT

A VISITATION MODEL FOR SELECTED CAMPGROUNDS
IN THE PROVINCE OF QUEBEC

BY

Jacques Arthur Auger

The present study was undertaken to estimate
participation in selected campgrounds in the Province of
Quebec. In Quebec none of the Provincial Parks are used
exclusively for recreation and all of them are used
simultaneously for timber, mining and recreation output.
To achieve this, a systematic sampling survey was first made
in the city of Sainte-Foy located in the vicinity of the
city of Quebec. This was done by sending a questionnaire
to every fifth family for a total of 1700 households who
were asked to participate in the survey. The purpose was
to find out which socioeconomic characteristics have a
significant influence on the camping participation origi-
nating from this city. The 693 answers received were
processed by regression analysis. Although some variables
within the equation were statistically significant at the
.90 level of confidence, they did not account for a great
deal of the variation around the mean of the dependent
variable: the coefficient of determination (R2) for the

equation was 0.074, which indicated a poor goodness of fit.

Jacques Arthur Auger

The results thus obtained indicated that with this
type of model, it is very difficult to make an estimate of
future participation; yet the use of the "t" test and the
chi-square indicated quite clearly which socioeconomic
characteristics used could have a significant influence on
camping participation.

The process of analysis revealed that the younger
populations, the peOple with lower incomes and those with
larger families have a tendency to produce a large number
of camping days, whereas non-campers tend to belong to the
older age classes and have higher incomes and smaller
families.

In a second study with a different approach, a
sample was made of the participation recorded in three
different campgrounds as established by governmental
statistics for 1971. The purpose of this study was to
define the relationships between campgrounds participation
and different characteristics of the zones of origin of
the campground users. Three sets of data were collected:
1) the socio-economic characteristics of the zone of
origin; 2) the communication characteristics between the
zone of origin and the zone of destination; and 3) the
attraction characteristics of individual campgrounds.

These selected characteristics (independent varia-
bles) were then related to campground participation

(dependent variable) in statistical models.

Jacques Arthur Auger

By means of a stepwise regression analysis the size
and slope of the regression coefficients were estimated.
The coefficient of determination (R2) for each of the three
campgrounds ranged between .56 and .69. The equations
developed were then used to estimate the future participa-
tion in camping at the different campgrounds.

It was found through these equations for the years
1975 that the participation of the population will have
increased by 18%, 18%, 9% for each of the three campgrounds
respectively. For 1980, the predicted increases in
participation will be 35%, 26% and 15% for each of the

three campgrounds respectively.

A VISITATION MODEL FOR SELECTED CAMPGROUNDS

IN THE PROVINCE OF QUEBEC

BY

Jacques Arthur Auger

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements

for the degree of
DOCTOR OF PHILOSOPHY

Department of Resource Development

1974

ACKNOWLEDGMENTS

I wish to express my appreciation of the helpful
suggestions and co-operation of Dr. Daniel E. Chappelle,
Professor, Departments of Resource Deve10pment and Forestry
who directed the research. Special thanks also go to Dr.
Michael Chubb (Associate Professor, Department of
Geography), Dr. Milton Steinmueller (Professor, Department
of Resource Deve10pment), and Professor Sanford S. Farness
(Professor, Department of Urban Planning and Landscape
Architecture), all of whom served on my Committee.

Special thanks are also due to Dr. Josés Llamas,
Professor, Department of Civil Engineering at Laval Univer-
sity, who assisted in the statistical analysis of the data
and supplied helpful suggestions during the course of the
work.

I am particularly indebted to the Quebec Department
of Tourism, Fish and Game for granting me a leave of
absence in order to facilitate the completion of my studies

and the attainment of the degree of Doctor of Philosophy.

ii

TABLE OF CONTENTS

Chapter
INTRODUCTION . . . . . . . . . . . . . . . .
I REVIEW OF EXISTING LITERATURE . . . . . . .
The Demand Curve Models . . . . . . . . .
The Origin Center Models . . . . . . . .

The Site-Oriented Models . . . . . . . .

II ECONOMETRIC MODEL FOR PREDICTING THE
DEMAND . O O C O O C C O U C C C C C O O O 0

FOR CAMPING IN A PARTICULAR CITY . . . .

Origin Area . . . . . . . .
Size of Sample and Method of Sampling
The Variables . . . . . . . . . . .

Discontinuous Independent Variables .
Independent Cross Variables . . . . .
Quadratic Variables . . . . . . . . .
The Model . . . . . . . . . . . . . .
Level of Significance . . . . . . . .
Analysis of the Data . . . . . . . . .

III ECONOMETRIC MODEL FOR PREDICTING THE
DEW D O O . O O O O O O O O O O C O O O O 0

FOR DIFFERENT CAMPGROUNDS . . . . . . . .

The Variables . . . .
The Data . . . . .
The Attraction Index
Factor Analysis . .
Correlation . . .
Initial Factors .
Rotated Factors .

Communality . .
Eigenvalue . .

Factor Scores . . .
Proposed Application of the Factor

Analysis . . . . . .
The Variables . .
Activities . . .

iii

Page

13

14
21
25

32
32

34
34
35
38
4O
44
45
46
47

56

Chapter

The Physical Environment . . . .
Services . . . . . . . . . . . .
The Role of Factor Analysis . . . .
Factor Loading . . . . . . . .
Factor Score Coefficients . . .
Factor Scores . . . . . . . . .
Adjusted Scores . . . . . . . .
Campers' Preference Study . . . .
Activities . . . . . . . . . .
Weight . . . . . . . . . . . . .
Weighted Activity Scores . . . .
Adjusted Scores . . . . . . .
Calculation of the Attraction Index
The Model . . . . . . . . . . . . .
Analysis of the Data . . . . . . . .
Level of Significance . . . . . . .
Interpretation of the Results . . .
Method of Estimating Future
Participation . . . . . . . . . . .
General Application . . . . . . . .

SUWRY . O O O O C O O 0 O O O O O O 0

CONCLUSION AND SUGGESTION FOR FURTHER
STUDY 0 O O C O O O O O O O O O O O O O

BIBLIOGRAPHY . . . . . . . . . . . . . . .

WPENDICES O O O O O O O O O O O O O O O O

A.

Factors defined by the factor analysis
method, their values, communalities,

and factor score coefficients . . .
Participation expected at the three
campgrounds studied-years 1975 and
1980 O O O O O O O O O O O O O O O

Questionnaire used in the Ste-Foy
Study 0 I O I O O O O O O O O O O 0

List of Provincial Campgrounds . .

iv

Page

83
90
91
92
94
96
100
103
110
112
114
116
118
123
131
132
133

151
152

154

157
162

166

166

170

179

180

Table

10
11
12
13
14
15
16
17
18
19

20

LIST OF TABLES

Variables . . . . . . . . .

Mean Comparisons Between Campers

Non—Campers Subsamples .
Camping Participation . . .
Camping Participation . . .
Camping Participation . . .

List of Variables . . . . .

Varimax Rotated Factor Matrix

Factor Score Coefficients .
Factor Scores . . . . . . .
Standardized Factor Scores .
Adjusted Factor Scores . . .
Natural Characteristics . .
Services . . . . . . . . . .
Water . . . . . . . . . . .
Weighted Park Factor Scores
Raw Scores . . . . . . . . .
Activity Weights . . . . . .
Weighted Activity Scores . .

Standardization and Adjusted
scores 0 I O O O O O O O

Attraction Index . . . . . .

Activity

Page

40

51
52
53
54
80
92
95
98
100
102
106
106
107
108
111
113

115

116

119

Table Page

21 County of Argenteuil . . . . . . . . . . . . . 122
22 Values of Dependent and Independent

Variables . . . . . . . . . . . . . . . . . 125
23 Final Regression Equation . . . . . . . . . . 133
24 Projected Population . . . . . . . . . . . . . 143
25 Projected Income . . . . . . . . . . . . . . . 147
26 Estimated Number of Recreation Visits to

the Three Campgrounds during July and
August a o o O o o o o o o o o o o o o o o 152

vi

LIST OF ILLUSTRATIONS

Figure Page
1 Elements in Outdoor Recreation . . . . . . . . 8
2 Cross Variables . . . . . . . . . . . . . . . 42
3 Basic Measures for Coefficient of

Determination . . . . . . . . . . . . . . . 49

4 Distribution of Quebec Provincial
Campgrounds . . . . . . . . . . . . . . . . 58

5 Form Used by the Provincial Parks Service
During the Summer of 1971 . . . . . . . . . 64

vii

INTRODUCTION

Recreation and leisure'are two words that we hear
more and more in everyday conversation. Sociologists,
politicians, union leaders, clergymen and university
professors are greatly concerned with the notion of leisure
and recreation which is constantly assuming greater ‘
importance in the life of every citizen of industrialized
countries.

Furthermore, it seems that trends are such that in
future decades, recreation may become a field of activity
so important that it could have a primordial influence on
the economy and social order of the civilization of
tomorrow.

It would be interesting to analyze the evolution
of this phenomenon as well as the ultimate results of its
development, but we will have to limit the scope of our
investigation to working out some of the mechanisms of its
implementation, taking into account the economic and social
factors involved. These various concepts will be considered
time and again in the course of the present study which will
be restricted mainly to that part of outdoor recreation

which is called camping.

Let us first find a proper and apprOpriate
definition for recreation. In most cases, differences of
opinion on the notion of recreation is due to restricted
meanings in the definitions used by different authors.
For instance, in their definition of recreation, Clawson
and Knetsch noted that: "Recreation, as the word is used
in this book, means activity (or planned activity) under-
taken because one wants to do it."1 The Neumeyers,
distinguished sociologists who took a special interest in
the field of leisure, have defined recreation as follows:

" . . . any activity, either individual or collective,

pursued during one's leisure time."2
In the present study a definition of recreation
is proposed, which is as broad as possible and which would
be more satisfactory than those just mentioned. Such a
definition can be formulated as follows: Recreation is any
action, other than those answering a basic physiological
need such as eating or survival or basic psychological
needs, which satisfied the needs of the individual during
the time that he is engaged in this particular action. It
is important to note that such a definition is underlain by

the basic principle that a rational man always tries to

satisfy his needs.

 

1Marion Clawson and J. L. Knetsch, Economics of Outdoor
Recreation (Baltimore: John Hopkins Press, 1966), p. 6.

 

2Martin H. Neumeyer and Esther S. Neumeyer, Leisure and
Recreation (New York: Ronald, 1958). P. 17.

 

However, we must keep in mind that man's outlook may
be governed by two different approaches to the problem.
First, he may have a long-term motivation behind the
initiatives he takes; he is then probably trying to satisfy
future needs. A good example of such long-term planning is
when an individual, in order to develop his faculties and
perfect his mind, spends a good part of his early life
studying. He may prefer to do other things during these
years of his youth, yet he keeps on studying since he
hopes, in so doing, to satisfy future needs. In opposition
to the above, a large number of man's actions are
characterized by very short-term planning. This is
especially true when he attempts to satisfy his needs by
proceeding at once with the execution of his projects.

Such a short-term approach is well illustrated when a boy
watches a hockey match or a football game instead of
studying. He then satisfies his immediate needs while
watching the game, and his future is the least of his
cares. Some individuals succeed in combining work and
recreation; for instance, if writing this paper completely
satisfies the individual's actual needs and, at the same
time, contributes to a proper satisfaction of his future
needs, it is then considered a recreational experience as
well as a valuable work achievement.

The above discussion on the definition of recrea-
tion includes certain specific indoor or outdoor

activities which are considered as integral parts of

 

recreation. As mentioned above, this study deals
exclusively with outdoor recreation activities and consti-
tutes a rationale for including at this point, a discussion
on the characteristics of this particular type of
recreation.

One of the main requirements for most outdoor
recreation activities is the availability of certain types
of natural resources. Indeed, outdoor recreation
activities cannot very well be enjoyed without the presence
of bodies of water, forests, swamps, mountains or other
natural features. In this respect, outdoor recreation is
no different from any other use that man may have for
natural resources, such as farming, mining or timber
harvesting. Yet it is important to remember that there is
nothing in the physical landscape of any particular piece
of land that makes it a recreation resource. It is really
the combination of the natural qualities of the environ-
ment and the ability and desire of man to use these for his
personal enjoyment that makes a useful resource out of what
otherwise would be a meaningless combination of rocks,
soil and trees.

It may be said that outdoor recreation is very
rapidly becoming one of the major users of the nation's
natural resources. Some existing data indicate that the
present trend in this direction will likely increase

during the next decades. For example, estimates prepared

 

 

by Outdoor Recreation Resource Review Commission indicate
that the demand for recreation facilities will have doubled
by the year 2000, even if the participation of each citizen
presently engaged in this type of activity does not increase
above the present levels.3 However, it is preferable to be
very cautious with such predictions, because studies have
indicated that a saturation point has already been reached
for many activities and that a decline may be indicated
for the near future. A good example of such a phenomena
is the snowmobile market which, for the last couple of years,
has leveled up and may even decrease in the near future.

In the past, research in the field of demand for
outdoor recreation has been very scarce and irregular.
Such research has also been greatly influenced by many
fundamental problems resulting from the attitude of many
workers in the field of recreation. Chappelle, in a
state of the art paper takes a comprehensive look at these
basic problems.

The source of many problems in the field of
recreation is as mentioned by Chappelle: "It appears that
most recreation professionals hold strongly to the basic

tenet that all people "need" recreation, or even more

 

3U.S. Outdoor Recreation Resource Review Commission,
Outdoor Recreation for America (Washington, D.C.: U.S.
Government Printing OffIce, 1962), p. 30.

extreme,_that all peOple need certain types of outdoor
recreation activities and experiences."4

Such an affirmation appears quite groundless as it
does not seem to be based on facts and is not the result of
serious studies. In fact, this author was unable to find
scientific research that reached such a conclusion.

Also according to Chappelle several subtenets seem
to derive from the above such as: "the population's
'needs' for recreation must be fully satisfied" and "that
recreational services cannot be priced in the way other
goods and services usually are."5

Acceptance of the above premises by individuals
supposedly specialized in the field of recreation can give
a wrong orientation to research programs undertaken in that
field. Indeed, if these premises were true, what would
be the necessity of doing any research on the demand aspect
of recreation?

However, even if the two above-mentioned premises
were true, it would be absolutely impossible to satisfy
them completely on account of the limited quantity of
natural resources and the important concurrence between the
different potential users. It is the author's confirmed

Opinion that for the proper administration, distribution

 

4Daniel E. Chappelle, "The Need for Outdoor Recreation:
An Economic Conundrum," Journal of Leisure Research, 5
(Fall), p. 47.

5Ibid., p. 47.

and conservation of the natural resources available for
outdoor recreation purposes, a price must be attached to
almost every use required by the people who choose this way
of enjoying Quebec's natural resources. Thus it becomes
necessary for the individual who chooses to visit and use
public outdoor recreation areas to weigh the cost of other
goods or services (including other types of recreational
activities) that he might have acquired for the same amount
of money. There is a personal choice to be made which
depends in large measure upon his personal scale of values
and his preferences, but conditioned also by many social
factors inherent to that specific level of the society to
which he belongs.

Some key elements in outdoor recreation are shown
in the following chart which has been drawn from two

different figures; one by R. I. Wolfe,6

and the other by the
Government of Ontario.7

As stated above, conflicting demands exist between
industries and the people involved in outdoor recreation

concerning the use of the same resources. Furthermore,

experience shows that various segments of the population may

 

6R. I. Wolfe, "Perspective on Outdoor Recreation: A

Bibliographical Survey," The Geographical Review, (19),
p. 213.

7Ontario, Department of Tourism and Information,
Tourism and Recreation in Ontario. (Toronto: Department
of Tourism and Information of Ontario, 1970): pp. ll-25.

 

a a

 

mmmhfl COHU

*

 

 

 

 

 

 

 

 

   

 

mwnnfinofl
-mwhowu UGOHUMMMG wocmumflu Mammufl mwum 0:» mCHUMAxO
nomzuon :ofluaummEou muflaflnfimmmou< mo aua>wuomuuu¢ uo >uoucm>cH
b e _ + 115
maowuowwoum hammdw
wounSmmu Humans: :0
4IIIlIlllIlIIIIIlllllllllllIIlIlllIlllll. whommon mummuocH
V J r T ._. a
muonuo mcflawmm
9.3352 9.33m
mmauu oocmu amount mcawxm
a d‘ a mewafimu use: mmwfiflwww
mumwmm 30Hu pcwxmo3 ouflxmom mowuflcoEm uwnuo Hw>mua mcwucmm
mufiommmo Hmcofluowuap we) HNCOmmmm can ommwuooo o w .
p o t 11» Emwusoe muuomm
:mamuq :onmomcoo P|||allcluL
03.”
Immuomx uoouuso
mcaccmHm :« :oflummao
we? nwuumm 636.35
0 homo u fillllall'a Hw>mH :oHumoaum
mmwuaaaowu OWUMMMMWM moon: HmofimoHonuxmm ousmHmH no
pwucm+umlw0ma nonmam :ofiumuauaumuum HMfloom nodumnawmuuofioa
. maqmus mousuoswum canmmumofioo . ..
_ 9 woman .NUEOE who wEwu who:
mmwum comma QUHDOmuu _ wwnznmwm oawn0E0u5< "Ga muonmnu
Eouw “my wamowm mo _ »
coflmuuumucoo ummmuuq _ a coHumNfiamnua :oflmasmom 00:05 <
_ mowuzfiog uuommauua o . _ Lawn
\
_ a a
8:3 8i 112.6% I Illllll. JESSE). \\ 5330
\
/|uul. llllllllllllllllll l\

 

 

ZOHB<mmme KOODBDO ZH mBZWSMAm

A musmflm

have demands which differ considerably from one another and

are often conflicting within the field of outdoor recrea—

tion. For anyone having the responsibility of making such

recreational activities available, it is of primary

importance to take these factors into consideration before

determining the types of land and water resource develop-

ments to be undertaken by public as well as by private

developers.

It is of utmost importance to establish first some

definite objectives which are pursued in predicting demand

for outdoor recreation. These objectives can be listed as

follows:

1..

Identify and analyze present and potential outdoor
recreation issues and opportunities.

Propose and evaluate alternative approaches in pre-
dicting demand for outdoor recreation.

Acquire a proper understanding of the components and
interactions of the recreation system.

Evaluate the demand for different types of recreational
activities.

Develop statistical planning tools to assist in the
evaluation of the effects of proposed future develop-
ments.

Recommend governmental policy guidelines and

strategies for outdoor recreation developments.

10

For many years, workers in recreational planning
have repeatedly pleaded for "more data". More often,
however, the data requested and collected pertain only to
one narrow portion of the spectrum of outdoor recreation
without an appropriate knowledge of what was collected and
why such data were collected. In too many instances, data
were collected simply for the sake of having statistics or
simply to have the information, without knowing exactly
what to do with such data.

Now, with the aid of a model and its broader
framework, it is possible to see the interrelationship
between numerous elements of the problem. Such a model also
provides a comprehensive framework in which the data
collected can be used. Furthermore, it shows what data
might be needed and likely to be used within such a
structure. It can show the relative priorities in the
collection of different types of data. It also explicitely
indicates how the information will be processed and what
portion of the entire problem will be deve10ped and
explained.

It is important to be very cautious when developing
a prediction model, because the basic data used in the
model change very rapidly due to technological improvements
in this field of activity which is so important to our
society, and, as a result, the model can prove worthless

within a very short time.

11

Yet, this sort of difficulty should not prevent us
from developing models, and it is the author's opinion that
it is worthwhile to develop complete models for shorter
periods of time in order to have adequate adaptation to
changing conditions. It is further possible that the true
solution to this problem is that given by Chappelle when
he says: "It is important to have a continuous process of
model revision as conditions change."8 That can mean more
work and greater cost. However, according to Chappelle,
it is not an insurmountable job and it should be possible,
at least in theory, to develop simulation models for the
various subsystems which can be both spatially and
temporally dynamic.

Some years ago the institutional bases of society
were quite stable and it was then possible to make accept-
able provisions based on different socio-economic
variables. It was then quite normal to believe that these
variables would normally evolve according to a slow and
regular pattern. However, during the last few decades,
important changes have occurred in our basic institutions
and have completely changed the approach that we might have

taken a few years ago in solving any problem concerning

 

8Daniel E. Chappelle, Quantitative Analysis in a
Qualitative World: Modeling Forestry_System to Imorove
Decision-Making. Workshop on Computer and Information
System in Resource Management Decisions (Cleveland:
Society of American Foresters National Convention,
Septembre, 1971), p. 12.

 

 

 

12

socio-economic implications such as the one which is con-
sidered in the present study.

One of the major factors which make the building of
an outdoor recreation model extremely difficult is the
complexity and the rapid change taking place continuously
in our society. It is very difficult to know what will
happen to any particular component, even over a short
period of time. There are a great numer of additional
complications which are brought about by other factors
such as the ephemeral quality of a great amount of today's
goods as well as their increasing variety.

The above are only a few of the many different
changes that occur in today's society. These changes make
the work of researchers much more difficult than in the
past and can cause great complications in the building of a
model intended to analyze the needs of the society and make
accurate predictions for the future of such needs, which is

part of the present research program.

CHAPTER I

REVIEW OF EXISTING LITERATURE

As mentioned before, the problem of allocation for
outdoor recreation activities has been ignored by
researchers for a long time seemingly because the two
following assumptions were generally accepted. First, it
was assumed that every individual has a strong need for
recreation and, second, that this need must be completely
satisfied. Many professionals in the field of recreation
are still convinced that the above-mentioned assumptions
must be maintained. It must be said, however, that fewer
of these professionals accept such a point of view nowadays,
and that a great number of them are questioning its
validity; as a result of this, during the last few years,
valuable research programs have been initiated in order to
estimate the demand for outdoor recreation. These research
programs can be grouped according to the various character-
istics they might have in common.

In the following section, some techniques used in
estimating recreation demand are summarized. Many models
have been prepared by various authors. The most important

of these are presented below and can be grouped as follows.

13

14

The first group comprises models that estimate
consumers' surplus, while the second group estimates demand
originating from particular population centers or origin
centers. These models are defined according to the various
characteristics of a specific origin center which is the
area from which most of the users come. Finally, the third
group of models estimates the demand for a particular
recreation site according to the characteristics specific
to the site and not to the origin centers as in the two
other types of models listed above.

Before we start describing the different models, it
is important to mention one of the tools which have been
used to forecast future outdoor recreation requirements;
it is a time series of visits to specific recreation areas.
Occasionally such time series have been linearly extra-
polated and thus used to indicate the trend of annual
visits over a number of years. This method has the advan-
tage of being simple and easy to apply. However, it does
not take into account possible changes in different socio-
economic factors thought to be important in affecting

recreational participation.

The Demand Curve Models

 

Quite a few researchers are interested mostly in
determining the economic value of outdoor recreation,

because they feel the need of a common measure to compare

15

recreational activities with other more easily measurable
values of land uses.

Following the initial lead of Hotelling,1 Trice and
’Wood2 have suggested that the benefits which a person de-
rives from his use of a particular recreation area are
directly related to the distance he is prepared to travel
in order to visit that area. It may be inferred, therefore,
that the benefit is related to his travel cost. Trice and
WOod proposed that an analysis be made of the point of
origin of all people visiting a particular recreation area.
These points of origin could then be plotted on a map and
grouped into distance-zones around the recreation area.
From data thus obtained, the average cost of travel from
each zone to the recreation area could be calculated. The
average cost of travel for visitors from any point in the
most remote zone would set, as demonstrated above, a
"market" value for the recreation opportunities afforded by
the area. In this way, it could be reasoned that any
visitor from a zone located closer to the recreation area
would receive a certain amount of free recreation value,

or consumers' surplus, equal to the difference between the

 

1H. Hotelling, "Letter to R. A. PREWIT," in The .
Economics of Public Recreation (Washington, D.C.: National
Park Service, Department of the Interior, 1947).

2A. H. Trice and S. E. Wood, "Measurement of
Recreation Benefits," Lands Economics, XXXIV, 3 (1958)
pp. 196-207.

 

16

average cost of travel from the most distant zone and the
average cost of travel from the zone in which he lives.
According to this method proposed by Trice and WOod, the
economic value of this particular recreation area would be
equal to the aggregate consumers' surplus and this could be
obtained as follows: Calculate the difference between the
average cost of travel from the most remote zone or
"market zone" and the average cost of travel from each of
the other zones; multiply such difference by the number of
visitors from each of these zones; the sum of these
products over all zones are taken as the value of the
recreation area.

There is considerable disagreement concerning
application of consumers' surplus to the assessment of
recreational benefits, and a number of authors argue that
such a value is not the equivalent of benefits. Some of
their conclusions are quoted below.

Scott says: "The concept we seek, therefore, is
not the area under the demand curve for visits, which
would imply discrimination, but the largest profit rectangle
that can be inscribed, which would imply a single price."3

It follows that a valuation which is to justify a

recreational use of a piece of land must also be a

 

3A. D. Scott, "The Valuation of Game Resources: Some
Theoretical Aspects," Proceeding of the Symposium on the
Economic Aspects of Sport Fishing, Report No. 4 (Ottawa:
Dept. of Fisheries of Canada, May 1965), p. 27.

l7

simulation of a competitive rent. The valuer must not
assume that the hypothetical sale of recreational services
discriminates between customers; all must be assumed to
pay the same price per visitor-day.

Finally Coomber and Biswas mentioned that: ”This,
in effect, is a measure of total user benefits (consumers'
surplus being equivalent to net benefits) and not
comparable with normal market prices established for other
elements of, for example, a multi-purpose scheme which
includes recreational facilities."4

According to the author, such a comparison would be
possible if the consumers' surplus for other goods were
measured. However, the consumers' surplus could be used
in order to compare the economic value of the different
recreation developments. If we do so, the value of the
comparison is of interest because a common basis is used to
make the comparison. This method, however, does not suggest
any way to construct a demand curve which could be used to
estimate the consumers' surplus.

In order to enlarge on the method devised by Trice

and Wood and make it more practical, ClawsonS suggests a

 

4Nicholas H. Coomber and Asit K. Biswas, Evaluation of
Environmental Intangibles (Ottawa: Ecological Systems
Branch, Department of Environment, 1972), p. 26.

 

5M. Clawson, Methods of Measuring_the Demand and Value
of Outdoor Recreation (Philadelphia: Resource for the
Future, 1959), Reprint No. 10.

 

18

procedure which permits the building of a series of demand
curves in order to measure recreation benefits of an area.

This is done in the following manner: For any
given area already subdivided into a series of distance-
zones, as described above, Clawson built a demand curve for
each distance zone. In that curve, the abscissa is the
number of visitors who visit the area per 1000 of popula-
tion for each distance-zone. The ordinate is the total
average cost per visitor. This comprises the cost of travel
from the point of origin to the recreation area to which
is added, for each distance-zone, an "entrance fee". The
size of this fee is progressively increased with the
distance.

This approach permitted Clawson to estimate the
effect of a variable entrance fee on the number of visitors
from any of the points or origin. The other variables,
such as the socio-economic characteristics of the visitors,
are assumed to be constant.

Once the demand curves were obtained, as described
above, for each distance zone, Clawson built an integrated
demand curve which showed the relationship between the
number of visitors from the region around a recreation
area and the admission fee, by adding for each individual
zone the number of visitors corresponding to each level of
entrance fee.

Based on the same principles as outlined by Clawson,

but a little more sophisticated, the method of analysis

 

 

19

devised by Knetsch6 assumes that each distance group is
different and reacts differently to a change in entrance
fee. In order to achieve this, he proposes the following
formula: V = f(C, Y, S, G) where V = number of visits per
1000 population; C = cost paid by the park users for such
visits including the cost of travel, food, lodging as well
as the entrance fee; Y = income of the park users; S =
number, or the density, of the competing recreational areas
near the geographic area under study, and finally, G =
degree of congestion in the area.

Other authors using the same basic formula as
above have made interesting analyses of certain specific
activities within a recreation area. For example, in 1965
Wennergren7 estimated the demand for boating at various
Utah reservoirs, using the relation between average number
of trips per boat and average travel cost plus on-site cost.

A similar study by Smith and Kavanagh,8 in 1968,
concerned the empirical estimation of a Clawson-style demand

curve for a trout fishery center. This study was based on

 

6J. L. Knetsch, "Outdoor Recreation Demands and Bene-
fits," Land Economics, XXXIX, No. IV, (1963).

 

7E. Boyd Wennergren, Value of Water for Boating

Recreation, Bulletin 453 (Utah: Agricultural Experiment
Station at Utah State University, 1965).

 

8R. J. Smith and N. J. Kavanagh, ”The Measurement of
Benefits of Trout Fishing: Preliminary Results of a Study
at Grafham Water, Great Ouse Water Authority, Huntingdon-
shire," Journal of Leisure Research, 1, No. 4 (Autumn
1969).

 

20

data obtained at the Grafham Water, Hungtindon, England.
The basis of the Clawson approach in this instance was a
relationship between the distance from the place of living
to the fishery and the number of visits from zones
surrounding the fishery. Using this relationships, a total
demand curve for trout fishing was constructed.

A more direct approach was used by LaPage.9 He
determined a demand curve by simply asking a sample of
712 families, whether they would camp "much more," "much
less" or "about as much as they camp now" at standard
Family camping fees of $1.00, $2.00, $3.00, $4.00 or $5.00
per night. Then, by assigning an arbitrary multiplier to
the "much less" and "much more", the author constructed
a rough demand curve. This method seems very subjective,
and as LaPage noted: "At a later point in the interview,
each camper was asked whether he was not camping as much
as he would like and, if not, why not. The responses of
these campers to this latter question appear to be quite
inconsistent with their affirmation that they would camp
more often if the entrance fees were at the minimum level
of one dollar. Therefore such inconsistency in the answers
of the consumer indicates that this approach is unreliable

and should be used with great care.

 

9W. F. LaPage, "The Role of Fees in Campers Decisions,"
Research Paper NE-118 (Washington, D.C.: U.S. Forest
Service, 1968).

 

21

The Origin Center Models

 

An origin-center approach has often been used to
determine the demand for specific geographical areas.

As population increases and competitive pressures
for land preservation and development of recreation areas
continue to increase, the necessity to estimate the demand
originating from different population centers becomes
evident. These studies were made by using a new set of
models based almost exclusively on certain socio-economic
characteristics specific to the population of these areas,
such as: total population, income, age, number of children,
etc. Based on this principle, Gillespie and Brewer10
prepared a water-oriented recreation demand model by a
method of direct sampling of a metropolitan population.

An economic model was developed to assess the
socioeconomic characteristics influencing the demand for
water-oriented outdoor recreation in the St-Louis area.

The variables considered were assumed to have a significant
influence on the demand for water-oriented outdoor
recreation. For this reason, all variables were considered
in a first phase of the analysis. Using the ”t" test for
the b value of each variable, the variables not found to be

significant is different from a second were dropped and an

 

10G. A. Gillespie and Durward Brewer, An Econometric
Model for Predicting Water-Oriented OutdoorlRecreation
Demand (Missouri: Agricultural Experiment Station, Dept.
of Agricultural Economics, University of Missouri, 1969).

 

22

additional computer run was made considering only the
significant socioeconomic characteristics. Factors found
to be most significant were: annual family income,
education, sex, race, age and occupation. Covariance
analysis was used to isolate the influence of each
specific factor or characteristic for this type of
recreation. This model gives an estimation of the number
of recreation days produced by a specific origin area.

Based on the procedure outlined above, McCoy11
attempted to establish a relationship between certain
socioeconomic factors of households in a specific popula-
tion and the intensity of use of outdoor recreation by the
members of each household within the same area. He utilized
a stepwise regression analysis where, at each step, one
variable is added to the regression equation.

Least squares procedures have been used in order to
determine the values of regression coefficients. The
resulting coefficients may be employed to estimate future
participation when the values of the independent variables

were estimated for target years, if certain assumptions

are maintained.

 

11E. W. McCoy, "Analysis of the Utilization of Out-
door Recreation in Tennesseew—(unpublished Ph.D. Thesis,
University of Tennessee, Dept. of Agricultural Economics,
1966).

 

23

A different approach was used by Cicchitti gt 31.12

in an econometric analysis where they provided an empirical
model of the demand and supply factors for various outdoor
recreational activities.

According to the authors, in order to estimate
future demand for outdoor recreation activities, the use
of a single equation which does not take into account the
supply characteristics of the market can only be success-
ful if the market is one where supply can vary with demand.
This means that if demand increases for different reasons
such as increase in income, population, etc., supply
changes and adapts itself to changing conditions. In this
manner, within a relatively short period of time, a new
equilibrium can be reached between supply and demand.

The authors suggest that supply changes might be
followed by demand changes in a fashion known as recursive
structure. In such cases, the quantity demanded during
one specific year (T) is a function of the quantity
supplied during the preceding year (T—l). The model used
in their study is a disaggregated demand and supply model
for one specific individual in a definite time period.

The technique utilized in Cicchitti's model is

straight-forward. The underlying population was divided

 

12Charles J. Cicchitti, Joseph J. Seneca and Paul
Davidson, The Demand and Supply of Outdoor Recreation
(New BrunsWiCk: Bureau of Economic Research, Rutgers,
The State University, 1969).

 

24

into groups or classes based on such key socioeconomic
characteristics as race and age. Each population subgroup
was isolated for forecasting and simulation purposes.
Projections have been developed for each class for any
given year, and these projections have been aggregated to
provide forecasts for the given year.

A more global approach was used in 1962 by the
ORRRC.13

A separate projection was prepared for each
recreation activity for which current participation rates
were available within socioeconomic subclasses of the
population. The gross effects, from 1960 to target dates,
on participation rates of each socioeconomic factor were
estimated by reweighting the 1960 rates according to a
projected distribution for each of five factors: family
income, education, occupation, place of residence and age-
sex. In order to remove the multicollinearity effect
between independent socioeconomic variables, a multivariate
analysis was used.

The composite result of all factors acting together
was then estimated from the net effects mentioned above to

give a series of projected rates of participation by the

population in any activity.

 

13Outdoor Recreation Resource Review Commission,
Report No. 26, cp.cit., p. 11.

25

The Site-Oriented Models

 

Some models have been developed to estimate and
predict demand for different specific recreation sites. In
such models, only variables which characterize the
recreation site are used.

Some authors used a gravity model in order to
estimate the demand for outdoor recreation. Such a model
draws on the concept of population potential developed by
Stewart.14 It is based on the principle that any popula-
tion concentration exerts an influence that varies directly
with its size. However, as explained by Richardson,15
this influence is reduced by distance, a force field
surrounds each population cluster and, at any point within
the field, its intensity can be measured by dividing the
size of the population by distance.

16 formulated a probabilistic

Wennergren and Nielson
model and established a technique to estimate usage of
recreation sites. The only two variables used in this

model are distance and size of the site. From this simple

 

14J. Q. Stewart, "Empirical Mathematical Rules Con-
cerning the Distribution and Equilibrium of Population,"
The Geographical Review, 37 (1947).

 

15Harry W. Richardson, Regional Economics (New York:
Praeger Publishers, 1969), p. 39.

 

16Boyd E. Wennergren and Darwin B. Nielson, ”Proba-
bility Estimates of Recreation Demands," Journal of Leisure
Research, 11 (1970).

26

model, the authors attempts to formulate probabilities of
use by recreationists coming from various points of origin.

Wennergren and Nielson started with the basic
hypothesis that the utility-generating ability of a recrea-
tion area is inversely related to the distance between the
site and the place of residence of the users. A second
hypothesis used is that the size of a recreation area can
be an important datum because it is assumed that the
utility generated by a recreationist increases with the
size of the area.

In the case of a boating recreation area, the authors

developed the following equation:

 

_ a b
Pik " Sk / Dik
a b
2 Sk / Dik
K = 1
where:
Pik = Probable number of visits to a boating

site (i) from a given origin area (k)

th

Sk = Surface area of the k boating site

th h

Dik = Distance from the i origin to the kt
boating site

a = A parameter which reflects the effect of
the size of the area of the site on the

number of trips to the site

27

b = A parameter which reflects the effect of
the distance on the number of trips to the
site.

In the above equation, the use of only two
characteristics to reflect the boaters' utilization of a
given site greatly simplifies its application. The least
that we can say of the above technique is that it is far
from being realistic if applied to a real world situation.

A different approach was tried by Carlton S. Van
Doren. In this study the author was interested in
designing a suitable method for projecting attendance of
campers at the Michigan State Parks. Three steps are
included in the procedure. The first step consisted in
classifying each state park in terms of its attraction as
a trip destination. This was expressed in terms of
attraction indices. A second step uses regression analysis
to evaluate the relationship between attendance at the
park (camper-days) and the attraction indices.

The third step consisted in building a gravity
model in which the components are the attraction indices
and the number of camper-days produced in the 59 park
destinations by users from the 88 origin areas.

According to Van Doren, this travel model can be

used as a predictive tool for estimating use of existing

 

17C. S. Van Doren, "An Interaction Travel Model for
Projecting Attendance of Campers at Michigan State Parks:
A Study in Recreational Geography,“ (Unpublished Ph.D.
Thesis, Michigan State University, 1967).

28

or proposed recreational sites. However, even if the
attraction index is a very important independent variable,
and possibly the only one, a model based almost solely on
this variable is very incomplete, since it does not take
into account variables such as the socioeconomic factors,
etc. Another problem with Van Doren's approach is that he
restricted his variables to those only that are within the
limits of the campground or to the campground character-
istics.

Crampon18 used such a model as a tool to estimate
the number of persons likely to visit any specific
destination from any point of origin and also to determine
the significance of certain characteristics of these
visitors.

The concept on which Crampon's model is based is
that a specific and measurable relationship does exist
between the visitors from a specific point of origin who
are attracted by a given recreation area; this number of
visitors is taken as the dependent variable.

In Crampon's model, the independent variables
used are:

l) the magnitude of the population of a market (origin
area) and

2) the distance between the distination and that market.

 

18L. J. Crampon, "A New Technique to Analyze Tourist
Markets," Journal of Marketing, 30, (April 1966): PP. 27—
31.

 

29

The relationship between these variables can be

expressed by the following equation:

Vod = b Po Tod

where:

<
II

od The number of visitors from a given market

area (0) visiting a given destination (d)

P0 = The population of the market (0)
Tod = The travel distance between (0) and (d)
b = A constant

This model was applied to all the states in the
United States, including the District of Columbia. Each
of these were considered as possible origin and destination
areas for all the other states.

It must be remembered, however, that a model which
considers only the population of an origin area and the
distance between this area and the destination leaves much
to be desired, since it is assumed that there are no signi-
ficant differences in the tastes, incomes and age
distributions, etc., within a population of a given market
area.

19

J. B. Ellis constructed a model which simulates

recreational travel on the highways of Ontario. In this

 

19J. B. Ellis, A System Model for Recreational Travel
in Ontario, Report No. R.R. 126 (Toronto: Ontario Depart-
ment of Highway, 1967).

 

 

30

model, the author includes independent variables related
to the origin area as well as to the destination area.

The method used by Ellis for modeling recreational
travel as a system is based on the segregation of the
province into three classes of components: (1) A set of
origin areas, (2) A set of highway link components,

(3) A set of destination areas.

Each component is modeled by means of the following

equation.

Y.=G.X.
11] JhJ

th = The propensity to camp at a particular campground, of
an origin area.

th = The flow or attendance of campers in each component,
measured on a full season basis.

Gj = l as the demand for a campground is inversely
c+t

proportional to the out of pocket cost and time of
travel.
It could then be possible to reformulate the above
equation as:

c + t

In this systems model, from each origin area there
is a very large number of paths, all presenting an
individual resistance which is increased by a lengthy trip

and decreased by an attractive destination. Highway users

 

 

 

31

are postulated to distribute themselves over the system on
a basis of decreasing flow with increasing resistance.

The method of combining the above equations, one
for each component in the system, is based upon linear
graph methods.

However, since no socioeconomic variables are
included in the model, it is impossible to take into
account the influence of the variations of these different
variables. Ellis therefore assumes that there is no
difference between the socioeconomic characteristics of

different origin areas.

CHAPTER II
STATISTICAL MODEL FOR PREDICTING THE

DEMAND FOR AMONG RESIDENTS OF STE-FOY

At present, the Quebec Department of Tourism, Fish
and Game is actively engaged in planning and operating
extensive programs concerning the development, conservation
and preservation of land and water resources. Many of
these programs and projects include the development of
recreation facilities for the public.

So that it be possible to achieve an adequate
economic feasibility analysis intended to serve in the
overall planning of the development of recreation projects,
it is necessary that, in all provincial recreation areas,
the number of annual recreation-days be determined. In
this study, the main objective is to estimate present as
well as future outdoor recreation demand in the Province
of Quebec.

In order to obtain all the information required
for the feasibility analysis mentioned above, it was
decided to build a regression analysis model which would
use, as basic data, the responses obtained from a question-
naire addressed to some 1800 heads of households located in

the town of Sainte-Foy, a suburb of Quebec City.

32

 

 

33

‘ The method and the questionnaire used during this
survey will be described in detail below. Let us say for
the moment that each individual interviewed (through the
questionnaire) gave all pertinent data related to the
socioeconomic factors concerning his particular household.
When compiling the data obtained during this first survey,
it was discovered that the number of citizens of Sainte-
Foy who did not engage in any outdoor or camping activities
was so great, and their ratio in relation to the total
number of persons interviewed was so high, that the
analysis could not possibly bring out some very important
points that we considered essential to obtain significant
results in our analysis. For instance, when we used the
least squares method in order to estimate the different
regression coefficients bl to bx it was found that the
coefficient of multiple determination (R2) had a value of
0.075 which is much too low to give any confidence in the
projection.

It is, however, possible to determine the subpopula-
tion of individual recreation users and which variables
influenced the degree of participation and also, to
indicate whether its participation tends to increase or
decrease. It would be of interest to planners if one
could predict what type of individuals will participate in
the various outdoor recreation activities.

In order to complement the first analysis and

achieve the objectives that we had set forth to reach in the

34

present study, a second analysis was undertaken which used
an entirely different approach and yielded the desired
results.

In the two following chapters we will explain in

detail how these two models were developed.

Origin Area

 

Sainte-Foy, a city of 9,420 households, a suburb
of Quebec City, was selected as the typical origin area or
study unit for the present survey. Sainte-Foy is a white-

collar middle-class town containing few industries.

Size of Sample and Method of Sampling
The method used consists in selecting every fifth

family from a list1

of all the households in the city of
Sainte-Foy. The sample thus obtained comprises 1,804
sampling units or households representing 20% of the 9,420
households in the city. A questionnaire was mailed to the
head of every family thus selected.

The sample was designed to be large enough to give
confidence that the regression line obtained from the
sample is similar to the “true" regression line which

would be obtained from the entire population. It was felt

that a larger sample would probably have given additional

 

lAnnuaire Marcotte de Québec (Québec: R. L. Polk and
Co., Ltd, 1971).

 

 

 

 

35

precision to the results but its cost would have reached
far beyond the resources put at the disposal of the author.
The first returns of the questionnaire represented 41% of
the households contacted; a subsequent series of question-
naires sent to those who had not answered the first time

raised the proportion of replies to a total return of 48%.

The Variables

 

Two types of variables must be considered in the a
present study: 1) the independent variables which are the
characteristics attached to each individual household head
such as his occupation, his age, his salary, etc., and
2) the dependent variables which, in the present study,
will be expressed in terms of number of camping days per
household for different types of camping. In the present
study, data for three different dependent variables were
collected. These are:

1- Number of camping days in the province of Quebec.

2- Number of camping days in provincial campgrounds

in Quebec.

 

3- Number of camping days in six selected camp-
grounds located in the vicinity of Quebec City:
Mare du Sault, Stoneham, Villeneuve, Ile d'
Orléans, Montmagny and Beaumont.

The questionnaire was prepared in such a way that we

would obtain the maximum amount of data on the socioeconomic

 

36

conditions of the population. Most of these variables can
be considered as factors which can influence demand for
outdoor recreation. It is understandable, as we have seen
in the introduction, that these factors could be so
numerous that it would be too expensive and too difficult
to answer a questionnaire that would attempt to include all
of them. Furthermore, the multiplicity of the answers
would give so many variables that an adequate proper
analysis would not be possible. For the reasons mentioned
above, it was felt most important to keep the questionnaire
as short and as concise as possible. In order to achieve
this, it was decided to obtain only information believed

to be of prime importance. The basis on which such
Ivariables were selected was mainly the author's own judg-
ment and, also, studies which have been conducted by
various authors. Amongst these we can mention: The
Outdoor Recreation Resource Review Commission2 study,
Cicchitti's study on the demand and supply for Outdoor

3

Recreation and McCoy's study of Recreation in Tennessee.4

 

2Outdoor Recreation Resource Review Commission,
Report No. 26, op.cit., p. 16.

3Cicchitti, op.cit., p. 97.

4Edward Wayne McCoy, "Analysis of the Utilization of
Outdoor Recreation in Tennessee," (Unpublished Ph.D. Thesis,
University of Tennessee, August 1966): p. 25.

37

This process automatically reduced the number of
variables to those listed below:

1- Age of head of household

2- Education level of head of household

3- Household income level

4- Number of children in household

5- Occupation of head of household

The first four of these independent variables are
continuous variables because they can be easily measured
with an interval of higher scale, whereas the fifth one,
which describes the occupation of the head of the household,
must be considered a nominal scale. These classes have
been used in collecting data mostly for practical reasons
in order to make the questionnaire easy to answer. As an
example of a continuous variable we can see below how the
household income can be measured.

Income:

$1,999. or less

$2,000. to $3,999.

$4,000. to $6,999.

$7,000. to $10,999.

$11,000. to $15,999.

$16,000. or more

On the other hand, there is absolutely no basis on
which we could rate on an interval scale an individual

employed in the public service and another one employed in

 

 

38

the construction industry; we are therefore using a nominal

scale.

Discontinuous Independent Variables

 

For the purpose of the present study, the
occupational variable (a dummy variable) is assigned a
value by using the following procedure: the occupation
factor has been subdivided into ten different groups
corresponding to the divisions used by the Dominion Bureau
of Statistics. It was important to use such a classifica-
tion, in order to obtain from that government bureau the
estimates of the future disttribution of workers by
occupations. These groups are listed below:

- Employed in primary industry

- Employed in finished products industry

- Employed in the public service

- Employed as office worker

- Professional

- Artist

- Employed in the construction industry

- Machine operator

- Student

- Retired.

For any individual head of a household, the value
of his occupation is given the value of one in the table
set up for the regression analysis. The only exception is

for people classified as retired (group 10) whose

 

 

 

 

39

occupation is given a value of zero. The selection of this
group, to be given a value of zero, is purely empirical
and any other of the 10 groups could have been assigned
this value of zero which is needed as a basis for compari-
son with the other occupations in the course of the
regression analysis procedure.

When setting up the matrix with the data obtained
from the questionnaire and to be used as matrix data in

the regression analysis, each individual head of household

 

has a classification of one opposite his name in the
vertical column corresponding to his occupation. In all
other vertical columns corresponding to occupations other
than his own, the figure zero is inscribed as seen in
the table below. We can observe, in the same table, that
to conform with what was said above, only zero appears in
all the occupational vertical columns opposite the names
of retired heads of households.

The regression coefficients associated with the
occupation variable other than that of the retired head

of the household indicate the deviations of all the other

 

occupation intercepts from this zero base. For example

if the model is used to estimate annual camping days for a
family in which the head of the household is retired,

the occupation variable in this case would have a value of
zero and it would drop out of the system so that no
occupation parameters would enter into the equation for

this particular group of individuals. The effect of the

 

4O

 

o o---

0

3x

o--—

o o---

00

o o o a o o
. . . . . .
. . . . . .
. . . . ._. .
o o o o o o
o o a o o o
o a o o o o
o o o H o o

O O O C)"-

AHMGOHmmmmoumv

iamunwmuo
Audubnmm oaansmv
lumnuumc

AHMGOAmmomoumv

oaosmmsom

.
paozmmsom
oaonomsom
oaosmmsom

oaosmmsom

mam

N
H

soaum>ummoo

"mmHQMHHm>

 

poms Hopoz may cw moanmflum> mfifiso mo cowumucmmmumom

H mamflfi

41

retired class appears in the constant term. If the model
is used to estimate annual camping days for a family in
which the head of the household is a professional, the
occupation variable in this case would have a value of one
and the parameter b5 which would be determined for this
occupation would be added to the value of a. It would

then be entered as:

Y=a+b5(l)

or
Y = a + b5
where
b5 = parameter determined for the professional
class.

Independent Cross Variables

 

It is a known fact that some independent variables
interact with one another and that the resulting joint
effect influences the behavior of the dependent variables
in a different way than would do so each of the individual
independent variables. For this reason it was decided to
introduce other variables as cross variables.

The following figure illustrates how one independ-
ent variable can affect any number of other independent
variables through a phenomenon of interaction which gives

a resulting cross product as explained below:

42

FIGURE 2

Cross Variables

 

 

Dependent 5
Variable
Y1 4
~\\\\\\\ use
3 /
2
I
income
x1

In the above diagram, the effect of income X1 on
the dependent variable Y1 is not only a function of the
actual income of the different groups of people that form
the population of a city, but also a function of the age
of those people.

Hence, it is evident that independent variables
cannot always be used as such without due account being
taken of the effects that one variable can have on other
variables. In other words, in a regression analysis of
these data, the effects of those variables cannot be

considered as simple values and added to one another

43

because they are not additive. To illustrate this point,
let us consider that in the above figure the annual income
of the family and the age of the head of the household

have been introduced as cross variables. If the
coefficient of this cross variable is significant, it means
that the effect of one of these two social characteristics
is itself a function of the value obtained with the other
characteristic.

If we consider the relation existing between age
and income, it is evident that the same income has differ-
ent effects on the dependent variable according to the age
of the individual.

In the present study, the variables introduced
as cross variables are: (1) age of head of household;

(2) household income; (3) number of children in household.

The above three variables are expressed by:

1) x1 X3
2) x3 x4
3) x1 x3 x4

where:

Age of head of household

Household income level

><
u»
ll

Number of children in household

X
as
H

44

Quadratic Variables

 

After computer processing the available data, we
expect to obtain a series of curves which we know to be
curvilinear from previous experience. In order to
facilitate processing, it was assumed that some variables
do not have linear relationship with the dependent variable;
therefore, the variables are introduced in both regular
and quadratic form:

Y = a + blx1 + 01 xi
in which Y = The number of visitors from a specific popula-
tion center, a = The value of the Y coordinate at the point
of intersection of the curve and the Y axis, b1 = The slope
of the curve at this point, and c1 = sign that indicates
direction and the degree of curvature of the curve.
Theoretically the value of "a" should correspond to the
point of origin (coordinate 0-0) of the curve. In fact,

it is always above the point of origin, due to some
variables which are not included in the statistical model.

In the present study we have introduced, in
quadratic form, the four following variables:

x1 = age of head of household

x2 = education level of head of household

x3 = income level of household

x4 = number of children in household

45

The Model

To analyze the collected data, a multiple-
regression model is used. Such a model measures the
simultaneous influence of a number of independent variables
upon one dependent variable, and describes the average
relationship between these variables; this relationship is
used to predict values for the dependent variables.

As seen in the above section, the variables are
introduced into the regression analysis model as: simple
variables, quadratic variables, and cross variables. The

resulting equation has therefore, the following form:
1
Y = a + 2 b. X. i = 1, ..... , 21

in which
Y = number of camping days

= partial regression coefficients

age of head of household

= education level of head of household

XXXU‘

= household income level

><
«n h. u» h) id

= number of children in household

X

= 1 if employed in primary industry,

0 otherwise.
if employed in finished product industry,
otherwise.

if employed in public service

X
ll
0 H O l—‘

otherwise.

10

11

12

13

14

15

19

Id c> rd c> rd C: re c> id c: Ia c: rd

0

46

if employed in office work
otherwise.

if professional

otherwise.

if arts

otherwise.

if employed in construction industry
otherwise.

if employed as machine operator
otherwise.

if student

otherwise.

if retired

otherwise
. Xla = Quadratic variables
. x = Cross variables

21

Level of Significance

 

After carefully examining research of other authors

in the field of recreation and appropriate areas, it is

believed that with the present type of study, a level of

significance of 0.1 is appropriate, hence one may accept

47

the risk of being wrong 10% of the time. Little5 and

6 in two studies of the same kind, have also used a

McCoy
level of significance of 0.1. So in the present study the
significance of the independent variables at a level of

0.1 is thus specified.

Analysis of Data

 

The S.P.S.S. multiple-regression analysis computer
program7 was used to estimate the regression coefficients.
The S.P.S.S. mutliple regression analysis is a computer
program with automatic stepwise addition of variables to
form a least squares equation. This program provides a
means of choosing independent variables which will lead
to the best possible prediction with the fewest independent
variables. Through the use of this stepwise-regression
analysis, an independent variable is chosen at each stage
amongst all the independent variables which are not included

in the equation. In this way, at every stage of the

 

5A. D. Little, A Study of Factor Affecting Pleasure
Travel to U.S., Report to the U.S. Travel Service
(Washington, D.C.: U.S. Department of Commerce, 1967),
p. 198.

 

6E. W. McCoy, "Analysis of the Utilization of Outdoor
in Tennessee," op.cit., p. 39.

7Norman H. Nie, Dale H. Bent and C. Hadlai Hull,
Statistical Package for the Social Sciences (New York:
McGraw Hill Book Company, 1970), pp. 208-244.

48

analysis, the variables of the preceding stage are
reexamined. The independent variable selected is the one
which will reduce to the greatest extent the unexplained
variance around the mean of the dependent variable. It
follows, therefore, that the selected independent variable
is also the one which will raise to the greatest extent
the coefficient of multiple determination.

In the present study, with the data obtained in
the 693 completed questionnaires submitted by households
of Sainte-Foy, the regression model was tried for the
dependent variable. All the variables listed above were
assumed to have a significant influence on the demand for
camping and were included in the models. In every case
the value of the coefficient of multiple determination
(R2) was found to be much too small to be of any practical
significance and leads one to conclude that the selected
independent variables account for very little variability
of the dependent variables. Such surprising results led
us to make a very careful study of the whole procedure in
order to find out what caused this model, carefully
elaborated, to be of so little practical value in this
particular analysis.

It seems at first sight that the main difficulty
is due to the fact that a very large proportion of the
population interviewed through the questionnaire did not
participate in camping in any way. Therefore we may

conclude that in the city of Sainte-Foy, since the majority

49

of the people have a dependent variable of zero or near
zero, the value of R2 cannot possibly be high enough to
make prediction on the future use of campgrounds.

In the survey made in Sainte-Foy, out of 693
households who answered the questionnaire, 516 did not camp
during the summer of 1972. The survey yielded a participa-
tion mean of .78 camping-day per household.

The following (Figure 3) also explains the basis
of the coefficient of multiple determination. In this
case, however, it is adapted to the particular case where
a great proportion of the individual observations has 0

as value for the dependent variable (camping nights).

FIGURE 3

Basic Measures for Coefficient of
Determination When a Great Proportion
of Observations are non Participants

add
00"”

6N0.UIC~JO

 

 

 

0 XX XXX XX XX X X X

15 25 as 45 55 ‘5 ’5 INDEPENDENT

VARIABLE

50

The R2, which is the explained variance divided by
total variance around the mean thus obtained, is very small.

It was believed that further interpretations of the
data obtained can bring some very interesting information.
It does not seem possible to predict whether an individual
will participate or not in the camping activity; conse-
quently, having determined the subpopulation of individual
campground users, it was proposed to determine which
variables influenced the degree of participation and whether
the entire recreation activity tends to increase or
decrease. It would be interesting if one could predict
which individuals are likely to participate in various
outdoor recreation activities, but since this seems
impossible, it seems promising to compare the difference
between the participating and non-participating individuals
for the different variables.

In the following analysis the "t" test is applied
to the interval-scale variables. In such case the null
hypothesis is that ii = Xi, where fl is the mean value of
the participants and that {2 is the mean value of the non-
participants.

The chi-square test is commonly used for testing
nominal scale variables. This test is used to verify the
hypothesis that the means or relative frequencies of the K
population means are not significantly different. In this

analysis a contingency table is constructed in order to

.Jua:

 

 

51

test the hypothesis that there is no relationship between
the individuals' occupation and their participation in
camping.

Table 2 gives the t values and the direction of

effect for the mean differences between campers and non-

 

 

 

 

campers.
TABLE 2
Mean Comparisons Between Campers

and Non-Campers Subsamples
Variables Campers Non-Campers t

Z 82 i 62
Age 41.452 75.386 44.229 93.882 - 3.4
Education 5.678 4.379 5.729 4.35 - .28
Income 13.160 1.650 13.700 1.920 - 44.77
Number of
children 3.492 2.688 3.461 2.195 + 1.526

 

From the survey made in the city of Sainte-Foy, we
know that there are 177 campers and 516 non-campers. With
a level of significance of 0.1 the critical value of 1.64
is used. The t test leads to the conclusion that there
is not sufficient evidence to indicate a relationship bet-
ween the degree of participation and the number of children

or the level of education of the participants.

52

The following table was constructed to verify the
hypothesis that there is no relationship between individ-

uals' occupations and participation in camping activities.

TABLE 3

Camping Participation

 

 

 

g
E.“-
X5 X6 x7 X8 x9 x10 X11 X12 x13 X14 TOTAL 1
Camping 3 13 22 33 9o 4 2 3 1 5 176 "n';
Non-Camping 7 20 52 69 263 6 14 13 1 74 519
woman 10 33 74 102 353 10 16 16 2 79 695

where

x5 = Employed in primary industry

x6 = Employed in finished products
x7 = Employed in public service

x8 = Employed in office work

x9 = Professional

x10 = Arts

 

xll = Employed in construction industry
x12 = Employed as machine operator
x13 = Student

x14 = Ret1red

The critical value of x2 for a = .10 and 9 degrees
of freedom is 14.684; hence, because of the magnitude of x2,

25.86, the hypothesis that the mean degree of participation

 

53

is equal for all the k occupations is rejected. This leads
to the conclusion that the frequencies of participation
differ by occupation.

It was then decided by a trial and error process
to group together occupations which did not show any
difference in the rate of participation and to sort out the
occupations which did show a difference. It was first
decided to eliminate five of the occupations because of
the small number of individuals belonging to these groups.
These are x5, x10, x11, x12 and x13. After many trials,
the chi-square test showed no difference between X6' x7,
X8 and x9 (see Table 4) and as a result x14 is sorted out

as the only one with a significantly different rate of

 

 

 

participation.
TABLE 4

Camping Participation

x6 X7 X8 X9 TOTAL
Campers 13 22 33 90 158
Non-campers 20 52 69 263 404
TOTAL 33 74 102 353 562

x2 = 4.35

The critical value of x2 for a = .10 with 3 degrees

of freedom is 6.251, therefore, because of the magnitude of

54>

4.35, the hypothesis that the relative frequencies of
participation of the k occupation are equal is accepted and
the conclusion is that the frequency of participation does
not differ significantly among the four occupations.
However, when the tenth occupation X14 is entered into the
test, the x2 obtained falls into the rejection region.

For more sensitivity, it was then decided to

aggregate X6 x7 X8 and x9 and to test against x14.

TABLE 5

Camping Participation

 

 

 

X6 X7 X8 x9 x14 TOTAL
Camping 158 5 163
Non-camping 404 74 478
TOTAL 562 79 641
x2 = 67.50

The critical value of x2 for u = .10 with 1 degree
of freedom is 2.706, therefore because of the magnitude of
67.50, the hypothesis of an equal participation is rejected.

The above analysis tends to indicate that a young
population with lower income and a larger family produces
a larger number of campers, whereas non-campers belong to
the older population with a higher income and smaller

families. The analysis of the binominal variables. using

‘o —-_

55

the chi-square method, further indicates that there would
be no significant difference between the rate of participa-
tion of people with different occupations, except for
retired people who show a clear tendency toward a lower

participation.

CHAPTER III

STATISTICAL MODEL FOR PREDICTING THE
DEMAND FOR CAMPING AT VARIOUS PROVINCIAL PARKS

Following the completion of the first study, it was
felt that positive results were needed in order to obtain
some valuable projections on the future demand for camp-
grounds in the Province of Quebec. For this reason, a
second study was undertaken in which groups of individuals
were used as observation units instead of individual house-

1 and Grubb and Goodwin.2

holds, such as suggested by Cheung
In so doing it was expected that any group of individuals,
such as those represented by the population of a county or
of any other geographic unit, should have much greater
stability than individuals. It is indeed most likely that
each of these groups will send at least a few visitors to
the campgrounds each year.

In this second model, the observation units selected

are the counties of origin of the citizens travelling to the

 

1H. K. Cheung, A Day-Use Park Visitation Model, Report
to the Canadian Outdoor Recreation Study Committee (Ottawa:
Department of Indian Affairs and Northern Development, 1970).

 

2H. W. Grubb and J. T. Goodwin, Economic Evaluation of
Water-Oriented Recreation, Report No. 84 (Texas: Agricul-
tural Experiment Station, 1968).

 

56

57

campgrounds; the independent variables are the calculated
average of the socioeconomic characteristics of all the
residents of any such county of origin. There are other
variables which are also included in this model, such as
the distance between the origin and destination points,
and others which will be described in a later part of the
present study.

In this model the dependent variable (the number
of camping nights) is calculated by adding, for a definite
period of time, the number of visitor/days that are
recorded at a given campground and which originate from the
different counties of the Province. Let us assume that
during a sampling period of 14 days, campground A, is
visited by x persons from a particular county. Since some
of the visitors remain at the site for more than one night,
the total number of camping nights for these people will be
greater than x. The same thing is bound to happen for
almost any other county in the province.

It was possible to obtain the data described above
for the three following provincial campgrounds: (1) La Mare
du Sault, in Laurentides Park; (2) Lac La Vieille and
(3) the Mont Orford Park (see Map No. l for location).

All of these campgrounds have been defined in the
framework of the P.P.B.S. system as destination campgrounds.

The Department of Tourism, Fish and Game, in describing the

58

(.0: .35.... £ 3:32ng .0 3:52

 

 

                 
  
         

.3

aficaoanEUU 3.2.30.5 .
.010... E 13.. upsemnEoU .

aunt; _:_ aucqm a:

.35..-

,: . °\
, Q
, .::.III_I, . . . Pk
o. 90
4 = . .22.... .. 3 _

.953? .2:

   

 

Simzsxuszsaox w.:.u.><._ us

 

1
E
E<4n3u851 >101-
. , . nquauin‘l ..
, . . fl ., .j; 5.31:3“
. :5: a. ’

sIISLJ-

I'D-is

 

 

 

 

 

 

 

 

 

 

59

different programs comprised in the framework of the
P.P.B.S.,3 has utilized three different types of campgrounds:

l - Overnight campgrounds, that are usually
located along major highways and mostly used
by travelers for only one night.

2 - Peri-urban campgrounds located in the
vicinity of urban areas. These are usually a
supplement to the urban-oriented recreation
facilities.

3 - Destination campgrounds; most campgrounds
of this type are located within the large
provincial parks and usually have a supplemen-
tal function to the park itself. Most
campgrounds of this third group are resource-
oriented.

We cannot be absolutely sure that the campgrounds

classified above in the third group are used exclusively
as destination recreation area by all visitors. Yet we do
not think that it is presumptuous to assume that the parks
themselves are used as destination areas by the great
majority of the park users for the following reasons:

(1) most provincial parks are located some distance from
the major tourist roads; (2) most provincial parks offer

outdoor life facilities of a very high standard and of

 

3Jacques Auger, Dossier Préliminaire des Programmes,
Report to the Treasury Board (Quebec: Ministére du
Tourisme, de la Chasse et de la Péche, 1972): p. 49.

60

great diversity. (For instance, we know that in these
parks, fishing and hunting are among the best in North
America. Rivers of very good quality for canoeing are
present in all the parks and there are many other types

of activities available to resource-oriented visitors of
the parks); (3) the Province of Quebec is so large and the
provincial parks so widely spaced that it is difficult for
a visitor to make a tour of these recreation areas.

The above-mentioned reasons make the campgrounds
located within provincial parks quite different from
overnight campgrounds which are always located very close
to important tourist roads and which have very little to
offer, except overnight facilities. These overnight camp-
grounds have more or less hotel accommodations and they
cannot be considered as attractive by themselves, but they
are usually located along the roads going to attractive
centers such as Montmagny, Beaumont, etc.

The previous map shows with different shapes the
location of the various types of campgrounds located
within the boundaries of the Province of Quebec, as well as
the three campgrounds which are studied in the present

research study.

The Variables

 

The provincial park authorities have collected data
on the number of people from various points of origin who

visited the different parks during a two-week survey

61

period in 1971. Taken along, these do not yield the infor-
mation needed for an economic evaluation of the
campgrounds. However, these data are important and must be
evaluated in economic terms in order to estimate the
recreational benefits of the campgrounds considered in
this study. The cost of travel and the acquisition of
recreational equipment as well as the rental of all
necessary services are the major costs incurred by park
users. The fact that the consumer allocates a portion of
his income to the use of recreation facilities means that
these activities have an economic value for this consumer.
The demand curve for any specific good or facility
is defined by Samuelson as follows: ". . . there exists
at any one time a definite relationship between the market
price of a good and the quantity demanded of that good.
The relationship between price and quantity bought is
called the "demand curve."4
If in the above-mentioned relationship the price of
a good is changed, the quantity demanded will change.
However, it is possible to shift the entire demand curve
by introducing a variation in different characteristics
such as income, market structure, or simply a change in
the desire of the people to spend their money on a

particular good. Such a change in demand means that

 

4Paul Samuelson, Economics: An Introductory Analysis
(New York: McGraw-Hill Co., 1967). P. 66.

62

different quantities will now be bought at the same
price.

It is believed that each of the factors listed in
the first chapter of the present study might affect in one
way or another the demand in any particular outdoor
recreation area.

Unfortunately it was found to be very difficult,
or in some cases impossible, to gather data bearing on
some of the variables. Therefore, for the purpose of the
present study some of those less important factors used in
the first study had to be eliminated. The author had to
be satisfied with only those variables that were available
from the statistics at hand and which are shown in the

general form of the visit-estimating equation shown below.

Y = a + b1 x1 + b2 X2 + b3 x3 + b4 X4 + b5 x5 +

b6 X6 + b7 x7
in which
Y = number of visitor/days from a particular county
of origin.
a = intercept
bi = partial regression coefficients

x1 = population of the county of origin of the
visitors
x2 = per capita income in the county of origin of

the visitors

63

x3 = round trip cost from the county of origin to
the campground, in dollars

x4 = average age of the heads of households in the
origin county

x5 = average age of the inhabitant in the origin
county

x6 = degree of urbanization of the origin county

x7 = substitution effect variable which reflects
the competitive effect of other campgrounds

. . 5
located near the or1g1n area.

The Data

For the present study, it was decided to use the
county as the observation unit, because this geographic
entity is the basis for all statistics available and thus
gives us a chance to estimate the influence of the above-
mentioned variables upon the number of visitors to camp-'
grounds. As mentioned above, data for these variables were
most readily available on a county basis. In some cases,
the campgrounds did not receive any visitors from certain
counties during the two-week period used for the present
survey. However, these counties could not be ignored since
they each represent a potential supply of visitors. Such
origin areas were entered into the analysis with a zero

value for the dependent variable.

 

5This variable will be dealt with extensively in a
later part of the present study (p. 67).

64

In order to have information on the number of
visitors from each county, the author was provided access
to data collected during the summer of 1971 by the
Provincial Parks Service. Each party entering the provin-

cial campgrounds was requested to complete the form shown

 

 

 

 

 

below.
FORM USED BY THE PROVINCIAL
PARKS SERVICE DURING THE SUMMER OF I97I
BOWERNEIENT DU QUEBEC
"BTU?! DU TOURISIEJ’E LA CHASSE [1' DE LA PECHE
MOTION GENERALE DEB PARCS
Damn-m CAMPING No
MAR. 01!. 0M: 01m 08A um: mu-IN—mu-oor
E] E] U D D U
m
mum-nun"? _— J
m m mm. nuuu- am mus-cm
D U D count-com m
TENT mm m "AV-m"
mnmcuunou - can nun-um“
amt-flu -x .8
m... 1:]... - .___-.
um» won - um 09 cm E SITE ”o ”"1me 3

Each form indicates the county of origin of each

group of visitors, the number of persons in the group and

 

the number of nights that each person spends in the park.

From the data thus collected, it was possible to obtain

the number of visitors per day from each county of origin.
The compilation was made for a sample period of two

weeks (the first week of July and the first week of

August) for each of the three campgrounds used in this

study. It was no easy matter to collect and classify data

 

65

because the classification system of the Department did not
lend itself to that type of use of the statistics accumu-
lated over a period of years. The present project proved
so valuable that the authorities of the Department of
Tourism, Fish and Game decided that, starting in 1973, new
forms would be used and a data bank would be formed in
order to make appropriate information more readily
available.

The data obtained were then matched with the

 

different characteristics of the county of origin for the
purpose of calculating an acceptable estimate of recreation
visits. For example, if, during the sample period from
county A, 200 parties originated totaling 567 visitor/

days to campground x1, the number 567 was entered as the

value for the dependent variable for county A. The

 

variables entered as independent variables are different
socioeconomic characteristics of the origin county. Data
inputs for independent variables used in the present study
are derived through calculation from raw data obtained
from a variety of sources.

The 1970 census has been used extensively to
determine the population (x1), the per capita income (x2),
the average age of the head of the household (x4) and the
average population age (XS).

The round-trip cost (x3) from the origin county to
the destination area has been calculated by measuring the

distance in miles from the approximate center of gravity

 

66

of the population within every county to the sample camp—
ground. The measure of these distances was taken along
the major highways most likely to be followed by the park
users in traveling to the campgrounds.

The distance found is then multiplied by $0.19,
which is the average cost per mile for travelling by car
within the Province of Quebec as calculated and published
by the Quebec Automobile Club.6

The degree of urbanization (variable X6) was con-

 

structed for each county. One of three values (1, 2 or 3)
was attributed to each county in the Province of Quebec.

A value of 3 was given to high density counties, a value of

 

6This cost has been calculated for a 1971 Chevrolet
Impala, as follows:

 

 

 

 

Item Average cost
variable per mile
1- Oil and gas $0.0345
2- Maintenance 0.0085
3- Tires 0.0075
$0.0505
Fixed cost Cost per year
1- Insurance fire $30.00
2- Insurance collision 145.00
3- Insurance 237.00
4- Driving permit, license fees '45.00
5- Depreciation (3 year period) 945.00
$1,402.00

For an average of 10,000 miles per year
A- Variable cost 5.05 X 10,000- $505.00

$1,907.00 or

19 cents/mile

 

67

2 to medium density counties, and a value of l to counties
with the lowest density of population, as follows:

1 - A high density county has been defined as a
county that comprises a major city such as
Quebec or Montreal with a population greater
than 100,000.

2 - A medium density county may be an area located
within 50 miles of a major city: Or a county
which includes a city having between 50,000
and 100,000 of population.

3 - A low density county is one which is located
in a rural area and includes no major city.

The substitution effect (variable X?) was
constructed for each county in the Province of Quebec. The
model thus obtained takes into account, for any campground,
the effect of all the competing campgrounds of the same
type within 200 miles of the center of each county. Data
used in this analysis indicated that a great majority of
the visitors originate within 200 miles of each sample
campground. The assumptions underlying this variable are
as follows:

(1) The larger the number of campgrounds located
within the ZOO-mile radius of any county and
offering the same kind of facilities and enter-
tainment, the less likely residents of that

county will visit any particular campground,

68

(2) The camping attraction index is an important
variable used to measure the attraction of a
site on the users. The number of visitors
attracted to a recreation area is generally
considered as directly proportional to the
value of the index. On the other hand, Renoux7
believes that the logarithm is a better indi-
cator of the importance of the attraction index.
To prove his point, he has demonstrated that
the attraction index of one very attractive
area is of less importance than the added
influence of many recreation areas of lower
attraction indices, even if the total of these
indices is close to the attraction of the more
attractive area.

(3) The distance between the campground and the
origin county also affects the number of visits
to recreation areas. The value of the number
of visitors to a particular recreation area
was assumed to be inversely proportional to
distance. The substitution effect variable was

determined as follows.

 

7Maurice Renoux, "Techniques Econométriques de
Prévision de la demande Touristique et Amorce de leurs
Integrations dans un Systéme Décisionnel" (Unpublished
Doctoral dissertation, Université Aix-Marseille, France,
1972), p. 408.

69

n
X3 = i=1 log10 Si
7i}—
where
Xj = The substitution effect for county j
n = The number of campgrounds located within a
ZOO-mile radius from the county of origin
si = The attraction index of the different camp-
grounds, as described below.
dij = The distance from the campground to the center

of gravity of the population for county j.

There are as many terms in the gravity equation as
there are campgrounds within 200 miles from the center of
county j (n equals the number of campgrounds within 200
miles from the center of gravity of county j). In our
calculations, the logarithm of any campground attraction
index has been weighted by dividing it by the distance,
in miles, from the campground to the center of gravity of
the county. Large numeric values associated with the
substitution effect variable are expected for counties which
have very attractive campgrounds in their vicinity, whereas
counties having few campgrounds in their neighborhood are

expected to have smaller numeric gravity values.

The Attraction Index

 

Many methods of measuring the attractiveness of a

recreation area have been suggested. Wenger and

70

Videback8

explored the usefulness of the eye pupillary
response as a measure of aesthetic reaction to photographs
of forest scenes.

Factor analysis has been used by Van Doren,9

10 11

Shafer and the Ontario Department of Highways to deter-
mine an attraction index. The following section contains a
description of the procedure observed in applying the
'factor analysis to this type of study. This is followed
by a description of the work done in the Province of Quebec,
where such a procedure will be used in order to obtain the
required attraction index for this specific study.

In the present research program, factor analysis is
used purely as a tool intended to help derive a useful

attraction index from data obtained in the course of two

different surveys. In the first one, abundant information

 

8Wiley D. Wenger Jr., Richard Videback, Pupillapy
Response as a Measure of Aesthetic Reaction to Forest
Scenics Report No. 1, Project K, 10-272 (Washington, D.C.:
U.S. Department of Agriculture, Bureau of Economic Research,
Sept. 1968).

 

9Carlton S. Van Doren, "An Interaction Model for Pro-
jecting Attendance of Campers at Michigan State Parks: A
Study in Recreational Geography" (Unpublished Ph.D. disser-
tation, Michigan State University, 1967).

loElwood L. Shafer, J. F. Hamilton, Elizabeth A. Schmidt,
"Natural Landscape Preferences: A Predictive Model,”
Journal of Leisure Research, 1, No. 1 (Winter 1969).

 

llOntario Department of Highways, A System Model for
Recreational Travel in Ontario, Report No. R.R. 148
(Toronto: Department of Highways of Ontario, 1967).

 

71

was collected on the characteristics of different camp-
grounds. In the second one, made by the Department over a
period of several months, data were collected on a sample

of 700 campers who have used the camping facilities.

Factor Analysis
Cooley and Lohnes define multivariate analysis as:
'. . . generally considered to include those statistical
procedures concerned with analyzing multiple measurements
that have been made on a number N of individuals. The
important distinction is that the multiple variates are
considered in combination, as a system."12
Factor analysis, a multivariate statistical method,
has that unique characteristic of making it possible to
consider a large number of interrelated variables and
reduce them to a smaller number of factors. As defined by
Schwartz:
". . . this technique attempts to determine the number
and nature of the underlying factors affecting the
relationship between a set of variables. This tech-
nique maintains, in effect, that the variables can be
added and studied together rather than separately.
Two advantages of utilizing this technique are

immediately apparent. First, the combined influence
of the most widely different variables can actually

 

12William W. Cooley and Paul H. Lohnes, Multivariate
Procedures for the Behavioral Sciences (New York: John
Wiley and Sons, 1962), p. l.

 

72

be studied. Secondly, this technique limits the
number 8f variables with which the research must
c0pe."l
Factor analysis follows different specific steps
which are namely: (1) Correlation, (2) Initial Factors,

(3) Rotated Factors, (4) Communality, (5) Eigenvalue,

(6) Factor Scores.

Correlation

 

It is first necessary to build a correlation matrix
which is an m by m rectangular array of correlations bet-
ween pairs of individual variables. This matrix is
instrumental in determining the coefficients of the linear

combinations which produce the factors.

Initial Factors

A factor may be defined as a linear combination of
variables in a data matrix, the variables thus combined
being of such a nature that, whenever any change occurs in
one of them, an identical or corresponding change occurs in
the second or any other related variables; it is then clear
that if such two variables are used as independent variables
to explain a dependent one, then, only one of these two
should be used.

The above-mentioned correlation matrix may be

factored by constructing a set of new variables on the basis

 

13Ronald D. Schwartz, "Operational Techniques of a
Factor Analysis Model," The American Statistician,
(October 1971), p. 38.

73

of the interrelations exhibited in the data. In so doing,
it is possible to define the new variables either as
mathematical transformations of the original data, or in
the form of inferential assumptions about the structuring
of variables and about their source of variation.

The former approach, which used defined factors, is
called principal-component analysis and has been defined

by Nie gt a1 as a:
". . . method of transforming a given set of
variables into a new set of composite variables or
principal components that are orthogonal
(uncorrelated) to each other. No particular
assumption about the underlying structure of the
variables is required. One simply asks what would
be the best linear combination of variables-best in
the sense that the particular combination of
variables would account for more of the variation
in the data as a whole than any other linear
combination of variables. The first principal
component, therefore, may be viewed as the single
best summary of linear relationships exhibited in
the data. The second component is defined as the
second best linear combination of variables, under
the condition that the second component is orthogonal
to the first. To be orthogonal to the first
component, the second one must account for the
proportion of the variance not accounted for by the
first one. Thus the second component may be defined
as the linear combination of variables that accounts
for the most residual variance after the effect of
the first component is removed from the data. Subse-
quent components are defined similafiy until all the
variance in the data is exhausted."

This solution, however, does not yield a framework

which is easily interpreted; the configuration of this

 

14Nie, Bent and Hull, "Statistical Package for the
Social Sciences," 0p.cit., p. 216.

74

factor structure is not unique. One factor solution can
be transformed into another without violating the basic
assumptions or the mathematical properties of a given

solution. As mentioned by Nie gt 31:

". . . there are many statistically equivalent ways
to define the underlying dimensions of the same set
of data. This indeterminancy in a factor solution is.
in a way, unfortunate because there is no unique and
generally accepted best solution. On the other hand,
not all the statistical factor solutions are equally
meaningful in theoretical terms. Some are simpler
than others; some are more informative than others;
and each tells us something slightly different about
the structure of the data. Therefore, one is left to
choose the best rotational method to arrive at the
terminal solution that satisfies the theoigtical and
practical needs of the research problem."

Rotated Factors

 

In choosing a reference frame for the interpretation
of the variables in terms of factors we must decide on the
criterion by which to make the choice. According to

16 it is necessary to rotate the factors' axis so

Thurstone
that there will be as many zero or near zero loadings as
possible in the factor analysis matrix. In other words,
it seems unlikely that, if we use many observation units,

all the variables involved should be important for all the

observation units. It seems much more likely (unless the

 

15Nie, Bent and Hull, "Statistical Package for the
Social Sciences," op.cit., p. 212.

16L. L. Thurstone, Multiple Factor Analysis (Chicago:
University of Chicago Press, 1940), p. 93.

 

75

observation units are very similar) that some groups of
variables should be concerned with some of the observation
units and other groups with other observation units.

The numerical values of the factor loading are, of
course, dependent on where we put the factorial reference
frame. For every position in which we like to place the
reference axis we get a new set of loadings. That means
360 different positions if we vary by only one degree at a
time, and, as it is obvious that we may vary by minute
fractions of a degree, an infinite number of such loadings
are possible.

We thus have, mathematically, a great number of
solutions to the factor problem and for every position in
which the axes of references are placed, we get a new set
of loadings. However, since we are interested in a unique
solution, it is possible to define different criteria which
can be used to indicate whether or not we have rotated the
factor axes to a satisfactory position. According to
Adcock17 there are basically two such criteria: (1) The
explanation should be as simple as possible. (2) It must
be consistent with other explanations which we accept.

The loadings or numbers in a given row of the factor
analysis matrix represent both regression weights and

correlation coefficients of factors to evaluate a given

 

17C. J. Adcock, Factorial Analysis for Non-
Mathematicians (Melbourne: University Press, 1954), p. 29.

 

 

76

variable. Each variable has a factor loading on each factor

and the loading can be considered as a simple correlation

between a variable and a factor. More specifically, we

could represent it by using the following general formula:
Z. = a. F + a. F + .... + a. Fm + d. U.

3 11 1 12 2 3m 3 J
(j=1' 2' .... n)

where
23. = Variable j
P = A common factor
Uj = A unique factor
ajl’ dj = Regression weights
Communality

 

The total variance of a variable accounted for by

the combination of all common factors, designed by h2, is
usually referred to as the communality of the variable.

The variance of a variable accounted for by a
factor is given by the square of the respective factor
loadings. It is then possible to express the proportion
of the variance in one variable accounted for by all

common factors as follows:

hi = ? a21i (i=1, 2, . . . m)
l=l
where
hi = Variance accounted for by variable 1
m = The number of factors
aii = The square of the respective factor

loadings, for variable 1 in each factor.

 

77

Eigenvalue

 

The total variance of one factor accounted for by
the combination of all common factors is usually referred
to as eigenvalue. The variance of a variable j accounted
for by factor i is the square of the respective factor
loadings. It is therefore possible to calculate the total
amount of variance accounted for by a factor, by adding
the square of the loadings in each column; we may say that
the variance accounted for by a particular factor is

expressed as follows:

2 n 2 .

K = Z a. (j=l,2,...n)
l j=l 31

where

Ki = Variance accounted for by factor 1

ail = The square of the respective factor
J loadings for the different variables in

factor 1

n = The number of variables

Factor Scores

 

After the terminal factor matrix is obtained, it
might be necessary to build composite scales that represent
the theoretical dimensions associated with the respective
factors. By means of a factor score matrix, it is possible
to assign factor scores to each observation.

A factor score can be defined as the relative
importance of a factor in one observation. For example,

in one of the campgrounds observed, one of the factors could

78

be more important than the same factor in a second camp-
ground, hence there is a factor score for each factor and
each observation (campground). Thus, it shows that one
observation possesses the general characteristic particular
to one factor to a higher or lower degree than the other
observation in the model. It is then possible to develop
an index showing the importance of a particular character-
istic for each observation.

To obtain the factor score, it is necessary to make
a conversion of the factor loading obtained in previous
operations into convenient weights which, when multiplied
by the standardized real value of the variable, insure that
each one contributes in proportion to its importance. This
weight or factor score coefficient is determined from the

following formula:

a = s1 R—1
where
a = The factor score coefficient matrix
S1 = The rotated factor matrix such as explained
above
R—1 = The inverse of the correlation matrix

The factor score coefficient represent the
regression weights of each standardized variable to con-
struct a factor index.

The factor score for the first observation can then

be obtained by solving the following equation:

m
Q1 =‘ i=1 J1i 211

where
01 = The factor score for the lst observation
jli = The factor score coefficients defined

from the previous equation for the
different variables in the first
observation.

Zli = The standardized value of the different
variables in the first observation.

As a factor score is obtained for each observation
and each factor, it is then possible to make a comparison
between the various observations and to determine to what
degree a particular observation possess the characteristics
particular to a specific factor such as it has been defined

by the factor analysis.

Proposed App1ication of the Factor Analysis

As stated above, the present study applies the
factor analysis method only to the state-owned campgrounds
in the Province of Quebec which were identified as
destination campgrounds above (see page 57). The Department
of Tourism, Fish and Game is the owner and administrator
of 27 destination campgrounds distributed throughout the
province. In the present study, these campgrounds have
been visited and thoroughly studied in order to gather the
necessary information to carry out this particular analysis
and thus supply all the information needed to arrive,

through factor analysis, at an attraction index.

80

The Variables

 

According to professional judgment, a total of 48
characteristics were selected as representative of the
general attractiveness of a campground. We may find these
characteristics inside a campground, or outside but within
reasonable distance from the campground.

These different characteristics were then grouped
under three major headings: (l) The facilities and
services; (2) The physical environment and (3) The
recreational activities. These characteristics are the
variables that will be used in our survey. They are listed

below according to the grouping described above.

TABLE 6

List of Variables

 

A - Facilities and services:

1 - Showers 7 - Playground

2 - Flush toilets 8 - Swimming pool
3 - Laundry 9 - Roads

4 - Electricity 10 - Type of sites
5 - Grocery ll - Shelters

6 - Restaurant

Table

6

81

(Cont'd)

B - Physical environment:

 

l - Elevation

2

05h)

ooqmm

Overlook

Falls

Virgin Forest

Shade

Vegetation

Quietness of surroundings
Flies and pests

Hiking trails

C - Activities:

 

1- Swimming

2
3

10

Angling

Organized activities
Hiking

Organized sports (team)
Hunting

Canoeing

Tennis

Golf

Boating

10

11

12

13

14
15
16

17

11
12
13
14
15
16
17
18
19

Quality of water
for fishing

Presence of body
of water

Offshore composition

Foreshore
composition

Water pollution
Turbidity
Offshore composition

Boat rental

Ski

Sun bathing
Badminton
Horseback riding
Sailing
Yachting

Water skiing
Camp fire

Bicycling

 

The variables listed above have been inventoried

and graded according to a system devised by the author.

82

In the following section we discuss the grading system used
for each variable. Different sources of information have
been used as guidelines. However, in many cases, there is
a great deal of subjectivity and it was often necessary to

rely on professional judgment.

Activities

 

Different types of grading systems were considered;
however, after many trials, it was decided to use a
trichotomous grading system. In this system, three
different values can be given to any activity; such a
system is based on the assumption that the absence of the
activity has no value and a grade of zero is then attri-
buted; if the recreation activity is made available on the
campground, we then assume that it adds to the attractive-
ness of the area. It is then given a value of one, if the
activity is of average quality, and a value of two if it is
of a much higher quality.

In order to have the most possibly realistic
measure of the variables in the system described above, we
have also given numeral values to activities located
outside the campground. These values however vary with
the distance at which they happen to be from the recreation
area (campground). These variables are graded such as

shown in the following examples.

83

Activities within the campground:

 

 

 

 

211.12
Absence of the activity 0
Average quality of the activity 1
Good quality of the activity 2
Activities outside the campground:
Quality of activity Distance from Value
or facility recreation area
good 1 mile or less 2
good 1 to 3 miles 1
average 1 mile or less 1
average more than 1 mile 0

The Physical Environment
Different classifications have been studied in
order to determine to what extent they could help in deter-

mining the grading system needed for the different physical

environment characteristics. The Canada Land Inventory,18

19 20

the deVries and the ORRRC methods were carefully studied.

 

18Field Manual: Land Capability Classification for
Outdoor Recreation (Ottawa: Department of Forestry and
Rural DeveloPment, 1967).

19L. deVries, An Approach to the Recreational Capabil-
ities of Shoreland by Photo Interpretive Method (Toronto:
Lockwood Survey Corporation Limited, 1966).

200.8. Outdoor Recreation Resource Review Commission,

Potential New Sites for Outdoor Recreation in the North-
east (Washington, D.C.: Government Printing Office, 1962).

84

In the following section, the grading system used
for such natural resource characteristic is explained.

Viewpoint, Waterfalls, Virgin Forest, Trails:

Each of these natural characteristics is very
important and can have a great influence on the attractive-
ness of a recreation area. However a great deal of
subjectivity is involved in grading these variables and
what one individual would rate high can be rated in a
completely different manner by another individual. In
order to get around such a problem, a dichotomous scaling
system is used where a value of l is given if the

characteristic is present and 0 if it is absent.

Vegetation:

 

This characteristic is of great importance since
different types of vegetation and certain combinations of
the same add value of different importance to the
attractiveness of the campgrounds, because they increase
the enjoyment of the park users. A short study of the
different possible vegetation types existing in Quebec
indicates a great number of possible combinations. In
this study, the vegetation characteristics have been graded
according to four basic classes:

Barren 0

Evergreen 1

Mixed evergreen and deciduous 2

Deciduous 3

85

In this classification the deciduous forest has been
given the highest rating because it has been observed that
people in Quebec prefer the deciduous type of forest mainly
because of the smaller amount of flies and pests which are
very abundant and extremely unpleasant during the best part
of the summer season. It is possible that in other parts
of the country, where there are fewer pests, the same
persons would prefer a mixed type of forest. Another factor
which can play an important role is the fact that in Quebec
a deciduous forest usually grows in areas where the climate

is less rigorous.

Quietness of Surroundings:

This characteristic does not apply to any natural
resource variable as such, but to the overall quality of
the environment. It was graded according to the general
level of noise in the immediate surroundings for example;
if a campground is located very close to a heavily used
highway, or an airport where the level of noise is high,
it would be graded low.

The quality of the surroundings has been graded for
each park according to the following classification:

Very noisy 0

Moderately noisy 1

Quiet 2

86

Fishing:

The lakes and rivers in the Province of Quebec have
always been renowned for their fishing qualities. It is
therefore reasonable to assume that the quality of fishing
can influence, to a certain degree, the general attractive-
ness of the campground. For that reason the fishing
quality of bodies of water found at different campgrounds
or in their immediate vicinity are quantified in order to
be included in the attraction index. The bodies of water
present at each campground have been rated with the help of
Gilles Shooner of the Wildlife Division, Department of
Tourism, Fish and Game, according to the following
numerical scale.

No fishing possibilities

0
Poor fishing conditions 1
Medium fishing conditions 2

3

Good fishing conditions

Quality, as used in the above classification, is
based mainly on the type of sport fishing that can be
practiced at any particular site, for example, salmon
fishing and trout fishing being rated higher than pike or
wall-eyed fishing. It was however first assumed that there
was a relatively good chance of success, or a relatively
good chance of catching some specimens of the desirable

species.

87

Relief:

In Quebec, camping grounds are located throughout
the Province, and are associated with a great variety of
topographies. According to deVries,21 areas with high
relief and showing strong contrast are favored by a great
number of park users.

After consultation with professionals in the field

of recreation, the following grading is assumed:

0 - 100 feet 0
100 - 200 feet 1
200 - 700 feet 2
700 feet and above 3

Body of Water:

 

The types and shapes of the bodies of water present
in a campground area may be of great importance in grading
the park, since activities of different types will be made
possible on different bodies of water. A great number of
activities are usually possible on a lake, such as
swimming, sailing, water skiing, fishing, whereas a river,
even if it also has many possible activities, is usually
limited as compared to a lake. This characteristic is

graded according to the following classification:

21DeVries, "An Approach to the Recreational Capabil-
ities of Shoreland by Photo Interpretive Methods," op.cit.

88

No water
River

Lake

OJNI-‘O

Lake and River

5195.2:

This variable is of prime importance, since
swimming is one of the most pOpular outdoor activities
practiced by campers. In order to describe this variable,
the following three characteristics have been used: the
composition characteristic of the offshore, the composition
characteristic of the foreshore, and the length of the
beach.

Each of the above characteristics have been
discussed by deVries. However, some minor changes have
been brought in the following, such as the presence of
grass on the offshore. Each has been graded according to
the following classifications.

Composition characteristics of the offshore:

No offshore 0
Rock 1
Gravelly 2
Grass

kw

Sand

89

Composition characteristics of the foreshore:

No foreshore

Gravelly

0
Rock 1
2
Sand 3

Length of the beach:

No beach 0
1 to 100 feet 1
100 to 500 feet 2
500 to 1000 feet 3
1000 feet and more 4

Water quality:

 

The water quality of a body of water, when used for
outdoor recreation activities, is related to its pollution
and its natural turbidity. It is obvious that clear water,
free of any pollution, is superior and preferred by park
users; however the measurement of such characteristics is
very difficult and subjective. In order to quantify these
characteristics, the author relied on his own experience
and judgment, and graded them according to the following
classification.

Pollution of water:

No water
High pollution

Moderate pollution

le-‘O

No pollution

90
Turbidity of water:
No water 0
Turbid water 1
Clear water 2

Services:

The popularity of any campground among the park
users depends greatly upon the facilities and services
available at a recreation area. Facilities and services
have been defined in different ways by different
individuals.

For the purposes of this study, services can be
defined in terms of the utilities such as toilets, showers,
grocery, etc., and which may be found within or close to
the campground.

Different types of grading systems have been con-
sidered. However the most satisfactory and less subjective
one seemed to be a dichotomous scaling system, where a
value of one is given to the variable, if present in the
area, and a value of zero if it is absent. For instance, if
there are flush toilets on the campground, a value of one
is given to the variable, and if, on the contrary, these
are lacking, a value of zero is given.

Flush toilets:

Presence of flush toilets 1

Absence of flush toilets 0

91

The Role of Factor Analysis

 

The list of variables defined above includes out-
door activities variables. However, it was found, by a
trial and error process, that the factor analysis was
improved when these trichotomously-scaled activity variables
were deleted; the major problem encountered, when using the
activity variables in the factor analysis, was that the
different activities did not correlate with a common factor,
but instead correlate with different factors.

The results of the initial factor analysis conducted
after the deletion of the activity variables were not
entirely satisfactory. It was therefore decided to
eliminate some additional variables which have been found
of no useful significance when working out the factor
analysis. It was found that two of these variables, namely
the presence of toilets and showers, have no detectable
correlation with any particular factor used in the present
analysis. Another variable, the presence of a waterfall,
which occurs in only one of the provincial campgrounds, has
also been deleted, as a single observation is insufficient
to establish correlation.

After many runs through the factor analysis program,
it was found that the three-factor solution was the most
satisfactory. A short study of the variables grouped by
the factor analysis method shows that the first factor

groups together all the physical environment variables

92

except water which characterize the campgrounds; the

second factor groups all the water characteristics and,

finally, the third factor groups the different services

present in the park.

So the three factors were named

respectively, physical environment, water and service

factors.

Factor Loading:

 

The following table shows the factor loadings as

determined by the factor analysis for each of the 25
variables. The variables which are included in the factor

have been underlined in Table 7 for the sake of classifi-

 

 

cation.

TTEHHE 7

Varimax Rotated Factor Matrix

Factor 1 Factor 2 Factor 3

Physical en- Water Services

vironment
1 Laundry -0.l7592 -0.03640 0.90157
2 Electricity -0.00814 -0.26059 0.55269
3 Grocery -0.27877 -0.10993 0.90137
4 Restaurant -0.15249 —0.18563 0.86472
5 Playground -0.23713 0.25315 0.66013
6 Swimming pool -0.45083 -0.63489 0.28218
7 Roads -0.33978 -0.11592 0.53124
8 Type of site 0.37858 -0.09773 0.03240

93

TABLE 7 (Cont'd)

 

Factor 1 Factor 2 Factor 3

Physical en- Water Services

vironment
9 Presence of shelter 0.00700 -O.16900 0.54300
10 Relief 0.74720 0.08815 0.49215
11 Panorama 0.30623 -0.09414 0.34647
12 Virgin forest 0.87569 0.15642 -0.27372
13 Shade 0.78087 -0.-1351 -0.09737
14 Vegetation 0.75019 0.10913 ~0.08866
15 Quietness 0.67250 0.16841 -0.34129
16 Flies and pests -0.75126 -0.41917 0.20765
17 Hiking trails 0.87570 0.15642 -0.27372
18 Quality of fishing 0.77497 0.20221 -0.lO754
l9 Bodies of water 0.06685 0.82365 -0.02750
20 Offshore composition 0.06963 0.96309 -0.15690
21 Foreshore composition 0.06882 0.98228 -0.l7879
22 Water pollution 0.75234 0.54487 -0.09750
23 Turbidity of water 0.64370 0.56143 -0.25626
24 Length of the beach 0.07337 0.84756 -0.24898
25 Boat rental 0.26000 0.43579 0.30506

 

Following the results obtained above, it was found
necessary to build a measuring scale that would represent
the relative value of each variable corresponding to the
three factors selected. In order to obtain the desired
measuring scale, two successive steps were necessary. The

first one being to establish the factor score coefficients

94

that determine the correlation between variables and the
factors obtained above; the second one being to establish

the factor scores themselves.

Factor Score Coefficients:

 

Nie g; 2122 has outlined in statistical package
for the Social Sciences a subprogram in which they propose
to use the least square regression method to estimate the
factor score coefficients. They explain that such a program
defines the best set of coefficients for the variables in
such a way that the correlation between the composite
variable and the hypothetical factor is maximum.

The formula used in calculating the score

coefficient matrix is:

 

A = s' R"1
where
A = The factor-score coefficient matrix
s = The rotated factor structure matrix
R = The correlation matrix
For instance, the regression coefficient associated
with the jth factor is then given by:
m -1
ai1 = p21 rlp Slj
where
ajl = Regression coefficient associated with the
factor
22

Nie, Bent, and Hull "Statistical Package for the
Social Sciences," op.cit., p. 226.

95
r-l
1p = Element of inverse of the correlation matrix
slj = Element in jth column of the rotated factor
matrix
j = l, 2, . . ., K
1 = l, 2, . . ., m
The factor score coefficients thus obtained are

shown below in Table 8.

{RABIJB 8

Factor Score Coefficient

 

10

ll

12

13

14

Laundry
Electricity
Grocery store
Restaurant
Playground
Swimming pool
Roads

Site

Shelter
Elevation
Viewpoint
Virgin forest
Shade

Vegetation

FACTOR 1

-0.53825

0.03009

-0.64535

-0.11750

-0.68889

-0.76400

-0.78430

1.39230

0.23971

1.71969

0.23917

1.38254

1.59370

3.18497

FACTOR 2

-1.47252

-0.96861

-0.51069

-0.46813

0.56815

-l.964l9

-0.74037

0.08070

-0.35388

0.73783

-0.49275

0.36052

-0.11069

-0.11798

FACTOR 3
2.66338
1.03714
0.95394
1.62458
0.90274
1.86111
2.05202
0.16954
0.83972

-0.28186
0.66872

-2.22840

-l.10721

-0.77155

TABLE 8 (Cont'd)

FACTOR 1 FACTOR 2 FACTOR 3
15 Quietness 2.26452 0.30693 -3.47351
16 Flies and pests -2.77755 -0.94988 2.31977
17 Hiking trail 2.35697 0.62948 -2.45126
18 Fishing quality 3.26238 0.53754 -l.53450
l9 Bodies of water 1.55506 0.64168 -0.71194
20 Offshore composition 1.11117 1.30178 -1.21766
21 Foreshore composition 1.33342 1.34912 -l.21156
22 Water pollution 1.72460 0.71448 -0.95275
23 Turbidity of water 1.09641 0.37849 -l.33761
24 Length of the beach 0.59241 1.33228 -l.25978
25 Boat rental 0.60978 0.81500 0.61953

96

 

Factor Scores:

 

Once the factor score coefficients were obtained, it
was deemed sufficient to go to the second step and determine
the factor scores for all 27 provincial campgrounds used in
this model. The factor scores, which represent the relative
importance of one factor at one particular site compared
with the same variable in another site, were obtained
through the standard analytical method.

To build the factor scores scale, two methods can be
used.

These methods are described by Nie §E_al in the

following manner:

97

"It is customary to build factor scales employing
only those variables that have substantial loadings
on a given factor. It seems, however, that the
complete estimation has some advantages over the
first method. In the shorter method, the influence
of variables not included in the scale is not con-
trolled; they will affect the scale through this
intercorrelation with the variables used in the scale.
In the complete estimation method, on the other hand,
some variables are simply used as supression

variables to give the best estimate of the given
factor."23

In the present paper, because of the amount of work

involved in the complete estimation method, it was decided

to build the factor score scale by employing only those

variables which have substantial loadings on the given

factor. In other words, only the value of the variables in

each

factor with the highest loadings are used. For

example, if different variables have the following loadings

for three different factors, only variables 1 and 2 will be

used

tion

in order to determine the factor score for one observa-

in factor number 1.

 

Factor 1 Factor 2 Factor 3
Var. 001 0.88920 0.07829 ' 0.03230
Var. 002 0.78523 0.014023 0.005768
Var. 003 0.10210 0.67352 0.06342
I I I I
I I I I
I I I I
Var. 0070 0.32460 0.34210 0.04274
23

Ibid., p. 226.

98

Then, in order to determine the different scores,

 

the following formula is used by Nie eg'el:24
m _
Factor Score = 151 a (X1 _xi)
- sdi
where
a = The standard-score coefficient
x1 = The original value of the variable i
ii = The mean of the original values for a
variable i
sdi = The standard deviation of the original value
for a variable i
m = The number of variables.

The factor scores for the different campgrounds

considered in the present study are shown below in Table 9.

TABLE 9

Factor Scores

 

 

Factor 1 Factor 2 Factor 3

Physical

environment Water Services
1 Lac Dozois 22.46 4.23 -ll.29
2 Belle Riviére 18.71 6.04 -10.60
3 Moisie 17.22 4.63 5.02
4 Lac Albanel 19.01 .79 - 8.45

 

24Ibid., p. 227.

 

99
TABLE 9 (Cont'd)
Factor 1 Factor 2 Factor 3
Physical
environment 'Water Services
5 Riviere Chalifour - .65 4.39 -10.60
6 Lac d'Argenson 18.71 -6.61 -13.56
7 Lac Normand 15.66 7.75 -13.56
8 Lac Savary 19.01 7.75 -l3.56
9 Riviere Matane 2.60 -4.12 2.78
10 Riviere des Ecorces 8.60 -2.64 - 7.64
11 Lac Arthabaska 9.67 -l.85 -10.60
12 Lac Lajoie 20.23 6.90 -10.60
13 Lac La Vieille 23.69 5.42 - 7.64
14 Lac Monroe 12.96 7.75 4.48
15 Lac Bellevue 21.45 3.44 -13.56
16 Port Daniel 25.16 4.12 - 8.93
17 Lac Rimouski 23.94' 4.30 -l3.56
18 La Sablonniere 8.75 5.49 -13.56
19 Voltigeurs -27.66 -4.12 7.53
20 Mont Orford 3.13 6.78 8.77
21 Oka - 7.34 7.75 -10.60
22 Stoneham - 9.34 4.54 8.13
23 Mare du Sault 17.04 -2.65 8.19
24 Barriere John 28.46 -4.12 - 7.63
25 Etang a la Truite 10.06 5.35 -l3.56
26 La Loutre 15.26 7.07 5.80

27 Lac du Milieu 15.38 6.04 - 5.28

 

100

In the above table, factor scores show the relative
importance of one factor in the different campgrounds. For
example it shows that for the Lac Dozois campground, the
physical environment factor is more important with a value
of 22.46 than that of Belle Riviere which shows a value of

18.71.

Aejusted Scores
The above table lists the factor scores determined

for each of the 28 campgrounds selected. However, these
values have to be standardized with a mean of 0.00 and a
standard deviation of 1.0 for the following reasons, as
described by Racine:

"The method of standardization makes it possible

to compare values pertinent to different variables.

These values may follow differensscurves which could

differ widely from one another."

Table 11 shows these standardized values.

TABLE 10

Standardized Factor Scores

 

Factor 1 Factor 2 Factor 3
l Lac Dozois .71 .13 - .69
2 Belle Riviére .42 .59 - .60
3 Moisie .44 .23 1.23

 

25J. B. Racine, Modéles Graphiques et Modeles Mathé-
matiques en Géographie Humaine. Essai de Méthodologie
Expérimentale (Ottawa: Université d'Ottawa, Department de
Geographie, 1969), p. 42.

 

Table 10 (Cont'd)

4

5

1000401

10
11
12
l3
14
15
16
17
18
19
20
21
22
23
24
25
26
27

Lac Albanel
Riviére Chalifour
Lac d'Argenson
Lac Normand

Lac Savary
Riviere Matane
Riviere des Ecorces
Arthabaska

Lac Lajoie

Lac La Vieille
Lac Monroe

Lac Bellevue

Port Daniel
LaclRimouski

La Sablonniere
Voltigeurs

Mont Orford

Oka

Stoneham

Mare du Sault
Barriére John
Etang a la Truite
La Loutre

Lac du Milieu

101

Factor 1

 

.44
-1.05
.42
.18
.44
- .84
- .37
- .28

.54

Factor 2

 

.43
1.02
- .07
.10
.15
.45
-1.97
.78
1.02
.21
-l.60
-1.97
.41
.85
.59

Factor 3

 

- .30
.60

- .96
- .96
- .96
1.04
.23

- .60
- .60
- .24
1.25
- .96
- .39
- .96
- .96
1.63
1.78

1.70
1.70
- .23
- .97
1.41
- .05

 

102

Table 10 shows the same relative importance of one
factor as in Table 9 but presented on a different basis, as
explained above.

In Table 11 for convenience, we standardized around
a mean of 100 and by using a standard deviation of 40

eliminate the negative values.

 

TABLE 1 1

Adjusted Factor Scores ; = 100

s.d. = 40
Factor 1 Factor 2 Factor 3
l Lac Dozois 128.46 104.98 72.12
2 Belle Riviere 116.80 122.96 70.03
3 Moisie 117.60 109.18 149.28
4 Lac Albanel 117.60 70.38 88.00
5 Riviére Chalifour 44.11 106.78 76.03
6 Lac d'Argenson 116.20 129.18 61.61
7 Lac Normand 107.20 80.34 61.61
8 Lac Savary 117.60 80.84 61.61
9 Riviére Matane 28.24 20.77 101.69
10 Riviere des Ecorces 85.26 36.00 90.85
11 Arthabaska 88.80 43.99 76.03
12 Lac Lajoie 121.60 132.40 76.03
13 Lac La Vieille 132.07 117.21 90.42
14 Lac Monroe 98.80 140.77 150.06
15 Lac BelleVue 125.28 97.21 61.61

Table 11 (Cont'd)

Factor 1 Factor 2 Factor 3
16 Port Daniel 136.81 104.42 84.40
17 Lac Rimouski 139.22 106.00 61.61
18 La Sablonniére 86.99 117.98 61.61
19 Voltigeurs 9.91 21.16 165.27
20 Mont Orford 155.98 131.20 171.26
21 Oka 35.60 140.77 76.03
22 Stoneham 29.21 108.41 168.00
23 Mare du Sault 111.20 36.00 168.00
24 Barriére John 147.21 21.20 90.82
25 Etang 5 la Truite 90.00 116.39 61.21
26 La Loutre 106.42 134.99 156.49
27 Lac du Milieu 106.42 123.60 98.03

103

 

 

Campers' Preference Study

 

In the present study, we had sufficient data to
indicate the most desired characteristics requested by the
users of the campgrounds. These data were obtained through
a survey made by the research service among the campers of

26 Because of this, we are in a

the Province of Quebéc.
position to obtain an attraction index based on factors
which are themselves evaluated or weighed according to the

preference expressed by the users themselves.

 

26Service de la Recherche, Essai sur 1es usagers des
terrains de camping provinciaux du Quebéc (Quebec:
Mifiistére du Tourisme de la Chasse et de la Péche, 1971),
pp. 102-109.

 

 

104

The present approach is quite different from that of
most of the studies we have consulted and in which equal
weight was assigned to each factor. In those studies, the
authors assumed that there was no difference in the campers'
preference for different campground characteristics.

Such an unrealistic assumption was usually made because of
lack of information on the recreationist's preference for
the various activities.

During the summer of 1971, the Department of
Tourism, Fish and Game conducted a survey throughout the
Province of Quebec on camper's attitudes toward activities,
facilities and various natural characteristics.

This survey was conducted in 17 parks distributed
throughout the province, five in the Montreal area, five
in the Quebec City area, four in the Gaspe Peninsula, and
one in each of the following regions: Eastern Townships,
Drummondville, and Lake Saint-Jean (see Map No. 1). A
total of 698 interviews were conducted during the summer
period; these interviews were distributed as follows:

201 interviews from June 13 to June 30.

167 interviews from July 1 to July 15.

207 interviews from July 16 to July 31.

123 interviews from August 1 to August 18.

698

One of the survey questions consisted in asking the

user to list characteristics most desirable in campgrounds

105

he intends to use. Along the same line of thought, during
the same survey, the camper was asked to state those
characteristics he appreciates the least in the campground
he visits. This last question was used as a cross reference
permiting the analyst to verify answers obtained in reply

to the preceding question.

With all the answers, one single list was then
drawn-up. Such a list enumerated the different character-
istics preferred by interviewed campers. It was then
possible to sort out from this list the different character-
istics which could be grouped under each of the factors
defined earlier with the factor analysis method. These are
(1) natural characteristics; (2) services and (3) water.

Tables 12 to 14 below show the number of persons
that preferred each type of characteristics. Each figure
shown in these tables, represents the total of positive

answers from users.

106

TABLE 12

Natural Characteristics

 

User's preference listed in decreasing order

 

 

l Quietness of the environment 171
2 Beauty of the site 153
3 Absence of flies 109
4 Natural character 104
5 Lot of space available 30
6 Shade 23
7 Privacy of the sites _22
612
TABLE 13
Services
User's preference listed in decreasing order

1 Overall quality of the services offered 168

2 Good planning of the services offered such as
distance between individual campers 104
3 Order and cleanliness 96
4 Quality of welcome at reception 36
5 Access roads 28
6 Internal circulation 11
7 Security 8
8 Facility to obtain a camping site __3

 

107

TABLE 14

Water

 

User's preference listed in decreasing order

1 Presence of good swimming facilities 62
2 Closeness to water 44
106

 

The three tables above indicate that out of a total
of 1,172 users, 612 or 52.2% of the peeple interviewed
showed a preference for the natural characteristics, 454 or
38.7% for services and 106 or 9.1% for the presence of
water.

From the above, it was concluded that it is
possible to attribute weights or value of different
magnitudes to the different factor scores obtained for each
observation; this is a new element which permits a more
refined estimation of the desires of campers. With this in
hand, we need not be limited to an assumed equal weight of
unity for each factor score, as used by most researchers in
the subject area. On the contrary, these preference
rankings provide a measure of the extent to which one factor
is preferred to another by the campers.

In the above section, it has been shown from our
survey that the people going to campgrounds located in
provincial parks prefer natural characteristics in a pro-

portion of 1.349 individuals or 612/454 against 1.0 or

108

454/454 who prefer high quality services and .233 or
106/454 who are more interested in the presence of water.
It is important to note that this set of weights applies

to the average park visitor and that when these weights are
applied to a particular campground it is not certain that
visitors may be considered as average visitors.

The following table gives the weighted factor
scores for each factor and for each of the different camp-
grounds. To obtain the figures of Table 15, the adjusted
scores found in Table 12 are multiplied by the weights

which have been developed in the section above.

Tfidflflfi 15

Weighted Park Factor Scores

 

Weight = 1.349 Weight = .233 Weight = 1.00

 

 

Natural

Characteristics Water Services
1 Lac Dozois 173.29 24.46 72.12
2 Belle Riviére 157.56 28.65 70.03
3 Moisie 158.64 25.44 149.28
4 Lac Albanel 158.64 16.40 88.00
s Riviére Chalifour 59.50 24.88 76.03
6 Lac D'Argenson 156.75 30.10 61.61
7 Lac Normand 144.61 18.72 61.61
8 Lac Savary 158.64 18.72 61.61
9 Riviére Matane 38.09 4.84 101.69

 

109

Table 15 (Cont'd)

Weight = 1.349 Weight = .233 Weight = 1.00
Natural

 

Characteristics Water Services
10 Riviére des Ecorces 114.24 8.39 90.85
11 Lac Arthabaska 119.79 10.25 76.03
12 Lac Lajoie 164.04 30.85 76.03
13 Lac La Vieille 178.16 27.31 90.42
14 Lac Monroe 133.28 32.80 150.06
15 Lac Bellevue 169.00 22.65 61.61
16 Port Daniel 184.55 24.23 84.40
17 Lac Rimouski 187.80 24.70 61.61
18 La Sablonniere 116.01 27.49 61.61
19 Vbltigeurs 13.37 4.93 165.27
20 Mont Orford 210.44 30.57 171.26
21 Oka 48.02 32.80 76.03
22 Stoneham. 39.40 25.26 168.00
23 Mare du Sault 150.01 8.39 168.00
24 Barriére John 198.59 4.94 90.82
25 Etang a la Truite 121.41 27.12 61.21
26 La Loutre 143.56 31.22 156.49
27 Lac du Milieu 143.56 28.80 98.03

 

In Table 15 each column shows the relative value of
the characteristics of each campground, this indicates for

each site the interest of the users.

110

Activities:

 

The relative values of the different activities
associated with the parks was not included in the factor
analysis calculation but was calculated separately for the
reasons given on page 91.

This scale of values is based on the assumption
that the presence of an activity has a positive value and
that its absence has no value. The activities have then
been graded according to the following trichotomous scale:

No activity 0

Medium quality 1

Good quality 2

The following table shows the distribution and
relative value of each activity that is offered to campers

in the provincial campgrounds of the Province of Quebec.

TABLE 16

1:1].

Raw Scores

 

 

VI
.2
u
.— m U,
> U C
.— I— '-
u 0 ‘0
U o_ ._
“I U" 0'! W‘-
.U .0 .E .E .x .- u
E IE 5 -5 S ‘2 8 'E ‘8 .5 ‘° 3 .— '5 o :3
'3 .2 8‘ :5 8’ 5 S 5 '6 3 I; S 3 '3 3 3 m
WKOIOIUFU¢WWIW>JU
Mara du Sault 0 2 2 0 0 0 0 2 0 l 0 0 0 0 2
Stoneham 2 1 l 0 1 0 1 0 l 2 0 1 2 0 0 0 2
Mont Orford 2 l 1 2 0 0 l 0 2 l l l 1 l 0 0 2
Barricrc John 0 2 O 2 O 2 0 O O 0 0 O 0 O O 0 2
Etang Truite 0 2 0 2 0 2 2 0 0 2 0 1 0 0 0 0 2
Port Daniel 0 2 O 2 O 2 O O O 0 0 0 0 0 0 0 2
Lac Rimouski 2 2 0 2 1 2 2 0 0 2 0 2 0 l l l 2
Lac Monroe 2 2 l 2 l 0 2 0 0 2 1 2 0 l 1 l 2
Lac La Vieille 2 2 0 2 0 2 2 0 0 2 0 2 0 1 1 l 2
Lacujoie 11020020021122222
Lac Bellevue 2 2 0 2 0 2 2 0 0 2 0 l 0 0 0 0 1
Lac Normand 2 2 0 2 0 2 2 0 0 2 0 2 0 2 O 0 O
Lac du Milieu 2 1 0 2 0 0 2 0 0 2 0 2 0 0 0 0 2
Lac d'Argenson 2 2 0 2 0 0 2 0 0 2 0 2 0 2 0 0 2
Lac Albanel 0 2 0 2 0 0 2 0 0 2 0 1 0 0 0 0 1
Riviire Chalifour 1 2 0 2 0 0 2 0 0 2 0 l 0 0 0 0 2
M01818 1 2 0 1 0 0 0 0 0 0 0 1 0 0 0 l. 0
Lac Arthabaska 0 2 0 0 0 0 1 0 0 1 0 0 0 0 0 0 2
Ln Loutre 1 1 o o o o 2 o o 2 o 1 o 1 1 o 0
Belle Riviére 2 2 0 0 0 0 2 0 0 2 0 2 0 1 1 0 2
Rivitre den Ecorces 0 2 O 0 0 0 2 0 0 2 0 1 0 0 0 0 l
Riviira Matane 0 l 0 0 0 0 0 0 1 0 0 0 2 0 0 0 l
Oka 2 l 0 1 0 0 1 0 l 1 0 2 l 2 2 2 2
Voltigeurs l 0 0 1 2 0 l 0 2 1 0 l 1 0 0 0 1
La Sablonn1§re l 2 0 2 0 0 2 0 0 1 1 2 0 0 0 0 2
Lac Dozois 2 2 0 1 0 2 2 0 0 2 0 2 0 0 0 0 2
Lac Savary 1 2 0 1 0 2 0 0 0 2 O 0 0 0 2 0 2
Bollard des Ormeaux l l l l 0 0 2 0 0 2 0 2 2 2 2 2 1

 

112

Weights:

In the above-mentioned survey conducted by the
Department of Tourism, Fish and Game the preference of the
campers for each type of activity was also investigated.

In the survey conducted by the Department, the
users of camping facilities were asked to identify those
kinds of recreation activities they prefer at campgrounds.

This was an open type of question in which each
person gives his personal rating to each of three differ-
ent activities normally associated with campgrounds; these
opinions, were to be based entirely on personal taste. A
value of three was given to the first choice, of two to the
second choice, and of one to the third choice.

In the following table the number of points given
to each activity is distributed between the first three
columns according to the choice of individual campers. The
fourth column gives the total number of points according to
each individual activity. The fifth column is the total
number of points expressed in percentage; this value is then
used as the weight of each individual activity. In order
to arrive at this step it is necessary to make the assump-
tion that the values given to the different activities are

interpersonally valid.

TABLE 17

Activity Weights

113

 

10

11

12

13

14

15

16

17

18

19

20

Swimming
Fishing
Organized activities
Hiking

Organized sports
Excursion
Hunting
Canoeing

Tennis

Golf

Row boating
Skiing

Sun Bathing
Badminton
Horseback riding
Sailing

Yatching

water Skiing
Camp fire

Bicycle ride

Points

lst 2nd 3rd

 

798 332

492 144

144

93

90

54

30

42

30

45

21

42

24

27

12

12

12

94

98

56

56

72

52

44

92

38

26

36

22

16

16

12

10

8

10

90

65

44

30

13

15

16

13

20

17

24

Total

Points

1,220
701
282
221
159
125
118

99
94
94
83
77
62
58
31
30
27
25

25

22

3,523

% = weights

 

34.6

19.8

2.7

2.7

2.4

1.8

1.6

0.9

0.9

0.8

0.7

0.7

 

100%

 

114

Weighted activity scores:

The activity weights as calculated indicate the
activities most desired by the users of the campgrounds;
it is now possible to take into account the campers
preference for each different campground activity. The
activity weights calculated in Table 17 are used to deter-
mine the weighted activity scores for each activity offered
in the campgrounds studied. In order to obtain the figures
of Table 18, the recreational activity grades found in

Table 16 are multiplied by the weights shown in Table 17.

TABLE 18

115

Weighted activity Scores

 

 

m
.2
U
.- m 01
> t .E
75 o ‘c
u o. ._
m m m m 1.
c c
'0 'U -- o— J a— o
01 0 0 CD «I .c u 01 .x 1.
c 0| N N 0'1 c to u to 0'! c v: ._
°-£°-°*--.S'-.: 8 28.52.“-_.
E L g E «‘5 u 8 c u- m -— u 0 0. <1:
-— m m .x m I: c c - 3 «- c 1. -- u u E 1—
3 -- 1. .. L 3 to u o .1: :1 o to to m to o
m u. o z: o :l: u I- o x m m x m >- 3 u .—
Mare du Sault 00 40 07 13 00 06 00 00 05 00 02 00 00 00 00 01 .74
Stoneham 69 20 07 00 04 00 03 00 03 05 00 O2 02 00 00 00 01 1.16
Mont Orford 69 20 07 13 00 00 03 00 05 02 02 02 01 01 00 00 01 1.26
Barriére John 00 40 00 13 00 07 00 00 00 00 00 00 00 00 00 00 01 .61
Etang Truite 00 40 00 13 00 07 06 00 00 05 00 02 00 00 00 00 01 .74
Port Daniel 00 40 00 13 00 07 00 00 00 00 00 00 00 00 00 00 01 .61
Lac Rimouski 69 40 00 13 04 07 06 00 00 05 00 04 00 01 01 01 01 1.52
Lac Monroe 69 40 07 13 04 00 06 00 00 05 02 04 00 01 01 01 01 1.54
Lac La Vieille 69 40 00 13 00 07 06 00 00 05 00 04 00 01 01 01 01 1.48
Lac Lajoie 35 20 00 13 00 00 06 00 00 05 02 02 02 02 02 01 01 .91
Lac Bellevue 69 40 00 13 00 07 06 00 00 05 00 02 00 00 00 00 01 1.43
Lac Normand 69 40 00 13 00 07 06 00 00 05 00 04 02 00 00 00 00 1.46
Lac du Milieu 69 20 00 13 00 00 06 00 00 05 00 04 00 00 00 00 01 1.18
Lac d'Argenson 69 40 00 13 00 00 06 00 00 05 00 04 00 02 00 00 01 1.40
Lac Albanel 00 40 00 13 00 00 06 00 00 05 00 02 00 00 00 00 01 .67
Riviére Chali- 35 40 00 13 00 00 06 00 00 05 00 02 00 00 00 00 01 1.02
four
Moisie 35 40 00 06 00 00 00 00 00 00 00 02 00 00 00 01 00 .84
Lac Arthabaska 00 40 00 00 00 00 03 00 00 02 00 00 00 00 00 00 01 .46
La Loutre 35 20 00 00 00 00 06 00 00 05 00 02 00 01 01 00 00 .70
Bella Riviére 69 40 00 00 00 00 06 00 00 05 00 04 00 01 01 00 01 1.27
Rivibre Ecorces 00 40 00 00 00 00 06 00 00 05 00 02 00 00 00 00 01 .54
Riviére Matane 00 20 00 00 00 00 00 00 03 00 00 00 02 00 00 00 01 .26
Oka 69 20 00 06 00 03 03 00 03 02 00 04 01 02 02 01 01 1.17
Voltigeurs 35 00 00 06 09 03 03 00 05 02 00 02 01 00 00 00 01 .67
La Sablonniere 35 40 00 13 00 07 06 00 00 02 02 04 00 00 00 00 01 1.10
Lac Dozois 69 40 00 06 00 07 06 00 00 05 00 04 00 00 00 00 01 1.38
Lac Savary 35 40 00 06 00 00 00 00 00 05 00 00 00 00 02 00 01 .89
Ballard des 35 20 07 06 00 07 06 00 00 05 00 04 02 02 02 01 01 .98

Ormeaux

116

Adjusted scores:

 

The above table lists the weighted activity scores
determined for each of the 28 campgrounds selected. These
values are then standardized around a mean of 0 with a
standard deviation of 1; then, in order to make these
values comparable with values obtained previously for each
factor, it is necessary to adjust these values around a
mean of 100 and a standard deviation of 40., such as shown

in the following table.

TABLE 19

Standardized and Adjusted Activity Scores

 

 

 

 

Total

weighted

activity Scores Adjusted

scores standardized scores

2 = 1.00 i = 0.00 2 =100

s.d. = .358 s.d. = 1.00 s.d. = 40
Lac Dozois 1.38 1.06 142.41
Belle Riviére 1.27 .75 130.00
Moisie .84 - .45 82.00
Lac Albanel .67 - .92 63.25
Riviére Chalifour 1.02 .06 102.43
Lac Argenson 1.40 1.12 144.85
Lac Normand 1.46 1.28 151.20
Lac Savary .89 - .31 87.67

Riviére Matane .26 -2.07 17.26

Table 19 (Cont'd)

Riviére des Ecorces
Lac Arthabaska
Lac Lajoie

Lac La Vieille
Lac Monroe

Lac Bellevue

Port Daniel

Lac Rimouski

La Sablonniere
Voltigeurs

Mont Orford

Oka

Stoneham

Mare du Sault
Barriére John
Etang a la Truite
La Loutre

Lac du Milieu

Total

117

weighted
activity

score

S

 

x
s.d.

1.00
.358

Scores

standardized

i =
s.d. =

0.00
1.00

 

.54
.46
.91
1.48
1.54
1.43
.61
1.52
1.10
.67
1.26
1.17
1.16
.74
.61
.74
.70
1.18

-1.28
-1.51
- .25
1.34
1.51
1.20
-1.09
1.45
.28

- .92
.73
.45
.45

- .73
-1.09
- .73
- .84
.50

Adjusted

scores

2 =100

s.d. =

40

 

48.84
39.68
90.03
153.69
160.44
148.01
56.42
158.00
111.20
63.20
129.21
118.00
118.00
70.81
47.52
70.81
66.42
120.00

 

118

Calculation of the Attraction Index

The index of camping quality for each campground is
a weighted combination of its scores on the individual
variables. It is however important to note that the pre-
vious studies apply to the average park visitor, and that
when the value defined for the average visitor is applied
to specific campgrounds it is necessary to make the
assumption that such value is interpersonally valid. Only
then, can we add up these different values. The composite
index for each park is the sum of the four scores, as shown

in Table 20 below.

119

 

mm.mae 55.06H mo.omH om.~m m~.mma mouse: 683
mm.m4¢ mm.mma mv.om Hm.sm ma.mpa maanmfi> an own
mm.omm mo.om no.6» mm.om vo.eoa maonmq one
ma.mv~ mm.mm no.6h m~.oa mh.maa mammnmanu< can
mm.~o~ vm.m¢ mm.om mm.m 4~.4HH mmonoom moo mnmn>em
om.HmH o~.ha mm.aoa 4m.¢ mo.mm manna: mumn>em
4m.m~m em.hm Hm.ae ~8.wa sm.mma anm>mm one
ea.mem o~.ama HG.HG ma.ma Hm.qsa oceanoz own
Hm.mmm mm.4qa HG.HG oa.om mh.mma comcmmna.o own
vs.~mm m4.~oa mo.oa mm.v~ om.mm usomflamno wnmn>nm
m~.e~m m~.mm oo.mm oq.ma se.mma Hmamnae one
om.mav oo.~m m~.mva qs.m~ 4m.mma mango:
¢~.~mm co.OMH no.8“ me.m~ mm.hma mnmu>8m maamm
m~.~Hv. H¢.~¢H ~H.~a o¢.v~ m~.mha mflonoa own
.HflBOB mmfluflxrflunud mTOH>me kum3 mOHumHHOUOMHmnu

Housumz

 

xoocH cowuomuuué

om mqmdfi

120

 

 

mm.omm oo.o~a mo.mm om.m~ Gm.mva unflaflz so one
mm.nmm ~¢.Go m5.mma -.Hm mm.m¢a mausoq ma
mm.om~ Hm.o> H~.He NH.>N H4.H~H munsna mamum
hm.aem ~m.a¢ mm.om 4m.4 mm.mma anon mumnnnmm
H~.hmm Hw.oe oo.mea mm.m Ho.oma nasmm so 6982
mo.omm oo.mHH oo.moa m~.m~ ow.mm ausmcoum
mm.va~ oo.maa mo.ea om.~m No.m¢ mac
m¢.H¢m H~.mma mm.aha hm.om vw.oa~ 690690 6202
ah.o¢~ o~.mm h~.mon mm.¢ am.ma mnsmmnuao>
Hm.eam o~.HHH HG.HG 54.nm no.6HH mnmeaeoflnmm no
oa.~m¢ oo.mma HG.HG on.¢~ om.hma nxmsoaam omq
mm.m5m ~5.om ov.vm m~.v~ mm.5ma Hmflcmo upon
n~.oo¢ Ho.mwa Hm.ae mm.- oo.mma ms>maamm own
.HJNBOH. wwwnuflerfluud mmuwarhmm kumg mOHumHHmuOMHMSU

Housumz

Ac.ucoov om manna

121

The last column of Table 20 shows that the camp-
grounds are graded according to the value of their
recreative attraction index. For example Mont Orford with
the highest value is the most interesting to visit.

Let us consider at this point a case where we use
the attraction index to define the substitution effect for
one particular county (see page 67, paragraph 4). Let us
take the county of argenteuil with the six campgrounds that
are located within a radius of 200 miles measured from the
center of the county.

Table 21 lists these six campgrounds along with
their respective distances and attraction indices. It is
followed by the numerical value of the equation mentioned
above. This gives the substitution effect index for this
county. The same formula was applied to all the other
counties of the province and the substitution effect indices
for these counties are listed in Table 22.

The value of these indices indicates how much the
citizens of one county will be inclined to remain in the

vicinity of their point of origin when they go camping.

122

TABLE 21

Campgrounds located within a radius of 200 miles
from the center of the county of Argenteuil

 

 

 

Campgrounds Distance . Attraction Index
Lac La Vieille 144 449.49
Lac Savary . 144 326.64
Lac Lajoie 48 390.95
Lac Munroe 60 476.58
La Sablonniere 60 316.31
Oka 36 274.85

 

From these data the value of the substitution
variable for the county of Argenteuil is determined by means

of the following gravity formula.

 

 

 

 

 

 

n
Xij = i i 1 1“ho Si
13
Xi' = 10910 449.49 + 10910 326.64 + 10910 360.95+
3 144 144 48
10910 476.58 + 10910 316.31 + 10910 274-35 = 245.52
60 60 36
Xij = 0.24552 = Substitution effect variable.

For instance, the people of Argenteuil with an
index of 0.24552 would be expected to camp closer to home

than the people of Abitibi. These people, with an index of

123

.10887 will have to find the needed facility outside of the

immediate vicinity of their county.

The Model

 

In the previous section we have defined the
different variables and explained how these were arrived
at. The 1971 data plotted on an arithmetic scale give a
curvilinear distribution of the points. It was observed
that when plotting the same data on a double logarithmic
scale the points had a tendency to group themselves along
a straight line. This type of curve proved to be more
useful for further analysis.

Such an equation, where we used the double-
1ogarithm form, is a special case of the more general
technique of transformation of variables to achieve
straight-line relationships. In that instance, the
dependent variable Y and the independent variables xl to X6
are transformed into log Y and log x1 to log x6 and a
linear regression equation was then calculated, using the
transformed value in place of the original data. Such an
equation where we used the double-logarithmic values is
linear in its parameters but curvilinear in its variables;
hence there is continuous change in the dependent variable
depending on the level at which a particular independent
variable happens to be entered into the equation. A
double logarithmic transformation was used to obtain the

equation form shown below for each campground. In this

124

equation, the dependent variable is the number of visitors
originating from the different counties, and the independ-
ent variables represent the values of the different
socioeconomic variables that characterize the counties
located within 200 miles of each sample campground. The
data used in this analysis indicate that a great majority
of the visitors originate within 200 miles from each
destination area. The form of the equation is:27
log (Y + 1.0) = a + b1 log x1 + b2 log x2 + b3
log x3 + b4 log x4 + b5 log x5 + b6 log x6 + b7

 

109 x7
where

Y1 = The number of visitor-days from a particular
county of origin

x1 = The population of the county of origin of the
visitors

x2 = The per capita income in the county of origin
of the visitors

x3 = The approximate cost of the round trip from the
county of origin to the campground

x4 = The average age of the heads of households in
the county of origin

27

In this equation we added 1.0 to the value of each
dependent variable for each observation. Such a constant
is necessary as in some cases there is no visitation
originating from a particular county and the logarithm of
zero is undefined.

125

x5 = The average age of the inhabitants in the
county

x6 = The degree of urbanization of the county of
origin

x7 = The substitution effect variable constructed
to reflect the competitive effect of other
campgrounds
(this was extensively explained in page 67
above).

The following table indicates the values of the

dependent and independent variables for each of the 72

counties of the Province of Quebec.

126

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ocean. mn.~m nu.o¢ a on.¢oo md.noo oo.-o onnd aqaoo oooo 000 0000 oaoum nod
Amman. NN.0~ ~¢.~¢ A o~.co~ mn.h¢~ on.~nH oo- nméoo 0000 000 #000 ouauco>ucom no
Ougwm. oo.o~ oo.h¢ H mw.h¢o on.¢wo m~.ono coda unwoo «H00 000 0000 Howauuom um
wodnn. o~.w~ HN.0€ a m~.ono ow.on~ oo.¢no Omoo Nuwoo @000 N00 muoo oundBOOuHom In
Nnfinm. No.5N ~0.m¢ N m¢.omo Nn.nno on.nwo OMAN admoo Node mac odoo muonudnauom no
amomw. o~.o~ Ho.n¢ H ¢O.H¢o wo.an~ ch.wno ohma smooo Owoo ooo «moo mun-on In
wnnon. -.n~ ~o.n¢ g ow.-o o~.~oo w~.ono owfia onwoo waoo «Ho 0000 uomum no
owomn. mm.om oo.¢¢ g Nn.oNo mm.¢on an.o¢o Odo“ cameo mnoo 400 #500 «menwnuH4 In
Nnnano ms.o~ mh.oe A 0H.Ono 00.noo on.noo Owen nomoo oooo 000 0600 stuucwwu< nu
hwwofi. n®.n~ ~m.n¢ a om.sma on.~oo dw.om~ owha oodao «Odo «00 0000 «nauw94 1~
95 9V ”V “SH 03 “OH 10“ 3V d H 1 1
I," am“ 3A 1"". W430 I’O Elam 9A 0 m P. e
39 es 991 n o n da d o
as J P1 ene .43u lu Wlu 11 n 1 W
o 3 e e u 1 I. J D. 1.1 D. J D. 3 e I. .I
3|. 5 05 P 09 99 JD. 95 e 0 E J
19 9 I29 1A3 A3 3A1 a .4 J a
n 11w.“ MW”. PRU .u..d u. m: M P
H w mm a... m d a d n d as m J a n mozcaou
m .M 5 (€an pUD ”“33 S”? ml. 0- H S
m Wm: mm Mmm mmm N m
u. m. a a .1. a .1.
m D. mvcsotmanu ucotomm_c

mo_no_to> ucoocodovc_

mo_nm_to> ucoocodouc_ can ucoccodoo 0:» mo mo:_m>

NN m4m<h

c. omaa>oo 10 Conan:
mo_nm_to> unoccOQuo

21227

#0000.
50NON.

mmhmfi.
omoLN.
oNMLN.
a25ng.
nmoas.
MAL5N.
woman.
mmnwm.
nnmwm.
moamm.
«mama.
omenw.
622mm.
.mNmLN.
anonm.

000N0.

00.0N
00.0N
Nn.00
50.nN
N0.0N
0H.nN
00.0N
0N.0N
00.0N
00.5N
N5.5N
0~.0N
00.0N
00.0N
05.5N
00.0N
0N.5N

00.0N

00.00
05.00
00.50
00.H0
00.00
N0.00
00.00
00.00
00.00
00.50
H0.n0
00.50
00.00
00.00
0N.50
0H.00
00.n0

00.H0

00.000
Nn.0N0
00.000
00.N50
00.N50
00.NON
00.05N
N0.~00
Nn.0N0
00.000
00.000
00.000
N0.00~
00.000
Nn.550
00.000
0N.0mo

00.nN0

00.000
00.050
00.000
00.H00
00.~00
0N.00N
N0.~5N
00.5HH
0N.~00
N0.00~.
0N.000
0H.50~
00.00~
Nn.550
N~.0N~
00.000
N0.00~

0N.m50

00.N50
00.050
00.000
N0.~H~
0N.0H~
0N.N0~
00.55H
N0.000
00.N00
0N.000
00.000
0N.m50
00.5N0
00.000
0H.N00
0~.000
Nn.000

00.N50

0~0~

005a

000a

000N

00~N

005a

00NH

OnHH

005a

0000

000N

0000

000N

000N

05-

000a

00NN

0NHO

00000
00N00
00000
0HOH0
00000
00H00
00000
HON00
00000
0H000
00000
00000
000H0
00000
00~00
H0~00
N00~0

00NNO

0000
N000
NH00
0000
N000
0000
0000
0N00
N000
0000
N000
0000
0000
N000
0000
0000
0000

0000

000

000

000

N00

00N

000

000

000

000

000

N00

000

000

0N0

N00

000

500

050

0000
0a00
0000
0000
N000
0000
0000
0000
0000
0000
0000
0000
N500
5000
#000
0000
0000

0000

ouuowaon
o-u>uunn
covmcfiucsz
~09:

auocwuau
umo=0100nc0
unmnwdnau
oaCoucoum
vcoaasua
uoumonouoo
mocwmucoznxaon
coudaoo
«Eauaou«no
haamsoouuso
unoao “396 .7230
one Muo>ofluunu
cacadaano

xlneuno

-0N

:5N

unN

10N

uNN

wHN

uoN

1N0

1H“

 

Ae.uc660 - u3m<5

112$!

00N00.

0000N.
000NN.
5000N.
0NO0N.
N0000.
NO0NN.
0N000.
0000N.
N5000.
00000.
0000N.
00000.
0000N.
00NNO.
.NN000.
0000N.

0N550.

00.0N

N0.0N
00.5N
00.0N
05.0N
00.0N
N5.0N
00.0N
00.0N
00.5N
00.5N
00.0N
00.0N
00.0N
00.0N
N0.0N
00.0N

00.0N

0N.00

00.50
00.50
50.00
N0.00
00.00
00.50
00.00
00.00
N0.00
00.00
0N.00
00.00
00.00
00.00
0N.00
00.00

05.00

00.5N0

00.000
00.N50
00.000
0N.000
05.000
0N.00N
00.000
00.N00
00.000
N0.550
N0.000
00.000
00.5N0
00.000
N0.000
00.N50

00.000

N0.000

00.N00
00.500
00.000
00.000
0N.000
N0.00N
05.00N
00.000
N5.000
05.000
00.0N0
0N.050
00.N50
00.550
00.050
00.000

00.000

0N.050

00.5N0
00.000
00.050
00.000
N0.000
0N.000
N5.000
N0.000
00.0N0
00.000
00.0N0
00.000
N0.050
00.000
No.000
00.0N0

00.500

000N

0000
00N0
0000
0550
000N
0000
0NNO
0500
0500
00N0
000N
000N
000N
0Nn0
000N
0NNO

0000

000NN

00N00
00N00
00000
00000
00000
00N00
05N00
00N00
00N00
00N00
0N000
05000
00000
00000
00000
00000

00N00

050N

0000

0000

0000

0000

0000

0000

0000

0000

5000

0000

0000

5000

0000

0000

0000

0000

0N00

000

000

000

000

000

000

000

000

000

000

000

000

000

000

000

000

0N0

000

0000

0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0500
5000
0000
N000
0000
0000

0000

Anamon «LHVLuauucoz

accouoaucoz
hc0daucoz
E0auucoz
00:080nu0z
00ucnmwx
udewaauux
canon:
wmdoc0xnaz
ouw0c00u00
00000.0
n0>wa
co0unaomn<.0
o0u0au000

unono
cuohcuc0u0 and
umm sconuucwnm u¢0
u006090

oxnauaoaux

100
-00
100
:00
1N0
:00
:00
100
:00

:50

 

00.0:000 NN 000<5

11291

0000N.
0005N.
505N0.
00NON.
00NON.
00NON.
N0000.
0500N.
000N0.
NN000.
N000N.
0000N.
NON00.
0500N.
0N000.
N000N.
0500N.

50000.

00.0N

00.0N

00.0N

00.0N

00.5N

00.0N

50.NN

00.5N

0N.5N

00.0N

N0.5N

00.0N

55.0N

00.5N

00.0N

00.0N

00.0N

00.0N

00.50
00.00
05.00
00.00
00.00
00.00
00.00
00.00
00.50
05.00
00.00
N0.00
00.00
00.00
N0.00
00.00
N0.00

0N.00

00.000
00.000
N0.000
00.000
00.NNO
00.0N0
00.000
0N.000
N0.000
00.550
00.000
00.000
00.000
00.000
00.000
0N.000
00.000

0N.000

00.050
N0.000
00.000
00.000
00.050
00.050
00.000
00.N50
00.000
0N.000
00.000
00.000
00.5N0
0N.N00
00.500
N0.000
00.000

N0.550

N0.000
00.050
00.000
00.500
00.050
N0.000
00.000
00.050
N0.000
0N.000
N0.000
N0.000
05.000
00.5N0
00.000
00.500
00.000

00.000

0000

050N

0050

000N

050N

0000

0500

000N

0000

0000

00NN

000N

050N

0N50

0000

0N00

0N00

0000

00000
00000
00000
00000
0N000
00000
05N00
00000
50000
00000
00000
00000
00000
50000
00000
00000
00N00

00000

0000
5000
0N00
00N0
00N0
00N0
0000
N000
0000
0000
0000
0NNO

0N5
0000
0000
0000
NN00

0N00

000

N00

000

000

0N0

N00

000

000

000

000

N00

000

0N0

000

000

000

000

000

0000
0N00
0N00
0000
0N00
0000
0000
0000
0000
N000
N000
0000
0500
N000
0000
0000
0000

0000

nomcc0aom
oxoouauonm
vuomuosm
600unuzauc0om
coonuuc0um
onucwun»=-ucgam
hqcosmam
o0002.60
5565-56-6uaapam
063 308.30
0:630000
560060000
000030

usucuuom
ou0u=om
aaoc0mam
uw0oouz

0000>uo0maz

1N0

:00

100

:00

100

150

 

.Au.uc660 «N u3m<p

130

N000N.
000NN.
0000N.
0500N.
5000N.
50000.
NON00.

00000.

05.0N
00.0N
00.0N
00.5N
5N.0N
00.0N
05.0N

00.0N

00.00
50.50
00.00
00.00
00.00
00.50
05.00

00.00

00.0N0
00.0N0
00.0N0
00.000
N0.000
N0.550
0N.050

N0.000

00.000
00.000
0N.050
00.050
N5.000
N0.000
00.050

00.000

00.500
N5.000
0N.050
00.000
00.000
00.050
N0.NON

N0.550

00N0

0000

000N

0000

000N

0000

0000

0000

00000
50000
00000
55000
00000
00N00
05000

00000

0000

0000

0500

0000

0500

0000

0000

0000

000

000

000

000

000

000

000

500

0000

0000

N000

0000

0000

0000

0000

N000

oxnuau»
60003
noumauuo>
58.3:
occonouuoa
consoum0swa
630:0800m0ewa

vmoumcuum

1N5
:05
:05
I00
I00

:50

 

00.0c060 NN 040<5

131

Analysis of the Data

 

The S.P.S.S. multiple-regression analysis computer
program28 was used to estimate the regression coefficients.
The S.P.S.S. multiple regression analysis is a computer
program with automatic stepwise addition of variables to
form a least squares equation. This program provides a
means of choosing independent variables which will provide
the best prediction possible with the fewest independent
variables. Through the use of this stepwise-regression
analysis, an independent variable is selected at each
stage from all independent variables not presently in the
equation. The independent variable selected for inclusion
is that variable which will reduce to the greatest extent
the unexplained variance around the mean of the dependent
variable. It follows, of course, that the selected
independent variable is the one that provides the largest
increase in the coefficient of multiple determination

(R2) .

 

28Nie, Bent, and Hull "Statistical Package for the
Social Sciences," cp.cit., p. 208.

132

Level of Significance

 

In this study we apply a level of significance of

29 30 in two studies of

0.1 such as used by Little and McCoy
the same kind.

The significance level, which is predetermined, is
a restricting criterion which will terminate the selection
of the variables. To check the above, the significance of
the observed value of F, which is the variance ratio, is
calculated to explain the variance of the last variable
included against the F distribution. That provides a
judgment on the contribution made by each variable as
though it had been the most recent variable entered,
irrespective of its actual point of entry into the model.
Any variable which provides a non-significant contribution
is not added to the model.

31 stepwise regression

According to Draper and Smith,
is the best of the variable selection procedures. However,
stepwise regression can easily be abused by including

irrelevant variables in the equation, since the inclusion

 

29Little, "A Study of Factor Affecting Pleasure Travel
to U.S.," op.cit., p. 198.

30Edward Wayne McCoy, "Analysis of the Utilization of
Outdoor Recreation in Tennessee," op.cit., p. 39.

31M. R. Draper and H. Smith, Applied Regression
Analysis (New York: John Wiley & Sons, Inc., 1966),
p. 172.

 

 

133

of even one irrelevant variable will almost certainly
increase R2 due to chance fluctuation. It is therefore

necessary to use sensible judgment in the initial selection

of the variables.

Interpretation of Results

 

For the campgrounds studied in this model we obtain
three equations, one for each campground. The different
equations are shown in the following table. These equations
show only the variables which have been found significant

in the least square analysis.

TAIHJB 23

Final Regression Equation

 

 

 

Campgrounds Equations R2 value S value
La Vieille log(Y+1.0)=0.43892 + 1.11236 log x .67973 0.49919
-2.04462 log x3+l.11790 log x6-1.7 302
109 X7
Mont Orford log(Y+l.0)=l.68015 + 1.70128 log x1 .55988 0.65651

-1.30490 log X3-1.18291 log X7

La Mare du Sault log(Y+l.O) - 15.94958 + 1.09780 log .65031 0.49720
X1 -2.l7l93 log x3-8.52690 log x4

 

In these equations:
Y = The estimated number of visitor-days from a
particular county of origin at the different

campgrounds

134

x1 = The population of the county of origin of the

visitors

x3 = The cost of the round trip from the county of

origin to the campground

x4 = The average age of heads of households in the

county of origin

X6 = The degree of urbanization of the county of

origin

x7 = The substitution effect variable constructed

to reflect the competitive effect of the other
campgrounds.

In the above equations, the coefficients of the
population of the county of origin and the coefficient of
the variable that reflects the nature of the county of
origin indicate that each successive increase in value of
those variables results in ever smaller increases in visits,
assuming that all other variables remain constant. However,
the coefficient of the cost of the round trip, the
regression coefficient of the average age of the heads of
households, and the regression coefficient of the
substitution effect variable, indicate that each successive
increase in value of these variables results in ever smaller
decreases in visits to the campgrounds, when all other
variables are taken as constant.

In a subsequent analysis a check was made to

discover if the different equations of Table 23 are

135

statistically different from one another. For example, a
check was made for a significant difference between the
equation of Lac La Vieille and that of Mont Orford.

The method of achieving this is as follows:
10) The coefficients in the first equation are replaced by
the coefficients in the second equation.
20) With the new coefficients in place we calculate the
projected values for each county.
30) Using the original value found for each county and the
second set of values obtained after substitution of the
coefficients for the same counties, we run a least squares
analysis and obtain the following equation:

X = a + by

where

X Estimated values before substitution

y = Values after substitution
40) It is a known fact that if the two sets of results are
identical we will get a regression line at exactly 45° with
an origin at 0.
50) We then verify whether the coefficients a and b of the
equation shown in step No. 3 are significantly different
from 0 and l which we normally have in any linear curve at
45°. In doing so, if we find that:

a) a = o and b = l, we conclude that the equations are
identical.

b) a f o and b = l, we conclude that the equations are

parallel.

136

c) a = o and b # 0, we conclude that even if the
origin is the same the equations are different.
d) a f o and b # 0, we conclude that the equations are
different.
60) In order to verify this we used the well known "t"
test below:

t

II
D)
I
9)
£13
8.
(‘1'
ll
(7‘
I
0‘

where a and b are the two coefficients obtained in the
equation shown in step 3.

Using the above principles and applying the "t"
test we may now proceed to verify if the equations of
Table 23 are statistically different. Let us compare
first the equation of Lac La Vieille with that of Mont
Orford using the data below computed from the least squares

analysis.

137
n = 33
2 x1 = 819.26 2 y1 = 976.84
2 xi = 67351.19 2 yi = 102651.95
E = 24.83 § = 29.60
2 - _ _
2 x1 [(2 x1) (x1)] — x00 — 47012.
x x2 - [(2 y ) (§ )1 = s = 73730 38
1 l 1 ll '
XY- ‘= =
Z l 1 [(2 x1) Yl)] $01 51923.28
b = 501 = 51923.28 = 0.72
S11 73730.38
a = £1 - [(b) (21)] = 24.83 — (.72 x 29.60) =
3.51
Re31dua1 sum of squares Roo = S00 - b $01 = 47012 -
(0.72 x 51923.28) = 9.62
In the above:
n = number of counties
x1 = Value obtained before substitution
y1 = The value obtained after substitution

We just verify the null hypothesis Hl : a = 0 against

alternative a # 0 at a level of significance .01.

where Sa is the estimated standard error of a, given by:

 

 

138
-2 A
Sa = (1 + x ) Roo = (1 + 876.16 ) 9.62 = 0.11
(n 511) n-Z (33 73730.38) 31
t = 3.51 - 0 = 31.91

0.11

The null hypothesis is then rejected, as 31.91 is
much larger than 1.64, which, in conformity with table of
t distribution, is the value of "t” with a level of
significance of .10 when n = 33.

We now verify the null hypothesis H2 : b = 1 against
the alternative b # l at a level of significance .01.

t = b-Bo

Sb

where Sb is the estimated standard error of b given by:

 

 

 

 

sb = Roo / n-2 = 9.62_4_31 = .0065
s 73730
11
t = 0.72 - 1 = 43.07
'.0065

The null hypothesis is then rejected as 43.07 is
much larger than 1.64 which in conformity with table of
t distribution is the value of t with a level of signific-
ance of .10 when n = 33. We then conclude that the
equations are different.

Let us now see if there is a significant difference
between the equation for Mont Orford and that for La Mare
du Sault.

Here again, we obtain two sets of estimations as

explained above from which we have the following:

 

139
n = 55
2x2 = 373.14 zyz = 2548.43
x2 = 12931 28 y2 = 2335880 17
2 2 ' z 2 '
x2 = 46.35 Y2 = 6.78
y2 - [( y2) (" )1 = s = 2217760 40
z 2 2 Y2 00 °
x2 - [( x ) (i )1 = s = 10401 40
z 2 2 2 2 11 °
x - "" =
X x2y2 [(2 2) (Y2)] - 801 116273.58
b = 801 = 116273.58 = 2.47
800 47011
a = §2 - b i2 = 46.35 — 11.18 (6.78) = 29.45
Re81dua1 sum of squares R00 = S00 - b 801, =
2217760.40 - 11.18 (116273.58) = 938757.43

In the above:

n = number of counties
x2 = values obtained before substitution
y2 = values obtained after substitution.

We first verify the null hypothesis Hl : a = 0 against

the alternative a # l at a level of significance .01

Sa

where Sa is the estimated standard error of "a" given by:

 

 

 

140
sa = (1 + 33) R00 = (1_ + 45:96 ) 938757.41 =
(n 511) n—2 (55 2217760.40) 53
17.4
t = 29.45 - o = 1.06
17.4

 

The null hypothesis is then accepted and we may
conclude that the origins are similar.
We now verify the null hypothesis H2 : b = 1 against

the alternative b # 1 at a level of significance .01.

 

 

 

 

t = b - Bo
Sb
where Sb = Roo / n-2 = 938757.40/53 = 0.089
Soo 2217760.4
t = 2.47 - l = 15.30
.089

The null hypothesis is then rejected, as 15.30 is
much larger than 1.64, which, in conformity with table of
t distribution, is the value of t with a level of
significance of .10 when n = 55. We then conclude that
even if the origin is the same, the equations are different.
From the above, it must be inferred that the
equations corresponding to the three campgrounds under

study are not similar.

Data Used in Projecting Visitation Estimates for the Years

1975 and 1980

 

The estimation of future recreation participation

at the campgrounds studied requires the projection of

141

values for the independent variables included in the
estimating model.

In this type of model, two variables are liable to
change over the period of time considered in the present
study: (a), the population of the county of origin x1, and
(b), the mean income per capita of the county of origin x2.
Other variables likely to change in the same lapse of time
are significant in none of the equations. However, since
no projection of the future value of these variables is
available on a county basis in Canada, it was necessary,
in this study, to estimate changes expected to occur in the
next ten years. It was first assumed that the average age
of the head of household and the average age of the
population would likely change very slowly, and that over a
period of ten years, such variables would remain relatively
stable. It was therefore decided that the 1971 values would
be used in the projections. The statistics show that in
the last ten years both the population and the income
distribution of the population have changed considerably.
It was therefore necessary to estimate changes in these two
variables for the forthcoming decade. The following method
was used.

It was first assumed that there is a simple linear
relationship between the value of these two variables and

time. Using this relationship, a linear equation was then

142

deve10ped with the use of the least squares method for each

of the 72 counties in the Province of Quebec.

Yi = a - b xj
where
Y1 = population of the origin county
xj = year for which we are interested in
estimating the population
a and b = coefficients

From the equations thus obtained estimates of
future papulation and future income for the 72 counties
were calculated. The corresponding regression coefficients
are shown in the following table, along with the projection

for 1975 and 1980.

143

TABLE 24

Projected Population of the Counties of Origin

 

County Regression Partial re- Projected Projected

slope gression population population
coefficient in 1975 in 1980
(a) (b) (00) (00)
Abitibi 1153.200 .800 1246.0 1212.0
Argenteuil 343.600 -4.000 279.0 259.0
Arthabaska 475.400 3.500 531.4 548.9
Bagot 224.400 1.000 240.4 245.4
Beauce 663.400 -2.100 629.8 619.3
Beauharnois 530.800 -0.900 516.4 511.9

Bellechasse 283.2 -5.9 188.8 159.3 '
Berthier 292.2 -2.9 245.3 231.3
Bonaventure 460.6 -2.9 414.2 399.7
Br6me 148.0 -0.6 138.4 135.4
Chambly 1472.0 63.6 2489.6 2807.6
Champlain 1211.4 -10.6 1041.8 988.8
Charlevoix est 175.8 -l.2 156.6 150.6
Charlevoix ouest 160.6 -2.0 128.6 118.6
Chateauguay 335.2 19.2 642.4 738.4
Chicoutimi 1686.6 -7.2 1571.4 1535.4
Compton 264.6 -5.6 175.0 147.0
Deux-Montagnes 338.0 8.0 466.0 506.0
Dorchester 377.4 -5.2 294.2 268.2
Drummond 612.4 3.7 671.6 690.1
Frontenac 332.2 -5.9 237.8 208.3

144

Table 24 (Cont'd)

 

County Regression Partial re- Projected Projected
slope gression population population
coefficient in 1975 in 1980
(a) (b) (00) (00)
Gaspé-est 458.9 -5.3 374.1 347.6
Gaspé-ouest 231.0 -6.0 135.0 105.0
Gatineau 539.8 0.4 546.2 548.2
Hull 814.6 17.3 1091.4 1177.9
Huntington 158.2 -0.4 151.8 149.8
Iberville 191.4 0.8 204.2 208.2
Joliette 475.4 2.6 517.0 530.0
Kamouraska 293.6 -3.5 237.6 220.1
Labelle 306.4 -O.4 300.0 298.0
Lac Saint-Jean est 476.6 -1.7 449.4 440.9
Lac Saint-Jean ouest 661.0 -7.6 539.4 159.6
Laprairie 303.6 21.3 644.4 1252.3
L'Assomption 397.6 14.8 634.4 708.4
Lévis 532.6 7.5 652.6 690.1
L'Islet 267.6 -3.0 219.6 204.6
Lotbiniére 330.2 -5.5 242.4 214.7
Maskinongé 225.2 -1.3 204.4 197.9
Matane 379.4 -8.5 243.4 200.9
Matapédia 401.6 -11.5 217.6 160.1
Mégantic 623.4 -5.8 530.6 501.6
Missisquoi 312.8 2.4 351.2 363.2
Montcalm 202.4 -l.0 186.4 181.4
Montmagny 280.4 -1.5 256.4 248.9

Table 24 (Cont'd)

.145

 

County

Montmorency
Montréal
Napierville
Nicolet
Papineau
Pontiac
Portneuf
Québec
Richelieu
Richmond
Rimouski
Riviére-du-Loup
Rouville
Saguenay
Saint-Hyacinthe
Saint-Jean
Saint-Maurice
Shefford
Sherbrooke
Soulanges
Stanstead

Témiscamingue

Regression

slope
(a)

274.6
19462.4
120.8
335.0
356.0
216.8
542.6
3378.8
399.0
456.4
706.4
439.2
273.2
816.4
477.6
404.6
1178.6
577.2
837.0
108.0
385.2

654.0

Partial re-

gression

coefficient

(b)

65.4

3.1
38.1

2.2

Projected
population

in 1975
(00)

244.2
23739.2
117.6
283.8
282.4
188.0
496.2
4425.2
519.0
370.0
610.4
343.4
322.8
1426.0
512.8
438.2
1089.0
644.4
1072.2
111.2
362.8

554.8

Projected

population

in 1980
(00)

234.7
25075.7
116.6
267.8
259.4
179.0
481.7
4752.2
556.5
343.0
580.4
346.8
138.3
1616.5
1277.7
488.7
1061.0
665.4
1145.7
112.2
355.8

523.8

 

 

146

Table 24 (Cont'd)

 

County Regression Partial re- Projected Projected
slope gression population population
coefficient in 1975 in 1980
(a) (b) (00) (00)

Témiscouata 333.8 -9.6 180.2 132.2
Terrebonne 1053.8 26.3 1474.6 1606.1
vaudreuil 295.0 6.9 405.4 439.9
Verchéres 266.4 6.5 370.4 402.9
Wblfe 201.4 -4.5 129.4 106.9

Yamaska 175.6 -2.6 133.4 120.4

 

 

147

TABLE 25

Projected Income of the Counties of Origin

 

County Regression Partial re- Projected Projected
slope gression income in income in
coefficient 1975 1980
(a) (b)
Abitibi 748.030 78.252 2000.06 2391.32
Argenteuil 817.121 97.238 2372.93 2859.12
Arthabaska 691.818 56.258 1591.95 1873.24
Bagot 602.424 46.293 1343.16 1574.58
Beauce 453.788 66.084 1511.13 1841.55
Beauharnois 1085.909 76.014 1292.23 2682.77
Bellechasse 469.091 41.678 1135.94 1344.33
Berthier 580.909 50.245 1384.83 1636.05
Bonaventure 406.818 61.643 1393.11 1701.32
Brame 615.757 54.371 1485.70 1757.55
Chambly 1034.394 161.888 3624.60 4434.04
Champlain 758.333 117.308 2635.26 3221.80
Charlevoix est 603.106 70.766 1735.37 2089.19
Charlevoix ouest 527.121 54.930 1406.00 1680.65
Chateauguay 713.485 175.489 3521.30 4398.75
Chicoutimi 684.848 149.510 3077.01 3139.71
Compton 501.818 80.874 1795.80 2200.17
Deux-Montagnes 611.818 140.489 2859.64 3562.09
Dorchester 426.212 48.146 1196.55 1437.28
Drummond 793.484 70.105 1915.16 2265.69
Frontenac 448.182 58.357 1381.89 1817.59

 

148

Table 25 (Cont'd)

 

County Regression Partial re- Projected Projected
slope gression income in income in
coefficient 1975 1980
(a) (b)

Gaspé-est 423.636 65.210 1467.00 1793.05
Gaspe-ouest 492.121 99.930 2091.00 2590.65
Gatineau 1354.000 62.000 2346.00 2656.00
Hull 787.424 131.678 2238.53 3552.66
Huntingdon 716.061 67.273 1792.49 2128.19
Iberville 871.515 62.587 1872.91 2185.84
Joliette 652.727 74.580 1846.01 2218.91
Kamouraska 360.909 65.629 1410.97 1739.12
Labelle 482.121 58.776 1422.54 1716.42
Lac Saint-Jean est 897.864 96.790 2428.50 2930.45
Lac Saint-Jean oeust 739.242 47.937 1506.23 1745.92
Laprairie 711.515 127.972 2759.07 3398.93
L'Assomption 564.394 156.119 3062.30 3842.89
Lévis 643.636 115.979 2499.30 3079.19
L'Islet 411.212 66.224 1524.32 1684.64
Lotbiniére 348.333 63.077 1693.56 1672.95
Maskinongé 614.697 60.559 1583.64 1886.44
Matane 375.909 71.014 1512.13 1867.20
Matapédia 583.000 44.000 . 1287.00 1507.00
Mégantic 595.000 122.308 2551.93 3163.47
Missisquoi 957.727 58.042 1886.43 2176.61
Montcalm 462.576 69.476 1574.19 1921.57

1..

I.“

 

 

149

Table 25 (Cont'd)

 

County Regression Partial re- Projected Projected
slope gression income in income in
coefficient 1975 1980
(a) (b)
Montmagny 511.515 61.049 1488.30 1793.54
Montmorency 600.000 100.000 2200.00 2700.00
Montreal 1246.061 116.119 3103.96 3684.56
Napierville 504.697 69.790 1621.34 1970.29
Nicolet 385.455 75.315 1590.49 1967.07
Papineau 620.152 78.951 1883.37 2278.12
Pontiac 714.394 56.119 1612.30 1892.89
Portneuf 717.273 75.804 1930.14 2309.16
Québec 856.515 126.049 2873.36 3503.48
Richelieu 841.670 93.287 2334.26 2893.98
Richmond 873.940 111.573 2658.84 3216.97
Rimouski 473.182 85.664 1843.81 2272.13
Riviére-du-Lcup 565.000 70.000 1685.00 2035.00
Rouville 500.455 156.084 2997.80 3778.22
Saguenay 344.545 231.993 4056.43 5216.40
Saint-Hyacinthe 798.636 63.287 1802.23 2127.66
Saint-Jean 1287.727 49.196 2074.86 2320.84
Saint-Maurice 750.606 126.189 2769.63 3400.57
Shefford 824.242 68.706 1923.54 2267.07
Sherbrooke 991.818 78.182 2242.73 2633.64
Soulanges 628.485 103.182 2279.40 2793.01
Stanstead 820.606 54.266 1688.86 1960.19
Témiscamingue 982.727 74.195 2169.85 2540.82

150

 

Table 25 (Cont'd)
County Regression Partial re— Projected Projected
slope gression income in income in
coefficient 1975 1980
(a) (b)
Témiscouata 500.909 50.629 1310.97 1504.12
Terrebonne 718.939 106.189 2417.96 2948.91
Vaudreuil 722.727 187.273 3717.49 4655.46
Verchéres 630.758 153.217 3082.23 3848.31
Wolfe 368.788 68.776 1469.20 1813.08
Yamaska 385.758 68.986 1489.53 1834.46

 

151

Method of Estimating Future Participation

As seen above we now have a particular equation
which was developed for each of the three sample campgrounds.
These equations may now be used to estimate the participa-
tion of the citizens from any of the counties located within
a radius of 200 miles of each of the three campgrounds
mentioned above. This radius of 200 miles was used in the
present study because the data indicate that most visitors
to campgrounds originate within that distance.

In order to estimate the participation for the
years 1975 and 1980, the values for the independent
variables for each county were inserted in each of the
three equations developed above (page 135).

The results obtained through the solution of each
equation for each county and a summation of these results
gives an estimate of the total expected visitation for
the campgrounds. The following tables show the estimated
number of recreational visits to each of the three camp-
grounds during the peak of the summer season, that is,
duringthe months of July and August. In order to estimate
the number of visitations during these two months, the
estimates for the two-week survey period during the same
months were used. It was then possible to estimate the
participation during that eight-week period, assuming that

the survey is representative of that period.

152

TABLE 26

Projected Number of Recreation Visits
to the Three Campgrounds During July and
August From Within a 200 Miles Radius

 

 

1972 1975 1980
Mont Orford 34780 40900 46933
Lac La Vieille 4152 4879 5211
La Mare du Sault 9340 10188 10758

 

General Application

 

In the Province of Quebec, none of the Provincial
Parks is used exclusively for recreation and all of them
are used simultaneously for the production of timber,
mineraly and recreation. Furthermore due to sharp
competition within these parks, it is very important to
estimate the future demand for recreation in order to
obtain the value of the area for recreational use. Such a
value, ones established for recreation, may be compared
with that already established for the alternative uses of
the same area, namely: timber or mineral production.

For the population, the estimation of future demand
is especially important, in three different interrelated
aspects of outdoor recreation activities. First it is most
important to know the quantity and quality of the land,
water and other natural resources that will be needed in

the future for outdoor activities. If this information is

 

153

not known some time ahead, we may be assured that the most
valuable land, if not all the land in provincial parks, will
have been acquired by the forest and mining industries and
that it will be extremely difficult to acquire areas for
public use, when the need arises to establish recreation
facilities. It is also possible that the water, air, and
even the soil, could be spoiled, through pollution caused

by industrial activities, to a much greater extent than by
any use that could be made of the same areas for recreation
purposes.

A second field in which future needs must be
estimated as early as possible is that related to the
recreation activities which are accessory to camping, such
as fishing, hunting, hotels, etc. . . . It is impossible
to imagine that a government could build a new recreation
site when it is needed, and make available to the users
abundant fishing, hunting or hiking facilities, if this
has not been foreseen some time ahead of actual construc-
tion.

Finally the forecast must also include the major
problem of availability of manpower for both the construc-

tion and operation of new campgrounds and recreation sites.

“t.

“

SUMMARY

The present study was undertaken to estimate camp-
ground participation in the Province of Quebec both in
government and private campgrounds. In order to attain the
result intended, a survey was first made in the city of
Sainte-Foy, a suburb of the provincial capital. This was
done by sending a questionnaire to every fifth family in
the city, for a total of 1,700 households. The purpose was
to find out which socioeconomic characteristics have a
significant influence on the camping participation
originating from this city. A total of 693 answers were
received and subsequently analyzed by regression analysis.
The results thus obtained indicated that with this type of
model, it is very difficult to estimate further participa-
tion; yet it indicated quite clearly which socioeconomic
characteristics are likely to have a significant influence
on camping participation.

In this analysis, the “t" test is applied to the
interval-scale variables to verify the null hypothesis that
i1 = fi, where 21 is the mean value of the participants and
Y2 is the mean value of the non-participants. Upon going
through the analytic procedure, it was found that the

younger populations, the peOple with lower incomes and

154

 

155

those with larger families have a tendency to produce a
larger number of camping days, whereas non-campers belong
to the older populations with higher incomes and smaller
families.

The chi-square test is applied to the nominal
scale variables to verify the hypothesis that frequencies
of participation are not influenced by the occupation of
the participants. The analysis of the data indicated that
there is no significant difference between the rate of
participation of the people with different occupations,
except for retired persons who show a clear tendency toward
a lower participation.

In a second study we used the basic data on the
participation of the population in different campgrounds
as established by the governmental statistics for 1971.
The purpose of this study was to define the relationships
between campground participation and the socioeconomic
characteristics of the people of each county.

Factor analysis is used to define the attraction
index of the different campgrounds. It is a multivariate
statistical method which makes it possible to consider a
large number of interrelated variables and reduce them to
a smaller number of factors.

By means of a stepwise regression analysis, the
size and slope of the regression coefficients have been

estimated for three campgrounds. In order to achieve this,

156

the variables having the greatest influence on the degree
of participation in campground attendance were first
identified.

Then a double logarithmic equation was filled to
the data obtained for three different campgrounds. An
equation was developed which was later used to estimate
future participation in camping at the three campgrounds.
Using this equation with forecasted values for the
independent variables for the years 1975 and 1980 indicated
that the participation of the population is expected to
increase by 17.60% and 34.92% for Mont Orford, 17.53% and
25.51% for Lac La Vieille, 9.02% and 15.21% for La Mare du

Sault as compared with the 1971 participation.

 

CONCLUSIONS AND SUGGESTIONS
FOR FURTHER STUDY

The present study is the first attempt to implement
an econometric analysis in the field of outdoor recreation
in the Province of Quebec. The use of campgrounds has
developed very rapidly in recent years for both state-owned
and private enterprises. As a result of this accrued
activity, abundant data and statistics have been accumulated
but have never been used for an analysis of the phenomenon
and for an adequate forecast of the future use of camp-
grounds.

It is the author's belief that this first attempt
has been successful in pointing out the most prominent
factors which govern the use that the public presently
makes of existing campgrounds. The present study went
further and used these factors to estimate the future
participation of the public in campground activities for
1975 and 1980.

However, there are some shortcomings which are
listed below along with some recommendations for further
studies.

1) Plenty of data concerning the campgrounds understudy
had been collected, but none had been properly classified

and made readily available. That is why we had to limit

157

u,

158

our compilation to two-week sample periods for each camp-
grounds. Due to this limitation, we had to make the
assumption that for the estimated number of recreational
visits each week is representative of the month to which it
belongs.

In the near future, the Department of Tourism,
Fish and Game will use new registration forms which will
permit the building of an improved data bank. This one
would be not only most valuable to make readily available
all information needed for future studies of this kind,
but more complete, and more dependable, because it would be
based on a sample covering a full season instead of a
few weeks.

2) For the present study, future values of the different
socioeconomic variables were needed. Since no projection
of these variables was available on a county basis in
Quebec, it was necessary to make our own estimates of the
changes expected to occur in the next ten years.

Such projections could be improved through close
cooperation between the Dominion Bureau of Statistics and
the Department of Tourism, Fish and Game. Thus, it would
be possible in the future to obtain better estimates on a
county basis.

3) It is important to be very cautious when developing
a prediction model because basic data used in the model
change very rapidly due to technological improvements and

other exogenous events. However, the author believes that

 

159

it is worth while to develop complete models for short
periods of time in order to have adequate adaptation to
changing conditions. It is further possible to have a
continuous process of model revision as conditions change;
that may mean more work and greater costs, although not an
insurmountable task. It is even possible to develop
simulation models for the various subsystems which could
be dynamic in both time and space.

In addition to the analysis made in the present
study, there are many aspects of recreation demand that
require further investigation. For instance the present
analysis deals with conditions which may be considered as
entirely static because it was based on a sample survey
made for different campgrounds at one point in time during
which the consumer had at his disposal a specific supply
of recreation Opportunities. It has been assumed that the
availability of facilities will gradually increase with the
demand.

In this study those possible effects that a change
in supply may have over a period of time on the participa-
tion of the peOple in recreational were not taken into
account. It is possible that the future supply of outdoor
recreation opportunities will be adequate to meet the
estimated demand, nevertheless we are not sure that the
supply will expand to meet the increased demand which was
demonstrated in this study. Since this aspect of the

question could not be included in the present study because

160.

of a lack of data, it is felt that as time goes on, more
and more data will be made available,_and any future
analysis of recreational activities is bound to come to
grips with the importance of the supply factors. That is
why it is believed that any further studies should make
use of the trends indicating the increasing amount of
facilities, this would permit to obtain more precise
results in future estimations.

In the present study a statistical relationship
has been established between the actual participation of
the people in campground activities and certain socio-
economic variables. After formulating this relationship,
the best possible scientific computations were made of the
future levels of the socioeconomic variables, and these
were used as a basis for predicting future participation.

In so doing it was assumed that in the future the same
relationship between these variables and the participation
will maintain.

It is suggested that, in a future study, a similar
statistical relationship be established between participa--
tion and supply. After proper processing of the data some
very interesting and useful information may result from such
a research program.

It must be remembered that in the present study most
of the variables were socioeconomic ones and thus impossible
to change or control. This new aspect of the question is

most interesting, especially for future planning, because

 

161

the availability of supply is one of the only variables on
which the decision maker has some control and thus may lead
to a better use of campgrounds.

Along the same line of thought, the principle of
"learning by doing" may be investigated to see how the
increasing availability of recreational facilities will
affect the ability of the individual through greater F“
opportunities to practice. This is especially true of most 4
sports and is bound to affect the eventual demand for this
sort of facilities.

In the first part of the present thesis, the study
was based on the socioeconomic characteristics particular
to individuals within a single city. In the second part,
the study was based on the socioeconomic characteristics
particular to the population of specific geographical units
(counties). It would be of great interest to use as the
basis for a future study the characteristics particular to
groups of individuals that make up the population of a
country. The groups could be professional, ethnical,
financial, religious . . . etc.

The results of such a study could add greatly to
the knowledge of the needs of the population and be very
useful in estimating the future demand for outdoor recrea-

tion.

BIBLIOGRAPHY

BIBLIOGRAPHY

BOOKS

 

Adcock, C. J. Factorial Analysis for Non-Mathematicians.
Melbourne: University Press, 1954.

 

Clawson, Marion and Knetsch, J. L. Economics of Outdoor
Recreation. Baltimore: John Hopkins Press, 1966.

 

 

Draper, M. R. and Smith, H. Applied Regression Analysis.
New York: John Wiley & Sons, Inc., 1966.

 

Ferguson, G. A. Statistical Analysis in Psychology and
Education. New York: McGraw-Hill Book Company, Inc.,
1959.

 

 

Neumeyer, Martin H. and Neumeyer, Esther S. Leisure and
Recreation. New York: Ronald, 1958.

 

 

Nie, Norman H., Bent, D. H. and Hull, C. H. Statistical
Package for the Social Sciences. New York: McGraw-
Hill Book Company, 1970.

 

Richardson, Harry W. Regional Economics. New York:
Praeger Publishers, 1969.

 

Samuelson, Paul. Economics: An Introductory Analysis.
New York: McGraw-Hill Book Company, 1967.

 

Thurstone, L. L. Multiple-Factor Analysis. Chicago:
University of Chicago Press, 1940.

 

PERIODICALS

 

Crampon, L. J. "A New Technique to Analyze Tourist Market."
Journal of Marketing, XXX (April, 1966).

 

Knetsch, J. L. "Outdoor Recreation Demands and Benefits."
Land Economics, XXXIX, No. IV, (1963).

 

Merewitz, L. "Recreation Benefits of Water Resource Develop-
ment." Water Resources Research, II, No. 4, (1966).

 

162

 

163

Stewart, J. Q. "Empirical Rules Concerning the Distribution
and Equilibrium of Population." The Geographical
Review, No. 37, (1947).

 

Schwartz, Ronald D. "Operational Techniques of a Factor
Analysis Model." The American Statistician,
(October, 1971).

 

Shafer, Elwood L., gt 31. "Natural Landscape Preferences:
A Predictive Model." Journal of Leisure Research, 1,
No. 1, (Winter, 1969).

 

Smith, R. J. and Kavanagh, N. J. "The Measurement of
Benefits of Trout Fishing: Preliminary Results of a
Study at Grafham Water, Great Ouse Water Authority,
Huntingdonshire." Journal of Leisure Research, 1,
No. 4, (Autumn, 1969).

Trice, A. H. and Wood, S. E. "Measurement of Recreation
Benefits." Lands Economics, XXXIV, No. 3, (1958).

 

Wennergren, B. E. and Neilson, D. B. "Probability Estimates
of Recreation Demands." Journal of Leisure Research,
(1970).

Wolfe, R. I. "Perspective on Outdoor Recreation: A
Bibliographical Survey." The Geographical Review,
(1968).

 

PUBLIC DOCUMENTS AND REPORTS

 

Auger, Jacques. Dossiergpréliminaire des Programmes.
Québec: Service de la Recherche, Québec Ministere du
Tourisme, de la Chasse et de la Péche, 1972.

 

Canada Department of Forestry and Rural Development. Field
Manual: Land Capability Classification for Outdoor
Recreation. Ottawa: Canada Land Inventory, Canada
Department of Forestry and Rural Development, 1967.

 

Cheung, H. K. A Day-Use Park Visitation Model. Working
Progress Report on the Canadian Outdoor Recreation
Demand Study. Ottawa: Research Division, Canada
Department of Indian Affairs and Northern Development,
1970.

Cicchitti, C. J., Seneca, Joseph J. and Davidson, Paul.
The Demand and Supp1y_of Outdoor Recreation. Bureau
of Economic Research. New Brunswick: Rutgers, The
State University, 1969.

 

 

164

Coomber, Nicholas H. and Biswas, Asit K. Evaluation of
Environmental Intangibles. Ottawa: Ecological Systems
Branch, Canada Department of Environment, 1972.

 

Ellis, J. B. A System Model for ReCreational Travel in
Ontario: A Progress Report. Report No. R.R. 126.
Toronto: Ontario Department of Highway, 1967.

Gillespie, G. A. and Brewer, Durward. An Econometric Model
for Predicting Water Outdoor Recreation Demand.
Missouri: Agricultural Experiment Station, Dept. of
Agricultural Economics, University of Missouri, 1969.

 

La Page, W. F. The Role of Fees in Campers' Decisions.
Research Paper NE-118. Forest Service, U. S. Depart-
ment of Agriculture, 1968.

Little, Arthur D. A Study_of Factors Affecting Pleasure
Travel to U.S. Washington, D.C.: U.S. Travel Bureau,
Department of Commerce, 1967.

Ontario Department of Highway. A System Model for
Recreational Travel in Ontario. Report No. R.R. 148.
Toronto: Ontario Department of Highway, 1967.

 

Ontario Department of Tourism and Information. Tourism and
Recreation in Ontario. Toronto: Ontario Department
of Tourism and Information, 1970.

 

Prewit, R. A. The Economics of Public Recreation: An
Economic Study of the Monetary Evaluation of Recreation
in the National Park. Washington, D.C.: National Park
Service, U.S. Department of the Interior, 1947.

Québec, Ministere du Tourisme, de la Chasse et de la Péche.
Essai sur les Usagers des Terrains de Camping
Provinciaux du Quebec. Québec: Service de la
Recherche, Québec Ministére du Tourisme, de la Chasse
et de la Péche, 1971.

Racine, J. B. Modeles Graphigues et Modéles Mathématiques;
Ottawa, Ontario: Department de Géographie, Universite
d'Ottawa, 1968.

Scotte, C. C. Criteria for Evaluating the_guality of Water
Based Recreation Facilities. Report No. 1. Raleigh:
Water Resources Research Institute, North Carolina
State University, 1967.

U. S. Outdoor Recreation Resource Review Commission. Out-
door Recreation for America. Washington, D.C.:
Government Printing Office, 1962.

-41

 

165

U. S. Outdoor Recreation Resource Review Commission.
Potential New Sites for Outdoor Recreation in the
Northeast. Report No. 8, Washington, D.C.: Government
Printing Office, 1962.

Wenger, D. Wiley Jr. and Videback, Richard. Pupillary
Resource as a Measure of Aesthetic Recreation to
Forest Services. Report No. 1, Project K, Washington,
D.C.: U. S. Department of Agriculture, 1968.

 

 

Wennergren, Boyd E. Value of Water for Boating Recreation.
Bulletin 453: Utah Agricultural Experiment State,
Utah State University, 1965.

OTHER SOURCES

 

"Annuaire Marcotte de Quebec." R. L. Polk and Co.
Ltd., Quebec, 1971.

Chappelle, Daniel E. "A Resource Economist Looks at
Recreation Research." Departments of Resource
Development and Forestry, Michigan State University,
East Lansing, 1971. (Mimeographed)

Clawson, M. "Method of Measuring the Demand and Value of
Outdoor Recreation." Resource for the Future, Reprint
No. 10, 1959.

De Vries, L. "An Approach to the Recreational Capabilities
of Shoreland by Photo Interpretive Method.” Lockwood
Survey Corporation Limited, Toronto, 1966.

McCoy, E. W. "Analysis of the Utilization of Outdoor
Recreation in Tennessee." Unpublished Ph.D. disserta-
tion, University of Tennessee, Dept. of Agricultural
Economics, 1966.

Renoux, Maurice. "Techniques Econométriques de Prévision de
la Demande Touristique et Amorce de leurs Intégrations
dans un Systeme Décisionnel." Unpublished doctoral
dissertation, Université Aix-Marseille, France, 1972.

Van Doren, C. S. "An Interaction Travel Model for Pro-
jecting Attendance of Campers at Michigan State Parks:
A Study in Recreational Geography." Unpublished Ph.D.
dissertation, Michigan State University, 1967.

 

APPENDIX A
FACTOR DEFINED BY THE FACTOR ANALYSIS
METHOD, THEIR VALUES, COMMUNALITIES AND

FACTOR SCORE COEFFICIENTS

 

166

 

 

 

 

 

 

 

 

 

 

 

 

 

wo<a

.m~\Nc\N¢ u when zouhdwzuv

~b\NO\No

mt<zoz wJHL

w44m~10hu<m wm>442< mJJwIDOZ

nmvnn.eu cmnen.e mo~m~.ou o~cm¢>
11 ..1111.111.-11 111111 11-.1. 1 1- - ,;mwmmmucw11 animu.eu mo:¢Mnmn11:1;2;1mwnamp11
nnc-.o1 eceo..e mmeem.eu emoa4>
snow..ou eonau.e nckem.e1 cweaa>
naemk.eu nn:_v.e1 :Ecmc.o1 m~o¢<>
ckc-.e1 :m:a..:1 memc.o1 ewaa<>
o-h0.o1 Emmso.cu mdnom.o1 nmom<>
1 .1111 ,11::1..1 --1-1.11 11--.:111wwwwwwm11.-1mccnn.e 5e3-.ou . -o¢<>
Nos~m.e csmcm.c Enema.ca -e¢<>
emsme.e1 o-_n.c1 mmemm.e omem<>
cemm~.e reuse.e em¢-.o1 ouea<>
emnmm.e nonsn.e cmenc.eu o_om¢>
mmsmm.e ~nm~n.= mmeom.o1 esoa<>
1 1-- 11111 1 11 -11 -11,.- :11 -,1::ecammub1. osmow.a _m¢nm.eo odom<> 11
~o~ce.e mcec:.c n_m~e.eo 3_oa<>
~3n~e.c eersn.e coocm.e1 nsem<>
Emmeo.c1 ncmoe.e cmxmn.o -oa<>
m:m.~.e ¢_mm~.c somo~.e1 -em<>
ocsc~.o1 ~m~¢~.o meomm.e oqom<>
1 2111:1;11111 111- .1. -1- -c¢~.e «cone.e omom~.e ooem<>
111. ............ .1 nmdowaen coccneo mmwmm.e -111. 1uooa<>
eeE¢~.en mmmnc.e cmmmm.e sooma>
onkcn.ou mesnm.= nmemc.e oooa<>
emcee.ea ~c-:.e nsmmn.e mocm¢>
n.1nn.e1 ocmmc.o ensem.e eeom¢>
.111 .-111-1 -1 :e1w,.r-mw¢u1.1111rceoeu zohoqu - .1
mzosecxues the: xaeoeu seisozHcc oz_w: xsmpcx «chock .\
/

 

1657

.euees , . anew
o.o~ ~.-
c.5m m.em

won 230 m4) no boa

 

~nonn.o omoa<>

mensm.c amoz<>

vvmo~.c -c1¢>

~:-m.o owca<>

- 1 1- -1 .1--;-....m1s...... 82.;

~a~mo.c «Nom<>

nonmo.o n~o¢<>

:cnmo.o -oa<>

:moon.o -oa¢>

mmnr~.o owoa<>

i: ,- -11- rohomwo 1-1;. o~oz<>

omwnm.o odoz<>

wcaqo.o -ca¢>

nmcor.o c_cm<>

aowmm.c c~a¢<>

mmmcm.o n~cz<>

1 ; - ,;1:1--1 11-1.1-1Wmmmmne11- Niece)

Nrnm~.o -cz<>

_-_e.= odom<>

Nanmc.e acou¢>

roomm.c moom<>

ncmon.o nooe<>

awwnn.n . 1 . m- . :111 1. ,,1. ~m~oo.o, oocm<>
an-c.n m . --1.r¢w~ndd. ; 1 mcomm>
encam.a L -m¢m.o coaa<>
m34<>zuon echoed tbuoczoxxou wum<~m<>

 

 

wo<n

NN\NO\NO

.N~\NO\N= u uhqa ZO—hdwaue wtdzoz anu
LJJw—zapucu ww»;<2< uddwzboz

'y.-._/\

 

1158

I .’I 0.1 it.

.‘II'I 1". 1".

‘6'

 

 

 

 

 

 

 

 

 

 

 

N~\No\mo

 

m441_1ohu<u wm>4dz< wdqmtnoz

cemen.e aemne.e coec~.e omea<>
1 - 1 1 1- 1 mummw.cu cmkom.e ennee.e mmem<> 11
omcm~.ou «316m.e eenco.e p~e¢e>
emhoe.en emsem.o en~m~.c omem¢>
o~o-.c1 mmwmo.e Naoce.o m~e¢<>
ooom~.eu memco.c nccce.e cmea<>
cmeme.eu nonmm.e mmoce.e n~ea<>
1111-111111-.1111. cmee~.e1 .mmem.e nose~.e -em<>
i -n-.eu meom~.e eenkm.e -e¢<>
mc~e~.e s_o_¢.c1 emsme.e1 c~o¢<>
awncn.c1 Ecnc_.e em~e6.e asea<>
ooooc.c1 neces.e o_em~.e oseaa>
eneac.e1 .mnse.e1 ~mems.e esem¢>
11 11 11 - 1111 111 1 1 1 11 - mmmenmw1.11~wom~.e111 comemwm11 .11- ouem<> 1
eeccm.o 3_:ce.ou nmcom.e «deaem111.1.
mswce.e nsmme.e cmece.e n_ea<>
eencm.e eecc..e1 ecsoe.e Ndea<>
ocmne.e nseae.eu ammkm.e “sea<>
smsnm.c men...eu mermm.c1 esca<>
11. ..... 11 1 1 - 111 1 1 1 1 - mwmmw.e an3nc.eu mmene.eu oecuc>
neeecaq-;. nonnm.e. nzkwMum11- 1 -..meommM11 1:1
neecm.e nonn~.e1 ca~m_.e1 aeom<>
~n_eo.e arse~.e1 sem-.e1 ceeme>
oemmm.e amec~.eu 31660.61 meea<>
emeec.e scene.e1 mane~.e1 cecc<>
||-I1- 1 111 -41 Il-l111 1 1 1 1-1-1 mlgpoeu m .rohfl- - w. 1-101314... 11111..
1- xixeaz aceoeu smeeeox x<z_a<> 1\
/
.mexuexme u LPea ecsecwzo. wzezoz Luau

 

 

 

 

 

169

 

 

 

 

 

 

 

nmosc.e eemsm.e oecec.e o~em<>
mecm~._1 mmmnn._ semcm.e o~em<>
scenn.~1 cemem.e .ccce._ emem<>
me~m6.e1 ocese.c ec¢-.~ owea<>
cm-~.~1 Nscsn.~ ~3nnn.. mmea¢>
oc-~.~1 nesem.~ ~_-~._ e~ea<>
¢e_.~.ea nc_co.e ocmmm.. nmea<>
-11 emnmm.~1 emkmm.e mn~c~.n -ca<>
c~m~¢.~1 nccmc.e eccmn.m -e¢¢>
secsn.m mocca.eu mmee~.~1 e~e¢¢>
~mne3.n1 noeen.e ~mec~.m odeac>
mm_-.e1 nok~_.ec eccm_.n o_ea¢>
.-e_.~1 ace_~.eu eenom._ e_ea<>
11111mmmwwnww mnecm.e cmmmm._ odea<>
mencc.e mkmc¢.e1 ezcnm.e c_em<>
cm~o~.e1 ankne.e acc-._ nsca<>
meonm.e nmmmm.e1 secnm.e Nsem<>
cmcc~.e seeme.e emmcn.~ ~_eae>
memme.m smee~.e1 cmeme.e1 ezeae>
1-1126n.1 aieca.~1 eesce.ou oeem<>
«knee.e m_nom.m11 cmmmo.eu .111mmqmmm111
mmcmc.~ n_nce.eu em-~.e1 eeeac>
samme.e ccein.e1 mmmcc.c1 oeea<>
edene.~ Boner.e1 oeeme.e meea<>
nnmce.~ mmmee.~1 mmmnm.en eeem<>
11IEI111111 w. echoed - ~-;xeeuim1 .deHu<m11i:1111111111;
1.1 223211.88 wcoom "ESE 11A
.mkxwoxm: u ups: z:_hqhxu. w1<zoz m4~u

w44y_13hu«u mm>4<z< MJJMZDOZ

\
:l|\

«u uo<a ~h\NO\No

 

 

APPENDIX B
PARTICIPATION EXPECTED AT THE THREE

CAMPGROUNDS STUDIED - YEARS 1975 AND 1980

2170

MARE DU SAULT

 

 

conre 1975
ABITIBI “— 225397"
ARGENIEUIL 1.2968
ARTHABASKA 17.5212
77’783667"" " ’ ‘3?4755"
BEAUCE 25.2300
BEAUHARNOIS 5.3145
”‘mé'éfii‘é'éfii‘S'sff" ”"' " 3“.?9'5'5"“"""
BERTHIER 2.9235
BONAVENIURE 0.6141
"”Efifii' "“*"""*” ’m6w0‘
CHAMBLY 62.1009
CHAMPLAIN 39.4691
'""ER£8LEVbix"E' — ’ ’ 2.9933
CHARLEVOIS 0 2.8912
CHATEAUGUAY 8.1031
""611'1‘6‘6011111 _.-. _ "233.9826
COMPTON 1.0196
oeux-MONT 3.9618
'"“MDOREHESTER"" " "——“""'”—”' "'""9.8468‘ '
DRUHMONO 17.2727
FRONIENAC 2.4105
7“”élébf'ésf"" "' ’7'"“ "”0.44f7
GASPE 00551 0.4125
GATINEAU 2.6430
"""’fiUEE‘”””"""—""'” ”” “" '10.1688'””"
HUNIINGDON 0.5365
IBERVILLE’ 1.9000
”"""'.101.'i€'f'f€““ _ ' " ’ " 6.2739
KAMOURASKA 3.3094

1980
2.4638
1.1951

18.1382
3.5552
24.7463
5.2693
2.8913
2.7284
0.5913
0.5955
70.9536
37.2710
2.8679
2.6461
9.4428
228.1058
0.8420
4.3366
8.8951
17.7814
2.0791
0.4068
0.3131
2.6537
11.0623
0.5287
1.9409
6.4473

2.0655

171

 

 

 

 

 

 

 

LABELLE 1.2574 1.2482
"“‘LA’E’S’T‘JEKNE—"w' " -2.-- “"502‘1'0’98" 49.1301
LAC 51 JEANO 32.8789 8.6668
LAPRAIRIE 12.5525 26.0427
‘"”2C”15§6887168"””"'-' "'_"'""“”m”1277287" " 14.3690
LEVIS 76.2828 81.0405

L ISLET 3.7019 3.4245
——'"E0?BIEYE§E"”’ ”"“2“""’"‘ 9?6921”‘”’ 8.5115
MASKINONGE 5.1441 4.9784
MATANE 0.7862 0.6383
MATAPEDIA ’““’ “"”‘"0:5277"2—" 0.3733
MEGANTIC 19.3350 18.1788
MISSISQUOI 2.4930 2.5868
”""”fi§fif€NDfi—‘"’ "” ’"”‘ ' ” ’“1:3677 1.4245
MONTMAGNY 2.5122 2.4370
MONTMORENCY 15.4513 14.8264
’"’fi§fif§€4£"‘ " ”"‘”'”” "”'172?4875"' 187.4391
NAPIERVILLE 1.6680 1.6525
NICOLET 3.6853 3.4580

""" TDKfi'fNEAU " ’—" _'" "M ”679560 " 0.8707
PONTIAC 0.2696 0.2555
PORTNEUF. 40.0647 38.8244

”' QUEBEC "’ 749‘72341‘ ‘”“‘ “’ 1614.7637
RICHELIEU 9.5121 10.2592
RICHMOND 4.8684 4.4799
RIMGUSKT’ ““‘“““3:9483 “ '* ‘ ' 3.7356
RIVIERE 08 L 2.9920 3.0303
ROUVILLE 3.8542 1.5153
SAGUENAY 21.1988““‘””*"“‘ 24.3196

57 HYACINTHE 7.7165 21.0636
SAINT- JEAN ' 4.6645 4.7932

51 MAURICE 34.0985 ”'3321371
snerrono 11.2865 “11.6913

1172

 

 

 

sneaunooxe 18.0361
SOULANGE ’ 0.4715 ’"
STANSIEAD 2.9347
TEMISCAMING 0.9088
’” TEMISCOUATA "‘”““"’" "”0T9799”“"'

TERREdONNE 17.6157
VAUDREUIL 3.5322

"‘TEECFER’E’S"‘ M“ _ ”"2"" " 514—323” " M
WOLFE 1.5732
YAMASKA 1.1242

 

19.4069
'0.4761
2.8726
0.8532
0.6971
19.3407
3.8687
5.9340
1.2813

1.0042

 

 

173

LAC LA VIEILLE

 

 

 

COMTE 1975
4811181 "7 "WWW" 'w'7"8.1393—""m
ARGENTEUIL 6.6734
ARTHAUASKA 1.4157
"—‘8'480’1'm77 7“ ’ h” ”7 ”“7" “6:93—74 ' 7
854005 1.8865
UEAUHARNOIS 9.3147
’“E‘EFL'E‘EFXS’SE'M __ ""'"’”"" "6’.‘2’456” W
dERTHIER 1.2996
80N4v5~1085 0.2411
"””88085 """" ”' ‘ "' 7' '"0.3476
CHAMBLY 36.7917
CHAMPLAIN 13.9716
' 684815v01x 5 ' “0.3681
CHARLtVUIS 0 0.4122
CHATEAUGUAY 12.9714
0810001181 “ ' "4:2219‘ "‘“ '
COMPTON 0.7891
0504—8081 3.2051
”'"6086855158”"_"" ’"””"”“"’ ”6t9112”“‘””
DRUMMOND 2.0298
FRONTENAC 0.8361
”“‘”'0‘4585' 55‘1““ 2“ " ““'” "'0";‘8’7'35' ‘
GASPE 00551 0.2352
GATINEAU 48.4184
HULL ‘_ 49.3410
HUNTINGDON 1.5424
18ERVILLE 2.0883
JOLIETTE 2.5650
KAMOURASKA 1.1719

-_..

1980

75.7713

6.1435
1.4662
0.9591
1.8499
9.2344
0.2026
1.2118
0.2320
0.3392

42.1109

13.1834

0.3525
0.3768

15.1463

4.1144
0.6499
3.5126
0.8220
2.0903
0.7197
0.8037
0.1778

48.6157
53.7369

1.5199
2.1339

-*~2.6368

0 o 7869

174-

 

 

 

 

 

 

 

 

 

 

 

 

 

L485LL5 48.2610 47.9034
”’_’—C40'5T_35485 0.3147“”””‘"”"""””" 0.3084
L40 51 JEANO 0.5369 0.1390
LAPRAIRIE 12.0389 25.2201
L 4550881108 "‘”‘ "3.9576 “"' “'” "' ‘ 4.4747
L591s 3.6813 3.9140
L 15L51 0.6901 0.6377
L018181585"' "'" ““0T6236'"’”" '" ” ’ 0.5467
HASKINONGE 1.0969 1.0611
841485 0.6352 0.5143
"“”“8X1335514”'m"‘ ‘ 0.2403"””“”"" "'" 0.1712
85648110 3.2728 3.0746
8155150001 2.5660 2.6638
'“‘"’8081E4L8””‘ ”’ '"“ '"“i?0581"""_"’ 1.0265
MONTMAGNY 0.7410 0.7185
MONTMORENCY 3.4894 3.3465
~'“"8681715711.“ "””‘”"‘“" "‘”2§372663‘”’”""" 308.0994
NAPIERVILLE 1.2421 1.2304
8100L51 1.0263 0.9622
_'"‘B'4‘8i'8'54"0"'"‘“_—”’ ‘ " "'73442’1'” "' W 6.7719
PONTIAC 6.5371 6.1900
PORTNEUF 1.2592 1.2197
"”’"005850”"“” '““’55?Z902"””"’"“"' 38.4194
81085L150 4.3556 4.7024
81088080 1.2974 1.1925
“'"”81800581 '””‘ ' 3.0995 ” "”’”"" ”" 2.9304
RIVIERE 00 L 4.4625 4.5204
ROUVILLE 3.1991 1.2422
SAGUENAY ‘”'41.6956 ‘ ” 47.9199
51 HYACINTHE 3.2741 9.0572
54181- JEAN 4.8823 ’ 5.0189
51 8408105 19.1662 '”“'"”””'“”' 18.6189

SHEFFORO 2.5494 2.6420

 

175

 

 

 

 

5858880085 17.2310 18.5591
500L4805"“ 0.9862 ’"" ' 0.9961
514851540 0.8614 0.8430
15815048180 26.5462 24.9022
—_"—15815000414 0.6845 3‘ 0.4848
1588580885 61.2200 67.2976
VAUDREUIL 4.2616 4.6732
958085855“"7_ 3.1717 "" 7' 3.4687
80L55 0.5200 0.4224
YAMASKA 0.5737 0.5116

-—-_-—_—.. .. .. _..__...._._.._-

 

 

176

MONT ORFORD

 

 

 

 

 

 

 

 

 

 

00815 1975
4811181 "‘ "‘__ 72.23of_"""_‘
480581501L 9.6202
ARTHAdASKA 62.8347
84001 ' “‘7 ‘777"_"72E094Z” '
854005 53.8886
85408488015 40.3625
85LL5084§557 7'"'“"“3?2404 "
85818158 7.0133
80841581085 3.2603
”"7386?" ' ' *2 ' _ _. ’7‘41763‘1’7' ’
08488L1 709.3501
08488L418 78.4901
“ERX§[51018 5 _”"——' 2.0942
CHARLEVOIS 0 1.5758
08415400041 62.1625
'" 0810607181 “”"""’”"""“‘””" 5315873
0088108 21.1970
DEUX-MONT 20.6602
”WEEKS???— " " ‘ “7 _ .2--- 15.3060" ' ’
08088080 82.5359
FRONTENAC 13.4861
04585 551“”"‘”"‘"' "Zfilé7i” '“
04585 00551 0.6407
04118540 21.8487
80LL ”'“72:0978"‘”"
8081180008 6.1248
185811LL5 22.7204
‘“" JOLIETTE 23.9312
KAMOURASKA 9.6041

1980
68.9092
8.4767
66.2952
16.6690
52.2978
39.8316
2.4148
6.3014
3.0750

39.5352

872.0771

71.8197
1.9599
1.3736

78.7937

51.5152

15.7563

23.7673

11.3667

86.3324

10.7235
3.6683
0.4178

21.9850

82.1504
5.9883

23.4835

24.9640

4.6257

177

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L485LL5 5.6329 5.5692
L40 57 JEANE “ 4.7944 7' ‘ 77—." ’ 4.6499
L40 51 JEANO 7.3308 0.9285
L48841815 79.2457 245.5651
L 4550881108 14.4194 ‘”"‘“ "7"""" 17.3987
LEVIS 13.9751 15.3489
L 15L51 4.6239 4.0981
L018181585 "”"_""“_"_”'5.8265 ' ' " 4.7643
MASKINONGE 5.0949 4.8427
841485 2.4088 1.7442
2 841485014""' ' " ’“”""‘” ’""1:0§§f ’“ 0.6449
85048110 56.0718 50.9622
8155150001 55.9840 59.2790
"‘"‘80870358‘""'“ "‘””””"'“““”‘“577“7‘“” 5.4680
808184081 5.6885 5.4264
80818085801 3.6989 3.4698
‘_“8087854[ ” ""”"’ ‘”“499€;7055“'”” " ” 5683.1602
NAPIERVILLE 4.6923 4.6249
8100L51 12.6085 11.4240
‘""82818540 "‘ "”"‘“‘”‘“” “"’"923137 ”“"”““"” ' 8.0586
PONTIAC 4.3760 4.0256
PORTNEUF‘ 15.6231 14.8804
005850 ““‘“"“ ”575.7656‘”““‘”“m“ 650.0137
81085L150 47.8780 53.8290
81088080 90.7090 79.7381
81800581' “ ““ 9.8052 ‘“‘ ‘“‘ ' 8.9990
81v1585 00 L 12.9004 13.1574
80011LL5 48.2267 11.3492
54005841 '"““ 137.1591 """””““”' ' 169.6845
51 814018185 56.5552 268.1169
54181- JEAN 72.6214 75.7516
51 8408105 185.2599 w"177.2305
58555080 262.2612 276.9766

SHERUROOKE

1178

1939.9924

 

 

-_ “—500L4805
514851540

TEMISCAMING

".775“ 815000—4— 771" 7 _- _.

TERREBONNE

VAUDREUIL

7" 958685855 ___1-

WOLFE

YAMASKA

__ -.-_———-.—

90.1338
17.2560

" "52385 '1'

192.8994

24.8829

— 37.25405

7.2321
5.2592

3:10 62 ’ ’

2173.2976
3.1539
87.1969
15.6486
3.0591
222.9463
28.6511
43.0475
5.2615

4.4149

APPENDIX C

QUESTIONNAIRE USED IN THE STE-FOY STUDY

179

REPONDRE PM UN (XI AUX QUESTIONS SUNANTE'S.
4) Donoquollo cougarlooooltuol‘ duchof
MINISTERS DU TOURISME, dolomlllo? m

05 LA CHASSE ET DE LA PECHE

19 am on moino

 

 

 

20 am I 29 ono
30 am I 39 ono
40 am I 49 on:
60 om I 69 mo
00 mo I 64 ono
05 ans ot plus
LES BESOINS DES 0” EBECO'S 5) Dono quollo cougorlo oo oltuo lo nlvoou d'lno-
SONT-ILS SAT'SFA'TS EN tructlon du chof do fomlllo?
MATIERE DE CAM PING 7
Etudoo prima'iroo 1
(us non-cmrwns oowmf AUSSI titration: A c: QUESTIONNAIRE) Etudes oocondoiroo ponlolloo 2
Etudoo oocondoiroo complflm 3
Etudeo collogioloo 4
Etudes lochniquoo 5
Etudoo unlvonitoitoo pon'iolloo a
etudoo unlvoto'itoiioo com 7
Stud“ pool-arm 3
FA¢ON DE IEPONDIE AU QUESTIONNAIRE: Autro 9
. N. ionolo do“. 910! d‘uno “9°". p. quootion. 0) Dom quollo cougorlo oo oltuo lo tovonu on~
o No lomolo Iain dono Io cooo contonont lo chiflro. nuol do lo tomlllo?
Molno do 82.999
Do 3.000 I 6.999
Do 0.000 I 7.999
Do 8.000 I 9.999
Do 10,000 I 14.999
16.000 in plus
1) Don quollo cotogorlo oo oltuo l‘ooeupotlon
1) as vouo ovoz an du complng our... rm do I I I do ch09 do Mum-7
1911. Indlquoo-on lo nombn do nulto.
1 Employo dono lo troitmm do lo motiIto
SI VOUS N’AVEZ PAS FAIT DE CAMPING. ALLEI mm (minoul. moduli chlmiquo. tonllo. one.) 0
A LA QUESNON " Employ‘ dono uno uoino do piodulto linlo (mononicion. ole.) 1
2) 8| mo om mllloo an tortoln do camping od- 5W N oonrieo 6” public (DOW. human. “'1 2
mlnlotto por lo gowomomom du Outboo. Tuvoll do bunou (mm. ogom do bunou. ole.) 3
2:23;“ " “7" “Wu“ " Mb“ I I I Pmlooolonnol (modocln. ovocol. wolooooor. m.) 4
Si vous N‘AVEZ PAS urn/st DE mmm as 3 m (”mm "mm' ”mm "c" 5
CAMPING DU aouvmvmavr. 411.52 A LA auzs- W m 9 mm 5
"ON 4- ' 096mm: do mochlno (woo. comm. typogtoplio, otc.) 7
a) on vooo .m utmu l‘un «1.. unolno o. oom- 6mm a
plug oulvonto. lndlquoo lo nombro do nulu Am 9
9...... dono diocun d‘oux.
I) Indlquoz lo nombro d'onfonto I chomo.
.) Comping provincial a. Stonohom I Stonohom. L_L_J (2‘
5 Aucun who! 1
b) Complng provlnciol llo d'OrlIono I St-Joon. Un onlom 2
u. d'Ofloom. Li_l__l a.” mm, 3
c) ComplngptovlnclolLIMotoduSoundomlo I | I “mm” 4
pore doo Louromldoo. Ououo onlomo 5
9 Clnq onlonto e
4!) Camping provincial Vinconnoo o Booumont. Lfi_.|___l Six «mm. 7
o) Comping provinclol Montmoony I Montmagny. L . 1.] Soul onlonto 3
13 Hui: onlonto ot plus 9
I) Con-plug provinciol Villonowo I Villonouvo. lfi'l—J

(VOIR AU VEISO)

'.0

 

APPENDIX D

LIST OF PROVINCIAL CAMPGROUNDS

l.
2.
3.
4.
5.
6.
7.
8.
9.
10.
ll.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.

180

Lac La Vieille
Mont Orford

La Mare du Sault
Lac Dozois

Les Pins Rouges
Lac Lajoie

La Sablonniere
Lac Monroe

Paul Sauvé
Voltigeur

Lac Normand

Lac Bellevue
Stoneham

La Loutre

Belle Riviere
Val-Jalbert

Lac du Milieu

Lac D'Argenson
Riviére Chalifour
Lac Albanel

Lac Arthabaska
Lac Rimouski
Riviére Matane
La Barriere Johan
Etang de la Truite
Port Daniel
Moisie