m...”

.—

3137: - L...

 

       
   

1 A v x (I ‘2‘ : .
”,":(1,I" 11323.? 1, . ,
;£‘i [‘1 J”.:' ' .1

gun-i

   

    

    
 
     
      
     

Fi‘; ll“ ..‘

11*
11‘. . u

  

"'1’-

luuwl w... i g

 

 
 

"3.15:.

  
 
 
 
 

2;: a1:
1131,.- .1 ;

as

,-.u- n

- .4
’W-v-\

    
 

  
 
 

“w.

a "a...

“r, :.

  
 

ins“ "

    
  
  

'17.

.3.

“WT?“ "

.,~’<<.

  
 
 

 

         
     
    

 
 

 
    

      
     

     
   

  
  
 
    
 
   
 

 

  
 
   

   

1?? ' ‘. ,
'5'” 3 I. in:
5” "‘11" 11- 1,1
.11}; “I"Loa“ :1”, If 5.31:“ ’11,! :.
‘1' 13:14 :; '- 1111111111
2:11.111 11 1.1-: 21:31- 1.11. 1
‘ : ';:.:": A”; I» " I. h. A “31.11310. -1‘
1% 11111331131"; *3“ "311- 111:1“ 11.31111111311 1111111112 1. 1"? “f; 431.
1 1- I" 1‘ -. ‘ .1) I '1" 5‘ 1 1;
11111. 1 . 1 11 111, 1111 1,1 11111
1’. g ' f0 1:, 1.1.11 1" 11"“ . n .;
1 ‘ ,1 11-1 311 1121 3'11 1- » ~ 1"
11:11 '1 a
11.1”} m:- .1 ',:
' g1. ‘ r |
‘11"11’ 1111’!" 1 1
' “41!: ' 1111‘“
“31:11:!”
.11 , 1

 

$1 EX:
1 I

.1133."

'3

 

ms IlllllllllHllllllllllHllllllllllllU’lthlllllHIIIHIUI

1293 02058 6479

This is to certify that the

dissertation entitled
Trip Expenditures of Recreational Boaters in Michigan
presented by

Hee Chan Lee

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Park, Recreation and
Tourism Resources

 

  

Ma' rpr essor

   

Date 3W4, M; /777

MSU is an Affirmative Action/Equal Opportunity Institution 0- 12771

 

 

 

LIBRARY
Michigan State
University

 

 

PLACE lN RETURN BOX to remove this checkout from your record.
TO AVOID FINE return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

 

 

 

DATE DUE DATE DUE DATE DUE
., ~ ‘r' ”~ 0 1.. ‘3
3W0 10;: W 0631: 21. $008
[JAR 1 1:23;

 

 

 

 

 

 

 

 

 

 

 

11/00 c/CRCJDmOmpfis-pjd

 

TRIP EXPENDITURES OF RECREATIONAL BOATERS
IN MICHIGAN

By

Hee Chan Lee

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

DOCTOR OF PHILOSOPHY

Department of Park, Recreation and Tourism Resources

1999

ABSTRACT
TRIP EXPENDITURES OF RECREATIONAL BOATERS IN MICHIGAN
By

Hee Chan Lee

This study estimated trip expenditures of registered pleasure boat owners in
Michigan in 1998, tested for differences in measures of boating activity and trip
spending across different storage segments, and estimated the regional flows of boater
spending on trips by storage type. Data were collected through sample surveys of
Michigan registered boat owners. An in-season survey was utilized to estimate
boating activity and trip spending. Surveys were sent out in nine waves every week
over the 1998 summer. At the end of the boating season, two distinct end-of-season
surveys were conducted with smaller samples to evaluate potential nonresponse bias
and use estimates from the in-season survey. The study developed methods to
estimate both annual use and spending in a single survey. Annual use per boat was
estimated by applying a logistic model to the average days of use obtained from the in-
season survey. The average use estimated from the wave surveys was not statistically
different than the average use estimated from the season-end survey.

A total of 652,000 active registered boats in Michigan logged an estimated
18.4 million days of boating in 1998. Boats averaged about 28 days of use. Owners
of active registered boats spent an estimated $635 million on trips within Michigan in
1998. The total was divided $292 million on day trips and $343 million on overnight

trips. A typical boater spent $23 a day on day trips and $60 a day on

overnight trips, averaging about $35 per day overall. Boaters keeping their boats at
marinas spent $76 a day on boating trips, while at the other extreme boaters storing the
boats at waterfront primary homes spent $20 a day. The test results showed that there
were significant differences in both the levels and patterns of use and spending by
storage segment. All northern regions were net gainers from boater trip spending and
earned a net gain of $120 million in 1998. The South inland region showed the
biggest net loss of boater dollars, as resident boaters in this region spent $78 million
more outside the region than the region received. Out-of-state boaters spent $35
million in Michigan in 1998, mostly involving use of seasonal homes in the state.
Refinements were made in methods used in previous boater surveys for handling of
zeros and missing data in trip spending reports, and separating day and overnight trips.
The composition of respondents was also compared to that of nonrespondents on
characteristics that are relevant to trip spending to show no conclusive evidence of

nonresponse bias.

ACKNOWLEDGEMENTS

I would like to thank Dr. Daniel J. Stynes, my academic advisor and
dissertation chairman. He was instrumental in all facets of my research and study
while at Michigan State University. Dr. Stynes has been a tremendous help in all
stages of the project, from encouraging me to pursuing the topic, to directing the
research design and data analysis, to carefully editing many drafts of the dissertation.
Perhaps most important, he has always been willing to talk and to listen throughout the
process.

Valuable contributions were also made by committee members. I am
especially indebted to Dr. Edward M. Mahoney for many months of sincere dedication
and valuable guidance during this study. I would also like to express my sincere
appreciation to Dr. Donald F. Holecek and Dr. Larry A. Leefers for their valuable
assistance, not only in terms of dissertation, but also in other aspects of my doctoral
program as well.

I wish to thank the faculty members in the Department of Park, Recreation and
Tourism Resources for their guidance and assistance during my program at MSU.
Special thanks go to Dr. Betty van der Smissen for her help and encouragement during
my time at MSU. Appreciation is also extended to Mr. Young Rae Kim for assistance
rendered in preparation and distribution of the survey questionnaire and data coding,
as well as his unending support and friendship.

Words can not express my appreciation and gratitude for the love, sacrifice,

support, and encouragement so generously given by wife, Bock Hee. My two sons,

Seung Kon and Eui Kon, also demonstrated much patience and understanding over an
extended period of time, when they did not receive the usual amount of attention and
guidance from their father.

Finally, I would like to dedicate this dissertation to my parents, Sang Wu Lee
and Jung Sook Choi and parents in law, Sang Doo Lee and Hye Soon Kim. Without

their love and moral support, it would not have been possible for me to complete this

study.

TABLE OF CONTENTS

LIST OF TABLES ...................................................................................................... viii
LIST OF FIGURES ...................................................................................................... xi
CHAPTER 1 -
INTRODUCTION .......................................................................................................... 1
Problem Statement .................................................................................................... 4
Estimation of Boater Trip Expenditures ............................................................ 6
Stratification by Types of Boat Storage ............................................................. 8
Regional Flows of Boating Activity and Trip Spending .................................... 9
Objectives ................................................................................................................ 11
Boater Trip Spending by Storage Segment ...................................................... 11
Test for Differences in Boating Use and Spending across Storage Types ....... 11
Regional Flows of Boater Trip Spending ......................................................... 12
Organization of the Study ....................................................................................... 13
CHAPTER 2
LITERATURE REVIEW ............................................................................................. 14
Studies of Traveler Expenditures ............................................................................ 14
Studies of Boating Use and Spending ..................................................................... 22
Forecasting Recreation Use ...................................................................... . .............. 27
Spatial Patterns of Recreational Boating in Michigan ............................................ 30
CHAPTER 3
METHODS .................................................................................................................. 33
Sampling Design ..................................................................................................... 33
Measurement ........................................................................................................... 38
Variables Measured ........................................................................... . .............. 39
Handling of Zeros and Missing Data ............................................................... 41
Data Analysis .......................................................................................................... 43
Formation of Storage Segment ......................................................................... 44
Estimation of Boater Trip Spending by Storage Segment ............................... 44
Test for Measures of Variables across Storage Types ..................................... 45
Estimation of Regional Flows of Trip Spending ............................................. 46
End of Season Surveys ............................................................................................ 49
Nonrespondent Survey ..................................................................................... 50
Independent Survey .......................................................................................... 50
CHAPTER 4
RESULTS .................................................................................................................... 53
Survey Responses, Weights, and Nonresponse Bias ............................................... 53
Survey Response Rate ...................................................................................... 53
Weights ............................................................................................................ 54

vi

Test for Nonresponse Bias ............................................................................... 60

Storage Segment, Boat Owners, and Boats ............................................................. 64
Storage Segment ............................................................................................... 64
Descriptions for Boat Owners .......................................................................... 65
Boat-related Characteristics ............................................................................. 66

Statewide Boating Use and Spending (Objective 1) ............................................... 70
Estimation of Boating Use by Storage Segment .............................................. 70
Estimation of Boater Trip Spending by Storage Segment ............................... 82

Test of Differences in Use and Spending by Storage Segments (Objective 2) ....... 92
Boating Activity by Storage Segment .............................................................. 92
Boater Trip Spending by Storage Segment ...................................................... 95

Regional Distribution of Boater Trip Spending (Objective 3) ................................ 96
Spatial Distribution of Boats and Boating Use ................................................ 96
Distribution of Boater Trip Spending ............................................................ 100
Regional Net Gains of Boater Trip Spending by Storage Segment... ............ 104
Regional Net Gains of Boater Trip Spending by Spending Category ........... 106
Out-of-State Boater Spending on Trips .......................................................... 106

CHAPTER 5
CONCLUSIONS ........................................................................................................ 108

Summary of the Study ........................................................................................... 108
Estimation of Boater Trip Spending .............................................................. 108
Test for Measures of Variables by Storage Type .............................. -. ............ 109
Estimation of Regional Economic Benefits from Boater Trip Spending ....... 109
Selected Methods to Improve Spending Estimates ........................................ 110

Study Limitations .................................................................................................. 111

Recommendations for Future Research ................................................................ 1 13

Implications ........................................................................................................... l 15

BIBLIOGRAPHY ...................................................................................................... 1 19
APPENDICES

Appendix A. Questionnaire (In-Season Survey) ........................................................ 126
Appendix B. Questionnaire (Season-End Survey) ..................................................... 130
Appendix C. Measures in Boats & Trip Spending by Wave ..................................... 133
Appendix D. Sampling Errors .................................................................................... 134
Appendix E. Average Trip Spending ......................................................................... 135
Appendix E. Total Trip Spending .............................................................................. 137
Appendix G. Regional Distribution of Trip Spending by Spending Category .......... 139

vii

LIST OF TABLES

Table 1. Michigan Registered Boats by Region of Registration & Boat Size: 1998 35

Table 2. Distribution of the Sample by Region of Registration & Boat Size .............. 37
Table 3. Variables Measured in the Wave Survey ....................................................... 40
Table 4. Missing and Zero Responses to Spending Question ...................................... 42
Table 5. Survey Response Rate (Season-end Surveys) ................................. '. .............. 51

Table 6. Sample of Active Boats by Region & Boat Size (Season-end Surveys) ....... 51

Table 7. Weights for Active Registered Boats (Season-end Surveys) ......................... 52
Table 8. Survey Response Rate .................................................................................... 54
Table 9. Response Rates by Region of Registration & Boat Size ............................... 55

Table 10. Completed Sample of Boats by Region of Registration & Boat Size .......... 56

Table 11. Percentage of Registered Boats Inactive (Sample) ...................................... 57
Table 12. Registered Boats Inactive in 1998 (Population) .......................................... 57
Table 13. Active Boats by Registration Region & Boat Size (Population) .............. 58
Table 14. Sample of Active Boats by Region & Boat Size ......................................... 59
Table 15. Weights for Active Registered Boats ........................................................... 59
Table 16. Boat Ownership by Number of Boats Owned ............................................. 60
Table 17. Differences between Respondents & Nonrespondents ................. .............. 62
Table 18. Comparison of Boat Type between Sample and Population ........................ 63
Table 19. Boat Classification by Storage Facility & Location .................................... 64

Table 20. Distribution of Storage Segments for Sample and Population .................... 65

Table 21. Boat Owner Characteristics by Storage Segment: 1998 .............................. 68

viii

Table 22.

Table 23.

Table 24.

Table 25.

Table 26.

Table 27.

Table 28.

Table 29.

Table 30.

Table 31.

Table 32.

Table 33.

Table 34.

Table 35.

Table 36.

Table 37.

Table 38.

Table 39.

Table 40.

Table 41.

Table 42.

Table 43.

Table 44.

Boat Characteristics by Storage Segment: 1998 .......................................... 69
Average Use Days and 90% Confidence Intervals by Wave ........ . .............. 72
Results of Logistic Models for Annual Use Days ........................................ 73
Variation of Annual Days of Use with Ending Dates .................................. 75
Average Days of Use "so far" by Storage Segment ..................................... 76
Boating Use by Storage Segment: 1998 .............. 78
Summary of Boating Activity by Storage Segment: 1998 ........................... 79
Mean & 90% Confidence Intervals on Use Estimates ................................. 82
Average Trip Spending by Storage Segment & Type of Trip (Summary).. 84
Average Trip Spending by Storage & Spending Location (Summary) ....... 86
Total Trip Spending by Storage Segment & Type of Trip: 1998 ................ 87
Percent of Boaters Spending Nothing on Recent Trip by Trip Type ........... 90
Percent of Boaters Spending Nothing on Recent Trip by Storage Type ..... 91
Boating Activity by Storage Segment: 1998 ................................. i ............... 93
Distance Traveled and Body of Water Used by Storage Segment: 1988 94
Trip Spending by Storage Segment: 1998 ................................................... 95
Distributions of Boats and Boat Days by Region & Storage: 1998 ............. 99
Distributions of Spending by Origin & Storage ($million): 1998. ............ 101
Distributions of Spending by Destination & Storage ($million): 1998 ..... 103
Regional Net gains of Spending by Region & Storage ($million): 1998.. 105
Out-of-State Boater Spending on Trips to Michigan: 1998 ....................... 107
Measures in Boats & Trip Spending by Wave .............................. ............ 133
Sampling Errors for Trip Expenditures (90% Confidence Level) ............. 134

Table 45. Average Trip Spending by Storage Segment & Type of Trip: 1998 ......... 135
Table 46. Average Trip Spending by Storage & Spending Location: 1998 .............. 136
Table 47. Total Trip Spending by Storage Segment & Type of Trip: 1998 .............. 137
Table 48. Total Trip Spending by Storage Segment & Spending Location: 1998 138

Table 49. Regional Net Gains of Trip Spending by Spending Item ($million): 1998139

LIST OF FIGURES

Figure 1. Sampling Regions of Michigan .................................................................... 36
Figure 2. Schedule of Mailings for the Wave Surveys ................................. ‘. .............. 37
Figure 3. Spending Questions ...................................................................................... 41
Figure 4. Average Days of Use by Wave During the Summer .................................... 71
Figure 5. Logistic Estimates of Boat Days ................................................................... 74
Figure 6. Boater Trip Spending by Spending Category: 1998 ..................................... 86
Figure 7. Boats, Boating Use & Trip Spending by Storage: 1998 ............................... 89
Figure 8. Boating Regions in Michigan ....................................................................... 97
Figure 9. Percent Distribution of Trip Spending by Storage Segment: 1998 ............ 103
Figure 10. Regional Net Gains of Trip Spending by Storage: 1998 .......................... 105

xi

CHAPTER 1

INTRODUCTION

US. resident travelers spent $408 billion on transportation, lodging, meals,
entertainment and recreation services, and incidental items within the US. in 1997
(Travel Industry Association of America, 1998). The 1997 spending accounted for 11
percent of all consumer spending in the US. However, the significance of recreation and
tourism to the nation's economy is even greater than this figure represents because, for
example, leisure travel involves considerable portions of consumer spending on
transportation, housing, clothing, and food. Travel and tourism is the nation's largest
export industry, third largest retail sales industry and one of the America's largest
employer. It is in fact the first, second or third largest employer in 32 US. states (TIAA,
1998)

Estimating the expenditures on recreation and tourism within a state or region is
an important component of comprehensive recreation and tourism planning. Estimating
direct spending is a necessary first step to obtain a clearer picture of traveler's impact on
state and local economies. Economic impact studies in recreation and tourism are
undertaken to determine the effect of specific activities in a given geographic area on the
income, wealth, and employment of that area's residents (Frechtling, 1994a). Estimates
of traveler expenditures are often used to justify future developments and evaluate past
management performance, as traveler spending has significant economic impacts on

regions that attract more visitors.

While some expenditures on recreation and tourism can be estimated from
existing secondary data, others can not. Boat purchases can be estimated using
government data, but restaurant sales attributable to boaters are not available from the
secondary data. The Consumer Expenditure Survey's estimate for spending of $116
billion on gasoline and oil in 1997 (U .8. Bureau of Labor Statistics, 1998) does not
distinguish between spending on local auto transportation, on auto travel away from
home, or fuel used in boats. Consequently, to estimate consumer expenditures on
recreation and tourism requires multiple methods and sources.

Recreational boating is a major recreation and tourism activity throughout the
US. In a nationwide survey, recreational boating was the eighth most popular sports
activity, afier fishing, in 1997 (National Sporting Goods Association, 1998). The
National Marine Manufacturers Association (1998) estimated that recreational boaters in
the US. spent roughly $19.3 billion on craft-related items and boat purchases in 1997,
showing a substantial increase from $3.4 billion in 1970 and an increase of 9% over
1996. Spending on boating trips, which is generally well above the expenditures on craft-
related items or boat purchases was not included in this figure. For example, in a 1981
Michigan boating survey (Stynes et al., 1983), the one billion of total spending was
divided 66% for trip spending, 24% for craft-related spending, and 10% for boat
purchases. In a 1995 California study (Rust and Potepan, 1997), the majority (69%) of
the $2.5 billion of total spending fell into trip-related categories.

These figures represent not only the extent of participation in recreational boating
in this country but also some measure of the impact these millions of boaters have on the

economy. Because of the growth in boating and its related activities, many states are

concerned with providing public access and facilities for the boater. In many' states in the
US, recreational boating is linked to the states' economic development as boating
generates a considerable amount of economic activity (Stoll et al., 1988). Recreational
boating supports a variety of industries including boat builders, boat dealerships, marinas,
repair services, and a wide range of retail sectors of the economy (Neely et al., 1998;
Stynes et al., 1983).

Boating in Michigan is an especially important recreational activity with
substantial spending and ties to various industries. Michigan ranked first among the 50
states in the amount spent on boat purchases in 1997 with $480 million being spent
(NMMA, 1998). Michigan also led the nation in the number of registered watercraft in
1997, with 957,105 boats including pleasure and commercial craft (NMMA, 1998). The
popularity of boating in Michigan is attributed to the relatively easy access to the state's
Great Lakes waters, thousands of inland lakes, and thousands miles of rivers.~

Surveying boat owners provides boating agencies and industries with information
required to meet the needs of boaters. Boater expenditure surveys have been the key
instruments for obtaining information to estimate spending and economic impacts of
boating. A boater expenditure study provides information to assist government agencies
as well as industry associations and boating businesses who need to develop product and
marketing strategies. Information obtained from such surveys includes spending patterns
of distinct user groups, location of boater spending, and determinants of boater
expenditures. .

Boater spending also reflects a wide spectrum of sociodemographic phenomena

associated with boat owners, as a great deal of boating activity is associated with seasonal

homes, retirement in northern communities, and the like. Patterns of boater spending are
influenced by craft type and size, storage type, and boater's socioeconomic
characteristics. Estimating total boater spending requires that the number of craft and
boat days be estimated. Amount spent on trips may not be reported accurately because of
potential recall difficulties boaters may have. Estimating direct spending is necessary to
estimate boater's impact on state and local economies. Boater spending thus affords a
good vehicle for testing improvements in research methods, as well as recreation

planning and management.

Problem Statement

Spending by boaters is divided into three major categories: 1) new and used boat
purchases, 2) craft-related spending, and 3) trip-related spending (Stynes et al., 1983;
Lipton and Miller, 1995). Craft-related spending includes expenditures for equipment,
insurance, repair, storage of the boat, and other related items not directly associated with
individual boating trips. Trip-related spending includes all spending in conjunction with
boating trip, which is the variable cost of taking a trip. Trip-related spending includes
expenditures on groceries, restaurant meals, auto and boat fuel, boating and other
recreational gear purchased on trips, and other expenses incurred on boat outings.

Craft-related spending and spending on new and used boats can be estimated
using a season-end survey. Boat owners are usually asked at the end of the season to
report annual craft-related spending and spending on boats acquired in the previous year.
Spending on new and used boat purchases may be also estimated from secondary data.

Stynes et al. (1983) and Lipton and Miller (1995) estimated boater spending on new and

used boat purchases using sales taxes collected by state agencies. Because of the variety
and variability of boating trips and expenditures, however, trip-related spending is more
difficult to estimate. A year-end survey may not be appropriate to capture boater
spending accurately because of recall biases on trip expenditures. Patterns of trip
spending depend on a variety of trip patterns (e.g., day trips vs. overnight).

In a 1981 study in Michigan (Stynes et al., 1983) new and used boat purchases
accounted for ten percent of the measured total boater spending. Trip and craft-related
spending made up the remaining 90% of the total. Compared with craft-related spending,
boaters spent more than two times as much on trip-related spending in 1981. Craft-
related spending primarily directly accrues to boating industries while trip-related
spending benefits a wide range of retail sectors of the economy. Trip-related spending,
especially, has far reaching impacts on coastal communities, reaching many sectors of the
local economy through both direct and indirect effects of boater spending.

The research problem presented by this study is to estimate trip expenditures of
recreational boaters in Michigan in 1998. This is the first statewide study of boater trip
expenditures in Michigan since 1981. The 1981 boating study (Stynes et al., 1983)
provided a wealth of information about boating expenditures of Michigan registered boat
owners, but left many unanswered questions about mainly the approach used to elicit
boater expenditures. In addition to the need to update estimates of boater expenditures,
there is a need for improved and more cost effective methods to estimate boater spending.
Information on boater expenditures would provide a stronger basis for planning facilities
and services and an indication of the spending impacts of recreational boaters on local

economies.

Estimation of Boater Trip Expenditures

Boater trip spending is usually estimated through sample surveys because trip
expenditures on boating can not be readily extracted from economic accounts. Given the
dispersed nature of boating activity, it is difficult to obtain comprehensive estimates of
use from on-site surveys. For most recreational boating activities household surveys are
frequently used to estimate levels of activity, characteristics of trips, and spending.
Given the convenient sampling frame offered by boat registrations, boater studies
generally use mailed surveys where the sampling unit is the boat or boat owner.
Unregistered small boats are generally not included in these studies.

Boater trip spending estimates are based primarily on self-reported data. An
important issue in estimating traveler expenditures via survey is recall bias. Evidence
indicating the presence of recall bias in expenditure estimates by travelers is abundant in
traveler studies (Rylander et al., 1995; Stynes and Mahoney, 1989; Ellerbrock, 1981;
Mak et al., 1977). The resolution of this spending estimation issue, however,hhas not
received much attention from researchers studying boater expenditures. Generally,
average expenditure estimates provided by sample of boats and trips are multiplied by the
number of use days to yield estimates for total expenditures during the season. Errors in
either will result in errors in estimates of total spending. .

Season—end surveys have generally been used to collect both boat use and boater
spending data. The preferred approach to estimate annual use seems to be an end-of-
season survey. However, individual trip characteristics and expenditures can be
measured more accurately during or shortly after the trip. This suggests an in-season

survey to estimate trip spending (Stynes et al., 1983). Previous studies have frequently

used different surveys to estimate use vs. trip-related spending. For example, the 1981
study conducted by Stynes et al. (1983) employed a wave (survey) approach to estimate
trip spending. Surveys were sent out in six waves during the season asking the boater to
report trip spending for their most recent trip. Since annual use is not easily estimated in
the middle of a boating season, the 1981 study used a season-end survey conducted in the
previous year for days of use.

End-of-season surveys are problematic in estimating boater trip expenditures
because of potential recall error. Designs that require both in-season and season-end
surveys increase costs. To estimate both annual use and spending appropriately in a
single survey has not been addressed in boating studies. Procedures are needed to
extrapolate annual use from in-season surveys, to measure spending for a recent trip to
reduce recall errors, and to sample in waves over the summer to capture seasonal
variation in trips.

The 1981 Michigan boater expenditure survey did not clearly define what a trip
was. This is problematic for boaters at waterfront homes who boat from their backyard.
Waterfront home owners may consider a day of boating around home a "trip" or they may
report a more extended outing where they stayed overnight or boated a greater distance
from home. Boaters at seasonal waterfront homes may consider traveling from their
permanent home to the seasonal home a "trip" or they may report a day outing for boating
from the seasonal home. A speculation is that boaters would be more likely to report
overnight or extended trips that may not be their most recent trips. This could be
particularly the case when they are asked to report spending, as no "trip" spending is

necessarily involved when boating from a waterfront home.

It could not be ascertained exactly how many respondents to the 1981 Michigan
boater expenditures survey had missing spending data or reported no spending. The
report was incomplete on this point and the original data were not available. The 1981
study assumed that blanks or missing values on boater trip expenditures were'zero values.
The present study improves on this by more clearly defining a "trip", making it easier to
explicitly report no spending via a checkbox on the questionnaire, and by distinguishing

between day outings and overnight trips.

Stratification by Types of Boat Storage

Individual segments are more clearly tied to particular management or marketing
strategies. Disaggregating boaters into segments also makes it easier to track changes in
spending that frequently are tied to a changing mix of boaters (Stynes, 1998). Many
boating studies conducted in various states divide the boating fleet into distinct segments
to describe and explain patterns of boating use and spending. This study stratifies
boating use and spending by the type of boat storage, as the greatest variation among
boaters is expected to occur across different types of storage.

The appropriate segmentation may depend upon the particular situation and
application, but some of the key variables for classifying the boating market are clear.
For general management and planning applications, craft type, boat length, and storage
type appear to be the most useful segmentation variables, particularly storage.
Discriminating between boats kept at waterfi'ont sites and boats trailered from
nonwaterfront homes is important to identify needs for access sites and launching

facilities as well as for managing conflicts between these two groups (Stynes et al.,

1983). Boats stored at waterfront sites need to be further segmented into a marina group
and boats stored at permanent or seasonal waterfront homes.

Stynes et al. (1995) argues that storage locations are the best predictors of where
boats are used and explain the types and amounts of use. Wu (1995) classified boats into
marina, second home, waterfront homes, and nonwaterfront homes as the basis for her
models to estimate boating activity in Michigan at the county level. She argued that
boating use and spatial patterns of use may be explained by storage. Further,iuse
estimates by storage type better meet the information needs of public and private sector
providers who frequently serve particular storage categories. Storage type tends to be
correlated with craft size. Boats stored at non-waterfront sites are primarily smaller craft
that can be trailered to launch sites, while marinas provide storage for larger power and

sailboats.

Regional Flows of Boating Activity and Trip Spending

Estimating the flows of boating activity and spending among different regions of
the state is important to support regional planning. To assess the regional economic
effects of boater spending requires that local spending by resident boaters be
distinguished from spending of non-local residents who are attracted to an area. A
questionnaire must be designed for boaters to specify their spending between origin and
destination to avoid potentially erroneous assumptions on the allocation of spending
among regions. For example, Lipton and Miller (1995) assumed that while transportation

and grocery expenditures occur in the county in which the boater lives, expenditures on

boat fuel and restaurant meals occur in the county where the boat is launched from, with
some exceptions.

The 1981 Michigan boating survey (Stynes et al., 1983) distinguished trip
spending between origin and destination to distribute boater spending among regions.
However, the allocation of boater spending to different regions of the state required a
number of simplifying assumptions because the 1981 survey did not distinguish boaters'
residences from storage location. There will be some misassignments due to boats stored
other than in the region of residence (e. g., boats stored at seasonal homes) This problem
will tend to underestimate spending in coastal areas and northern regions where craft are
often stored at second homes and marinas. Regionalization of trip spending, therefore,
requires information concerning location of boat storage, as well as location of boat
owner residence and boating use.

Storage segments are very useful to describe regional flows of boater spending.
Regional planning can be helped by examining which non-local boater segment could
contribute more to the region’s economy. For example, a region may want to attract
more boaters in a marina segment who spend more dollars than other segments, given the

region’s existing and potential carrying capacity.

10

Objectives
The objectives of this study are:
1. To estimate boater trip spending by storage segment in Michigan during the 1998
season,
2. To test for differences in measures of activity and trip spending across storage
segments, and
3. To estimate regional net gains of boater trip spending by storage type and

spending category,

Boater Trip Spending 11y Storage Segment

To estimate boating use and trip spending in Michigan, the study employed a
wave approach, sending surveys to independent samples of boaters in nine waves during
the boating season. Boat use and boater trip spending are hypothesized to vary based
upon the type of boat storage. Storage types are facilities where boaters keep their boats
during the season. Boat storage types are important because they are the best predictors
of where boats are used and explain the types and amounts of use and spending. Storage
type is correlated with craft size. Spending also varies between day and overnight trips.
Spending profiles of boaters should be distinguished between these two types of trips in

order to provide more accurate estimates of trip spending.

Test for Differences in Boating Use and Spending across Storage Types
The characteristics of a segment must be distinguishable from those of other

segments so that product or service offerings and appeals can be tailored to the segrnent's

unique characteristics. Differences in measures of trip and spending patterns,
demographic characteristics of boaters, and boat-related factors are tested across storage
segments. Comparing the segments provides insights into the most effective

management and marketing strategies for boating business.

Regional Flows of Boater Trip Spending

To assess the economic effects of boater spending upon regions of the state
requires the separation of local spending by resident boaters from the spending of non-
local residents who are attracted to the area. Spatial distributions of boats and boat days
by region and storage type are estimated to allocate boater trip spending to different
regions. Storage segments are most useful to describe regional flows of boater dollars
since boat storage is the best predictor of where boats are used and explain amounts of

use and spending.

In summary, this study is designed to estimate boater expenditures onboating
trips based on in—season surveys. To provide better descriptions for patterns of boater
spending, the study stratifies boating use and spending by the type of boat storage. Tests
for differences in measures of boating activity and trip spending are conducted across
different storage segments. To support regional planning, the study also estimates the

flows of boating activity and spending among different regions of the state.

12

Organization of the Study

The study consists of five chapters. The next chapter reviews previous travel
expenditure studies, boating studies relating to boating use and boater spending, literature
relating to modeling recreational use and spatial patterns of recreational uses. The third
chapter describes the methods used to collect and analyze the data. Tests for nonresponse
bias in the wave survey are also presented. The fourth chapter presents the results of the
models to estimate boating use and trip spending in different storage segments, tests for
the differences in measures of boating activity and spending, and estimates regional flows
of boater spending. The fifth and final chapter provides an overview of the results and

offers recommendation for improving the study.

13

CHAPTER 2

LITERATURE REVIEW

The literature relevant to this study of boater expenditures comes from four
subject areas: 1) studies of traveler expenditures, 2) studies of boating use and
expenditures, 3) relevant approaches for modeling recreational use, and 4) studies of

spatial patterns of recreational boating in Michigan.

Studies of Traveler Expenditures

Estimating travel expenditures is a necessary first step to estimate the economic
impact of nonresident travel to a state or region. Economic impact estimates are often
used to justify future developments and monitor past management actions (F rechtling,
1994b). The large magnitude of impact that results from small errors in expenditure data
suggests that methodologies used in gathering visitor data be examined and refined to
ensure valid and reliable estimates. Researchers have long raised concerns about the
accuracy of surveys that are commonly used to gather economic information (Deleeuw
and Hox, 1988). These concerns are not always easily addressed since survey costs and
immediate data requirements often drive travel research methodologies (Rogers, 1991).

Trip spending estimates in travel studies are based primarily on on-site or
household surveys. On-site surveys of travelers are conducted while they are in the area
under study. On-site surveys are superior to household surveys if one assumes decline in
respondent recall as the time elapsed between expenditure and interview increases

(F rechtling, 1994b). Given the dispersed nature of activities related to recreation and

tourism, however, it is difficult to obtain representative samples of visitors and trips from
on-site surveys. The difficulty in projecting sample results to the total population is also
not resolved in on-site survey method.

Household surveys are more frequently used to estimate levels of recreation and
tourism activity and characteristics of trips than on-site surveys. Recall errors are a prime
weakness of household surveys when used to estimate trip spending. A strength is that
sampling frames for household surveys are readily available, and it is a simple matter to
project sample results to the total population for absolute estimates, something on-site
surveys do not readily permit (Frechtling, 1994b). Three basic modes of household
surveys are mail surveys, telephone interviews, and face-to-face interviews.

The conduct of surveys in the household has been discussed elsewhere
(Frechtling, 1994b; Babbie, 1992; F erber, 1978). Mail surveys allow the largest sample
size within a given budget and permit respondents to consider their answer carefully. On
the other hand, mail surveys are the slowest of the three modes, have potential for recall
bias, and produce the lowest response rates. Face-to-face interviews have the virtues of
shorter elapsed time between interview and processing relative to mail and high response
rates. The drawbacks of this mode are the high cost of interviewing and potentially poor
interviewer supervision. Telephone surveys produce results more quickly and are
superior in minimizing lag between interview and processing, but do not permit lengthy
questions with many choices.

Travel spending estimates are based primarily on self-reported data (Howard et
al., 1991). One of the main issues confronting tourism planners and researchers is how

accurately travelers recall the expenditures related to their trips. Evidence indicates that

15

response error (the difference between actual and reported expenditures) may be
substantial (Rylander et al., 1995; Stynes and Mahoney, 1989; Mak et al., 1977). In their
comparison of travel expenditures derived from post-trip survey questionnaires and trip
diaries, Mak et al. (1977) found that recall survey respondents significantly '
underestimated their expenditures relative to the diary. While the diary approach may
have been biased, it appears that the shorter recall period helps better capture traveler
expenditures.

Stynes and Mahoney (1989) found the post-trip recall estimate of respondents
attending a national conference to be 20 percent less than the expenditure estimates
provided by respondents during the conference. The conclusion was that recall errors in
the post-trip survey lead to the underestimates of spending. Rylander et al. (1995) tested
for the presence of recall bias in mailed survey questionnaires using data collected from
visitors at recreation sites. In line with previous studies (Frechtling, 1987; Stynes and
Chung, 1986; Ellerbrock, 1981), Rylander et a1. (1995) indicated that recall bias was
observed. They recommended procedures that obtain complete responses either during or
immediately upon the completion of a respondent’s trip.

In a study of the impact of the amount of elapsed time between an intercity trip
and the report of the trip on reported trip volume, Meyburg and Brog (1981) found that
the longer the elapsed time, the smaller the proportion of actual trips reported. For
example, more that 4 percent of actual intercity trips were unreported six to nine months
later, and 13 percent were unreported nine to twelve months after they occurred. It is fair
to say that if there is underreporting of trips, there must be underreporting of total

expenditures across trips.

16

The 1991 and 1996 National Survey of Fishing, Hunting, and Wildlife-Associated
Recreation (U SDI, USFW, and UDBC, 1993; 1998) employed a wave analysis to reduce
the recall period over which respondents had to remember their activities and
expenditures on fishing, hunting, and wildlife-associated recreation. The 1991 and 1996
FHWARs conducted interviews about every four months. The previous F HWARs used a
12-month recall period which resulted in greater reporting bias (U SDI, USFW, and
UDBC, 1998). Research on recall bias found that the amount of activity and '
expenditures reported in the 12-month recall FHWAR was over-estimated in comparison
with the amount reported in shorter recall periods (U SDI, USFW, and UDBC, 1998).

However, expenditures made 3 or 4 months prior to a survey may be still more
difficult for respondents to remember, than expenditures in the month immediately prior
to a survey. Based on the on-going Consumer Expenditure Survey (CES), Nelson (1996)
argued that recall biases on various consumer expenditure categories may lead to
underreporting in the early months of each interview period. For example, only 24.5
percent of purchases of boys’ and girls’ footwear are reported as taking place three
months prior to the interview, while 45.9 percent are reported as taking place in the most
recent month. The author argues that because it is highly unlikely that people tailor their
consumption patterns around their interview schedule, poor reporting seems to be the
cause.

One other source of direct comparison between short and long recall period
surveys is the US. Travel Data Center study of the 1973-1974 CES (1978). The CBS,
conducted by the US. Bureau of the Census, obtained expenditure information on travel

purchases by personal interview once every three months over a two-year period. It was

17

estimated that US. residents spent an average of $2.7 billion on foreign travel per year
during the survey period. The US. Department of Commerce Bureau of Economic
Analysis (BEA) annually estimates U.S. travel spending in foreign countries using self-
administered questionnaires distributed to returning residents, keeping the elapsed time
between expenditures and report to a minimum. BEA estimated an annual average of
$7.2 billion in consumer expenditures for foreign travel for 1972-1973. US. consumer
spending on foreign travel, then, may have been underestimated by more than 60 percent
through the CES because of recall bias associated with the long elapsed time between
expenditures and interview (cited from F rechtling, 1994b).

Rylander et al. (1995) found that trip complexity interacted with the recall bias
created by the passage of time in obtaining mailback questionnaires. Howard et a1.
(1991) documented a similar finding that recall accuracy is affected by two dimensions of
travel time: both the elapsed time between the trip and post-trip data collection and the
duration of the trip itself. Rylander et al. (1995) also found that groups with greater
spending tended to report less accuracy. It would be fair to say that the more complicated
spending items and larger spending amounts are subject to more recall bias than the less
complicated items and smaller amounts.

Nonresponse bias has been a concern with mail questionnaires. Nonresponse
bias, simply described, is the differences between the answers of respondents and
nonrespondents. Nonresponse bias is especially a concern when low response rates are
obtained (Stewart et al., 1993; Donald, 1960). This is especially true when the
characteristics of the nonrespondents are substantially different from that of the

respondents on characteristics of interest to the study. The low response rate in mail

18

surveys produces trip volume bias. There is evidence that nonrespondents to mail
surveys tend to be less mobile in terms of the number of trips than respondents
(Woodside and Ronkainen, 1984; Hunt and Dalton, 1983). Stewart et al. (1993) also
found that nonrespondents were more likely to have lower trip spending.

Generally, the composition of respondents is compared to that of nonrespondents
on characteristics that are relevant to the study in order to address whether
nonrespondents are different from respondents (Lawton and Parasuraman, 1980). If no
significant differences are observed between the two groups, the absence of nonresponse
bias is inferred. If significant differences are observed, caution should be appended to the
research conclusion to account for the possible bias, or the effects of nonresponse bias
should be adjusted for accordingly (Lambert and Harrington, 1990; Armstrong and
Overton, 1977; Daniel, 1975).

Few studies thoroughly assess the difference between nonrespondents and
respondents in mail surveys. Many nonresponse analyses have been based on limited
information available from secondary sources, samples of nonrespondents, or respondent-
nonrespondent comparisons of demographic profiles (Becker, 1984). Note that this
method does not directly test for nonresponse bias on the survey items, since the only
inferences that may be substantiated concern the demographic, sociological, or
performance characteristics themselves (Robert et al., 1970). This information may be of
little value in determining the nonresponse bias associated with the key variables of the
study, but can be useful when budget and/or time constraints preclude another follow-up

mailing.

l9

When information on nonrespondents is limited, nonresponse bias can be assessed
using follow-up mailings assuming that late respondents are a reasonable substitute for
nonrespondents and early and late respondent comparisons provide a reasonable
approximation of true nonresponse bias (Ellis et al., 1970). Research support for these
assumptions has been inconclusive within the tourism and recreation literature. Some
studies have reported no significant differences between early and late respondents
(Gitelson and Drogin, 1992; Dolsen and Machlis, 1991; Becker and Iliff, 1983; Hammit
and McDonald, 1982). Others have reported significant differences between early and
late respondents (Choi et al., 1992; Woodside and Ronkainen, 1984).

To identify the basis for these conflicting findings, Rylander et a1. (1995) tested
two hypotheses that late respondents (potential nonrespondents) are not different from
nonrespondents and results of wave analysis are no different from respondent-
nonrespondent comparisons. The authors concluded that the results of the follow-up
scheme were quite different from the comparisons of respondents to nonrespondents. A
nonresponse analysis based only on the three waves (one initial mail and two‘follow-ups)
would have led to erroneous conclusion of no significant nonresponse bias in the sample,
whereas significant differences between respondent s and nonrespondents were observed.

Lambert and Harrington (1990) suggested that samples be drawn from
nonrespondents after the planned follow-up mailings are completed to determine the
presence and direction of nonresponse bias. The authors recommended that a condensed
version of the questionnaire that contains key variables derived from analysis of the first
two follow-ups of questionnaires be sent to a sample of the nonrespondents for detection

of bias. If bias is not detected, researchers have increased confidence in the conclusions.

20

If bias is detected, the effects of nonresponse bias should be estimated and adjusted for
accordingly.

A variety of ways have been offered to deal with the potential problem of
nonresponse bias. Armstrong and Overton (1977) reviewed estimating methods and
described and tested the subjective and extrapolation techniques. An example of
subjective estimates technique involved selecting a panel of experts or judges, having
them identify survey items they believe to be subject to nonresponse bias, and state the
direction of the bias based on at least two response waves. Using group consensus on the
direction of the bias for selected items, valid predictions were reported (Armstrong and
Overton, 1977).

Another approach to rectify nonresponse bias is to weight the sample results
logically in order to adjust for nonresponse (Tanfer, 1993; Fuller, 1974). Statistical
weighting techniques also have been developed and presented in the literature as
especially useful when the response rate is not uniform across population subgroups
(Mandell; 1974). In a 1994 boating study in Michigan (Stynes et al., 1995), weights were
assigned for each boat size class, region and type to expand the final completed sample to
the population of active registered watercraft. Others have developed statistical models
to handle nonresponse in sample surveys (Kott, 1994).

The extrapolation methods, as presented by F ilion (1976) and Churchill (1988),
involves estimating the value of a population parameter by a linear extrapolation based
on the cumulative response rate over successive waves of replies. The logic of the
procedure is based on the purpose of surveys which is to estimate population figures

while correcting for nonresponse bias, rather than to estimate nonresponse bias for its

21

own sake. A nonlinear extrapolation model was presented by Daniel (1975) in his review
of ways to handle nonresponse in sociological surveys, and Zimmer (1956) developed an

extrapolation model based upon the response-nonresponse probability function.

Studies of Boating Use and Spending

Many studies have investigated the spending patterns of boaters, usually for the
purpose of documenting the economic impact of boating on community, region or a state
(Stynes et al., 1983; Lipton and Miller, 1995; Neely et al., 1998). Since variation in
boating activity and spending is great across distinct segments of boaters or boats,
boating studies are usually based upon segmentation to help describe and explain patterns
of boating use and spending. Distinct segments are also more clearly tied to particular
management or marketing strategies. The appropriate segmentation may depend on the
particular situation and application, but there exist some segmentation variables that have
been commonly used in boating studies.

Warner (1974) found that craft type (motor vs. sail) and length of craft are most
influential variables affecting boater expenditures using data collected from marina
boaters in Michigan. In a 1981 Michigan boating study, Stynes et al. (1983) divided craft
into four types of open, cabin, sail, and pontoon and Open craft was again broken down
into smaller and larger. The 1981 study also classified marina boats by type and length
of craft. In a survey of Californian boat owners (Public Research Institute, 1996), boats
were broken down into three size classes of under 16, 16-25, and over 25 feet, and boats
under 16 feet were again divided into jet-propelled, sail, and other. A 1985 survey of

Delaware registered boat owners (Falk et al., 1987) segmented boats by size class.

22

To reflect the particular geographical setting of Michigan boating, many
Michigan boating studies (Recreation Resource Consultants, 1972, 1975; MDNR, 1979;
Stynes and Safronoff, 1982) estimated boating use by the type of water body: Great
Lakes and inland lakes. Use in inland lakes was further divided inland lakes and
river/stream (Talhehn et al., 1988; Stynes et al., 1995). To assess the scope of
recreational boating and the contribution of boater spending in Oregon, Neely et al.
(1998) divided craft into registered recreational boats, commercial recreational boats, and
nonregistered recreational boats. The inclusion of nonregistered boats was to describe
the share of recreational boating activity taking place in nonmotorized craft such as
inflatable rafts, kayaks, and drift boats.

Storage type has been advocated in recent boating studies as the basis for the
primary segmentation of boats used in Michigan. Stynes et a1. (1995) argued-that storage
locations are important because they are the best predictors of where boats are used and,
along with boat size, explain the types and amounts of use. Wu (1995) classified boats
into one of four segments: marina, second home, waterfront homes, and nonwaterfi'ont
homes as the basis for models that estimate county levels of boating use. Ina 1994
Maryland boating study (Lipton and Miller, 1995), trip spending was broken down by
various categories: whether the boat was trailered or kept in the water; whether it was a
sailboat or powerboats; and by various size class within these groups.

Many boating studies surveyed in this section have used year-end surveys to
collect trip spending data (e. g., Neely et al., 1998; Public Research Institute, 1996; Lipton
and Miller, 1995; Falk et al., 1987; Sommerson, 1976; Warner, 1974). A potential

problem with these studies is recall error. Thus, responses to expenditure categories are

23

likely only rough estimates. To overcome this problem, Stynes et al. (1983) employed a
mailed survey sent out in six waves over the boating season to estimate per day trip
spending more accurately. Boaters were asked to report personal trip expenditures on
their most recent trip to reduce recall errors. Since annual days of use cannot be
estimated directly from the wave surveys, however, the 1981 study used annual estimates
of use obtained from a 1980 Michigan boater survey (Stynes and Safronoff, 1982).

Follow-up schemes are generally employed in boating surveys to achieve a higher
response rate. Follow-up mailings and reminder post cards are frequently used. In a
1981 Michigan boating study (Stynes et al., 1983), follow-up mailings were sent to
persons not responding within 10 days after the initial mailing. An additional follow-up
mailing was sent if no response was received within 14 days after the first follow-up
mailing. Multiple mailings provided a return rate of nearly 67%. In a more recent
Michigan boating study (Stynes et al., 1995), a second complete mailing was sent by
certified letter to all subjects who had not yet responded, three weeks later after the initial
questionnaires were mailed by first-class mail. Around 2,000 responses were received
within three weeks of the initial mailing and another 2,277 after the follow-up mailings
for overall response rate of 70%.

In a 1995 Oregon boating survey Neely et al. (1998) implemented a four-part
mailing procedure based on Salant and Dillman (1994). The mailing consisted of a cover
letter, a questionnaire, and a reminder post card. Finally, a follow-up questionnaire was
mailed to each addressee who had not yet submitted a completed survey. The Mailings
were sent over the course of four successive weeks for overall response rate of 71%. In a

1994 Maryland boating survey (Lipton and Miller, 1995), a postcard was mailed

24

reminding boaters to return the survey forms, one week later after the initial mailing to
obtain a 46% response rate. A second postcard was mailed two weeks later to
nonrespondents, and new survey forms were sent to boaters who had not responded after
another two weeks. After all follow-up mailings, overall response rate was 60%.

Only a few studies have estimated the regional distribution of boater
expenditures to support regional planning of boating. The 1981 Michigan study (Stynes
et al., 1983) allocated boater spending to different regions of the state. Trip spending was
allocated to regions based upon travel patterns measured in the 1980 boater survey
(Stynes and Safronoff, 1982). Average boater spending was split between the origin and
destination region. These average spending figures per boat day were then multiplied by
boat days from the origin-destination matrix of Stynes and Safronoff (1982) to yield a
statewide spending origin-destination matrix for boat trip spending.

To determine counties having a greater concentration of boating expenditures than
others, the Maryland study (Lipton and Miller, 1995) allocated spending among counties.
Since the questionnaire did not ask about spending location, the authors had to make
several assumptions about the county where the spending occurred. They assumed that
expenditures on some items such as transportation to launch site and groceries occur in
the county in which the boater lives. Expenditures on other items such as boat fuel and
dry storage were assumed to occur in the county where the boat launch from. However,
if boaters indicated that they went ashore during their boat trip in a county other than the
starting county, lodging and restaurant meals were allocated to the county where shore-

based purchases were made.

25

Boater spending is generally divided into three major categories: 1) new and used
boat purchases, 2) craft-related spending, and 3) trip-related spending (Stynes et al.,
1983; Lipton and Miller, 1995). Craft-related spending includes expenditures for
equipment, insurance, repair, storage of the boat, and other related items not directly
associated with individual boating trips. Trip-related spending includes all spending
related to boating trip such as expenditures on groceries, restaurant meals, auto and boat
fuel, and boating and other recreational gear purchased on trips. Spending on boats and
craft-related items primarily accrues to boating industries, while trip-related spending
benefits a wide range of retail sectors of the economy.

Trip-related spending is generally greater than the expenditures on craft-related
items or boat purchases. Stynes et al. (1983) estimated that Michigan's registered boat
owners spent over one billion dollars on boating in 1981. T rip-related spending made up
66%, while craft-related spending accounted for 24% of the total. New and used boat
purchases accounted for 10% of the measured spending. A Maryland study (Lipton and
Miller, 1995) estimated annual boater spending at about $1 billion in 1993. Trip- and
craft-related spending were $438 million (43%) and $428 million (42%), respectively.
New and used boat sales accounted for $144 million (14%). In a California study, Rust
and Potepan (1997) estimated that trip- and craft-related spending amounted to $2.5
billion in 1995. The majority of this spending fell into trip-related categories (69%).
Neely et al. (1998) estimated that Oregon's registered boat owners for recreational
purpose spent $858 million during the 1995 boating season. About 50% of the total was
spent on boat purchases, while trip- and craft-related spending made up 35% and 15% of

the total, respectively.

26

Forecasting Recreation Use

Forecasting plays an important role in most organizations since virtually all
planning and decision-making must rely on assumptions about the future (Stynes, 1982a).
Choosing a particular forecasting model is a complex decision. In comparing the relative
performance of different forecasting techniques, Makridakis (1986) noted that no study
has shown a clear superiority of one method over another and there is not any single
method which consistently outperformed the remaining methods. Fildes and Lusk (1984)
and Witt and Witt, (1995) also argued that the "best" method from the various forecasting
competitions is seldom identified. It seems clear that results relating to the relative
performance of different forecasting techniques cannot be generalized fi'om other
industries to tourism and recreation.

Forecasting methods are usually divided into qualitative and quantitative
techniques. Qualitative methods directly incorporate human judgement, while
quantitative methods generally employ formal mathematical models. When quantitative
models are difficult to apply to situations where variables are hard to quantify and
relationships are poorly understood, qualitative techniques are often used (Stynes,
1982b). The Delphi method of forecasting is the qualitative method that have attracted
the most attention in the tourism and recreation literature (Moutinho and Witt, 1995;
Kaynak and Macaulay, 1994; Var, 1984). This technique obtains expert opinion about
the future through questionnaire surveys of a group of experts in the field and is
particularly useful for long-term forecasting.

Two of the most widely used quantitative methods to forecast recreation and

tourism use are time-series analysis and causal models. Time-series methods estimate

27

use by extrapolating from use counts, such as visitor days, recreation occasions, permits
,or some other measures of participation. The method determines future values for a
single variable through a process of identifying a relationship for past values of the
variable (Witt and Witt, 1992). A problem with forecasting by extrapolation is that any
alteration in the trend is likely to generate poor forecasts, as it presupposes that the
factors which were the main cause of growth in the past will continue to be the main
cause in the future.

The lack of good time series data in outdoor recreation has restricted the use of
the time series method primarily to simple trend extension (Stynes, 1982b). Selection of
an appropriate functional form must be based upon an examination of the historical
pattern in the data series and assumptions about the grth process. Various functional
forms in time-series methods are surveyed in Witt and Witt (1992). Linear functions
assume use grows at a constant rate over time. Exponential fimctions assume the rate of
grth in use is directly related to the number of use. Both exponential and linear
functions are unbounded and can lead to absurd results if projections are made too far
into the future.

Logistic firnctions conform more closely to growth processes where constraints to
grth or saturation effects are encountered. In the logistic model, growth starts out
slowly, increases to maximum growth rate, and then slows down again, eventually
approaching a saturation level (Stynes, 1982b). Product life cycle curves follow the
logistic trend with an eventual decline at the end of the cycle (Howard and Crompton,
1980). Stynes and Szcodronski (1980) found that for long-range projections many

recreation activities follow trends similar to the product life cycle. In simple trend

28

extension, forecasters must use their judgement in deciding how far into the future a
given forecasting model may accurately project. Simple trend extension is generally not
recommended for forecasting more than five years into the future (Stynes, 1982b).

There are a number of more sophisticated time series methods which can be found
in Archer (1980), Wheelwright and Makridakis (1980), and F rechtling (1996). Evidence
shows that simple time-series models seem to perform as well as complex time-series
models in forecasting tourist arrivals (Chan et al., 1999; Chan, 1993). Choy (1984)
suggested that time-series methods are more likely to perform better than causal models
for projection of two years or less. For long-range forecasts, causal models may give
better forecasts.

Causal models forecast future recreation use by identifying relationships between
use and a set of demographic, socioeconomic, and environmental variables. These
relationships are usually identified via econometric analyses and then applied to forecasts
of the independent variables to predict future levels of recreation use (Stynes, 19823). A
major advantage of the causal approach is that it explicitly takes into account the impact
on demand of changes in the causal variables. An additional advantage with causal
forecasting is that it provides several statistical measures of the accuracy and significance
of the forecasting equations (Witt and Witt, 1992). Causal models, however, require
considerable user understanding in order to develop the correct relationships, and
therefore is generally more difficult to use that time-series methods (Witt and Witt,

1992).

29

Spatial Patterns of Recreational Boating in Michigan

T Five major studies that provided a description of the spatial patterns of
recreational boating use in Michigan were conducted by Michigan Waterway Division
(1965), Chubb and Chubb (1975), Stynes and Safronoff (1982), Talhelrn et al. (1988),
and Stynes et al. (1995). These studies estimated boating origin-destination patterns
using surveys where boaters reported the number of days and location they boated during
the season. The findings from these studies provide information on the spatial
distribution of boating use. The basic spatial patterns of boating use and flow of
recreational boats have been fairly stable over the years.

The studies showed that: 1) boats registered in southeastern Michigan counties
generate the majority of boat days in the state, 2) boating opportunities and resources are
unevenly distributed across the state, 3) the Upper Peninsula, northern Lower Peninsula,
coastal counties and lake areas provide relatively more boating opportunities and thus
attract a greater share of boat days from outside the regions, 4) the majority of boat days
in southern Michigan counties are accounted for by boats registered in the county or
nearby counties, and 5) a comparatively high percentage of boat days in northern
Michigan counties are accounted for by boats registered in southern counties.

[The 1994 Michigan boating survey (Stynes et al., 1995) provides the most current
information on statewide boating use at the county level. The 1994 study adjusted the
registration counts by size and county to reflect where boats are stored during the boating
season since registration statistics are a misleading indicator of the location of use.

Distinct allocation schemes were used to allocate boats within each region and segment

to individual counties. For example, boats in Great Lakes marina segments were

30

distributed according to the county's share of seasonal marina slips in the region. Finally,
various boat use parameters estimated by segment from the survey were applied to the
distribution of boats by segment for each county.

Stynes et al. (1995) argued that this approach yields much more reliable estimates
at the county level than would be obtained through direct cross tabulations of variables by
county using the survey data set which are as seen in the previous studies. Boats stored
in southeast and inland south regions generate about half of Michigan boat days, and
these regions receive 45% of boat days. Northeast and northwest regions received 17%
of total boat days and generated about 15% of total boat days in Michigan. An evident
south-to-north pattern was found as boat owners residing in the southern part of the state
stored their boats in northern counties.

Spatial patterns of recreational boating were also estimated using a model. The
RECSYS (Michigan Recreation System) was one of the earliest attempts to model
recreational travel flows for use in planning purposes. RECSYS predicts the spatial
distribution of recreation demand by simulating the movement of recreation users from
origin areas to destinations over the highway travel network (Ellis, 1964). This
simulation model assumes that recreational trips to a destination from any origin is some
function of a time-distance factor and the attractiveness at the destination. Using
RECSYS and boating use data from the 1965 survey (MWD, 1965), Chubb (1967)
predicted use at various destinations. Compared with the 1965 survey, Chubb found that
the RECSYS simulation retained a 19% standard deviation.

Using the 1994 Michigan boating survey, Wu (1995) developed a system of

models for estimating boating use in Michigan counties. The system of models consists

31

of boat allocation, trip generation and trip distribution models. Registered boats were
classified into four different storage segments. Boats in each storage segment are then
allocated to the counties where they are stored using a set of allocation models. A trip
generation model is used to predict number of boat days in the county of storage. Then
those boat days are distributed to the destination counties by trip distribution models for

boats at each storage segment.

32

CHAPTER 3

METHODS

Data for this study were collected through sample surveys of registered boat
owners. An in-season survey was utilized to estimate boating activity and trip spending
during the 1998 summer. Surveys were sent out in nine waves every week over the
summer. At the end of the boating season, two distinct end of season surveys were
conducted with smaller samples to evaluate potential nonresponse bias and use estimates
from the in-season survey. The boat registration file was used as the sampling frame to
adjust the sample to known population characteristics, and to expand estimates from the
sample to active boating fleet.

The methods chapter will detail the procedures used to collect and analyze the
data presented in this study. The chapter is divided into four sections: 1) sampling
design, 2) measurement, 3) data analysis, and 4) end-of-season surveys. The first three
sections deal with survey methods and procedures for data analysis associated with the
in-season survey. Survey methods, survey response rates, and weighting procedures
related to the two season-end surveys are discussed in the fourth sections. Survey
response rates, weighting procedures, and test for nonresponse bias for the wave survey

will be presented in the results chapter.
Sampling Design

The study population consists of all recreational watercraft with valid Michigan

registrations as of July 1, 1998. The computer file of registered watercraft maintained by

33

Michigan’s Secretary of State provides a convenient sampling frame although it includes
some expired registrations and a number of boats that were inactive in 1998. After all
non-pleasure boats and expired registrations were deleted from the file, a total of 751,012
pleasure craft with valid Michigan registrations in 1998 were obtained and used as the
sampling list. The sampling list contains the name and address of the boat owner, county
of registration, length of boat, and make of craft.

The population was stratified by five boat size classes and nine sampling regions
(Table 1). This stratification follows that of the 1994 study (Stynes et al., 1995) except
for the inclusion of personal watercraft (PWC) as a separate category and the merging of
two Upper Peninsula regions into a single region. Boat size strata are boats less than 16
feet, 16 to 20 feet, 21 to 28 feet, and greater than 28 feet. PWC are separated as a fifth
"size" strata. Boats less than 16 feet account for 43% of the total pleasure craft with valid
registrations, followed by boats 16 to 20 feet (33%). PWC and boats 21 to 28 feet
contribute equally (11%) to the total fleet. Boats larger than 29 feet account for 3% of
the total. Sampling regions are mapped in Figure 1. The study intentionally oversampled
larger craft and regions with smaller population sizes to have adequate subsarnples to
make estimates by size class and region.

A sample of 3,300 boats was selected using a systematic sampling procedure with
random start for each stratum (Table 2). The sampling unit is the boat, not the boat
owner. Boaters owning more than one registered boat were asked to report only for the
boat that was sampled. The length and make of the boat were printed on the mailing

label to identify the boat for which the survey was requesting information. By matching

34

Table 1. Michigan Registered Boats by Region of Registration & Boat Size: 1998

 

SIZE OF BOAT (FEET)
PWC <16” 16-20' 21-28' 29+ TOTAL %

 

 

Southeast Michigan 35,448 81,535 86,136 37,325 10,647 251,091 33%
Southwest Michigan 1 1,237 63,800 42,887 10,240 1,549 129,713 17%
West Central Michigan 12,579 48,017 32,093 8,880 2,407 103,976 14%

Thumb Region 9,545 39,851 30,321 9,775 1,493 90,985 12%
Northeast Michigan 1,843 14,558 9,288 2,801 173 28,663 4%
Northwest Michigan 3,448 30,843 18,089 5,283 847 58,510 8%
Straits 1,326 11,351 6,360 2,156 369 21,562 3%
Upper Peninsula 2,076 19,146 8,037 1,708 267 31,234 4%
Out of State 4,155 13,564 12,214 3,894 1,451 35,278 5%
TOTAL 8 1,657 322,665 245,425 82,062 19,203 751,012 100%
% 11% 43% 33% 11% 3% 100%

 

a. Excludes PWC'S.

survey responses with the registration information, it can be verified that subjects
reported for the boat that was sampled.

Surveys were sent out in nine waves. The first mailing to 900 registered boat
owners was sent on July 24, 1998. Subsequent mailings were sent to groups Of 300
subjects each week starting on August 3, August 10, August 17, August 24, August 31,
September 7, September 14, and September 21, 1998, respectively. Mailings were
stopped after September 20 because recreational boating activity in Michigan declines
significantly with the approach of cool weather. Mailings each week were sent out in
three micro-waves (i.e., Monday, Wednesday, and Friday) to avoid over-representing
weekday or weekend trips in the subjects’ report Of their “most recent trips”. A total of

3,300 questionnaires were sent (Figure 2).

35

 

    
    
  
   
 

Straits

Northwest ’ Northeast

 

 

Thumb

 

West Central

Southwest Southeast

Out-of-State

 

 

 

Figure 1. Sampling Regions of Michigan

36

Table 2. Distribution of the Sample by Region of Registration & Boat Size

 

SIZE OF BOAT (FEET) _
PWC <16" 16-20' 21-28' 29+ TOTAL %

 

 

Southeast Michigan 83 210 153 131 102 679 21%
Southwest Michigan 49 145 101 93 45 433 13%
West Central Michigan 51 135 102 64 51 403 12%
Thumb Region 47 122 96 59 48 372 l 1%
Northeast Michigan 38 75 52 53 22 240. 7%
Northwest Michigan 58 130 1 1 1 75 46 420 13%
Straits 26 58 49 37 24 194 6%
Upper Peninsula 39 88 59 52 36 274 8%
Out of State 44 92 70 43 36 285 9%
TOTAL 435 1 .055 793 607 410 3,300 100%
% 13% 32% 24% 18% 12% 100%

 

a. Excludes PWC'S.

900 ..

600 1

300.: i
l

0 . . . .
Jul. 24 Aug. 03 Aug 10 Aug. 17 Aug. 24 Aug. 31 Sep. 07 Sep. 14 Sep. 21

# of Questionnaires

 

1 Mailing Date

 

Figure 2. Schedule of Mailings for the Wave Surveys

37

F ollow—up mailings were not made to reduce survey costs and to avoid over-
representing late season trips. The same sampling proportions by region of registration
and size class were used for each mailing to yield representative samples in each wave,
after weighting. As the primary purpose of the surveys was to Obtain spending estimate
for a recent trip and to estimate boating use up to each point in time, it was important to
Obtain a sample representative not just of boats, but also of trips throughout the season.
Follow-ups surveys would bias the sample more toward end-of-season trips. While this
bias could be handled by an additional weighting procedure, the added cost of follow-ups

were not justified in terms of potential improvement in the estimates.

Measurement

A self-administrated instrument was chosen for this study. A business reply
return address was printed on the questionnaire. The questionnaire was four pages long
in booklet form (Appendix A), similar to those used in previous boater surveys. The
questionnaire was developed by making adjustments to the 1981 (Stynes et al., 1983) and
the 1994 (Stynes et al., 1995) Michigan boater survey instruments. A map presenting
Michigan’s 83 counties was inserted in the booklet to help respondents identify the
counties where they stored and used their boats. A cover letter accompanying the
questionnaire explained the survey, noted that participation was voluntary and explained
procedures for assuring confidentiality of the responses. Questionnaires were numbered
to keep track of the dates the questionnaires were sent out, and to match survey responses

with the registration information.

38

Variables Measured

 

1)

2)

3)

4)

The variables measured by the questionnaire fall into five groups (Table 3).
Boat characteristics are used to form subgroups for which spending is estimated.
Information on where the boat is kept during the boating season and the use of
marinas and launching facilities is required to construct storage segments. Boat
length and type of boat are also important to describe boats. To estimate the
number of active boats, a question related to inactive boat is added.

Boating use "so far" in 1998 is asked to estimate annual days of use per boat.
Boating use is broken down into total days of boating, days on Great Lakes
waters, and additional days not underway. Boating use on Great Lakes waters is
defined as any days the boat was underway on the Great Lakes and connecting
waterways, including lakes and rivers that provide access to the Great. Lakes.
Additional use is defined as days the boat was in the water but used at the dock.
Information on the most recent boating occasion is necessary to describe
boating patterns. Length of stay is required to convert spending to a per day
basis, and to split spending between day trips and overnight. Boating destination
is reported by county and body of water. Party size is also reported.

Boater trip spending is asked for the most recent trip. A checkbox is included to
measure whether or not any Spending occurred on the trip. Trip Spending is
estimated in eleven categories to identify which sectors of the economy benefit
from boater spending. The spending questions split trip spending between “near
home-within 20 miles” and “away from home” to estimate the flows of boater

spending among regions.

39

Table 3. Variables Measured in the Wave Survey

 

Variable Questionnaire
Number

 

1. INFORMATION ABOUT BOAT 1 .’ 7

Number of boats owned, Craft type, Length, Years owned, County
where boat is kept, Type of storage facility (permanent residence,
cottage or second home, public marina, commercial marina, owned
space in marina or dockaminium, yacht/boat club, and other),
Location of storage facility (a waterfront site with access to the
Great Lakes, an inland lake waterfront site, a river or stream
waterfront site, and a nonwaterfront site), Type of facility during
the non-boating season, Boat inactive.

2. USE ofBOAT 8 — 9

Days used, Days used on the Great Lakes, Days not underway.

3. MOST RECENT BOATING OCCASION 10 — 15

Descriptions of the most recent boating occasion, Date of the
occasion, County of use, Great Lakes use, length of stay, Party size.

4. SPENDING ON THIS MOST RECENT OCCASION 16 - 17

Whether or not to spend, Spending on Boat fuel, Temporary
dockage, Launch fee, Repair & maintenance spending related trip,
Marine supplies, Restaurant, Groceries, Auto gas, Shopping,
Recreation, and Other expenses, Spending evaluation.

5. CRAFT OWNER INFORMATION 19 — 23

County, State, and Zipcode of residence, Age of owner, Household
size, Level of income, County of seasonal home.

 

5) Boater characteristics are used to describe boat owners who registered their
boats in Michigan. Variables are age of boat owner, level of income, years boat
owned, household size, whether the boater owns a seasonal second home, and

residence location.

40

Handling of Zeros and Missing Data

Previous surveys have encountered difficulties in distinguishing missing
spending data from no spending. A speculation is that boaters would be more likely to
report overnight or extended trips that may not be their most recent trips. This could be
particularly the case when they are asked to report spending, as no trip spending is
necessarily involved when boating from a waterfront home. The questionnaire was
designed to clearly define what a trip was and to make it easier to explicitly report no

spending via a checkbox (Figure 3).

 

16. Did you spend any money on this boating occasion? Cl YES Cl NO (skip to question 18)
1
16a If Yes, please report how much money you spent within each of the following categories. Report
spending near your permanent or seasonal home in the first column and spending away from home
(more than 20 miles from home) in the second column. If you are boating from a waterfront home,
report spending on any items bought specially for this outing (e.g. boat gas, groceries etc.). If boating
away from home. report all expenses on this trip.

SPENDING ON YOUR MOST RECENT BOATING OCCASION
(enter zero ifyou did not spend anything in a particular category)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

NEAR HOME AWAY FROM HOME
BOAT EXPENSES
Boat Fuel and oil 5 $
Temporary dockage $ $
Pump-out & launch fees $ $
Repair and maintenance $ 5
Marine supplies 5 S
PERSONAL EXPENSES
Restaurant meals and drinks S S
Groceries & take out food & drink $ $
Auto gas and oil 3 $
Shopping & souvenirs $ 5
Recreation & entertainment 5 $
Other expenses $ $

 

 

 

Figure 3. Spending Questions

41

Missing values related to expenditure variables require special attention. Average
spending estimates can be quite sensitive to how missing data are treated. To handle this
problem, the study employed a filter question asking the boater to answer "yes" or "no" to
a question of whether spending occurred or not before reporting the amounts spent. This
procedure is expected to separate out zero expenditures from missing values by assuming
that a “no” on the checkbox (# 16) and blanks on the second questions (# 16a) indicate
true zeros, and blanks on the first and also blanks on the second questions indicate
missing values.

Of the 394 respondents who left the spending question (16a) blank, 368 (93%)
checked "no" to question 16. Only 26 (7%) of the respondents left both 16 and 16a blank

(Table 4). These were treated as missing and eliminated from the Spending analysis.

Table 4. Missing and Zero Responses to Spending Question

 

SPEND ANY MONEY ON THIS OUTING ?

 

Yes No Blank ‘ TOTAL
SPENDING REPORTED
Non-zero reports 492 22 a 63 577
Blank 0 368 26 394
TOTAL 492 390 89 97 1

 

a. Twenty two respondents who reported "no" spending on their most recent trips wrote "$0" on the
second question.

42

Data Analysis

The returned questionnaires were coded as they were received using an Access
program designed for data entry. Ranges were specified to define valid responses for
each variable in the file. Based on this procedure, the probability of entry/coding errors
could be reduced. Extensive and time-consuming procedures for data cleaning were
followed to check illogical response values.

The questionnaire for the wave survey asked about the most recent trip. Five
boaters who reported more than 30 days on the most recent boating outing were
eliminated from the spending analysis. Respondents spending more than $2,000 a trip or
more than $500 a day (17 cases) were also eliminated from the spending analysis as
analysis Showed that these trips are very unique and severely impact (20% more per day
spending when included) on mean spending. Some boaters apparently reported annual
dockage expense instead of the temporary dockage fee. If a dockage fee was greater than
$30*number of nights on this outing, it was replaced with zero (26 cases).

The study defines a boat day as any day or portion of a day spent actually in the
water under power or sail. The questionnaire also asked to report days of additional use
at the dock or mooring without being underway. The additional days were not included
as a boat day because there was no evidence indicating that an additional day was
involved in actual boating activity. Around 20% of active registered boats in Michigan
reported that their boat was involved in at least one day of additional use "so far".
However, only 0.4% of the total fleet reported that the boats were used at the dock or

mooring without getting underway on the most recent trips.

43

Formation of Storage Segment

Boats were assigned to one of four segments based on "type of storage facility"
(permanent residence, cottage or second home, public marina, commercial marina, owned
space in marina, and yacht club ) and "type of storage location" (Great Lakes waterfront,
inland lake waterfront, river or stream waterfront, and nonwaterfront sites). Boats kept at
"other" types of storage facilities, and cases with missing storage information are
excluded from the classification. The resulting segments are (1) marina: boats stored at
one of public marina, commercial marina, owned space in marina, and yacht club, (2)
waterfront primary home: boats stored at permanent residence located in waterfront sites
(3) waterfront seasonal home: boats kept at cottage or second homes located in waterfiont
sites, and (4) nonwaterfront home: boats stored at cottage, second homes, or permanent

residence located in nonwaterfront sites.

Estimation of Boater Trip Spending by Storage Segment

For a given segment, total trip spending per boat is computed as: days of use per
boat * spending per day = days on day trips * average spending on day trips + days on
overnight trips * per day spending on overnight trips. Total trip spending for each
storage segment is then obtained by multiplying the total trip spending per boat by the
number of active boats for the given segment. Total trip spending for each stOrage
segment is then summed over the segments to yield statewide total spending. Note that
spending is split between spending on days and overnight trips to apply distinct spending

profiles to each type of boating trip and provide more accurate estimates of trip spending.

The parameters that must be estimated are the average days of use per boat and the
average per day spending on boating trips.

A boat day is defined as any day or portion of a day spent actually in the water
under power or sail. Annual days of use can not be obtained directly from the wave
surveys as boaters were asked to report activity in mid-season. Annual use per boat was
estimated by extrapolating from in-season estimates to year-end totals. Average days of
use "so far" in the year were estimated for the fleet as a whole at nine points during the
summer. A logistic curve was fit to these nine points to estimate a relationship between

days of use and the point in season where use was measured.

Test for Measures of Variables across Storage Types

Differences in boating activity and spending across the four storage types were
tested. The variables selected to test boating activity included boat days, Great Lakes
boat days, percent of overnight trips, boat days per trip, and type of boating destination
between the Great Lakes and inland lake. The variables associated with spending on trip
included percent of trips with no spending, average spending per trip, and spending per
day. Statistical tests were also applied to test for differences in boat owner- and boat-
related variables (age, number of boats, household income, seasonal home ownership,
craft type, size of boat, years of boat owned, storage location). One-way AN OVA was
employed to test interval-level variables (those for which means are compared), while x2
tests were calculated for nominal-level variables (those where distribution of responses

were compared).

45

Estimation of Regional Flows of Trip Spending

In assessing the economic effects of boater trip spending upon region of the state,
spending of non-local residents who are attracted to the area must be identified. The net
gains and losses to each region of the state from boater spending can be estimated based
on boater origin-destination patterns and the per day spending estimates. Boaters who
live in region A and boat in region B represent a loss of dollars to region A and a gain to
region B. The net flows for any region R are computed by estimating (1) spending by
origin (region of residence) = all spending on boating trips by residents of region R and
(2) spending by destination region = all boater trip spending that takes place in region R.
The difference between the two is the net gain or loss to the region. Notice that local
boater spending (spending by boaters on trips within the region where they live) is
included in both the origin and destination spending estimates and cancels out when the
two are subtracted.

Spending by origin region is estimated by multiplying the number of active craft
registered in the region by the average days per boat and then the average spending per
boat day. This calculation is carried out separately for each storage segment to capture
regional differences in the numbers of craft, days of use, and spending across storage
segments. Similarly, spending by destination region is computed for each storage
segment from the estimates of boat days taking place in each region. Boat days within
the region are multiplied by the average spending per day for each segment. For boaters
on trips originating from outside the region, only spending "away from home" is
included. Their "near home" spending is allocated to the region where they live. For

boating activity from within the region both the "near home" and "away from home"

46

spending is included. For boaters using seasonal homes, the "near home" spending is
allocated to the region where the boat is stored, i.e., the region where the seasonal home
is located.

This procedure takes into account differences in use patterns and spending across
storage segments, but assumes no regional variation in either the average days of use or
per day spending within a given storage segment. Regional flows are also computed for
individual spending categories (food, gas, etc.) capturing which items tend to be bought
at home or at boating destinations.

The procedure to estimate the regional flows of boater spending can be
operationalized as follows. Total trip spending by origin is calculated by the following
equation:

(1) ORIGINf = r,* *d, *s,,
where

ORIGIN I." = trip expenses by boaters in storage segment i who live in region k,

r," = number of boats in segment i owned by residents of region k,
d, = average days of use of boats in segment i, and
s, = average spending per day for segment i.

Spending by destination is composed of spending "near home" by residents of the
region and spending "away from home" by boater from other regions. For boaters in the
seasonal home segment, "near home" spending is assumed to take place in the region

where the boat is stored. For spending by residents of the region:

(2) DESTRf = r," *d, *s},

47

where

DESTR,‘ : trip spending by residents in storage segment i in residence region k,

r," = number of boats in segment i owned by residents of region k,
d, = average days of use of boats in segment i, and
s,1 = average "near home" spending per day for segment i.

For spending occurred in destination region:
(3) DESTD,‘ = 1),." * sf,
where
DESTDf = spending occurred in destination region k by boater in storage segment i.
D." = number of total use days of segment i in destination region k, and
s. = average "away from home" spending per day for segment i.

Total Spending of storage segment i by destination region k is expressed as:
(4) 01551;." = DESTRf + DESTDf.

The regional flow of boater trip spending for a given segment i is the difference
between DEST," and ORIGIN}?
(5) NETGAINi" = BEST," - ORIGINi".
A negative value of NE T GAIN 1‘ means that region k is a contributor of boating trip

dollars to other regions in the state for segment i, whereas a positive value indicates that
the region benefits from trip purchases by boaters in segment i who live in other regions,

including out-of-state.

48

Regional net gains of boater dollars for Spending category j, NE T GAIN i , is
expressed as follows:

(6) NETGAINJI‘.’=rk*d*s;.+Dk*sf-rk*d*sj, sj=31.+512.,

where

r" = number of boats owned by residents of region k,

d = average days of use of boats,

D" = number of total use days in destination region k,

s: = average "near home" spending per day for spending category j,

s 12 = average "away from home" spending per day for spending category j, and
s, = average spending per day.

Regional flows of boater monies in different spending categories can help to identify

which sectors of the region’s economy benefit from boater spending.

End of Season Surveys

Two separate year-end surveys were employed to evaluate nonresponse bias
(nonrespondent survey) and to compare the in-season use estimates with the end-of-
season survey (independent survey). The same format and questions were used for these
two surveys and the questionnaire was condensed by selecting the variables associated
with boat and boater characteristics and boating use (Appendix B). Questionnaires were
numbered to match survey response with the registration information. For the
independent survey, this procedure especially was necessary to design follow-up mailing

to those who not responded.

49

Nonrespondent Survey

 

For the nonrespondent survey a total of 500 boats were taken from boaters not
responding to the wave surveys by October 1. The samples were selected from the 1,530
nonrespondents who were included in the first through seventh waves of the in-season
survey. Mailings to nonrespondents were sent on October 12, 1998. The year-end
survey of 500 nonrespondents resulted in a total of 82 responses for a responSe rate Of 16
percent. Twelve boat owners indicating that their boats were not used in Michigan
waters in 1998 were omitted to result in a sample of 70 boats for the analysis used to test

nonresponse bias (Table 5).

Independent Survey

 

Another independent sample Of 500 boats was sampled from the population of
751,012 recreational craft with valid Michigan registration. Mailings to this sample were
sent on October 12, 1998. F ollow-up mailings were sent to persons not responding
within three weeks after the initial mailing to increase the response rate. This survey
resulted in 201 responses out of 500 samples for a response rate of 41 percent. This
response rate was slightly higher than the in-season survey, and substantially higher than
the nonrespondent survey. Twelve boat owners (6%) reported that their boats were not
used in Michigan waters in 1998 (Table 5). Deleting these inactive boats resulted in a
sample of 189 active boats to estimate annual days of use per boat.

The two end-of-season samples were combined to estimate days of use. A total
sample of 259 active boats were obtained from the two end-of-season surveys. The

distribution of the sample by boat Size and region is presented in Table 6. In the same

50

Table 5. Survey Response Rate (Season-end Surveys)

 

 

 

 

Nonrespondent Survey Independent Survey
N % % N % %
Total Deliverable Total Deliverable
Total Questionnaires Mailed 500 100.0% 500 100.0%
Not Deliverable 0 0.0% 9 1.8%
Delivered 500 100.0% 100.0% 491 98.2% 100.0%
Returned Surveys 82 16.4% 16.4% 201 40.2% 40.9%
Active Boats 70 14.0% 14.0% 189 37.8% 38.5%
Inactive Boats 12 2.4% 2.4% 12 2.4% 2.4%
Nonresponse 418 83.6% 83.6% 290 58.0% 59.1%

 

Table 6. Sample of Active Boats by Region & Boat Size (Season-end Surveys)

 

 

 

SIZE OF BOAT (FEET)
PWC < 16' ‘16 - 20' 21 - 28' 29' + TOTAL %

Southeast Michigan 4 5 15 12 9 45 17%
Southwest Michigan 2 6 8 9 7 32 12%
West Central Michigan 4 6 1 1 6 8 35 14%
Thumb Region 3 2 9 5 7 26 10%
Northeast Michigan 2 6 7 6 2 23 9%
Northwest Michigan 0 8 l 1 8 10 37 14%
Straits 0 3 2 2 13 5%
UP 2 5 4 5 4 20 8%
Out of State 3 7 l 1 3 28 11%
TOTAL 20 48 82 56 53 259 100%
% 8% 19% 32% 22% 20% 100%

 

51

Table 7. Weights for Active Registered Boats (Season-end Surveys)

 

 

 

SIZE OF BOAT (FEET)

PWC < 16' 16 - 20' 21 - 28' 29' +
Southeast Michigan 8,656 13,033 5,129 2,901 1,123
Southwest Michigan 5,488 8,498 4,788 1,061 210
West Central Michigan 3,072 6,396 2,606 1,381 286
Thumb Region 3,108 15,924 3,009 1,824 203
Northeast Michigan 900 1,939 1,185 435 82
Northwest Michigan - 3,081 1,469 616 _ 80
Straits - 3 ,024 947 1,006 175
UP 1,014 3,060 1,795 319 63
Out Of State 1,353 1,549 992 1,211 344

 

manner as for the in-season survey, weights were applied to adjust the active sample to
the active population of registered watercraft in 1998 (details in the result chapter) (Table

7).

52

CHAPTER 4

RESULTS

Survey response rate, weights, and test for nonresponse bias for the wave surveys
are presented first in this result chapter. Corresponding to the three objectives of the
study, the results are presented in three sections. In section two, total days of use and trip
spending are estimated on a statewide basis and also by storage segment. Four storage
segments are constructed. Sampling errors for spending estimates are reported. The
estimates of annual days of use from the wave survey are compared to those from the
year-end survey. In section three, differences in the variables associated with spending
and use, demographics, and boat-related characteristics across storage segments are
tested. In section four, regional flows of boater trip spending are estimated for ten
subregions of the state. Spatial distributions of boat and boat days are estimated. The

regional flows are presented by storage segment and spending category.

Survey Responses, Weights, and Nonresponse Bias

Survey Response Rate

 

A total of 1,170 responses out of 3,300 samples were received by October 30,
1998 (Table 8). The response rate was 34.3 percent of the deliverable surveys (1.7
percent of mailings were returned as undeliverable). About 10 percent (113 responses) of
the boat owners returning usable surveys indicated that their boat was not used in
Michigan waters in 1998. Another 27 boaters indicated that they did not wish to

participate in this survey and 2 boaters returned blank surveys. After omitting inactive

53

Table 8. Survey Response Rate

 

TOTAL QUESTIONNAIRE MAILED

 

N % of Total % of Deliverable

Total Questionnaires Mailed 3,300 100.0%
Not Deliverable 57 1.7%

Delivered 3,243 98.3% 100.0%

Returned Surveys 1,113 33.7% 34.3%

Active Boats 971 29.4% 29.9%

Inactive Boats 113 3.4% 3.5%

Non-usable Questionnaire 29 0.9% 0.9%

Nonresponse 2,130 64.5% 65.7%

 

craft and unusable returns, a sample of 971 boats are available for the analyses.

Rates of response are slightly lower than average in southeastern Michigan and
the Upper Peninsula and slightly higher in three northern regions in Michigan (Table 9).
Owners of smaller boats less than 16 feet including PWC were less likely to respond than
owners of boats greater than 16 feet. Weighting procedures adjust for differences in

response rates, as well as the different sampling rates in each strata.

Weights

Because the sample was drawn disproportionately across regions and size
categories and the response rate was not uniform across regions and size class, weights
are needed to adjust the final sample of completed boater surveys to the population.
Completed sample sizes range from 198 in southeastern Michigan to 71 in the straits

region (Table 10). Weights are calculated by dividing the number of active boats by

54

 

region and size categories in the population by the corresponding cell counts in
completed samples. A total of 751,012 pleasure craft were registered in 1998 in
Michigan. This number, however, includes unknown number of boats that were inactive
in 1998. Thus, total number of active boats that were involved in any day of boating

activity must first be estimated to obtain appropriate weights.

Table 9. Response Rates by Region of Registration & Boat Size

 

 

Mailings Returns Undeliverable Response Rate "

SAMPLING REGION
Southeast Michigan 679 202 8 30%
Southwest Michigan 433 142 2 33%
West Central Michigan 403 134 5 ' 34%
Thumb Region 372 124 8 34%
Northeast Michigan 240 94 7 40%
Northwest Michigan 420 157 1 l 38%
Straits 194 75 4 39%
UP 274 82 4 30%
Out of State 285 103 8 ~ 37%
BOAT SIZE CLASS

PWC 435 87 8 20%
< 16' 1,055 259 15 25%

16 - 20' 793 346 20 45%
21 - 28' 607 258 9 43%
29' + 410 163 5 . 40%
TOTAL 3,300 1 1 13 57 34%

 

a. Response rate = return / (mailings - undeliverable).

55

Table 10. Completed Sample of Boats by Region of Registration & Boat Size

 

 

 

 

SIZE OF BOAT (FEET)
PWC < 16' 16 - 20' 21 - 28' 29' + TOTAL %
Southeast Michigan 13 44 62 46 33 198. 18%
Southwest Michigan 6 38 44 33 20 141 13%
West Central Michigan 10 28 49 27 19 133 12%
Thumb Region 12 24 30 31 19 1 16 1 1%
Northeast Michigan 9 18 22 32 8 89 8%
Northwest Michigan 12 36 51 34 22 155 14%
Straits 5 19 19 18 10 71. 7%
UP 10 23 20 16 14 83 8%
Out of State 9 19 40 16 14 98 9%
TOTAL 86 249 337 253 159 1,084 100%
% 8% 23% 31% 23% 15% 100%
Active Boats

Respondents reported whether the boat has been put in the water in Michigan in
1998. Inactive craft are defined as boats that had not been put in the water in'Michigan
as of the survey date. Boaters that had not used their boat at the time being asked were
assumed to be unlikely to use it by end of the season. About a tenth of all boaters
returning a survey indicated they had not yet used the boat (Table 11). The percentage of
boats that were inactive ranged from 20 percent for boats under 16 feet in length to 2
percent for personal watercraft.

A raw (unweighted) rate of activity for each size class was applied to the number
of registered craft to yield a population of active craft. The rate of inactivity by size class

in Table 11 was applied to the number of registered boats in the corresponding size in

56

Table 1. About 99,000 boats out of 751,000 registered craft in Michigan were estimated
to be inactive in 1998, 13 percent of all registered boats in Michigan in 1998.

Consequently, the number of active boats was estimated to be 651,623 (Table 12).

Table 11. Percentage Of Registered Boats Inactive (Sample) a

 

 

Inactive Boats Active Boats Total Returns % Inactive
PWC 2 84 86 2%
< 16' 50 199 249 20%
16 - 20' 36 301 337 11%
21 - 28' 17 236 253 7%
29' + 8 151 159 5%
TOTAL 1 13 971 1,084 10%

 

a. Figures in the last column represent unweighted percentages. Assumes that rate of inactivity within
size classes does not vary by region.

Table 12. Registered Boats Inactive in 1998 (Population)

 

 

Inactive Boats Active Boats Total Boats % Inactive
PWC 1,899 79,758 81,657 2%
< 16' 64,792 257,873 322,665 20%
16 - 20' 26,218 219,207 245,425 11%
21 - 28' 5,514 76,548 82,062 7%
29' + 966 18,237 19,203 5%
TOTAL 99,389 651,623 751,012 13%

 

57

Weights for Active Boats

The distribution of boats in the population by size and county of registration is
known, so the study can adjust for the disproportionate sampling to provide estimates that
will represent the active registered boating fleet as a whole. Weights are assigned for
each boat size class and region to expand the final completed active sample to the active
population of registered watercraft. These weights are derived by dividing the cell counts
in Table 13 by the number of completed active surveys in each corresponding category
from Table 14. Resulting weights are presented in Table 15 . Total numbers of active
boats are then obtained by applying these weights to the completed active sample. This
procedure yields a population of 652,000 active pleasure craft with valid registrations.

The sample matches the active boat population by region of registration and Size.

Table 13. Active Boats by Registration Region & Boat Size (Population)

 

SIZE OF BOAT (FEED

PWC < 16' 16 - 20' 21 - 28' 29' + TOTAL %

 

Southeast Michigan 34,624 65,163 76,935 34,817 10,1 1 1 221,649 34%

Southwest Michigan 10,976 50,989 38,306 9,552 1,471 1 1 1,293 17%
West Central Michigan 12,286 38,375 28,665 8,283 2,286 89,895 14%
Thumb Region 9,323 31,849 27,082 9,118 1,418 78,790 12%
Northeast Michigan 1,800 1 1,635 8,296 2,613 164 24,508 4%
Northwest Michigan 3,368 24,650 16,157 4,928 804 49,906 8%
Straits 1,295 9,072 5,681 2,011 350 18,409 3%
UP 2,028 15,301 7,178 1,593 254 26,354 4%
Out of State 4,058 10,840 10,909 3,632 1,378 30,818 5%
TOTAL 79,758 257 ,873 219,207 76,548 18,237 651,623 100%
% 12% 40% 34% 12% 3% 100%

 

58

Table 14. Sample of Active Boats by Region & Boat Size

 

 

 

SIZE OF BOAT (FEET)
PWC < 16' 16 - 20' 21 - 28' 29' + TOTAL %

Southeast Michigan 12 37 56 45 32 182 19%
Southwest Michigan 5 32 39 29 20 125 13%
West Central Michigan 10 22 45 27 19 123 13%
Thumb Region 12 16 27 30 19 104 11%
Northeast Michigan 9 14 20 29 6 78 8%
Northwest Michigan 12 30 44 31 19 136 14%
Straits 5 15 19 17 10 66 7%
UP 10 18 15 14 12 69 7%
Out of State 9 15 36 14 14 88 9%
TOTAL 84 199 301 236 151 971 100%
% 9% 20% 3 1% 24% 16% 100%

 

Table 15. Weights for Active Registered Boats

 

 

 

SIZE OF BOAT (FEET)

PWC < 16' 16 - 20' 21 - 28' , 29' +
Southeast Michigan 2,885 1,761 1,374 774 316
Southwest Michigan 2,195 1,593 982 329 74
West Central Michigan 1,229 1,744 637 307 120
Thumb Region 777 1,991 1,003 304 75
Northeast Michigan 200 83 l 415 90 27
Northwest Michigan 281 822 367 159 ' 42
Straits 259 605 299 1 18 35
UP 203 850 479 114 21
Out of State 451 723 303 259 98

 

59

Weights for Boat Owners

When describing boat owners as compared to boats, the sample must be adjusted
for multiple boat ownership. The 652,000 active boats were owned by 458,000 boat
owners. Boaters owning more than one registered boat would have a greater Chance of
being selected in the sample than owners of a single boat, so cases were weighted
inversely to the number of boats owned. About 70 percent of the boat owners having at
least one active boat registered in Michigan in 1998 own single boats, while 23 percent

own two boats and 8 percent more than two (Table 16).

Table 16. Boat Ownership by Number of Boats Owned 3

 

 

Number of Boats Owned % Number of Owners
1 boat 69% 316,281
2 boats 23% 103,340
3 boats 6% 28,219
4 or more boats 2% 9,859
SUM 100% 457,698

 

a. Unit of analysis in this table is the boat owner. The sample of boats was weighted inversely to the
number of boats owned by each respondent.

Test for Nonresponse Bias
As use and spending varies considerably with boat size and storage, and possibly
region (e.g., shorter season up north), the existence of nonresponse bias in these variables

would threaten the external validity of survey results. Since weights were applied to

60

adjust the final sample to the known region and boat size distribution in the population of
active craft, nonresponse bias would not be a problem for these variables. For example,
PWC'S had lower response rates, but this bias iS corrected in weighting. The Weights also
should correct for nonresponse bias in storage, as storage is closely related to boat size.
The remaining biases after weighting are likely to be related to inactivity and bias in
spending estimates due to possible biases in reporting of the most recent trips.

While the survey explicitly clarified inactive craft in the questionnaire, it is likely
that owners of inactive boats were less likely to return the survey. This would bias the
estimates of the number of active craft upward, which would inflate overall use and
spending estimates. The sample from the season-end survey of non-respondents was
compared to the sample of respondents in the wave surveys. A hypothesis tested is that
key variables of respondents are no different from those of nonrespondents. The x2 test
was used to identify differences between respondents and nonrespondents. The selected
variables compared are inactivity, boat type, age of boat owner, number of boats owned,
household income, and seasonal home ownership in Michigan (Table 17). The variables
are all ordinal-level.

There are no significant differences between the respondents and the
nonrespondents in any of the variables selected for the test. The null hypothesis of no
difference between respondents and nonrespondents can not be rejected at the 0.01 level
of significance. As a result, the test provides no conclusive evidence of nonresponse
bias.

Since the number of nonrespondents is small, the statistical tests for nonresponse

bias still may only compare late or reluctant respondents to early respondents.

61

Table 17. Differences between Respondents & Nonrespondents

 

 

Respondents Nonrespondents X” P value
INACTIVITY RATE N=1,084 N=82 x2=l-41 0.235
% of Boat Inactive 10.4% 14.6%
TYPE OF BOAT N=962 N=53 x2=7.36 0.195
Inboard “ 32.5% 35.8%
Outboard 29.9% 35.8%
Sail 8.7% 15.1%
Pontoon 15.7% 7.5%
Canoe/row 4.4% 1 .9%
PWC 8.7% 3.8%
AGE OF BOAT OWNER N=97l N=70 x2=542 0.367
younger than 41 16.6% 17.1%
41 — 50 24.8% 32.9%
51 — 60 26.1% 28.6%
61 — 65 10.8% 4.3%
66 — 70 9.1% 8.6%
older than 70 12.7% 8.6%
BOATS OWNED N=796 N=62 x3=1.24 0.539
1 45.7% 43.5%
2 32.8% 29.0%
3 or more 21.5% 27.4%
HOUSEHOLD INCOME N=870 N=58 x3=529 0.381
under $20,000 6.0% 1.7%
$20,000-$39,999 17.8% 13.8%
$40,000-SS9,999 19.4% 29.3%
$60,000-S99,999 32.3% 34.5%
$100,000-S 149,999 12.8% 10.3%
over $150,000 11.7% 10.3%
SEASONAL HOME IN MI N=967 N=68 x2=0.48 0.488
yes 66.5% 70.6%

 

In/Outboard is included.

62

Respondents were directly compared to the population as an alternative for determining
nonresponse bias. Since the sample of respondents is weighted to reflect differential
sampling rates across region and boat size categories, the variable of boat type was
compared between the respondents and the population. Note that this method does not
directly test for nonresponse bias on the survey item. The distribution of boat types for
the population was directly computed from all 751,012 recreational boats with valid
Michigan registration in 1998 and compared to that of the respondents (Table 18). The
result shows that there is a fairly close agreement between the two distributions,
indicating no nonresponse bias on this variable.

The extent of the bias in reporting of spending related to recent trips is unknown.
The procedures employed in this study likely reduced this bias, but did not eliminate it.
If longer trips with larger spending were more likely to be reported, spending estimates
will be biased upward. Trips might be better represented if they were directly sampled
via on-site survey, although problems in obtaining representation across the wide range

of potential boating sites would likely create even greater potential for bias.

Table 18. Comparison of Boat Type between Sample and Population

 

 

Sample Population

Inboard a 22% 22%
Outboard 42% 43%
Sail 4% 5% '
Pontoon 10% 10%
Canoe/Row 9% 9%
PWC 12% 1 1%
Number of Observations 971 751,012

 

a. In/Outboard is included.

63

Storage Segment, Boat Owners, and Boats

Storage Segment

 

Four storage segments were constructed based on "type Of storage facility" and
"type of storage location". Boats are classified first based on the type of storage facility
(Table 19). The majority (64%) of the fleet are stored at permanent homes and 26% are
kept at seasonal homes including cottages. Only 7% of the total registered boats are kept
at marinas including yacht/boat club. Boats are also classified by the type of storage
location. The majority of the fleet are stored at nonwaterfront sites (43%) or inland lakes
waterfront sites with no Great Lakes access (39%). Boats stored at waterfront sites with

access to the Great Lakes account for 16% of the total craft registered in Michigan.

Table 19. Boat Classification by Storage Facility & Location

 

 

Type of Boat Storage by Facility and Location %
TYPE OF FACILITY
Permanent residence 63.8
Cottage or second home 25.7-
Public marina 1.4
Commercial marina 2.5
Owned space in marina 0.8
Yacht/boat club 1.8
Other 3.9
100.0
TYPE OF LOCATION
A waterfront site w/ access to the Great Lakes & connecting waters 15.6
An inland lake waterfront site (no Great Lakes access) 39.3
A river or stream waterfront site (no Great Lakes access) 2.2
A nonwaterfront site 42.9
100.0

 

a. The sample of boats was weighted to the number of active boats registered in Michigan.

The resulting segments are marina, waterfront primary home, waterfront seasonal
second home, and nonwaterfront home (Table 20). F orty-two percent of the fleet
registered in Michigan are stored at nonwaterfront sites. Boats kept at waterfront primary
and seasonal second homes account for 28% and 23%, respectively. Boats kept at marina

account for 7% of the total fleet.

Table 20. Distribution of Storage Segments for Sample and Population 3

 

 

W Population
Number of Boats % Number of Boats %
Marina 166 18.0 44,758 6.9
Waterfront Primary Home 263 28.5 180,964 27.8
Waterfront Seasonal Home 208 22.6 151,030 23.2
Nonwaterfront Home 285 30.9 274,871 42.2
SUM 922 100.0 651,623 100.0

 

a. Case with missing storage facility or storage location information were excluded from the analysis, but
population estimates were adjusted to yield the total number of registered active boats within each

segment.

Descriptions for Boat Owners
Boat owners’ socioeconomic characteristics are compared among the different
storage segments (Table 21). Boat owners are considerably older than Michigan’s

population and have significantly higher incomes. The median age of boat owners is 50

65

and the median household income is just over $60,000 a year. Sixty-five percent of all
boat-owing households have no children residing in the household. About 70% of the
boat owners own a single boat, While 23% own two boats and 8% more than two.
Approximately 30% of boat owners own seasonal homes.

Owner characteristics vary significantly with type Of boat storage. Age, number
of boats, household income, and seasonal home ownership differ significantly across
storage types at the 0.01 level. Around 45% of the boat owners in waterfront seasonal
home segment are over 60 years of age, compared with 20% of boat owners storing their
boats at nonwaterfront home. Almost half of the boat owners who keep the boats at
waterfront homes own more than one boat, While only 20% of the boaters in marinas and
nonwaterfi'ont homes possess more than one boat. Boat owners in marina segment have
the highest incomes, followed by boaters storing their boats at waterfront homes. The
average numbers of children and adults in the household also differ among the segments

at the 0.05 level.

Boat-related Characteristics

 

The active registered fleet is made up mostly of smaller craft (Table 22). About
85% of all active boats are 20 feet or less in length. Only 3% of registered craft are over
28 feet. More than 40% of all registered boats are outboards, followed by inboards
(including in/outboard) which takes up 22% of the fleet. PWC makes up 12% of the fleet
and pontoon 10%. More than 50% Of the registered active boats have been owned 5
years or less, While 8% of the boats have been owned more than 20 years. MOre than

80% of boats are kept at inland lakes waterfront or nonwaterfront sites. Boats stored at

66

waterfront sites with access to the Great Lakes account for 16% of the fleet. Only 2% of
active registered boats are kept at river or stream waterfront sites.

Boat type and size vary Significantly with type of boat storage. All measured boat
characteristics are significantly different among the segments at the 0.01 level except
years of boat ownership. About 65% of boats kept at nonwaterfront sites are less than 16
feet long, including PWC. More than three fourths of boats kept at marinas are longer
than 20 feet and 17% of boats at waterfront primary home are larger than 20 feet. Almost
60% of boats at marinas are inboards (or in/outboard), while 60% of boats stored at
nonwaterfront sites are outboards. Pontoon and outboard boats are used most popularly
by boat owners in waterfiont primary home. About 83% of boats stored at marina are
connected with the Great Lakes, While more than three fourths of boats at waterfront

homes are located in inland lakes.

67

Table 21. Boat Owner Characteristics by Storage Segment: 1998 a

 

 

 

 

 

 

Total Marina Waterfront Waterfront Non- F or x2 P
Primary Seasonal waterfront Value
Home Home Home
column percent

AGE OF BOAT OWNER N=922 N=166 N=263 N=208 N=285 x2=65.97 0.000
younger than 41 25% 20% 19% 7% 35%
41 — 50 26% 28% 21% 25% 28%
51 — 60 20% 20% 27% 24% 16%
61 — 65 11% 17% 9% 23% 7%
66 -— 7O 7% 7% 7% 9% 7%
Older than 70 10% _7_% 16% 12% My
100% 100% 100% 100% 100%

CHILDREN IN HOUSE =922 N=166 N=263 N=208 N=285 x2=21.07 0.049
0 65% 62% 67% 70% 63%
1 12% 15% 14% 14% 10%
2 14% 16% 11% 12% 17%
3 6% 6% 8% 2% 7%
4 or more %’ 1_% fl 1_% 3%
100% 100% 100% 100% 100%

ADULTS IN HOUSE N=922 N=166 N=263 N=208 N=285 x2=20.97 0.031
1 12% 21% 9% 14% 10%
2 74% 73% 80% 69% 72%
3 10% 1% 6% 12% 12%
4or more &I 2% a) 5_°/2 2°19
100% 100% 100% 100% 100%

NUMBER OF BOATS N=756 N=137 N=212 N=172 N=235 78:81.11 0.000
1 69% 78% 51% 52% 81%
2 23% 17% 36% 34% 14%
3 or more 8% 5% 12% 14% _5_%
100% 100% 100% 100% 100%

HOUSEHOLD INCOME N=828 N=149 N=240 N=177 =262 x2=88.43 0.000
under $20,000 8% 8% 7% 3% 10%
$20,000-$39,999 19% 1 1% 17% 19% 22%
$40,000-$59,999 21% 1 1% 14% 22% 26%
$60,000-$99,999 33% 39% 45% 26% 30%
$100,000-$ 149,999 12% 22% 9% 16% 1 1%
over S 150,000 6% 10% 7_% 14% 2X1
100% 100% 100% 100% 100%
SEASONAL HOME (MI) N=918 N=165 N=263 N=205 N=285

29% 11% 9% 89% 18% F=219.4 0.000

yes

 

a.

68

Unit of analysis in this table is the boat owner. The sample of boats was weighted inversely to the
number of boats owned by each respondent. .

Table 22. Boat Characteristics by Storage Segment: 1998 a

 

 

 

 

 

 

 

Total Marina Waterfront Waterfront Non- X2 P
Primary Seasonal waterfront Value
Home Home Home
column percent

SIZE OF BOAT N=922 N=166 N=263 N=208 N=285 x2=505.9 0.0000
PWC 12% 3% 14% 13% 12%
< 16' 41% 9% 34% 36% 53%
16-20' 33% 12% 35% 41% 31%
21 - 28' 12% 45% 15% 10% 5%
28+ .32) 31% a 9% 9°40
100% 100% 100% 100% 100%

TYPE OF BOAT N=919 N=166 N=262 N=206 N=285 x2=558.5 0.0000
Inboard 5% 25% 7% 4% 2%
In/Outboard 17% 33% 16% 2 1% 12%
Outboard 42% 12% 28% 37% 59%
Sail 4% 20% 4% 7% 1%
Pontoon 10% 7% 2 1 % 14% 1%
Canoe or Row 9% 0% l 1% 4% l 1%
PWC 12% w 14% 13% 12%
100% 1 00% 100% 100% 100%

YEARS OWNED =922 N=166 N=263 N=208 N=285 78:14.89 0.2481
Less than 3 23% 24% 21% 18% 27%
3 to 5 29% 27% 34% 24% 28%
6 to 10 24% 24% 25% 26% 21%
11 to 20 17% 18% 12% 22% 16%
More than 20 8% 7_°/_o 8% 10% $46
100% 100% 100% 100% 100%

STORAGE LOCATION N=922 N=166 N=263 N=208 N=285 X2=1 ,236 0.0000
Great Lakes 16% 83% 18% 23% 0%
Inland lakes 40% 12% 77% 76% 0%
River or stream 2% 5% 5% 2% 0%
Nonwaterfront site 42% 0_% 0% % 100%
100% 100% 100% 100% 100%

 

a. Unit of analysis in this table is the boat. The sample of boats was weighted to the numbér of active
boats registered in Michigan.

69

Statewide Boating Use and Spending (Objective 1)
Estimation of Boating Use by Storage Segment
Boating use is defined as the number of days a boat was underway in Michigan
waters in 1998. Annual days of use per boat is distinguished between day and overnight
trips. It was assumed that a day trip and a one night trip involve 1 boat day, while
overnight trips of two or more nights, the same numbers of boat days as the nights is
assumed. Total boat use is estimated by multiplying the average days of use per boat in

1998 by the total number of active craft registered in Michigan.

Annual Boating Use

Annual days of use can not be obtained directly fiom the wave surveys, as boaters
were asked to report activity in mid-season. Days of use were measured up to the time
the boater was surveyed. Procedures were needed to extrapolate from the mid-season use
estimates to season-end totals. Average days of use "so far" in the year were estimated at
nine points during the summer. The average use per boat in 1998 can be estimated by
extrapolating from these nine data points to the end of the season. A logistic curve was
used to estimate a relationship between days of use and the point in season when use was
measured.

The logistic function is well suited to the growth processes Where constraints to
grth or saturation effects are encountered at the beginning and end of the time period.
In the logistic model growth starts out slowly, increases to a maximum grth rate, and
then slows down again, eventually approaching some upper limit. Boating activity in

Michigan starts slowly during June and then increases rapidly during July and early

70

August, falling again as the end of the boating season approaches in late August and early
September.

The nine waves, covering nine weeks during the survey period from July 24 to
September 20, provide nine points of the average days of use from the beginning of the
survey to each point in time (Figure 4). The average use days grew from 18 as of July 24
to about 30 days by end of the season (Table 23). Since no measures were made for early
season use, it is important to select an appropriate starting point that should be added to
the existing data points in order for the logistic model to reflect the beginning of the
season. A knowledge of the use patterns during the season needs to be combined with
the nine empirical observations. The beginning of the season was assumed to be May 15.

The logistic equation is estimated with ten points by adding a zero for May 15.

 

25 WW
1

20 WWW W WW

 

 

Use Days

15 ;W ,

10 lWWW--W WWW. WW WWW WW . .. l

 

l 5 W--- -WWW- -W- . W _ ._ ___._W___

 

 

l 3 5 7 9 Jul. 13 15 17 Sep. 21 23 25

Figure 4. Average Days of Use by Wave During the Summer

71

Table 23. Average Use Days and 90% Confidence Intervals by Wave

 

 

Wave Lower Bound Mean Upper Bound
1 16.1 18.0 19.9
2 15.9 18.9 21.8
3 19.1 23.3 27.5
4 13.7 17.2 20.7
5 20.0 24.6 29.2
6 22.1 26.8 31.6
7 19.4 23.1 26.8
8 18.3 22.5 26.8
9 25.2 29.7 34.1

Overall 20.8 22.0 23.3

 

The logistic function is expressed as follows:

a7 l

7 =—————,
() ' k+(ab)’

where t is time and (i, is the estimate of use days at time t. The parameters that should be

estimated are k, a, and b. One of the parameters (b) in the parenthesis should be
restricted as being less than 1. A saturation level is determined by l/k. The inflection
point of the logistic curve can be obtained from the following calculation: inflection =
log, (ea / b), where e is the exponential. The function is nonlinear in the parameters and
therefore must be estimated using a nonlinear estimation procedure. The model was

estimated with the data in Table 23.

72

The estimated logistic equations for annual boating use fit the data quite well in
terms of the goodness of fit measures based on the t—values on the coefficients and R-
squares (Table 24). The saturation level is 28.4 days (l/k = 1/0.035). The inflection

point is 7.5 days [loge (ea / b) = log. ((25509 / 0.130)] and the corresponding wave is 7

(Figure 5). The average use per boat in 1998, al , can be estimated by extrapolating the
logistic curve to the end of the season. The annual use estimate of boat days can be

estimated by selecting an appropriate end-date and plugging this end-date into the model.
The end of the season was set at October 19 (t = 23) and the corresponding average days

of use was 28.2 days per boat:

~ 1
8 d = =
( ) ’3 0.035+(5.509*0.13)23

 

Table 24. Results of Logistic Models for Annual Use Days a

 

 

Lower Bound Me_an Upper Bound
Coefficient t-value Coefficient t-value Coefficient t—value
k 0.042 11.33 0.035 12.80 0.030 15.22
a 5.367 9.51 ‘ 5.509 8.68 5.359 1 10.85
b 0.135 8.79 0.130 8.57 0.132 10.59
R2 0.981 0.985 0.987

 

73

"r-“t‘rrrActual Use Days

9' --— Estimated Use Days (Upper Bound)
Estimated Use Days (Mean)
35 " “fl *0“— Estimated Use Days (Lower Bound)

 

30

 

25
20 y

15

Use Days

1 l 0

 

O . _,__ .. __ _ _ _ . -WLL _W.W .4

l 3 5 7 9 l 1 13 15 17 19 21 23 25
(May (May (Jun. (Jun. (Jul. (Jul. (Aug. (Aug. (Sep. (Sep. (Oct. (Oct. (Nov.
15) 29) 12) 26) 10) 24) 10) 24) 7) 21) 05) 19) 02)

Waves

Figure 5. Logistic Estimates of Boat Days

Logistic curve should be not sensitive to ending date as the curve becomes
asymptotic in this range. To determine whether the ending date was chosen .
appropriately, the variation around the end-point chosen is reported (Table 25). The
estimates of annual days of use do not vary with further selection of ending dates,
showing result is not sensitive to this choice. The upper and lower bounds around the
average use estimate, which were estimated using the 90% confidence intervals on the
sample average use "so far", are also depicted in Figure 5. It, however, should be noted

that these intervals around the mean are quite conservative due to two error sources from

74

Table 25. Variation of Annual Days of Use with Ending Dates

 

 

Alternative for Ending Date Upper Bound Mean Lower Bound
22 (Oct. 13) 32.4 28.1 23.3
23 (Oct. 20) 32.6 28.2 23.4
24 (Oct. 27) 32.7 28.3 , 23.5
25 (Nov. 03) 32.7 28.4 23.6
26 (Nov. 10) 32.8 28.4 23.6

 

the sampling error related to upper and lower bounds of the data used for the model and

the errors related to econometric specification.

Boating Use by Storage Class

The annual days of use for each storage segment is obtained by computing the
ratio of the projected annual average for all boats to the sample average use "so far"
(weighted mean on sample). This ratio is applied to each subgroup mean to extrapolate
to the end of season. This procedure assumes that the temporal patterns of use over the
season do not vary significantly by storage class and the samples for each storage class
are distributed evenly over waves. The even distribution by storage segment over waves
was achieved by using matched samples and there were no significant differences in
response patterns by storage class (See Appendix C). The expansion factor (E) of 1.28

was obtained from the following equation:

E _ Pr0jected Average Use _ 28.2 = 1.28

 

9 — _
( ) Sample Average Use 22.0

75

Average days of use vary substantially with storage segments, from 32 days for
marina boats to 15 days for boats stored at nonwaterfront homes (Table 26). The annual
use by segment is estimated by multiplying the average use by segment i obtained from

the in-season survey by the expansion factor, E:
(10) (Z. =d, *E,

where

A

d 1- = annual estimate Of use days per boat in segment i,

d; = annual days of use per boat obtained directly from the in-season survey, and

E = expansion factor.

Total days of use for ith segment is calculated by multiplying the annual use estimate per

boat for segment i by the total number of active boats in segment i.

Table 26. Average Days of Use "so far" by Storage Segment

 

 

Projected Mean Standard Standard ' % Error 3
( ‘91 ) (d, ) Deviation Error
Marina 40.5 31.62 28.92 2.27 12%
Waterfront Primary Home 38.8 30.30 28.99 1.81 10%
Waterfront Seasonal Home 29.2 22.77 21.37 1.50 1 1%
Nonwaterfront Home 18.7 14.59 13.48 0.81 9%
OVERALL 28.2 22.02 22.86 0.76 6%

 

a. Percent of sampling errors at the 90% confidence interval.

76

Boating Use between Day vs. Overnight Trips

Boaters were asked to report expenditures on the most recent boating occasion.
For boaters whose most recent trips were day trips, the spending they reported is per day
spending, whereas for boaters Whose most recent trips were overnight trips, the spending
reported is for the entire trip. Since boaters reported spending on a trip basis, but use was
estimated on a boat day basis, a conversion factor is needed to convert spending on trips
to a per day basis. The procedures for the conversion can be described using the
following formulas.

First, total days of use per boat for a given segment i can be defined as follows
(suppressing i for brevity hereafter until it is necessary):
(11) D=[td *l+(l—-td)*a]*T,
where for a given segment i
D = total boat days of use per boat during the season,
T = total number of trips per boat during the season,
rd = percent of trips that are day trips (l—td = percent that are overnight trips),'and
a = average length of stay on overnight trip (note that boat day for a day trip is 1.0).

The parameters, D, rd, and a, in equation (11) can be estimated from the survey
for each segment. Solving equation (11) for T:

D

(12) T: td +(1—td)*a°

 

The percent of days on day trips and overnight trips can be calculated as:

T*td
D

 

(13-1) P4 =

9

77

T*(l—td)*a
13-2 = ,
( ) p. D

 

where for a given segment i

pd = percent of days on day trips and

p0 = percent of days on overnight trips.

The parameters used to compute days on day and overnight trips for each storage

segment are presented in Table 27. For example, for the fleet as a whole, the annual use

per boat was 28.2 (D). Annual trips per boat were divided 92% day trips (rd) and 8%

overnight trips (l-td). An average length of stay on overnight trip was 5.1 days. Plugging

these values into equation (12) yields: T = D/[tar+(1-td)*a] = 28.2/[92%+8%*5.1] = 21.2.

The percents of days on day trips and on overnight trips may be computed as: pd = T *td/D

= 21.2*92%/28.2 = 69%,p0 = T*td/D = 21.2*8%*5.1/28.2 = 31%. The same procedure is

used to calculate trips for each storage segment.

Table 27. Boating Use by Storage Segment: 1998

 

 

Marina Waterfront Waterfront Non- TOTAL
Primary Seasonal waterfront
Home Home Home

Annual Days of Use per Boat (D) 40.5 38.8 29.2 18.7 28.2
Boat Days per Overnight Trip (a) 5.8 4.5 3.8 5.4 5.1
Percent of Day Trips (Id) 78% 98% 95% 75% 92%
Percent of Overnight Trips (l-td) 22% 2% 5% 25% 8%
Annual Trips per Boat (I) 19.7 36.3 25.6 8.9 21.2
Days on Day Trips (pd) 38% 92% 84% 36% 69%
Days on Overnight Trips (po) 62% 8_% 16% 64% 31%

100% 100% 100% 100% 100%

 

 

78

Fleet Totals

Annual use of 28.2 days per boat was multiplied by 652,000 active pleasure craft
with valid Michigan registrations in 1998 to yield a total of 18.4 million boat days in
Michigan in 1998 (Table 28). Dividing the total boat days by an average boat days per
trip of 1.3 yields a total of 13.8 million boating trips in 1998. About 12.7 million (92%)
of them were day trips and 1.1 million (8%) were overnight trips. Total boat days on
overnight trips is 5.7 millions (5.1 boat days per overnight trip times 1.1 million

overnight trips).

Table 28. Summary of Boating Activity by Storage Segment: 1998

 

 

 

 

 

 

 

Marina Waterfront Waterfront Non- TOTAL
Primary Seasonal waterfront
Home Home Home -

Number of Boats (000's) 45 181 151 275 652
(%) (7%) (28%) (23%) (42%) (100%)
Average Days of Use per Boat 40.5 38.8 29.2 18.7 28.2
Total Boat Days (000's) 1,813 7,023 4,404 5,136 18,376
(%) (10%) (38%) (24%) (28%) (100%)
Total Boat Trips (000's) 883 6,573 3,900 2,436 13,792
(%) (6%) (48%) (28%) (18%) (100%)
Total Trips on Day Trips (000's) 690 6,445 3,719 1,829 12,683
Total Trips on Overnight Trips (000's) 193 128 181 607 1,109
Total Days on Day Trips (000's) 690 6,445 3,719 1,829 12,683
Total Days on Overnight Trip (000's) 1,123 578 685 3,306 5,692
Day Trips 78% 98% 95% 75% 92%
Overnight Trips 22% % 5_% 25% 8%)

100% 100% 100% 100% 100%
Days on Day Trips 38% 92% 84% 36% 69%
Days on Overnight Trips 62% % 16% 64% 31%

100% 100% 100% 100% 100%

 

79

Boats stored at nonwaterfront homes accounted for 42% of active craft in 1998.
However, the greatest contributor of boat days to the statewide total was boats stored at
waterfront primary home, making up 38% of the total days of use. Boats stored at
marinas were the most active as they averaged 41 days of use per boat, divided as
follows: 38% on day trips and 62% on overnight trips. Boats kept at nonwaterfront
homes were least active averaging 19 boat days per boat, but their distribution of boat
days between day and overnight trips was similar to that of marina boats. Boats stored at
waterfront primary and seasonal homes averaged 39 and 29 days in 1998, respectively.
These boats took a much higher percentages of day trips (92% and 84% for waterfront
primary and seasonal homes, respectively).

Boats stored at waterfront permanent homes were the greatest contributor to total
trips. They accounted for 48% of the total, followed by waterfront seasonal home
segment which took 28% of the total trips. Ninety-two percent of all boating trips
involved a day trip. Compared with boats stored at waterfront homes, those stored at
marinas and nonwaterfront sites are less likely to involve a day trip. Seventy-eight
percent of trips by marina boats and 7 5% of trips by nonwaterfront boats were day trips,
whereas more than 95% of the trips by boats stored at waterfiont primary and seasonal

homes were day trips.

Comparison of Wave vs. End-of-Season Surveys on Use Estimates
In this section the wave survey is evaluated. For the representativeness of each of
the nine wave surveys, the sampling ratios by boat size class and region of the subsample

being sent at each wave were exactly the same as those of the overall sampling scheme.

80

To test if the composition of any wave survey is significantly different from other wave
surveys on characteristics of interest, F -ratio and )8 test are used. A hypothesis of no
differences among the wave surveys for a given variable is tested. The variables tested
are: boat length and type, storage type, and the characteristics associated with trip and
spending. If the variables are not significantly different over the wave surveys, one
would say that the surveys are representative. The results indicate that for all the key
variables there are no significant differences across the nine wave surveys (Appendix C).
Each wave survey was representative.

Boaters averaged 28.2 days on the water in 1998. The annual days of use was
estimated from the wave surveys using the logistic curve. Since estimating annual boat
days is essential to estimating total spending accurately, it is important to know if the
annual use estimate is in fairly close agreement with annual days of use estimated from a
season-end survey. Total use days per boat were also estimated from the end of season
survey. The approach is to compare the confidence intervals for mean use days estimated
from the wave surveys and season-end survey, respectively. One can conclude that the
two estimates of use days are not statistically different, if the confidence intervals
overlap.

From the year-end survey, boaters averaged 30.3 days per boat on Michigan’s
water during the 1998 season (Table 29). The mean value for the season-end survey is
larger than the average use derived fi'om the logistic curve (28.2). Based on the 90
percent confidence intervals, however, the mean use days estimated from the season-end
survey is not significantly higher than average use estimated from the wave surveys, as

the confidence intervals overlap. The results indicate that the wave approach was

81

Table 29. Mean & 90% Confidence Intervals on Use Estimates

 

 

Days of Use from Days Of Use from
Wave Surveys " Season-End Survey b
Upper Bound 32.6 33.3
Mean 28.2 30.3
Lower Bound 23.4 27.3

 

3. Since the logistic models were applied to the data set which includes mean days of use and confidence
intervals on them, the logistic estimates around the mean may not be symmetric. Because of two error
sources from upper and lower bounds of the data used for the model and the econometric specification
itself, the resulting confidence interval around the mean would be very conservative.

b. The sample of boats was weighted to number of total active boats. Standard errors thus were
computed using the standard deviations item the sample weighted and the number of cases.

successful in estimating annual days of use. The estimate from the season-end survey,

however, is also subject to error.

Estimation of Boater Trip Spending by Storage Segment
Average Trip Spending
An average trip spending per boat day may be computed as a weighted average of
the average spending per day on day trips and overnight trips. Trip spending per day per
boat for segment i on spending category j is estimated using the following equation:
(14) r70. = pd, *mdl.j +pm. *moij ,
where
272,]. = spending per boat day in spending category j by segment i,
mdij = spending per boat on day trips in spending category j by segment i,
moi,- = spending per boat on overnight trip in category j by segment i, and

pat), p0,- : percents (weights) of days on day and overnight trips in segment i.

82

Statewide trip spending by recreational boaters is obtained by summing Mi,- over

segment i ’s and spending category j’s: ‘

(15) “21.23%, i=1,2,3,4, j=1,2,...,11,

where M is total statewide trip spending by registered boaters in Michigan during the
1998 boating season.

The average registered boater spent $975 on boating trips in 1998. Spending
varied from $788 for boaters storing their boats at waterfront primary home to $3,087 for
marina boaters (Table 30). A typical boater spent $35 a boat day in 1998. Boaters
storing their boats at marinas spent $76 a boat day which is more than two times as much
on trip Spending comparing with an average boater. Boaters storing their boats at
nonwaterfront sites spent $44. Boaters boating from waterfront seasonal homes spent
$29 per day. Boaters boating from waterfront primary homes Spent $20 a day. A typical
boater spent $23 a day on day trips and $60 a day on overnight trips. The highest
Spending per day was for trips involving overnight stays by marina boats ($91 per day),
whereas the least spending per day was $15 on day trips by boats stored at waterfront
primary homes.

Sampling errors for estimates of trip expenditures are also reported. Sampling
error depends on the size of the sample and the degree of variation in the population.
How close a sample estimate is to the population value can be determined by confidence
intervals. Errors were reported using 90% confidence intervals. This means that the true
figure will lie within plus or minus this tolerance of the reported mean with ai90%
confidence level. That is, based upon sampling error alone, there is a 10 percent chance

the true spending figure will lie outside of this confidence interval.

83

Table 30. Average Trip Spending by Storage Segment & Type of Trip (Summary)

 

Marina Waterfront Waterfront Non- TOTAL
Primary Seasonal Waterfront
Home Home Home

 

(1) PER BOAT PER DAY TRIP SPENDING FOR DAY TRIPS

Mean 52.6 15.0 27.7 30.8 4 23.1
Lower Bound a 37.8 10.9 20.8 25.7 20.0
Upper Bound a 67.4 19.1 34.6 35.9 26.2
% Sampling Error a 28% 28% 25% 17% 13%
(2) PER BOAT PER DAY TRIP SPENDING FOR OVERNIGHT TRIPS
Mean 90.7 79.0 36.9 51.4 60.2
Lower Bound 75.6 51.1 15.4 40.4 52.2
Upper Bound 105.8 106.9 58.4 62.4 68.2
% Sampling Error 17% 35% 58% 21% 13%
(1+2) PER BOAT PER DAY SPENDING (W EIGHTED AVERAGE OF TWO TRIPS)
Mean 76.2 20.3 29.1 44.1 34.6
Lower Bound 64.8 16.0 22.5 39.3 31.6
Upper Bound 87.6 24.6 35.7 48.9 37.6
% Sampling Error 15% 21% 23% 11% 9%
TOTAL $ PER BOAT PER YEAR 3,087 788 849 823 975

 

a. Sampling errors for estimates of trip expenditures are reported with a 90% confidence level.

For the fleet as a whole, sampling error of trip expenditures is 9%, indicating that
the population mean of boater trip expenditure is contained in the interval $32 to $38 at
the 90% confidence level. Estimates for particular segments are subject to larger errors
due to the smaller sample sizes. A spending estimate for boaters staying overnight
launched from their seasonal waterfront homes has the largest errors of 58%, followed by
sampling errors for overnight users in waterfront permanent homes (35%). Except for
these two groups, trip spending errors within particular segments are all less than 30
percent. Spending of nonwaterfront segment on day trips marina group on overnight trips

has the smallest errors of 17%, respectively.

84

The error structure changes somewhat for segments by trip type and storage.
Sampling errors for estimates of trip expenditures between day and overnight trips,
however, are the same. For most applications, these levels of accuracy appear to be
adequate. Efforts to reduce recall errors and related problems would appear to be more
useful than increasing sample sizes. Trip spending here was estimated on a per boat day
rather than an annual or trip basis. Standard errors were computed using standard
deviations of population means and square roots of case numbers (Appendix D).

Food constitutes 37% of the trip spending, divided 21% for groceries and 16% for
restaurant (Figure 6). Boat fuel accounts for 26% of the total, followed by auto gas
which takes up 9% of the boater trip budget. Maintenance expense associated with
boating trips contributes 7% of boater trip Spending. Average trip spending by various
spending categories are reported in Appendix E.

Average trip spending can also be summarized on the basis of where spending
occurred- near home or away from home (Table 31). The $35 per day spent on boating
trips was divided evenly between spending near home and more than 20 miles from
home. Boaters storing their boats at a marina or nonwaterfront home spent more money
away from home (73% and 68%, respectively), whereas boaters who boated from
waterfront homes spent more near home (75% and 81%, respectively). Average trip
spending by various spending categories are reported in Appendix E. While boaters
purchased more boat fuel near home than away from home, their spending on food was
divided evenly between near and away from homes. Food was the largest item for boats
stored at a marina, waterfront seasonal home, or nonwaterfront Site, while boats kept at

waterfront primary home spent more money on boat fuel.

85

 

( Temporary DOCkage Pump-out/ Launch
4% ’ 1%
Marine Supplies Recreation Others
4% 4% 4%
Boat Fuel
260/0

   

Shopping

4%

     
  

Groceri
Restaurant es
2 1 %

16%

Figure 6. Boater Trip Spending by Spending Category: 1998

Table 31. Average Trip Spending by Storage & Spending Location (Summary)

Marina Waterfront Waterfront Non- TOTAL
Primary Seasonal waterfront

 

Home Home Home

AVERAGE SPENDING PER DAY
$
(1) "Near Home" Spending 20.5 15.2 23.7 14.2 17.5
(2) "Away Form Home" Spending 55.7 5.1 5.4 29.9 17.1
(1+2) Average Spending on Trip 76.2 20.3 29.1 44.1 34.6
%
"Near Home" Spending 27% 75% 81% 32% 51%
"Away Form Home" 13% fl fit 68% 49%

100% 100% 100% 100% 100%

 

Total Trip Spending

Total trip spending was estimated by multiplying spending per boat day by the
number of boat days. Boaters with registered boats in Michigan spent an estimated $635
million on trips in Michigan waters in 1998 (Table 32). Boats stored at nonwaterfront
sites accounted for 36% ($226 million) of the statewide total. The rest of the total was
divided about equally among boats stored at marinas, waterfront primary homes, and

waterfront seasonal homes.

Table 32. Total Trip Spending by Storage Segment & Type of Trip: 1998

 

Marina Waterfront Waterfront Non- Total
Primary Seasonal Waterfront
Home Home Home -

 

(1) TOTAL TRIP SPENDING FOR DAY TRIPS ($ MILLIONS)

 

 

Subtotal 36.3 96.9 102.9 56.3 292.4
Percent of Spending by Segment 12% 33% 35% 19% 100%
(2) TOTAL TRIP SPENDING FOR OVERNIGHT TRIPS ($ MILLIONS)
Subtotal 101.9 45.6 25.3 169.9 342.7
Percent of Spending by Segment 30% 13% 7% 50% 100%
(1+2) TOTAL STATEWIDE SPENDING ON TRIPS ($ MILLIONS)
Total 138.2 142.5 128.2 226.2 635.1
Percent of Spending by Segment 22% 22% 20% 36% 100%
Trip Spending by Trip Type ($ MILLIONS)
Trip Spending on Day Trips 36.3 96.9 102.9 56.3 292.4
Trip Spending on Overnight Trips 101.9 £8 E 169.9 342.7
138.2 142.5 128.2 226.2 635.1
Trip Spending by Trip Type (%) ’
Trip Spending on Day Trips 26% 68% 80% 25% 46%
Trip Spending on Overnight Trips 74% 32% 20% 75% 54%
100% 100% l 00% 100% 1 00%

 

87

Forty-six percent of trip spending is on day trips and 54% on overnight trips. Of
the $292 million spent on day trips, boaters Who stored their boats at waterfront homes
spent $208 million, divided about equally between primary and seasonal homes, whereas
boaters who kept their boats at marinas spent only $36 million. Of the $342 million on
overnight trips, however, boaters who kept their boats at waterfront primary and seasonal
homes accounted for only $71 million, while boaters who stored their boats at
nonwaterfront homes spent $170 million. Boaters who kept the boats at marinas spent
$102 million on overnight trips. Total trip spending by spending category is reported in
Appendix F.

Total Spending on boating trips is also reported by where the spending occurred in
Appendix F. The total spending of $635 million was divided about evenly between trip
spending near home and away from home. The distribution of Spending between the two
spending locations varies across storage types, but is in a fairly close agreement with the
distribution between day trips and overnight across storage segments. Day users spend
more near home and overnight boaters spend more money away from home. Boats
stored at marinas and nonwaterfront homes spent more money away from home (73%
for marina boats and 68% for boats in nonwaterfi'ont homes). On the contrary, boats
stored at waterfront homes spent far less away from home.

Differences in the contribution of each storage segment to the number of boats,
days of use, and trip spending are depicted in Figure 7. Boats stored at nonwaterfront
homes make up 42% of the fleet, but account for smaller percentages of boating activity
and spending. Marina boats are the smallest contributors. to the statewide fleet and

boating activity, but constitute the second largest portion of total trip spending.

88

m"1§6éts”iii=léei F

50% ' 'fi‘i .Total Boat Days m“? 'v -—- * *'--— - '****

l . , DTotal Trip Spending

 

 

20% .

10% ..

 

 

 

 

 

0%‘

M arina Waterfront Primary Waterfront Nonwaterfront
Home Seasonal Home Home

Figure 7. Boats, Boating Use & Trip Spending by Storage: 1998

Comparing to boating activity and the fleet, boats launched fiom waterfront primary
home make up a smaller contribution to boater spending on trips. Boaters boating from
waterfront seasonal homes account for same percentages of each to the fleet, boating

activity, and spending.

Handling of Zeros and Missing Values on Trip Spending

The differences in spending patterns between day and overnight trips can be
partly explained by the difference in percent of boaters not spending money on their most
recent trip between the two types (Table 33). About 45% of all trips involved no

spending. The difference in spending, however, is significant between two types of

89

Table 33. Percent of Boaters Spending Nothing on Recent Trip by Trip Type

 

 

Total Day Overnight t- P

Use Trip ratio Value
TRIP SPENDING N=921 N=764 N=157
Boat Fuel 51.1% 56.9% 17.1% 8.38 0.000
Temporary Dockage 97.1% 99.1% 85.1% 13.04 0.000
Pump-out & Launch 93.2% 94.9% 83.0% 1 1.59 0.000
Repair & Maintenance 95.7% 97.5% 85.4% 5.50 0.000
Marine Supplies 90.5% 92.6% 78.4% 6.48 0.000
Restaurant 81.4% 87.4% 46.1% 16.62 0.000
Groceries 69.5% 78.2% 18.2% 15.84 ' 0.000
Auto Gas 73.7% 80.9% 31.2% 12.79 0.000
Shopping 94.4% 99.3% 65.9% 16.64 0.000
Recreation 93.7% 97.4% 72.3% 1 1.10 0.000
Other Expenses 91.6% 94.2% 76.3% 6.76 0.000
TOTAL 44.6% 50.7% 8.7% 10.73 0.000

 

boating trips at the 0.01 level. Less than 9% of overnight trips involved no Spending,
while more than 50% of day trips involved no spending. Day trips are less likely to
require any trip expenses, as day users are more likely to boat near home. -

Boaters on day trips were less likely to spend money on shopping, temporary
dockage, repair, or recreation. For day trips, spending on boat fuel takes place most
frequently (43%), followed by spending on groceries (22%) and auto gas (19%). Boaters
away from home to stay overnight spend money on boat fuel (83%) and groceries (82%)
most frequently, followed by auto gas (69%) and restaurant meals (54%).

The likelihood of spending money on a particular trip is also significantly
different across the segments at the 0.01 level (Table 34). More than 60% of the boaters

who stored their boats at waterfiont homes spent nothing on their most recent trip, while

90

Table 34. Percent of Boaters Spending Nothing on Recent Trip by Storage Type

 

 

Total Marina Waterfront Waterfront Non- F- ‘ P
Primary Seasonal waterfront ratio Value
Home Home Home

TRIP SPENDING N=894 N: 148 N=259 N=206 N=281
Boat Fuel 52.3% 44.5% 63.0% 66.4% 38.5% 16.79 0.000
Temporary Dockage 97.0% 87.3% 99.3% 99.5% 95.6% 13.67 0.000
Pump-out & Launch 93.3% 84.2% 95.9% 99.2% 89.6% 24.38 0.000
Repair & Maintenance 96.0% 95.5% 98.6% 97.0% 93.7% 1.58 ‘ 0.193
Marine Supplies 91.4% 81.2% 95.3% 93.5% 89.1% 8.66 0.000
Restaurant 81.9% 55.4% 92.1% 90.4% 74.4% 38.35 0.000
Groceries 70.6% 42.9% 90.7% 80.6% 56.0% 42.04 0.000
Auto Gas 74.6% 61.5% 95.3% 88.0% 55.5% 59.64 0.000
Shopping 95.0% 82.0% 98.6% 99.5% 92.1% 23.86 0.000
Recreation 93.8% 91.8% 97.6% 97.1% 89.7% 5.27 0.001
Other Expenses 91.5% 90.8% 98.0% 98.1% 83.6% 18.50 0.000
TOTAL 45.9% 30.4% 61.0% 62.6% 28.9% 37.45 0.000

 

only 30% of the boaters keeping their boats at marinas and nonwaterfi'ont homes spent
nothing. The result indicates that boaters who store their boats at waterfront homes may
more likely take their boats out Without incurring any expenses on a particular boating
occasion. Spending frequencies on auto gas and food differ most significantly. Marina
boaters are more likely to purchase food, while boaters from nonwaterfront homes are
more likely to buy fuel.

The results indicate the importance of identifying the types of trips and obtaining
a representative sample of boating trips when estimating spending. Boaters still may
have been more likely to report overnight trips and more extended outings, but the
procedures adopted to handle zeros and missing values on trip expenditures in this study

likely reduced the bias associated with trip spending reports.

91

Test of Differences in Use and Spending by Storage Segments (Objective 2)

Michigan's registered boating fleet were divided into four segments: marina,
waterfront primary home, waterfront seasonal home, and nonwaterfront home. Spending
estimates were reported for each segment. Differences in boating activity and spending
across storage types are tested using one-way ANOVA for interval-level variables and x2
test for nominal-level variables. The null hypothesis is that the population means or
distributions are all equal. Variables associated with trip and spending are compared
across the segments.

While all measures reported were estimated from the weighted sample, F and x2
values were computed using the sample unweighted. Alternatively, F value can be
calculated based on the sums of squares for between- and within-groups obtained from
the sample weighted and the within-group degrees of freedom obtained by subtracting the
number of subgroups from the number of cases. Another alternative for F statistics can
be obtained by multiplying F value computed directly from the weighted sample by the
ratio of the number of cases to the number of population. Outcomes of these three

approaches are identical.

Boating Activity by Storage Segment

Average boat days, Great Lakes boat days, and percent of overnight trips all
differed across storage types Significantly at the 0.01 leVel (Table 35). On average, boats
kept at marinas (41 days) and waterfront primary homes (39 days) are used more
frequently than boats kept at waterfiont seasonal homes (29 days) and nonwaterfront sites

(19 days). About 80% of the days by boats at marinas (33 days) occur on Great Lakes,

92

Table 35. Boating Activity by Storage Segment: 1998

 

 

Total Marina Waterfront Waterfront Non- F P

Primary Seasonal waterfront Value

Home Home Home

BOATING ACTIVITY

Boat Days per Boat N=902 N=163 N=257 N=202 N=280
28.2 40.5 38.8 29.2 18.7 F=30.90 0.000

GL Days per Boat N=911 N=163 N=259 N=206 N=283
7.4 32.7 4.4 5.0 6.5 F=78.16 0.000

% Overnight Trip N=903 N=161 N=257 N=206 N=279
8% 22% 2% 5% 25% F=39.51 0.000

 

while only 10% of the days by boats at waterfront primary home (4 days) take place in
inland lakes. More than 20% of the trips by boats kept at marina and nonwaterfront
homes are overnight trips, while less than 5% of the trips by boats stored at waterfront
homes are overnight.

Distance traveled to destination or storage location is important in des'cribing trip
expenditures as boaters are more likely to stay overnight and spend more on longer trips.
The variables associated with distances (from home to destination, from storage to
destination, and from home to storage) differ significantly among storage types at the
0.01 level (Table 36). Boats stored at nonwaterfront sites are transported greater
distances (69 miles) to the locations where they are used than boats kept in other types of
storage. In general the locations where boats are kept and used are similar for boats

stored at marinas and waterfront homes.

93

Table 36. Distance Traveled and Body of Water Used by Storage Segment: 1988

 

 

 

 

 

Total Marina Waterfront Waterfront Non- F or x2 P
Primary Seasonal waterfront Value
Home Home Home
DISTANCE TRAVELED (one way distance miles)
Home to Destination N=812 N=151 N=258 N=138 N=265
75.8 54.7 28.5 149.7 78.2 F=100.0 0.000
Storage to Destination N=902 N= 165 N=258 N=205 N=274
45.7 28.8 31.2 27.0 69.1 F=74.84 0.000
Home to Storage N=827 N=150 N=263 N=138 N=276
57.7 47.2 31.5 150.5 35.9 F=212.4 0.000
BODY OF WATER USED N=901 N=162 N=257 N=202 N=280 x2=251.1 0.000
Great Lakes Only 22% 76% 16% 17% 19%
Inland Lakes Only 62% 18% 76% 73% 53%
Both GL and IL 17% 8% Q 10% 28%
100% 100% 100% 100% 100%

 

Great Lakes craft tend to be larger than craft used on inland lakes, suggesting
some likely differences in spending patterns. The majority of boats use either Great
Lakes (22%) or inland waters (62%) exclusively. Seventeen percent of boats use both
Great lakes and inland waters. Boaters' choices of destinations vary significantly with
type of boat storage. For boats kept at marinas, the majority of boats (76%) use Great
Lakes exclusively, while the majority of boats at waterfront homes use inland waters
exclusively (76% for primary homes and 73% for seasonal homes). More than a half of
boats stored at nonwaterfront sites are used on inland waters solely. Less than 10% of

boats stored at marinas and waterfront homes, respectively, use both Great Lakes and

94

inland waters, while 28% of boats kept at nonwaterfront homes use both Great Lakes and

inland waters.

Boater Trip Spending by Storage Segment

The variables associated with spending on trip (percent of boaters spending
nothing on their most recent trip, spending per trip, and spending per day) differ
significantly with different storage types at the 0.01 level (Table 37). Approximately
30% of boats at marinas and nonwaterfront sites reported no spending on their most
recent trips, whereas more than 60% of boats stored at waterfront primary and seasonal
homes spent no money on the most recent trips. Marina boaters spend more money per
trip than boat owners in other storage categories. Owners of boats stored at marinas
spend $156 per trip, followed by owner of boats at nonwaterfront sites who Spend $93 per
trip. At the other extreme owners of boats kept at waterfront primary homes spent $22 on

an average boating outing.

Table 37 . Trip Spending by Storage Segment: 1998

 

Total Marina Waterfront Waterfront Non- F P
Primary Seasonal waterfront Value
Home Home Home

 

%Zero Spending on Trips N=894 N=148 N=259 N=206 N=281
46% 30% 61% 63% 29% F=37.45 0.000

Spending per Trip N=915 N=163 N=262 N=206 N=284

$46 $156 $22 $33 92.9 F=27.12 0.000
Spending per Day N=909 N=161 N=26l N=205 N=282

$35 $76 $20 $29 $44 F=26.86 0.000

 

95

Regional Distribution of Boater Trip Spending (Objective 3)’

To assess the economic effects Of boater trip spending on regions of the state,
flows of money into and out of each region must be estimated. Spending by non-local
residents who are attracted to the area is of interest. The net contribution of boater
spending to each region can be determined by subtracting the amount of spending by
boaters living there from the amount of spending occurring in the region as a boating
destination. This procedure omits trip spending within the region by boaters who live

there.

Spatial Distribution of Boats and Boating Use

The questionnaire asked boaters to report their spending "near home" and "away
from home". To allocate boater spending to different regions of the state, three
functional regions are identified: (1) the origin region where the boater lives, 1(2) the
storage region where the boat is stored, and (3) the destination region in which the boat is
used. Storage type is the basis for assigning spending to one of these functional regions.
For boats stored at a marina, primary waterfront home, or non-waterfront home, spending
"near home" is assumed to occur in the region where boaters live. Spending "away from
home" is assumed to occur in the destination region. For boats kept at seasonal homes,
near home spending is assigned to the storage region and spending away from home is
assigned to the destination region.

Michigan’s ten boating regions are shown in Figure 8. The regions include six
coastal regions, two inland regions, two Upper Peninsula regions. The boating regions

were adapted from Stynes and Safronoff (1982) to separate inland and coastal counties

96

 

    
 
         
   
  
   
 

U.P. North

U.P. South

I, Northeast

 

Northwest

1

 

 

 

West Central

 

Southwest Southeast

 

 

 

 

 

 

Figure 8. Boating Regions in Michigan

97

and capture broad flows from inland to coastal areas and from south to north. The net
gain of boater spending to a region can be defined as the difference between boater trip
spending occurring Within the region and total spending on boating trips of bOaters who
reside in the region. To allocate the spending among regions where trip purchases
occurred and among regions of boater residence, the number of boats by residence
regions and storage regions and total boat days by destination region must be estimated
by storage segment.

The 1998 survey instrument asked for both the county Where the boat was kept
and where the boater lived during the boating season. Distributions of boats by region of
boater's residence (region of boat storage for seasonal home segment) and storage type
were estimated directly from the weighted sample (Table 38). The questionnaire also
asked for the county where the boat was used most recently, as well as for the total
number of days the boat was used in Michigan waters "so far". Distributions of boat days
by region of destination and storage type were also directly estimated from the weighted
samples. This assumes that the distribution of use based on county the boater used most
recently is representative of overall use and the temporal patterns of use over the season
do not vary significantly for a given storage segment.

Nearly 50% of total registered boats in Michigan are owned by boaters living in
the south inland region of Michigan. More than 70% of the boats kept at marinas are
owned by boaters living in southeast and south inland of Michigan. More than 20% of
the boats stored at waterfront seasonal homes are Operated by out-of-state boaters. The
south inland region is also used most frequently by boaters, making up 40% of the total

boat days. For boats stored at marinas southeast Michigan is used most fiequently (3 8%).

98

Table 38. Distributions of Boats and Boat Days by Region & Storage: 1998

 

 

Marina Waterfront Waterfront Non- Total

Primary Seasonal waterfront

Home Home Home
BOATS BY RESIDENCE REGIONS (000's)
Southeast 16 (37%) 17 (10%) 26 (17%) 32 (12%) 92 (14%)
East Central 2 (5%) 2 (1%) 11 (7%) 16 (6%) 31 (5%)
Northeast 1 (1%) 10 (6%) 2 (1%) 9 (3%) 22 (3%)
Northwest 1 (3%) 13 (7%) 1 (1%) 12 (4%) 28 (4%)
West Central 3 (6%) 3 (2%) 4 (2%) 11 (4%) 21 (3%)
Southwest 3 (6%) ll (6%) 2 (2%) 10 (4%) ‘ 27 (4%)
South Inland 15 (34%) 100 (55%) 61 (40%) 139 (50%) 314 (48%)
North Inland 0 (1%) 15 (8%) 3 (2%) 22 (8%) 41 (6%)
SouthUP l (1%) l (1%) 2 (1%) 7 (3%) 10 (2%)
NorthUP 0 (1%) 7 (4%) 7 (4%) ll (4%) 26 (4%)
OutofState _2_ (%) 9 L0%) 22%) § 1%) E M
Total Boats 45 (100%) 181 (100%) 151 (100%) 275 (100%) 652 (100%)

BOATS BY STORAGE REGIONS (000's)

Southeast
East Central
Northeast
Northwest
West Central
Southwest
South Inland
North Inland
South UP
North UP
Out of State
Total Boats

20 (44%)
2 (6%)
3 (6%)
2 (5%)
5 (12%)
4 (8%)
6 (14%)
1 (1%)
1 (1%)
1 (2%)
9 M

45 (100%)

18 (10%)
2 (1%)
10 (5%)
l4 (8%)
3 (2%)
11 (6%)
99 (55%)
16 (9%)
1 (1%)
6 (4%)
9 roar
181 (100%)

BOAT DAYS BY DESTINATION REGIONS (000's)

Southeast
East Central
Northeast
Northwest
West Central
Southwest
South Inland
North Inland
South UP
North UP
Total Use

693 (33%)
132 (7%)
206 (11%)
80 (4%)
250 (14%)
194 (11%)
149 (8%)
38 (2%)
37 (2%)
3_4 £210
1,813 (100%)

492 (7%)
35 (0%)
403 (6%)
425 (6%)
92 (1%)
379 (5%)

4,518 (64%)
440 (6%)
61 (1%)
1_78 M

7,023 (100%)

5 (3%)
5 (3%)
22 (15%)
18 (12%)
3 (2%)
7 (5%)
33 (22%)
36 (24%)
9 (6%)
13 (8%)
1 rear
151 (100%)

129 (3%)
113 (3%)
550 (12%)
509 (12%)
48 (1%)
219 (5%)

1,137 (26%)
1,141 (26%)

209 (5%)
350 (8%)

«W (100%)

25 (9%)
l6 (6%)
17 (6%)
l4 (5%)
11 (4%)
12 (4%)
129 (47%)
30 (11%)
7 (3%)
11 (4%)
1 are
275 (100%)

506 (10%)
233 (5%)
477 (9%)
394 (8%)
300 (6%)
195 (4%)
1,519 (30%)
895 (17%)
201 (4%)
414 83%)

67 (10%)
26 (4%)
52 (8%)
48 (7%)
23 (4%)
34 (5%)
267 (41%)
83 (13%)
17 (3%)

- 31 (5%)
3. an
652 (100%)

1,823 (10%)
510 (3%)
1,636 (9%)
1,408 (8%)
688 (4%)
991 (5%)
7,329 (40%)
2,510 (14%)
506 (3%)
.97_3 Eye)

5533 (100%) 18,376 (100%)

 

99

The north inland region and the south inland region are frequently used by the boaters
boating from their seasonal homes. However, comparing to the number of boats owned
by seasonal home users living in the south inland region (40%), the number of boats
stored in seasonal homes in the region is less (22%). On the other hand, the north inland
region contains far more boats stored at waterfront seasonal homes (24%) than the
number of boats owned by seasonal home users living in this region (2%). Comparing to
the numbers of boats by residence regions, more boats are used in northeast, northwest,

and north inland of Michigan.

Distribution of Boater Trip Spending

 

Trip expenditures are evaluated by "spending destination" (region where trip
spending occurred) and by "spending origin" (region of boater residence). Spending is
reported in eleven categories by "spending destination" and "Spending origin". Trip
spending is also divided between spending occurring in residence region (at home) and
spending occurred in destination region (away from home).

Statewide trip spending was distributed to origin regions based on the distribution
of boats by region where boaters reside (Table 39). Using equation (1) (in page 47) the
total trip spending of $50.47 million generated in southeast region of Michigan by boaters
living in the region and keeping their boats at marinas can be computed as: 16,348
(boats)*40.5 (days)*$76.2 (spending per day) = $50.47 million.

The majority of the boaters reside in south inland and southeast regions of
Michigan, which are responsible for generating more than 60% of the total spending on

boating trips, divided 46% ($292 million) for south inland and 18% ($112 million) for

100

Table 39. Distributions of Spending by Origin & Storage ($million): 1998

 

 

Marina Waterfront Waterfront Non- TOTAL
Primary Seasonal waterfront
Home Home Home
Southeast 50.47 13.65 22.25 26.12 112.48 (18%)
East Central 6.43 1.87 8.94 13.42 30.66 (5%)
Northeast 1.91 8.16 1.67 7.25 19 (3%)
Northwest 4.62 10.57 0.87 9.62 25.69 (4%)
West Central 7.86 2.59 3.16 9.32 22.93 (4%)
Southwest 8.91 8.77 2.01 8.36 28.04 (4%)
South Inland 47.56 78.52 51.48 114.04 291.61 (46%)
North Inland 1.3 1 1.83 2.83 18.5 34.47 (5%)
South UP 1.64 0.81 1.5 5.84 9.79 (2%)
North UP 1.12 5.47 5.73 9.43 21.75 (3%)
Out of State 6.34 0.26 27.75 4.31 38.66 (6%)
SUM 138.17 142.50 128.18 226.22 635.08 (100%)

 

southeast region of Michigan. The rest of the regions in Michigan generated 2% (south
UP) to 5% (east central and north inland Michigan) of all boater trip spending, while out-
of-state boaters generated 6% ($39 million) of the total spending on boating trips during
the 1998 boating season.

Statewide trip spending was also allocated to the regions where spending occurred
(Table 40). Spending that was reported "near home" was distributed to residence regions
using the distribution of the number of boats by region where boaters reside (for seasonal
home users, near home spending was allocated to the storage region). Spending "away
from home" was allocated to destination regions based on the distribution of boat days by
region where the boat was used. Near home and away from home spending were then

summed to yield a total spent in each region.

101

Using the equations (2), (3), and (4) (in page 47-48) the total trip spending of
$52.20 million that the southeast region received as a destination from boaters storing
their boats at marinas can be illustrated as follows. First, the average trip spending of
$20.5 per day spent near home by marina boaters (in Table 31) is multiplied by the total
number of boat days (16,348 boats * 40.5 days per boat) by marina boaters residing in
southeast of Michigan to yield $13.56 million. Second, the average spending of $55.7
per day spent away from home by marina boaters is multiplied by the total days of use
(693,252 days in Table 38) by marina boaters in southeast of Michigan to yield $38.64
million. The total spending the region received from marina boaters is: $52.20 million =
$13.56 million + $38.64 million.

The largest portion of the spending takes place in the south inland region, which
accounts for 34% of the total. However, this region receives less money than it
generates. On the other hand, three north regions, north inland, northwest, and northeast
of Michigan, received more money than they generated. Based on trip spending by
"spending destination" in Table 40, the distributions of boater spending by storage type in
each region are depicted in Figure 9. Marina boaters attracted to southeast Michigan
were the biggest contributors (56%) to boater dollars on trips that the region received,
whereas seasonal home users made up only 4% of the total. Three north regions
(northeast, northwest, and north inland of Michigan) benefited most from the boaters
using waterfiont seasonal home for their boating. Boaters who stored their boats at
nonwaterfront home Spent more money on trips than any other storage segments in east

central, south inland, north inland, and south and north UP's of Michigan.

102

Table 40. Distributions of Spending by Destination & Storage ($million): 1998

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

   

 

 

 

 

 

Marina Waterfront Waterfront Non- TOTAL
Primary Seasonal waterfront
Home Home Home

Southeast 52.2 12.73 3.97 23.53 92.44 (15%)
East Central 9.07 1.58 3.84 11.27 25.76 (4%)
Northeast 11.98 8.17 18.38 16.58 55.1 (9%)
Northwest 5.7 10.08 15.23 14.86 45.88 (7%)
West Central 16.04 2.41 2.38 11.98 32.81 (5%)
Southwest 13.22 8.5 6.23 8.53 36.48 (6%)
South Inland 21.09 81.8 28.61 82.07 213.56 (34%)
North Inland 2.48 11.11 31.08 32.7 77.37 (12%)
South UP 2.48 0.92 7.1 7.89 18.39 (3%)
North UP 2.21 5.01 10.62 15.42 33.26 (5%)
Out of State 1.7 0.19 0.75 1.38 4.03 (1%)
Sum 138.17 142.50 128.18 226.22 635.08 (100%)

0% 20% 40% 60% 80% l 00%

Southeast H” "UM ' i T A 4'
1 East Central a... -.
l Northeast i
Northwest 1 8 . . . . ‘
West Central its-a. v..;. ass-.5 11.? Th; -.~. sr «rm. gm

Southwest 113-.19.: . .. ..
. l

1 South Inland m . -- newsn
i l l l

Nonh Inland _ '

 

 

 

 

 

 

   

 

 

 

 

 

South UP . _'..‘_'r_".‘;i-“;u5 j
Out of State ,- . .. i ‘ '

 

 

 

 

 

’ I Marina .Watgrfronht Primary Home DWaterfrOnt Seasonal Home .NonwaterfrTrht Home

Figure 9. Percent Distribution of Trip Spending by Storage Segment: 1998

103

Regional Net Gains of Boater Trip Smnding by StoragSegment

Assuming that the average spending within storage types does not vary across
regions, we can measure the regional flows of boater spending by storage segment (Table
41). For example, the amount of $ 1 .73 million that the southeast region benefits from
marina boaters is the differences between trip spending occurring in the regiOn ($52.20
million) and spending generated in the region ($50.47 million). All five northern regions
including the UP are net gainers from boater Spending on trips. These regions earned a
net gain of $120 million from boater spending on trips in 1998. These gains resulted
mainly from boaters storing their boats at seasonal homes. Northeast and north inland
regions have the largest net gains from boater spending making up $80 million. West
Central and southwest Michigan also benefited from boater trip spending but the amounts
were small.

The biggest net loss of boater dollars is from the south inland region. Boaters
living in this region spent $78 million more than the region received. The loss is
primarily from boaters who kept their boats at a marina or seasonal home outside the
region and also boaters coming from nonwaterfront homes. Southeast Michigan also
showed a net loss, as the region sent $20 million more into other regions than it received.
Boats registered from out-of-state spent $35 million in Michigan. Seasonal home users
from out-of-state are important contributors to the spending that the state receives, as they
spent $27 million in Michigan. Boaters originating from out-of-state are important
because they represent new dollars tO the state. Figure 10 depicts the flows of boater
spending by storage type. Boats stored at marina and seasonal homes account for most of

the regional flows of boater Spending.

104

Table 41. Regional Net gains of Spending by Region & Storage ($million): 1998

 

 

 

 

 

 

Marina Waterfront Waterfront Non- Sum
Primary Seasonal waterfront

Southeast 1.73 -0.91 -18.27 -2.59 -20.05
East Central 2.64 -0.29 -5.1 -2. 14 -4.9
Northeast 10.06 0.01 16.71 9.32 36.1
Northwest 1.07 -0.48 14.35 5.24 20.19
West Central 8.18 -0.18 -0.78 2.65 9.88
Southwest 4.31 -0.27 4.22 0.17 8.44
South Inland -26.48 3.28 -22.88 -31.97 -78.05
North Inland 1.17 -0.72 28.26 14.2 - 42.9
South UP 0.84 0.1 5.6 2.05 8.6
North UP 1.09 -0.46 4.89 5.99 11.52
Out of State 464 -0.06 -27.01 -2.92 -34.63
i 40 W "FF-WWW W _ -- -_ WW _

D Waterfront Home

- Seasonal Home
a Nonwaterfront Home

 

 

 

 

 

 

($ millions)

 

 

 

'3
:W
l
l
l
l
l
I
1

 

1

Southeast
East Central
Northeas
Northwest
Southwest
North Inland
Out of State

Central West
South Inland

l
l
1

Figure 10. Regional Net Gains of Trip Spending by Storage: 1998

105

Regional Net Gains of Boater Trip Spending by Spending Category

The regional net gains of boater dollars spent on trip can also be computed by
spending category to help identify which sectors of the region's economy benefit from
boater spending (Appendix G). A negative Sign in the regional flows of trip spending
indicates that a net loss of boater spending within spending items, while a positive Sign
indicates a net gain of the spending. Four north regions, northeast, northwest, north
inland, and UP of Michigan, are the net gainers of boater trip spending. Expenses for
food (groceries and restaurant) and boat fuel categories spent by nonresident boaters are
primary contributors to this gain. Lodging expenses were not measured in the survey.
Previous studies (Warner, 1974; Stynes et al., 1983) report that boaters seldom stay

overnight in commercial lodging establishments.

Out-of-State Boater Spending on Trips

Although Michigan registered boats owned by non-residents of the state that were
used at least once in Michigan waters in 1998 constitute 5% of the state's registered fleet
in 1998, they are an important segment because their spending represent new dollars to
the state, creating income and employment Within Michigan. Out-of-state registered boat
owners spent approximately $35 million in Michigan in 1998 (Table 42). The two south
regions (south inland and southwest) received more than 50% of the total out-Of-state
boaters spent in Michigan. Northwest and UP'S, each received approximately 10% of the
total out-of-state boater spending. Boats stored at seasonal homes account for most of the

regional flows of boater spending from out-of-state boaters.

106

Table 42. Out-of-State Boater Spending on Trips to Michigan: 1998

 

 

Boating Region Trip Spending ($ million) %
Southeast 1.72 5%
East Central 0.01 0%
Northeast 2.53 7%
Northwest 3 .4 1 10%
West Central 1.15 3%
Southwest 5.72 17%
South Inland 12.06 35%
North Inland 0.99 3%
South UP 3.79 11%
North UP 3.27 9%
TOTAL 34.63 100%

 

107

CHAPTER 5

CONCLUSIONS

This study estimated trip expenditures of active registered pleasure boats in
Michigan in 1998, tested for differences in measures Of boating activity and trip spending
across different storage segments, and estimated regional flows of boater Spending on
trips by storage type and spending category. Several refinements were made in boater
survey methods to address concerns about recall error, nonresponse bias, use estimates,
definition of a trip, and handling zeros and missing data. The findings relating to the
study objectives are summarized here, limitations of the study are specified,
recommendations for future research, and implications for private and public sectors of

the economy are made.

Summary of the Study

Estimation of Boater Trip Spending

 

A total of 652,000 active registered boats in Michigan logged an estimated 18.4
million days Of boating in 1998. Boats averaged about 28 days of use. Owners of active
registered boats spent an estimated $635 million on trips Within Michigan in 1998. The
total was divided $292 million (46%) on day trips and $343 million (54%) on overnight
trips. Boats stored at nonwaterfront sites accounted for 36% of the total trip Spending and
the rest of the total was divided about equally among boats stored at marinas, waterfront
primary homes, and waterfront seasonal homes. A typical boater spent $23 a day on day

trips and $60 a day on overnight trips, averaging about $35 per day overall. Boaters who

108

kept their boats at marinas spent $76 a day on boating trips, while boaters at waterfront

primary homes spent $20 a day.

Test for Measures of Variables by Storage Type

There were significant differences in both the level and pattern of spending by
storage types. Owners of boats stored at marinas spent $156 per trip, while at the other
extreme boaters at waterfront primary homes spent $22 on an average boating trip.
About 30% of the trips by boats kept at marinas and nonwaterfront sites involved no
money spent on the most recent trip, while more than 60% of the trips by boats at
waterfront homes involved no cash outlays on the most recent trips. On average, boats
kept at marinas and waterfront homes are used more frequently than boats stored at
waterfront homes. Owners of boats kept at different types of storage were Significantly
different in terms of a number of socioeconomic variables. Owners Who stored their
boats at marinas had the highest average income. Boat type and size were also
significantly different across storage segments. Boats stored at marinas are more likely to

be large inboard boats.

Estimation of Regional Economic Benefits fiom Boater Trip Spending

Spending by boaters on trips to different regions of the state was estimated by
storage segment and spending category in order to provide information that supports
regional marketing and planning efforts. All northern regions (north inland, northwest,
northeast, and UP'S of Michigan) realized a net gain from boater spending on trips.

Expenses for groceries, restaurant, and boat fuel by nonresident boaters at seasonal

109

waterfront homes were the primary contributors to this gain. The South inland region
experienced the biggest net loss of boater dollars, as resident boaters in this region sought
boating opportunities in other regions of Michigan. Out-of-state boaters, especially those
owning seasonal waterfront homes in Michigan, were an important segment, spending

$35 million in the state on boating.

Selected Methods to Improve Spending Estimates

 

Since the study developed methods to estimate both annual use and spending in a
single survey, procedures were needed to measure spending for a recent trip to reduce
recall errors, to sample in waves over the summer to capture seasonal variation in trips,
and to extrapolate annual use from the in-season surveys. The study employed a wave
approach in which surveys were sent out in nine waves asking boat owners to report trip
expenditures on their most recent trips. Follow-up mailings were not made because it
was important to Obtain a sample representative not just of boats, but also of trips
throughout the season. F ollow-ups surveys would bias the sample more toward end-of-
season trips. While this bias could be handled by an additional weighting procedure, the
added cost of follow-ups were not justified in terms of potential improvement in the
estimates.

Annual use per boat was estimated by applying a logistic model to nine points of
the average days of use obtained from the waves in the in-season survey. The estimate of
annual use days was compared to an annual use estimated from a season-end survey. The
average use estimated fiom the wave survey was not statistically different than the

average use estimated from the season-end survey.

110

The study tried to refrne the treatment of zeros and missing values on spending by
defining what a trip was, making it easier to explicitly report no spending viaa checkbox,
and by distinguishing between day outings and overnight trips. This procedure was quite
useful in separating missing values and true zeros. The procedure also likely reduced the
bias associated with trip spending reporting, as boaters, especially at waterfront homes,
would be more likely to report overnight or extended trips that may not be their most
recent trips. Comparisons of percentages of zero spending on trips between day vs.
overnight trips and across storage segments suggested the importance of identifying the
types of trips and obtaining a representative sample of boating trips when estimating

spending.

Study Limitations

1. Nonsampling errors of many different types and sources affect the accuracy of
survey-based estimates of spending. Possible sources of nonsampling errors are
numerous. Recall errors may be the most significant problem of this type of error.
Errors in recalling spending can be reduced by surveying respondents very close to
when the spending takes place and having clear spending categories that help the
respondent recall different expenditures. In this study the respondents were asked to
report spending on the most recent trips to reduce recall bias through the wave
approach. Even though the wave approach reduces the time between survey and the
trip, boaters may still have reported incorrect amount of expenditures.

2. As use and spending varies considerably with boat size and storage, and possibly

region, the existence of nonresponse bias in these variables would threaten the

111

external validity of survey results. Since weights were applied to adjust the final
sample across region and boat size categories to the population of active craft,
nonresponse bias would not be a problem for these variables. The weights also
should correct for any bias in storage, as storage is closely related to boat size.
Although the test results showed that there was no conclusive nonresponse bias due to
nonuse, the number of samples used for the nonresponse bias test was small. Errors
due to disproportionate response for low-frequency users, non-users, or other classes
of respondents are also common in sample surveys. Local visitors and those not
spending money tend to have lower response rates to spending surveys, which will
bias spending estimate upward. It is likely that owners of inactive boats were less
likely to return the survey. This would bias the estimates of the number of active
craft upward, which would inflate overall use and spending estimates.

. The extent of the bias in reporting of recent trips is unknown. The procedures
adopted in this study likely reduced this bias, but did not eliminate it. It is expected
that longer trips with larger spending were more likely to be reported. Trips might be
better represented if they were directly sampled via on-site surveys, although
problems in obtaining representation across the Wide range of potential beating sites
would likely create even greater potential for bias.

. Another problem associated with analysis of Spending data is confusion between trips
with no spending and nonresponse to spending questions. Results can vary
significantly depending on whether cases with blanks or missing values in the

spending questions are discarded or treated as zeros. The study tried to improve on

112

this using selected methodological refinements, but the bias associated with trip
spending reporting may not be eliminated.

Lodging expenses were not measured in the survey as previous studies of Michigan
registered boat owners reported that boaters seldom stay overnight in commercial
lodging establishments (Warner, 1974; Stynes et al., 1983). While owners of boats
stored at marinas and waterfront homes rarely use lodging facilities, boat owners who
trailer their boats from nonwaterfront homes are more likely to spend on lodging to
stay overnight. Thus, trip spending by boater who store their boats at nonwaterfront
homes would be deflated.

An important share of recreational boating activity is expected to take place in
nonmotorized craft such as canoes and drift boats. However, smaller unpowered craft
is not included in the study population because these boats are not subject to
registration. This would bias the estimates of the number of active craft downward,
which would deflate overall use and spending estimates. Boats visiting Michigan
waters from out of state, but not registered in Michigan are not included in the study.

This will bias overall use and spending estimates downward.

Recommendations for Future Research
The sample size (971 cases) was not large enough to estimate the regional net gains
of boater spending by storage type, as the number of cells across storage and region
categories is 40. For more accurate regional analysis of boater spending, it is thus
necessary to increase sample size. If one succeeds to assign an adequate number of

samples to each cell combined by region and storage, regionalization of boater

113

spending could be extended to the county level. In this case, once statistics on
boating activity and spending are reported at the regional level, county estimates can
be easily generated by allocating regional activity to counties within the region, based
on number of registered craft or boating opportunity.

. By disaggregating the boating fleet into different types of boat storage, the trip
spending estimates became more meaningful and more applicable to Situations
involving planning and marketing efforts. While boat days can be divided into inland
lakes boat days and Great Lakes boat days, the study reported trip spending for
overall use days only{: For general management and planing applications, destination
type between inland lakes and Great Lakes appears to be the useful segmentation
variable. It is expected that there exist important differences in patterns and types of
boating activity between Great Lakes and inland locations}

. It is recommended that the accuracy of boater spending estimates be tested in one or
more coastal communities. This would provide a test of the generalizability of
statewide spending figures to a local area. Careful comparisons of boater spending
and business receipts could provide checks on the accuracy of reported figures. Since
one of the primary activity in boating is fishing and boats are frequently used by
anglers, creel survey estimates on fishing days and expenditures may be used to judge
the accuracy of boating use and boater expenditures. For example, if one‘knows the
percentage of fishing as a primary activity of boating and the percentage of anglers
who use their boats for fishing purpose, a judgmental decision on the accuracy of

boating use estimate can be obtained.

114

4. There is an important research question that could not be addressed in this study.
Boater expenditure studies tend to ignore collecting information on the
infrastructural, environmental, and social costs stemming from boating. Many
boating studies have investigated the spending patterns of boaters, usually for the
purpose of documenting the economic impact of boating. This tends to estimate the
positive spending effects on community, region or on a state. Further analyses are
needed to assess both the positive and negative effects of boating developments. This
can include an assessment of taxes resulting from boaters as compared with public

services provided.

Implications
/
1.' Owners of active registered boats spent an estimated $635 million on trips within
‘Michigan in 1998. Although days on overnight trips accounted for 30% of the total

days of use, trip spending on overnight trips accounted for 54% ($343 million) of the
total spending on trips. Thus, boating is an important part of Michigan's tourism
industry. About half of the $635 million dollars that registered boat owners spend on
trips would be considered tourism activity (overnight trips or more than 50 miles
away from home). This estimate does not include spending associated with the use of
smaller unregistered craft, boater spending on lodging, out-of-state boaters on craft
not registered in Michigan, spending by guests of boat owners on boating trips,

spending associated with boat rentals and cruises, or most of the spending of boaters

on stays at seasonal homes. A rough estimate of boating's contribution to tourism

115

spending in Michigan is about $500 million, roughly 5% of all tourism spending in
Michigan in 1998.

. Boaters and boating should be an important part of Michigan's tourism promotional
program. Boating fits very well with Michigan's current theme of "Great Lakes,
Great Times". The featuring of lighthouse's at Welcome Centers and in promotional
materials is also a natural connection to boating. Michigan's promotional programs
should build on the water and Great Lakes images, by also targeting particular water-
based activities and activity segments, such as boating. Local communities (CVBs,
Chambers) should incorporate boating as part of their promotional materials including
lists of boating facilities and services.

. Three of the boating storage segments are particularly important tourism generators:
marina, seasonal home, and nonwaterfront homes. These segments are tied to
particular programs of public and private organizations that serve boatersin
Michigan. The Michigan Department of Natural Resources along with its partners
operates an extensive set of boating access sites around the state. Access to Michigan
waters is particularly important for boaters trailering craft from nonwaterfront homes.
These boats accounted for over a third ($235 million) of the total trip spending in
1998. Three quarters of the trip spending by boaters from nonwaterfront homes
involved an overnight trip. Communities and businesses near popular boating access
sites can benefit economically by tailoring offerings and services to these boaters and
publicizing the access Sites and community offerings. The state should cOntinue to
maintain the access site program and protect the quality of waters and waterfront

facilities for boaters. The local economic benefits of these boaters can be used to

116

persuade local partners to assist in maintaining and managing access sites in their
area.

4/The marina boating segment supports a sizeable industry of private marinas, boat

1

‘./
repair Shops and boat dealers in the state. The trip spending measured here largely

accrues to other businesses in the community, showing that marinas indirectly support

a variety of businesses in the community. Marina operators can use the estimates of

spending for this particular segment to better document their impacts on the

community. Such information can be important in obtaining favorable policies as
well as in establishing partnerships with others to better serve boaters and the
community interests/1

5. Boating is closely linked with seasonal homes as a quarter of all registered craft are
stored at seasonal homes. Spending per trip for the seasonal home segment is lower
than for the marina and nonwaterfront home segments. However, we intentionally
did not count spending associated with the trip to the seasonal home or the normal
expenses associated with seasonal home stays. Boating is an important factor in
seasonal home ownership and use.

6. There are a host of management and policy issues associated with boats at seasonal
homes. Most (85%) of these boats are stored on inland waters. Many of Michigan's
inland lakes, particularly in southeast Michigan have experienced carrying capacity
problems and conflicts between riparians and boating access sites users. Most
recently concerns have been raised about personal watercraft. The information on

spending and use patterns provided here can contribute to the these discussions, at

least at a broader state or regional level. The figures demonstrate that both access

117

sites users and boaters from seasonal homes have important local economic impacts.
It is therefore important to develop management policies that accommodate both
groups in a given area.

The study provided information on the number of boats kept at seasonal and
permanent waterfront homes. This is important because there are a variety Of issues
relating to the amount and type of use of Michigan inland lakes. They include use
levels and limits, enforcement and access to Michigan's inland lakes. Some regions
of the state, particularly in southeast Michigan, have experienced carrying capacity
problems and conflicts between riparians and boating access sites users. Information
on spending and use patterns provided here can contribute to decision making related

to boating management and facilities on inland lakes.

118

BIBLIOGRAPHY

,,

Armstrong, 8., and Overton, T. (1977). “Estimating Nonresponse Bias in Mail Surveys.
Journal of Marketing Research. Vol. 14, August: 396-402.

Babbie, E. (1992). The Practice of Social Research. Wadsworth Pub. Co.

Baker, K. Hozier, G. and Rogers, R. (1994). “Marketing Research Theory and
Methodology and the Tourism Industry: A Nontechnical Discussion” Journal of
Travel Research. Winter. 3- 7.

Becker, R. (1984). “Gatekeepers of Methodology: A Reply to Brown.” Leisure Sciences.
Vol.6: 513-518.

Becker, R., and Iliff, T. (1983). “Nonrespondents in Homogeneous Groups: Implications
for Mailed Surveys.” Leisure Sciences. Vol. 5: 257-267.

Brog, W., and Meyburg, A. (1980). “Nonresponse Problems in Travel Surveys: An
Empirical Investigation.” Transportation Research Record 775: Travel Demand
Models: Application, Limitations and Quantitative Methods. Washington, DC:
Transportation Research Board, National Academy of Science.

Choi, S., Ditton, R., and Matlock, G. (1992). “Homogeneity across Mail Survey Waves:
A Replicated Study.” Journal of Leisure Research. Vol. 24, No. 1: 79-85.

Chris, P. (1995). “Increasing Survey Response Rate: Practical Instructions frOm the
Total-Design Method.” Cornell Hotel and Restaurant Administration Quarterly.
Vol. 36, NO. 4: 66-73.

Churchill, G. (1988). Basic Marketing Research. Dryden Press, Chicago.

Daniel, W. (1975). “Nonresponse in Sociological Surveys: A Review of Some Methods
for Handling the Problem.” Sociological Methods and Research. Vol. 3,
February: 29 1-307.

Deleeuw, E., and Hox, J. (1988). “The Effects of Response-Stimulating Factors on
Response Rates and Data Quality in Mail Survey: A Test of Dillrnan’s Total
Design Method.” Journal of Oflicial Statistics. 4: 24 1-250.

Dillrnan, D. (1991). "The Design and Administration of Mail Surveys.” Annual Review of
Sociology. Vol. 17: 225-249.

Dolsen, D., and Machlis, G. (1991). “Response Rates and Mail Recreation Survey

Results: How Much Is Enough?” Journal of Leisure Research. Vol. 23, No. 3:
272-277.

119

Donald, M. (1960). “Implication of Nonresponse for the Interpretation of Mail
Questionnaire Data.” Public Opinion Quarterly. Vol. 24, Spring: 99-1'14.

Ellerbrock, M. (1981). “Improving Coupon Conversion Studies.” Journal of Travel
Research. Spring: 37-38.

Ellis, R., Endo, C., and Armer, J. (1970). “The Use of Potential Nonrespondents for
Studying Nonresponse Bias.” Pacific Sociological Review. Vol. 13, Spring: 103-
109.

Falk, J ., Graefe, A., and Swartz, D. (1987). The 1985 Delaware Recreational Boating
Survey: An Analysis of Delaware-registered Boaters. DEL-SG-06-87. Newark,
Delaware: University of Delaware. \Jr

A,
Q Filion, F. (1975). “Estimating Bias due to Nonresponse in Mail Surveys” Public Opinion
Quarterly. Vol. 40, Winter: 482-492. 12, {51‘

Filion, F. (1976). “Exploring and Correcting for Nonresponse Bias Using Follow-ups of
Nonrespondents.” Pacific Sociological Review. Vol. 19, July: 401 -408.

Frechtling, D. (1994a). “Assessing the Economic Impacts of Travel and Tourism-
Introduction to Travel Economic Impact Estimation.” In B. Ritchie and C.
Goeldner (eds.) T ravel, Tourism and Hospitality Research (pp. 325-332). New
York: John Wiley and Sons Inc.

F rechtling, D. (1994b). “Assessing the Economic Impacts of Travel and Tourism-
Measuring Economic Benefits.” In B. Ritchie and C. Goeldner (eds.) Travel,
Tourism and Hospitality Research (pp. 333-351). New York: John Wiley and
Sons Inc.

Fuller, C. (1974). “Weighting to Adjust for Survey Nonresponse.” Public Opinion
Quarterly. Vol. 38, Summer: 239-246.

Fuller, C. (1974). “Weighting to Adjust for Survey Nonresponse.” Public Opinion
Quarterly. Vol. 38, Summer: 239-246.

Gitelson, R., and Drogin, E. (1992). “An Experiment on the Efficacy of a Certified Final
Mailing.” Journal of Leisure Research. Vol. 24, No. 1: 72-78.

Hammitt, W., and MaDonald, C. (1982). “Response Bias and the Need for the Extensive
Mail Questionnaire Follow-ups among Selected Recreation Sample.” Journal of
Travel Research. Winter: 207-2 16.

Howard, D., Lankford, S., and Havitz, M. (1991). “A Method for Authenticating Pleasure
Travel Expenditures.” Journal of Travel Research. Spring: 19-23.

120

Humphrys, C. (1991). Carrying Capacity for Boating on Michigan's Inland Lakes. East
Lansing, Michigan: Michigan State University.

\-_-.. Kinnunen, R., and Schwartz, J. (1994). Upper Peninsula of Michigan Lake Superior 1992

Transient Boater Marketing and Economic Survey. Ann Arbor, Michigan:
Michigan Sea Grant Publications. MICHU-SG-94-204. VT;

*\

Kott, P. (1994). “A Note on Handling Nonresponse 1n Sample Surveys.” Journal of the
American Statistical Association. Vol. 89, No. 426. 693- 696. 5," 4'. a. ._

Lambert, D., and Harrington, T. (1990). “Measuring Nonresponse Bias in Customer
Service Mail Surveys.” Journal of Business Logistics. Vol. 1 1, No. 2: 5-25.

Lankford, S., Buxton, B., Hetzler, R., and Little, J. (1995). “Response Bias and Wave
Analysis of Mailed Questionnaires in Tourism Impact Assessments.” Journal of
Travel Research. Spring: 8-13.

Lawton, L., and Parasuraman, A. (1980). “The Impact of the Marketing Concept on New
Product Flaming.” Journal of Marketing. Vol. 44, Winter: 19-25.

Lipton. D. W, and Miller, 5. (1995). Recreational Boating in Maryland: An Economic
Impact Study. UM- SG- MAP- 95- 02. College Park, Maryland: Maryland Sea Grant
Extension Service, University of Maryland.

Mak, J ., Moncur, J ., and Yonamine, D. (1977). “How or How Not to Measure Visitor
Expenditures.” Journal of Travel Research. Summer: 1-4.

Mandell, L. (1974). “When to Weight: Determining Nonresponse Bias in Survey Data.”
Public Opinion Quarterly. Vol. 38, Winter: 247-252.

Meyburg, A., and Brog, W. (1981). “ Validity Problems in Empirical Analyses of Non-
Home-Activity Patterns.” Transportation Research Record 80 7: Travel Demand
Forecasting and Data Considerations. Washington, DC: Transportation
Research Board, National Academy of Science.

National Marine Manufacturers Association. (1999). Boating Statistics.
http://www.nmma-rnedialink.com.

National Sporting Goods Association. (1999). Sport Participation. http://www.nsga.org.

Neely, R., Johnson, R., and Young, B. (1998). The Business of Boating Recreation in
Oregon. ORESU-G-98-OO3. Corvallis, Oregon: Oregon Sea Grant, Oregon State
University.

Nelson, J. A. (1996). “The Frequency of Consumer Expenditures: An Empirical
Analysis.” Journal of Consumer Aflairs. Vol. 30, No. 2: 322-347.

121

Oppennann, M. (1998). “Destination Threshold Potential and the Law of Repeat
Visitation.” Journal of Travel Research. Vol. 37, November: 13 1-137.

Pindyck, R., and Rubinfeld, D. (1981). Econometric Models and Economic Forecasts.
McGraw-Hill. '

Roger, J. (1991). “A Non-Technical Perspective on Data Collection Methodologies for
Travel Survey: A Discussion Paper.” Journal of Travel Research. Winter: 43-47.

Rowland, M., and F orthofer, R. (1993). “Adjusting for Nonresponse Bias in a Health
Examination Survey.” Public Health Reports. Vol. 108, No. 3: 380-386.

Rust, E., and Potepan, M. (1997). The Economic Impact of Boating in California.
California Department of Boating and Waterways. California. Y2.

Rylander, R., Propst, D., and McMurtry, T. (1995). “Nonresponse and Recall Bias in a
Survey of Travel Spending.” Journal of Travel Research. Spring: 39-45.

Schott, R. (1995). The Impact of Great Lakes Recreational Boating on the Economy of
Michigan. MS. Thesis. East Lansing, Michigan: Michigan State University.

Spotts, D., and Mahoney, E. (1991) “Segmenting Visitors to a Destination Region Based
on the Volume of Their Expenditures.” Journal of Travel Research. Winter: 24-
31.

Stewart, S. (1990). A Pilot Study of Transient Boaters in Three Michigan Ports. MS.
Thesis. East Lansing, Michigan: Michigan State University.

Stewart, W., Anderson, S., Fesenmaier, D., and Lue, C. (1993). “Highway welcome
Center Surveys: Problem with Nonresponse Bias.” Journal of Travel Research.
Winter: 53-57.

Stoll, J ., Bergstrom, J ., and Jones, L. (1988). “Recreational Boating and Its Economic
Impact in Texas.” Leisure Science. Vol. 10: 51-67.

Stynes, D. (1983a). “An Introduction to Recreation Forecasting.” In S. Lieber and D.
F esenmaier (eds.) Recreation Planning and Management (pp. 87-95). State
College, PA: Venture Publishing.

Stynes, D. (1983b). “Time Series and Structural Models for Forecasting Recreation
Participation.” In S. Lieber and D. Fesenmaier (eds.) Recreation Planning and
Management (pp. 105-119). State College, PA: Venture Publishing.

Stynes, D. (1998). Guidelines for Measuring Visitor Spending.
http://www.msu.edu/user/stynes.

122

Stynes, D., and Chung, W. (1986). “Resistant Measures of Recreation and Travel
Spending.” Abstract 1986 Symposium on Leisure Research. Alexandria, VA:
National Recreation and Park Association.

Stynes, D., and Holecek, D. (1982). Michigan Great Lakes Recreational Boating: A
Synthesis of Current Information. MICH-SG-82-203. Ann Arbor, Michigan:

Michigan Sea Grant Program, ‘6;

4"

Stynes, D., and Mahoney, E. (1989). “Measurement and Analysis of Recreational and
Travel Spending.” Abstract 1986 Symposium on Leisure Research. Alexandria,
VA: National Recreation and Park Association.

Stynes, D., and Safronoff, D. (1982). 1980 Michigan Recreational Boating Survey.
MICHU-SG-82-202. Ann Arbor, Michigan: Michigan Sea Grant Program.

Stynes, D., Brother, G., Holecek, D., and Verbyla, D. (1983). Recreational Boating
Spending Patterns and Economic Impacts of Michigan Registered Boat Owners.

MICHU-SG-83-210. Ann Arbor, Michigan: Michigan Sea Grant Program. n J

Stynes, D., Wu, T., and Mahoney, E. (1995). 1994 Michigan Boating Survey. East
Lansing, Michigan: Michigan State University.

Sudman, S., and Kalton, G. (1986). “New Developments in the Sampling of Special
Populations.”AnnuaI Review of Sociology. Vol. 12: 40 1-429.

Talhelm, D. Jordan, 8., and Holecek, D. (1988). 1986 Michigan Recreational Boating
Survey East Lansing, Michigan: Michigan State University.

Talhelm, D., Mahoney, E., Bishop, G., Lee, H-C., and Wu, T. (1998). Great Lakes
Marina Census and Marina Needs. Research Report 548. East Lansing,
Michigan: Michigan Agricultural Experiment Station, Michigan State University.

Tanfer, K. (1993). “National Survey of Men: Design and Expectation.” Family Planning
Perspectives. Vol. 25, No. 2: 83-86.

Travel Industry Association of America (1998). Research and Statistics.
http://www.tia.org/research.

USDA, USFWS, and USDC. (1998). 1996 National Survey of Fishing Hunting and
Wildlife-Associated Recreation-Michigan.

Viswesvaran, C., Barrick, M., and Ones, D. (1993). “How Definitive Are Conclusions

Based on Survey Data: Estimating Robustness to Nonresponse.” Persbnal
Psychology. Vol. 46: 551-567.

123

Warner, T. D. (1974). An Analysis of Recreational Boating Expenditures. A Study of
Michigan Boaters. MS. Thesis. East Lansing, Michigan: Michigan State
University.

Witt, S., and Witt, C. (1992). Modeling and Forecasting Demand in Tourism. Academic
Press.

Woodside, A., and Ronkainen, I. (1984). “How Serious Is Nonresponse Bias in
Advertising Conversion Research?” Journal of Travel Research. Spring: 34-37.

Wu, T. C. (1995). A System of Models for Estimating Recreational Boating Use in
Michigan Counties. Ph.D. Dissertation. East Lansing, Michigan: Michigan State
University.

Yuan, M., and Yuan, S. (1996). “Sixteen versus Nine Expenditure Categories in Tourism
Survey: Is There a Difference?” Journal of Travel Research. Spring: 59-62.

124

APPENDICES

125

Appendix A. Questionnaire (In-Season Survey)

1998 MICHIGAN RECREATIONAL BOATING SURVEY

1. How many boats do you own that are currently registered in Michigan?

 

 

PLEASE ANSWER THIS SURVEY FOR THE BOAT INDICATED ON
THE MAILING LABEL

 

 

 

DESCRIPTIVE INFORMATION ABOUT YOUR BOAT

2. Type of boat El Inboard C1 Sail, unpowered D Canoe or Row
(check one) : CI Inboard/outboard D Sail, with power CI Personal watercrafi_(e. g. Jet ski)
El Outboard El Pontoon Cl Other, please specify

 

3. Boat length (feet)

 

4. How long have you owned this boat?
years

5. Where are you keeping this boat during the 1998 boating season?

a. County where the boat is kept

 

b. Type of facility (check one): c. Is this boat currently stored (check one)
E1 Permanent residence D On land
El Cottage or second home D In a dry stack facility
1:] Public marina E] In the water (wet slip, mooring or dockside)
El Rented space in a commercial marina E1 Attached to or on a larger boat
E1 Owned space in marina or dockaminium CI Other (please specify)

 

CI Yacht/boat club
D Other (please specify)

 

d. Is this location (check one):

D A waterfront site with access to the Great Lakes & connecting waters
[3 An inland lake waterfront site (no Great Lakes access)

D A river or stream waterfront site (no Great Lakes access)

1:] A non-waterfront site

6. Where do you store this boat most of the time during the non-boating season? (check one)

 

D A permanent residence El Yacht / boat club
E] A cottage or second home D Other rented space
C1 Marina [3 Other storage (please specify)

7. Has this boat been under power or sail in Michigan waters in 1998?

 

E1 NO :9 If No, do you plan to use the boat in Michigan this year? CI YES 1:] NO

1:1 YES IfNO, why not?

 

If your boat has not been put in water in 1998, skip to question 19.

 

 

 

126

USE OF YOUR BOAT THIS YEAR

8. How many days has this boat been operated in Michigan waters so far in 1998?
days of boating in 1998
83. How many of these days was the boat operated on the Great Lakes or connecting waters?
(That is, Lakes Huron, Superior, Erie, Michigan, and St. Clair: the St. Mary's. St. Clair and Detroit
River, and lakes and rivers that provide direct access to the Great Lakes).

days of boating on the Great Lakes

9. How many additional days (NOT COVERED ABOVE) has the boat been used this year when it was in the
water, but not underway (e.g. used at the dock or mooring for lodging, repairs, entertaining, fishing etc.).

days of additional use at the dock or mooring.

MOST RECENT BOATING OCCASION

 

This section asks about the most recent outing on which your boat was used (include trips to a boat stored
at a marina even if the boat was not taken out under power or sail).

 

 

10. Which of the following best describes your most recent boating occasion: (check one)

Transported my boat to a launch site

Boated from my waterfront permanent home

Boated from my waterfront seasonal home

Traveled to where the boat is stored (e.g. a marina or yacht club) and boated from there
Traveled to where the boat is stored (e.g. a marina or yacht club) and used the boat at the dock
without getting underway

D Other, please specify

DECIDE]

 

l 1. When did this most recent use occur?
day/month

12. Where did this most recent use occur? County , Body of Water

 

 

13. Was the boat used on the Great lakes or connecting waters on this occasion?

Cl YES Cl NO

14. Were you away from your permanent home or seasonal residence overnight on this boating occasion?

ClYES 13 NO

 

14a. If Yes, how many nights were you away from home? (nights)

 

 

 

15. How many people used the boat on this occasion? adults

children (17 or younger)

127

 

SPENDING ON YOUR MOST RECENT BOATING OCCASION

16. Did you spend any money on this boating occasion? 1:] YES 13 NO (skip to question 18)
1
16a. If Yes, please report how much money you spent within each of the following categories.
Report spending near your permanent or seasonal home in the first column and spending away
from home (more than 20 miles from home) in the second column. If you are boating from a
waterfront home. report spending on any items bought specially for this outing (e.g. boat gas,
groceries etc.). If boating away from home, report all expenses on this trip.

SPENDING ON YOUR MOST RECENT BOATING OCCASION
(enter zero if you did not spend anything in a particular category)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

NEAR HOME AWAY FROM HONIE
BOAT EXPENSES
Boat Fuel and oil S $
Temporary dockage $ $
Pump-out & launch fees $ $
Repair and maintenance $ $
Marine supplies $ $
PERSONAL EXPENSES
Restaurant meals and drinks $ $
Groceries & take out food & drink 8 $
Auto gas and oil $ $
Shopping & souvenirs $ 3
Recreation & entertainment $ $
Other expenses $ $

 

 

17. Was your spending on this trip outing (check one):
Cl More than you usually spend C1 Typical C1 Less than usual

BOATS ACQUIRED IN THE PAST THREE YEARS

18. Have you purchased or otherwise acquired a new or used boat within the past three years‘
(1996, 1997 or 1998)?

El YES Cl NO (If No, skip to question 19)

1

If YES, please complete the following table for each boat purchased since January 1996.

 

 

 

 

Year Make TYP? (see Length Price Bought Bought from
Bought question 2) (feet) new 5” (check one)
use
E1 new El- boat dealer
E] retail store
D used CI other boater
I] new 13 boat dealer
El used El retail store
1:] other boater
El new 13 boat dealer
[:1 used 13 retail store
CI other boater

 

 

 

 

 

 

 

 

 

128

INFORMATION ABOUT YOU AND YOUR FAMILY

This information is requested to provide a profile of registered boat owners and to identify boating patterns for
different subgroups of boaters. 19. Please give the county, state, and zip code of your permanent residence.

 

 

County State Zip Code

20. Age of the boat owner years
21. How many people, including yourself, reside in your household?
Adults Children under 18 years of age

22. What was your annual household income in 1997 ? (check one category below)

13 Under $20,000 [:1 $60,000-$99,999
Cl $20,000- $39,999 Cl $100,000 -$149,999
13 $40,000-$59,999 Cl Over $150,000

23. Do you currently own a seasonal home, condominium or cottage in Michigan?

D YES ¢ In what Michigan county is it located?
CI NO County

 

THANK YOU VERY MUCH FOR YOUR HELP WITH THIS SURVEY.

TO RETURN YOUR COMPLETED QUESTIONNAIRE, FOLD IT SO THE RETURN ADDRESS
SHOWS AND THEN TAPE OR STAPLE IT TOGETHER. MAIL 1T FROM ANY U.S. POSTAL BOX.

 

9 l " ‘03 NO POSTAGE
NECESSARY
1r MAILED
IN THE

UNITED STATES

 

 

 

 

BUSINESS REPLY MAIL

FIRST CLASS MAIL PERMIT NO. 941 EAST LANSING, Ml

 

 

 

POSTAGE WILL BE PAID BY ADDRESSEE

DEPARTMENT OF PARK, RECREATION
AND TOURISM RESOURCES
MICHIGAN STATE UNIVERSITY

131 NATURAL RESOURCES BUILDING
EAST LANSING MI 48824-9902

llllllllllllllllllllllllllllllllllllll lllllllllllllll

129

Appendix B. Questionnaire (Season-End Survey)

1998 MICHIGAN RECREATIONAL BOATING SURVEY

1. How many boats do you own that are currently registered in Michigan?

 

 

PLEASE ANSWER THIS SURVEY FOR THE BOAT INDICATED ON
THE MAILING LABEL.

 

 

 

DESCRIPTIVE INFORMATION ABOUT YOUR BOAT

2. Type of boat CI Inboard El Sail, unpowered CI Canoe or Row
(check one): Cl Inboard/outboard D Sail, with power 13 Personal watercrafie(e.g. Jet ski)
C] Outboard El Pontoon 13 Other ( please specify)

 

3. Boat length (feet)

 

4. How long have you owned this boat? years

5. Where did you keep this boat during the 1998 boating season?

 

a. County where this boat is kept (see map for county names)

b. Type of facility (check one): c. Was this boat currently stored (check one):
Cl Permanent residence Cl On land
El Cottage or second home El In a dry stack facility
El Public marina E] In the water (wet slip, mooring or dockside)
[:1 Commercial marina 1:] Attached to or on a larger boat
El Owned space in marina or dockaminium C] Other (please specify)

 

El Yacht / boat club
Cl Other (please specify)

 

d. Is this location (check one):

C] A waterfront site with access to the Great Lakes & connecting waters
El An inland lake waterfront Site (no Great Lakes access)

E] A river or stream waterfront site (no Great Lakes access)

D A non-waterfront site

6. Where do you store this boat most of the time during the non-boating season? (check one)

13 Permanent residence El Yacht / boat club
1:] Cottage or second home 1:] Other rented space
El Marina D Other storage (please specify)

 

7. Has this boat been under power or sail in Michigan waters in 1998?

 

El NO :9 If No, do you plan to use the boat in Michigan this year? El YES 13 NO

El YES If NO, why not?

 

If your boat has not been put in water in 1998, skip to question 10.

 

 

 

I30

USE OF YOUR BOAT THIS YEAR

 

8. How many days has this boat been operated in Michigan waters so far in 1998?
days of boating in 1998
8a. How many of these days was the boat operated on the Great Lakes or connecting waters?

(That is, Lakes Huron. Superior. Erie, Michigan, and St. Clair; the St. Mary’s, and Detroit
Rivers; and lakes and rivers that provide direct access to the Great Lakes)

days of boating on the Great Lakes

9. How many additional days (NOT COVERED ABOVE) has the boat been used this year when it was in the
water, but not underway? (e.g. used at the dock or mooring for lodging, repairs, entertaining, fishing, etc.)

days of additional use at the dock or mooring
BOATS ACQUIRED IN THE PAST THREE YEARS

10. Have you purchased or otherwise acquired a new or used boat within the past three years (96, 97, 98)?

 

 

 

[3 YES El NO (If No, skip to question 11.)
0
If YES, please complete the following table for each boat purchased since January 1996.
Year Make TYP? (see Length Price Bought Bought from
Bought QUCSUOU 2) (feet) “6“” or used (check one)
Cl new D boat dealer
Cl retail store
D used El other boater
Cl new El boat dealer

El used El retail store

El other boater
El new El boat dealer
El used El retail store
CI other boater

 

 

 

 

 

 

 

 

 

 

ANNUAL EXPENSES FOR THIS BOAT

ll. Estimate the amount of money spent in 1998 to operate and maintain this boat (shown on the label).
Report expenses only for the boat that you have described above. DO NOT include spending for
consumable items used on boating trips or transportation to and from boating areas (for example, auto fuel,
food, bait and lures).

 

 

 

 

 

Boat Equipment (e.g., motors, trailer, .

anchors, sails, fishing, water-ski, safety & $———— Seasonal SIlp rental or S
electronic equipment, etc.) Dry stack Storage

Repair & Maintenance (e.g., to hull, .

motor, trailer, mast, sails, galley, deck, 3 Put m and Haul out fees 3
shaft, prop, docks, etc.)

Boat Insurance $ Off-season storage $

 

 

 

12. How much money was spent on fuel for this boat in 1998? S

 

131

INFORMATION ABOUT YOU AN]; YOUR FAMILY

This information is requested to provide a profile of registered boat owners and to identify boating patterns for
different subgroups of boaters (see map to identifi/ county names for questions 13 and l7).

13. Please give the county, state, and zip code of your permanent residence.

 

 

County State Zip Code

14. Age of the boat owner years
15. How many people, including yourself, reside in your household?
Adults Children under 18 years of age

16. What was your annual household income in 1997 ? (check one category below)

Cl Under $20,000 Cl $60,000-$99,999
Cl $20,000- $39,999 Cl $100,000 -$l49,999
Cl $40,000-$59,999 [:1 Over $150,000

17. Do you currently own a seasonal home, condominium or cottage in Michigan?

Cl YES '39 In what Michigan county is it located?
Cl NO County

 

 

THANK YOU VERY MUCH FOR YOUR HELP WITH THIS SURVEY.
TO RETURN THE QUESTIONNAIRE, FOLD IT SO THE RETURN ADDRESS SHOWS,
TAPE OR STAPLE IT TOGETHER, AND MAIL IT AT ANY U.S. POST OFFICE MAILBOX.

9 I '1 103 NO POSTAGE
NECESSARY
IF MAILED
IN THE

UNITED STATES

 

 

 

 

 

 

 

 

BUSINESS REPLY MAIL

FIRST CLASS MAIL PERMIT NO. 941 EAST LANSING, MI

 

 

 

POSTAGE WILL BE PAID BY ADDRESSEE

DEPARTMENT OF PARK, RECREATION
AND TOURISM RESOURCES
MICHIGAN STATE UNIVERSITY

I31 NATURAL RESOURCES BUILDING
EAST LANSING MI 48824-9902

132

 

wmmd om. TR o\ohm foov o\omm o\oov {one {cow face £ch o\omv face

manz suz mwuz Suz wouz canz mwuz zuz wemuz Eonz 3225 Sofia e.
:3 3.5x . :8 as :8 :3 s3 _ :8 :3 as $3 $8

Nouz Enz Suz ewuz mouz cauz ~sz whiz memnz omouz SE. an ac e.

a: as. s: :9. :3. $8 a: s; s: as? 28: acoeoaeéoz

osc— fccm o\oom £ch gem o\o_m o\cmm o\.._ N .xcmm $8 08o: 353% E99033
{one fawn o\on £on o\ovm o\oom {one foam o\omm {ohm 2:0: infirm Eotbfi?

 

 

 

 

 

 

 

 

 

sh so 9% 5 gm :2 .E. Q.» so as 2E:
:3 setup anz puz mmuz mwuz mouz wwuz an Ruz mmmnz Nguz 8555 50m
:8. $8. :52 $8. $02 :8. s8. :52 $2: :52
:2 ea 3: ea SN. ea: :2 em. :2 $2 0?.
s: <1 <1 _ :2 gm Sn :3 en e3 so 3.5850
:2 $2 .5 $3 e: .5 ea «.2 e... _ so. 898..
e: e: an ER Sm es. an em an e... 2%
s: as; 3:. as. :2 :3 sum :3 SQ. some geese
s: :2 :8 es: cs: :3 SS. $2 $8 $8 6.8583385
83 8.2an wouz quz mwuz Suz Mauz Nouz Suz owuz wemnz weonz 38 no mat.
:52 see $8. $8. :8. see :2: see :8. .22
§ sslm em fl flow .Nm em fl Axum § +3
:3 s: s: :2 so. so 5 $2 $2 $2 .wN - .N
:2 :9. as? as: :2 $3 :5 :3 $2 see. .8 - 2
exam :3 $2 $8 so”. :2 $2 $3 $2. $9. .2 v
:2 es s: ea 32 ea: :2 ea. $2 $2 0?.
on; $.8an wouz Suz mwuz Suz wouz monz qu2 T2 aemnz :ouz some 535438
a NT: o w a e m e m m _ EB
0235

 

963 3 @55on 3:. a. Seem E 8:582 .mV 2an

963 3 @562on nth a. 38m E $5322 .0 iguana

I33

Appendix D. Sampling Errors

Table 44. Sampling Errors for Trip Expenditures (90% Confidence Level)

 

 

N Mean Standard Error %Error

DAY TRIPS

Marina 85 52.6 9.0 28%

Waterfi'ont Primary Home 238 15.0 2.5 28%

Waterfront Seasonal Home 191 27.7 4.2 25%

Nonwaterfront Home 213 30.8 3. I 17%

Day Trips All 727 23.1 1.9 13%
OVERNIGHT TRIPS

Marina 58 90.7 9.2 17%

Waterfront Primary Home 15 79.0 17.0 35%

Waterfront Seasonal Home 13 36.9 13.1 58%

Nonwaterfront Home 62 51.4 6.7 21%

Overnight Trips All 148 60.2 4.9 13%
ALL TRIPS BY STORAGE

Marina 143 76.2 7.0 15%

Waterfront Primary Home 253 20.3 2.6 21%

Waterfront Seasonal Home 204 29.1 4.0 23%

Nonwaterfront Home 275 44.1 2.9 1 1%
ALL TRIPS BY TRIP

Day Trips 727 23.1 1.9 13.4%

Overnight Trips 148 60.2 4.9 13.3%
TOTAL 875 34.6 1.8 9%

 

134

Appendix E. Average Trip Spending

Table 45. Average Trip Spending by Storage Segment & Type of Trip: 1998

 

Marina Waterfront Waterfront Non- Total
Primary Seasonal waterfront

 

(1) PER BOAT PER DAY TRIP SPENDING FOR DAY TRIPS

Boat Fuel 18.0 7.3 7.1 8.5 8.0
Temporary Dockage 0.0 0.0 0.1 0.1 0.0
Pump-out & Launch Fees 0.1 0.2 0.0 0.4 0.2
Repair & Maintenance 0.2 1.2 2.8 3.3 1.9
Marine Supplies 3.7 0.7 1.7 1.1 1.2
Restaurant 13.5 2.5 3.4 3.9 3.6
Groceries 8.7 2.1 9.3 5.9 5.1
Auto Gas 2.5 0.5 2.3 4.9 1.8
Shopping 4.3 0.2 0.0 0.1 0.3
Recreation 1. 1 0.3 0.5 0.4 0.4
Other Expenses 0.5 0.1 0.7 1.9 0.5
Overall Mean 52.6 15.0 27.7 30.8 23.1
Lower Bound " 37 .8 10.9 20.8 25.7 20.0
Upper Bound a 67.4 19.1 34.6 35.9 26.2
% Sampling Error 3 28% 28% 25% 17% 13%
(2) PER BOAT PER DAY TRIP SPENDING FOR OVERNIGHT TRIPS
Boat Fuel 22.7 28.4 3.8 7.8 12.4
Temporary Dockage 12.2 8.4 0.4 0.9 3.9
Pump-out & Launch Fees 0.9 1.8 0.3 0.6 0.7
Repair & Maintenance 1.2 1.1 8.2 3.1 3.1
Marine Supplies 3.9 1.3 1.5 1.2 1.8
Restaurant 17.7 12.8 8.1 6.7 9.6
Groceries 15.1 9.4 11.0 12.2 12.3
Auto Gas 3.4 8.1 1.5 7.0 5.7
Shopping 8.8 2.3 1.3 3.3 4.0
Recreation 1.7 5. 1 0.0 4.4 3 .4
Other Expenses 3.1 0.2 0.8 4.3 3.2
Overall Mean 90.7 79.0 36.9 51.4 60.2
Lower Bound 75.6 51.1 15.4 40.4 52.2
Upper Bound 105.8 106.9 58.4 62.4 68.2
% Sampling Error 17% 35% 58% 21% 13%
(1+2) PER BOAT PER DAY SPENDING (WEIGHTED AVERAGE OF TOW TRIPS).
Boat Fuel 20.9 9.0 6.5 8.1 9.3
Temporary Dockage 7.6 0.7 0.1 0.7 1.2
Pump-out & Launch Fees 0.6 0.3 0.1 0.5 0.3
Repair & Maintenance 0.9 1.2 3.6 3.2 2.3
Marine Supplies 3.8 0.8 1.6 1.2 1.4
Restaurant 16.1 3.3 4.1 5.7 5.4
Groceries 12.7 2.7 9.5 9.9 7.3
Auto Gas 3.0 1.2 2.2 6.2 3.0
Shopping 7.1 0.3 0.2 2.2 1.5
Recreation 1 .5 0.7 0.4 3.0 1.3
Other Expenses 2.1 0.1 0.7 3.4 1.4
Overall Mean 76.2 20.3 29.1 44.1 34.6
Lower Bound 64.8 16.0 22.5 39.3 31.6
Upper Bound 87.6 24.6 35.7 48.9 37.6
% Sampling Error 15% 21% 23% 11% 9%

 

a. Sampling errors for estimates of trip expenditures are reported with a 90% confidence level.

135

Table 46. Average Trip Spending by Storage & Spending Location: 1998 -

 

Marina Waterfront Waterfront Non- Total
Primary Seasonal waterfront
Home Home Home

 

(1) PER BOAT PER DAY TRIP SPENDING NEAR HOME

Boat Fuel 6.4 7.7 5.9 3.4 6.0
Temporary Dockage 0.1 0.5 0.0 0.0 0.2
Pump-out & Launch Fees 0.2 0.3 0.0 0.1 0.2
Repair & Maintenance 0.1 1.1 2.8 1.3 1.5
Marine Supplies 2.7 0.7 1.6 0.7 1.1
Restaurant 4.2 2.0 2.4 1.3 2.1
Groceries 4.1 2.2 8.4 3.7 4.3
Auto Gas 1.4 0.5 1.7 2.3 1.4
Shopping 1.2 0.1 0.1 0.1 0.2
Recreation 0.0 0.2 0.1 0.3 0.2
Other Expenses 0.1 0.1 0.7 1.0 0.5
TOTAL 20.5 15.2 23.7 14.2 17.5
(2) PER BOAT PER DAY TRIP SPENDING AWAY FROM HOME
Boat Fuel 14.5 1.3 0.6 4.6 3.4
Temporary Dockage 7.5 0.2 0.1 0.6 1.0
Pump-out & Launch Fees 0.4 0.0 0.0 0.4 0.2
Repair & Maintenance 0.8 0.1 0.8 1.9 0.8
Marine Supplies 1.1 0.1 0.0 0.5 0.3
Restaurant 11.9 1.3 1.7 4.4 3.3
Groceries 8.6 0.5 1.1 6.2 3.1
Auto Gas 1.6 0.7 0.5 4.0 1.6
Shopping 5.9 0.3 0.1 2.1 1.3
Recreation 1.4 0.5 0.3 2.7 1.2
Other Expenses 2.0 0.0 0.0 2.4 0.9
TOTAL 55.7 5.1 5.4 29.9 17.1
(1+2) PER BOAT PER DAY SPENDING ON TRIP
Boat Fuel 20.9 9.0 6.5 8.1 9.3
Temporary Dockage 7.6 0.7 0.1 0.7 1.2
Pump-out & Launch Fees 0.6 0.3 0.1 0.5 0.3
Repair & Maintenance 0.9 1.2 3.6 3.2 2.3
Marine Supplies 3.8 0.8 1.6 1.2 1.4
Restaurant 16.1 3.3 4.1 5.7 5.4
Groceries 12.7 2.7 9.5 9.9 7.3
Auto Gas 3.0 1.2 2.2 6.2 3.0
Shopping 7.1 0.3 0.2 2.2 1.5
Recreation 1.5 0.7 0.4 3 .0 1.3
Other Expenses 2.1 0.1 0.7 3.4 1.4
TOTAL 76.2 20.3 29.1 44.1 34.6

 

I36

Appendix F. Total Trip Spending

Table 47. Total Trip Spending by Storage Segment & Type of Trip: 1998

 

 

 

 

 

 

Marina Waterfront Waterfront Non- Total
Primary Seasonal waterfront
(1) TOTAL TRIP SPENDING FOR DAY TRIPS (S MILLIONS) -
Boat Fuel 12.4 46.9 26.2 15.6 101.2
Temporary Dockage 0.0 0.1 0.2 0.2 0.6
Pump-out & Launch Fees 0.1 1.1 0.1 0.8 2.1
Repair & Maintenance 0.2 7.9 10.3 6.1 24.5
Marine Supplies 2.6 4.6 6.2 2.1 15.4
Restaurant 9.3 16.0 12.6 7.2 45.1
Groceries 6.0 13.3 34.5 10.8 64.5
Auto Gas 1.7 3.4 8.5 9.0 22.6
Shopping 3.0 1.1 0.0 0.3 4.3
Recreation 0.7 2.0 1.8 0.7 5.2
Other Expenses O_.3 E 23 a Q
TOTAL 36.3 96.9 102.9 56.3 292.4
Percent of Spending by Segment 12% 33% 35% 19% 100%
(2) TOTAL TRIP SPENDING FOR OVERNIGHT TRIPS (S MILLIONS)
Boat Fuel 25.5 16.4 2.6 25.9 70.4
Temporary Dockage 13.7 4.9 0.3 3.1 22.0
Pump-out & Launch Fees 1.0 1.0 0.2 1.9 4.0
Repair & Maintenance 1.4 0.7 5.6 10.1 17.8
Marine Supplies 4.4 0.8 1.0 4.0 10.2
Restaurant 19.8 7.4 5.5 22.0 54.8
Groceries 17.0 5 .4 7.5 40.2 70.2
Auto Gas 3.8 4.7 1.0 23.0 32.5
Shopping 9.9 1.3 0.9 10.9 23.0
Recreation 1.9 2.9 0.0 14.6 19.4
Other Expenses 3_5 fl (_)_6 _1_4._I fl
TOTAL 101.9 45.6 25.3 169.9 342.7
Percent of Spending by Segment 30% 13% 7% 50% 100%
(1+2) TOTAL STATEWIDE SPENDING ON TRIPS (S MILLIONS) '
Boat Fuel 37.9 63.4 28.8 41.5 171.6
Temporary Dockage 13.8 4.9 0.5 3.4 22.6
Pump-out & Launch Fees 1.1 2.1 0.2 2.7 6.1
Repair & Maintenance 1.6 8.5 16.0 16.2 42.3
Marine Supplies 6.9 5.4 7.2 6.1 25.7
Restaurant 29.1 23.4 18.1 29.2 99.8
Groceries 23.0 18.7 42.0 51.0 134.7
Auto Gas 5.5 8.1 9.5 32.0 55.1
Shopping 12.9 2.4 0.9 11.2 27.3
Recreation 2.7 4.9 1.8 15.3 24.6
Other Expenses E M 3_1 fl 2§_l_
TOTAL 138.2 142.5 128.2 226.2 635.1
Percent of Spending by Segment 22% 22% 20% 36% 100%
SUMMARY (S MILLIONS)
Trip Spending on Day Trips 36.3 96.9 102.9 56.3 292.4
Trip Spending on Overnight Trips w fig 25_.3 L99 3_42_.Z
138.2 142.5 128.2 226.2 635.1
Trip Spending on Day Trips 26% 68% 80% 25% 46%
Trip Spending on Overnight Trips E’é 32% 20% 75% 54%
100% 100% 100% 100% 100%

 

I37

Table 48. Total Trip Spending by Storage Segment & Spending Location: 1998

 

 

Marina Waterfront Waterfront Non- Total
Primary Seasonal waterfront
Home Home Home
(1) TOTAL TRIP SPENDING NEAR HOME (S MILLIONS)
Boat Fuel 11.6 53.9 26.2 17.7 109.4
Temporary Dockage 0.1 3.6 0.0 0.2 3.8
Pump-out & Launch Fees 0.4 1.9 0.1 0.5 3.0
Repair & Maintenance 0.2 8.0 12.4 6.4 27 .0
Marine Supplies 4.9 4.9 7.0 3.6 20.3
Restaurant 7.6 13.9 10.7 6.4 38.6
Groceries 7.5 15.1 37.0 19.0 78.6
Auto Gas 2.5 3.3 7.4 11.6 24.8
Shopping 2.2 0.6 0.3 0.6 ‘ 3.7
Recreation 0.0 1.4 0.4 1 .4 3.2
Other Expenses Q Q _2;9_ a Q
TOTAL 37.1 107.0 104.5 72.7 321.3
Percent of Spending by Segment 12% 33% 33% 23% 100%
(2) TOTAL TRIP SPENDING AWAY FROM HOME (S MILLIONS)
Boat Fuel 26.3 9.5 2.6 23.8 62.2
Temporary Dockage 13.6 1.4 0.5 3.2 18.8
Pump-out & Launch Fees 0.7 0.2 0.1 2.2 3.1
Repair & Maintenance 1.4 0.5 3.6 9.8 15.3
Marine Supplies 2.1 0.6 0.2 2.6 5.4
Restaurant 21.5 9.5 7.4 22.8 61.2
Groceries 15.5 3.6 5.0 32.0 56.1
Auto Gas 3.0 4.8 2.1 20.4 30.3
Shopping 10.7 1.9 0.5 10.6 23.7
Recreation 2.6 3.5 1.4 13.9 21.4
Other Expenses _3_.1 0_.1_ g2 £3 _6._3
TOTAL 101.0 35.5 23.7 153.5 313.8
Percent of Spending by Segment 32% 11% 8% 49% - 100%
(1+2) TOTAL STATEWIDE SPENDING ON TRIPS (S MILLIONS)
Boat Fuel 37.9 63.4 28.8 41.5 171.6
Temporary Dockage 13.8 4.9 0.5 3.4 22.6
Pump-out & Launch Fees 1.1 2.1 0.2 2.7 6.1
Repair & Maintenance 1.6 8.5 16.0 16.2 42.3
Marine Supplies 6.9 5.4 7.2 6.1 25.7
Restaurant 29.1 23.4 18.1 29.2 99.8
Groceries 23.0 18.7 42.0 51.0 134.7
Auto Gas 5.5 8.1 9.5 32.0 . 55.1
Shopping 12.9 2.4 0.9 1 1.2 27.3
Recreation 2.7 4.9 1.8 15.3 24.6
Other Expenses 3.8 Q 3_.1_ L6 2E
TOTAL 138.2 142.5 128.2 226.2 635.1
Percent of Spending by Segment 22% 22% 20% 36% 100%
SUMMARY (S MILLIONS)
Trip Spending Near Home 37.1 107.0 104.5 72.7 321.3
Trip Spending Away From Home m m 2y fl 3_13§
138.2 142.5 128.2 226.2 . 635.1
Trip Spending Near Home 27% 75% 81% 32% 51%
Trip Spending Away From Home 7_3_% 25% % fig w
100% 100% 100% 100% 100%

 

138

 

mod».-

 

N... . c... 2.2. 2...:- ....... 2... 2.8 2... 2....- ......N- .2...
8. .- ...... .8... 2... 2.... NN... o... 2.... n... ..N...- on..- .88.... 850
... .- 2 .. .2. .3 c...- ...... ...... ..N.. 2.. 2...- .N...- 8.8.8.. e .58....
....N- 2 .. NN... 8.. NN..- 2.... .2. MN. NNN m...- NN. .- 80 93.
m . .2- .N.N 2N .N.N. ....2- SN .2 N... 8.. MN. .- ..N..- 8:820
.N.- N... .N.. .N. N..N.- 2... .8. 2N 2... .2.- ..N- 8598..
... .- 2.... .2. 5.. .N.N- NM... 2 .. a... N... NN...- m... .- 82...... 8.82
N...- ...... 2.... 2.... 2..- a... 2 .. 2N .....N ..N...- .....N- 8888...). N. .88.
2....- N... N... 2.... 0....- m... 2... 2.... x... .N... 2.... .88.. 8 8280
NNN- .N.N 2... .3. MN..- N...N .N.N ..N... 2: 2....- .N...- .2... a...
20.0mm 0. $0. go 2.0 52
8... .N... 2...... NMNN .2; 2.... SN. 2..... 2... .N.N .....N. .50..
2 .o N... 2.... 2... ...N .N. m... .N. NNN .N. 2.... 888...”. 850
2... .N.N o... a. on... N... N... .2. m... N...N 2... 8:888. a. m2......
2.... .N... NN.. ....N ......N ...N N..N z... 2.... ..MN m... 80 2.2
2... 2N N... 2.2 .N.N. .N.N 2.. no... ....N. e... 8.: 8.880
N... 0.... .N.N .N. . NmNm . . .. N... .N... .N... 2.... .8... .5882
2... N. N... 8N N8... 3... 2.. 2... SN 8... .3 8:8... 8.8.).
MN... .N.N .... a... ...... SN 2... ...m 2.... N... o... 8888...). 8 :88.
8... 8.. ..... 8N ....N .N.N N...N m... ...N a... 8.. .88.. 8 8.8.80
.. .. ...N No... .2: MN... 2.... N... m . .N. .N.N. .N.. N...N .8... 8...
$5000 0252.... 0.2.3 .6 02.02.... .5:
8.... .N. . N .N.N. N... ......N .......N SNN ....N .....2 .....N 2.2 . -.<N0e
2 .. 8... .8... o... .N. . 2.... 8.. .8... a... 2... No... 888...”. 850
z... N... .8... SN 8.8 N...N N. .N .N.N N... N..N a... 8:888. e m8.82..
.N.N N..N 8.. .2. .N.N 2N ....N .N.N a... .3 2 .w 80 2.2
2... a... 2N N... 2..... 2.. .N... ..... N... NNN N08 8.820
2... .. .m 3.. o... 2...... w... .2 N..... .N.N ..... 2.2 .5282
.....N 3.... .m... N... .8... N... No... .8... ..N... .N.. z... 8.3%. 8.8...
.3 MN. N... .....N 3.2 N... .0. 2... ..N.. .N.N 8.. 8888...). e :83.
2.... w... .2. ..N.. NMN. m... ..N.. .N. N... ..N.. N... .253 a. 88.8..
2.... .N. NNN n... 2...... 2... MN. SN 2... SN .N.N .8... 8...
2.00.0 E 0252...... .5:
exam n— D n— : fina— :_ 652 5 $03 _EEUU “mo? ago _EEOU ammm
.8 50 £82 5.... 582 5.8. -58. .83 -582 -582 .8. -58.

 

203.00 mcmncomm .3 220225 at... .0 :28..ng _mcewom .0 2:25am».

.8. H28...... a... 2.88... N. 88.8... .5 8 8.80 82 .82.... .2. 2..-N

139