N81 NO. oS A STUDY OF DENTON 'S SHOPPING GOODS TRADE LOSSES TO DALLAS AND FORT WORTH THESIS Presented to the Graduate Council of the North Texas State College in Partial Fulfillment of the Requirements For the Degree of MASTER OF BUSINESS ADMINISTRATION by 223491 Roy C. Anderson, B. B. A. Longview, Texas August, 1953
75
Embed
A STUDY LOSSES TO THESIS North - Digital Library/67531/metadc699543/m2/1/high_res_d/1002603921...N81 NO. oS A STUDY OF DENTON 'S SHOPPING GOODS TRADE LOSSES TO DALLAS AND FORT WORTH
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
1. The Number of Respondents in EachIncome Level .9.0. . *. . . . 0. .. . .. . 28
2. Where the Members of the Households Work,The Number, and the Percentage Workingin Each Locationn.. . . .. 0.. . . . . . 29
3. The Number of Newspapers Taken by theRespondents . .. .. . . . . . . . . .. 30
4. Suggested Reasons Why People Go to Dallasand Fort Worth to Shop . . . . . . . . . 31
5. Where the Respondents Prefer to BuyApparel . . . . . . . . . . . . . . . . . 33
6. The Number of Purchases Reported for VariousArticles of Apparel, Where Purchased, andthe Number and Per Cent in Each Location . 34
7. Purchases of Furniture, Where Made, and theNumber and Per Cent in Each Location . . . 36
8. Purchases of Appliances, Where Made, and theNumber and Per Cent in Each Location . . . 37
9. Purchases of Automobiles, Where Made, and theNumber and Per Cent in Each Location . . 38
10. Total Purchases in Each Merchandise Class,the Number, and Per Cent in Each Location. 39
11. Purchases Made in Each Location Combiningthe Four Merchandise Classes, the TotalNumber and the Total Percentage . . . . . . 41
12. Suggested Reasons Why People Go to Dallasand Fort Worth to Shop by Income Groups 43
iv
13. Where Respondents Prefer to Buy ApparelbyIncoe Groups ............. 44
14. The Number and Per Cent of Purchases ofApparel in Each Location by IncomeGroups and the Total Number and Per Centin Each Location . . . . . . . .. *.. * .*.45
15. The Number and Per Cent of Purchases ofFurniture in Each Location by IncomeGroups and the Total Number and Per Centin Each Location . . . . . .. . . . . . . *. 46
16. The Number and Per Cent of Purchases ofAppliances in Each Location by IncomeGroups and the Total Number and Per Centin Each Location . . . . . . .. . *., *.. . 47
17. The Number and Per Cent of Purchases ofAutomobiles in Each Location by IncomeGroups and the Total Number and Per Centin Each Location . . . . . . . .. . . . . . 48
18. Total Per Cent of Purchases Combining theFour Merchandise Classes Made in EachLocation by Income Groups . . . . . . . . . 49
19. The Per Cent of Respondents in Each IncomeLevel Who Preferred to Shop in Dallasand Fort Worth . . . . . . . .. ..9 * * . .0 49
20. The Per Cent of Purchases in Each MerchandiseClass and the Total Per Cent Made inDifferent Locations by Those Who Preferredto Buy Apparel in Denton and Those WhoPreferred Dallas and Fort Worth . .* . . . 51
21. The Per Cent of Purchases by MerchandiseClass and the Total Per Cent of Purchasesin Each Location Made by the RespondentsWho Did and Those Who Did Not Work inDallas or Fort Worth -orth-.*.-.-.- - . . . 52
V
Table Page
Table
22. The Per Cent of Purchases by MerchandiseClass and the Total Per Cent of Purchasesin Each Location by the Households ThatDo and Those That Do Not Take Dallasand/or Fort Worth Newspapers . . . . . ..
vi
Page
CHAPTER I
INTRODUCTION
This study deals with retailing problems characteristic
of small cities located within the trade orbit of large
cities. Denton, Texas, is a small city with a population
of 21,3451 people. The city is in the trade orbit of Dallas,
Texas, which has a population of 432,927,2 and Fort Worth,
Texas, which has a population of 277,047.3 Denton is thirty-
eight miles northwest of Dallas and thirty-five miles north
of Fort Worth.
Large cities have numerous attractions that a small
city cannot offer. For example, Dallas and Fort Worth at-
tract people from Denton for such things as athletic events,
medical attention, shopping accommodations, recreation, pro-
fessional skills, business transactions, and the like. Be-
cause of its size, Denton cannot possibly offer the varied
attractions found in Dallas and Fort Worth.
Dallas and Fort Worth newspapers have wide circulations
in Denton. All Dallas and Fort Worth radio stations are
1Texas Almanac and State Industrial Guide, p. 86.2 1bid* 3lbidtp 87.
1
2
easily in range of Denton, and all television programs
seen and heard locally are televised through Dallas and
Fort Worth channels. Also, Dallas and Fort Worth offer
many more job opportunities than Denton.
All the above factors contribute to influencing the
shopping habits of some Denton residents since people be-
come potential shoppers when attracted to Dallas, Fort
Worth, or other cities for any reason. One can readily
see the difficulty Denton merchants might have in retain-
ing the city's retail trade. This study was designed to
show proportions of the trade losses of selected shopping
goods items and possible reasons for such losses.
Statement of the Problem
The primary objective of this study was to determine
the proportion of shopping goods trade Denton merchants are
losing to other cities. A secondary objective was to find
reasons for the loss of trade. Since Denton is within the
trade orbit of Dallas and Fort Worth, special emphasis was
placed on finding the percentage of trade going to these two
cities. The study deals only with residents dwelling inside
the city limits of Denton, and it is limited to the fields
of apparel, furniture, appliances, and automobiles.
3
Elements of the Study
Several questions will immediately arise in the readers
mind when considering this study. First, the reader will
want to know how much shopping goods trade Denton merchants
are losing to Dallas, Fort Worth, and other cities. Next,
the question will arise as to why Denton residents make
purchases elsewhere. When asking why Denton consumers pur-
chase elsewhere, the reader will be curious about the in-.
fluences and attractions Dallas, Fort Worth, and other
cities have to Denton shoppers. Next, the reader will ask
what products are most commonly bought elsewhere, and
whether or not these purchases are made by distinct groups
of shoppers. Also, the reader will wonder if consumers are
actually dissatisfied with shopping accommodations in Denton,
or if other influences guide their buying habits. Lastly,
the reader will want to know how Denton retail merchants
can combat the loss of trade.
The study makes an attempt to answer these questions
by determining what shoppers have done in the past, and then
conditions that exist which might influence them.
Scope of the Study
The study deals only with shopping goods, since con-
venience goods are usually purchased near the consumer's
place of residence. The nature of convenience goods
4
causes them to have little or no value toward the objectives
of the study.
No consideration was given to purchases made in Denton
by persons living in other towns. This would tend to off-
set the purchases made in other places by Denton residents,
but showing this was not a purpose of the study.
It was assumed that the significant part of the retail
trade being lost by Denton merchants was going to Dallas
and Fort Worth. Therefore, the percentage of the trade
loss was not determined for any particular cities other
than these two.
No pretension was made to state exactly how much trade
Denton is losing, and certainly no dollar volume was esti-
mated. Neither could it be stated definitely that certain
elements are present causing the trade loss, but indications
and influences as to why trade was being lost were derived
from the study.
Purpose of the Problem
This study should be of real interest to Denton retail
merchants since its purpose was to explore the trade losses
sustained by a small city located in the orbit of one or
more much larger cities. After understanding their position,
Denton merchants will be better qualified to decide the
best courses of action that should be taken to reduce trade
losses.
This study should also be of special interest to the
Denton Chamber of Commerce since it affords some classi-
fication data on the residents of the city.
Survey of Related Literature
Very little has been written on the subject of Denton's
retail trade movement. A thesis was written by John Brooks
on the topic of the retail trading area of Denton with the
application of Reilly's Law of Retail Gravitation. The
study determined the breaking points of the Denton retail
trading area and analyzed the characteristics of the area.
Other literature that has had an important bearing on
this study deals with techniques and methods that should be
used in making such an investigation. The books considered
to be exceptionally good by the author and of the most
value were Marketing and Distribution Research by Brown,
Statistical Techniues M in ng Research by Ferber,
Basic Methods of Marketing Research by Lorie and Roberts,
The Technique of Marketing Research prepared by the American
Marketing Association, and A Survey of a Retail Trading
Area by Jones.
Definitions of Terms
In a study of this nature there are always a number of
terms that should be defined to expedite the reader's.
understanding. The following paragraphs are devoted to
6
defining terms and words that have a particular meaning in
the study and which should be clarified.
One of the first terms used that should be clarified
is "shopping goods," which Maynard and Beckman described
as those goods in which
. . . the consumer desires to compare prices, qualitiesand styles at the time of purchase. The shopper wantsto compare the offerings of a number of stores andmake a decision on the basis of actual comparison ofthe various offerings.4
The term is used in this study to cover the fields of ap-
parel, furniture, appliances, and automobiles.
One should also have a knowledge of how the term
"retail merchants" is used. Maynard and Beckman describe
a retail merchant as "a merchant middleman who sells pri-
marily to the ultimate consumer."5 Their definition covers
the manner in which the term is used in this thesis.
The term "trade loss" is very important since this
study is principally concerned with Denton's trade loss.
The author uses the term in relation to the amount of shop-
ping goods trade done in other cities by Denton consumers.
The word "residence" also has a distinct meaning in
this study. A residence is considered to be the dwelling
place of one or more persons comprising an individual
4H. M. Maynard and T. N. Beckman, Principles of Market-jim, fourth edition, p. 27.
5I~a. p. 115.
7
household. Each separate dwelling unit within the city
limits of Denton is referred to as a residence.
When the word "household" is used, it refers to one
or more persons living in a single dwelling place. It
includes any persons who live together collectively under
one roof.
CHAPTER II
SOURCES OF DATA AND METHODS OF PROCEDURE
Sources of Data
The data used in this study were obtained from members
of the individual households in Denton, using the survey
method.
Methods of Procedure
The, sample ds .-- The methods of procedure used in
this study were designed to obtain information from Denton
residents that would yield a true picture of what their
purchasing habits actually are and possible reasons for
these habits. Every consumer was potentially a source of
valuable information, but because of the limitations of
time and resources, the author chose what appeared to be
the most applicable sample design as a means of obtaining
the necessary facts.
Seldom is it feasible for researchers to use simple
random sampling methods to determine characteristics of a
given population. For this study, the necessary information
was not available from which a simple random sample could
8
9
be taken; therefore, a more complex form known as cluster
sampling was used.
To justify the use of cluster sampling, the reader
should have some understanding of basic random sampling
theory since this theory holds true for both types of
samples. When a survey and a sample of the potential
respondents to determine facts about a population are used,
the respondents interviewed must be representative of the
entire population and every person in the population must
have an equal chance of being chosen. If a researcher was
looking for information on all consumers in the city of
Denton and used the telephone directory to compile the
list of persons to make up his sample, this sample would
not be representative since every person in Denton did
not have an equal chance of being chosen. However, if a
list of all the residents in Denton could be obtained and
the persons used in the sample were chosen at random after
determination of the necessary number in the sample, it
would be representative of the entire population. "When
a certain segment of a population has no chance whatever
of having any of its members included in a random sample,
that sample cannot be representative of the whole popula-
tion." In the first instance, those people who did not
1Robert Ferber, Statistical Techniques in MarketResearch, first edition, p. 70.
10
have telephones had no chance of being chosen, but with a
list of the total residents of the city, every person had
the same chance of being chosen, thus producing true ran-
domness.
One of the basic concepts of sampling found through
statistical investigation is that the characteristics of a
representative segment of a given population will resemble
the characteristics of the whole population with certain
limitations which can be stated as definite numerical values.
Within these limitations, this representative segment will
yield information as accurate as though the whole popula-
tion had been analyzed. This has enabled researchers to
determine characteristics of a Yhole group by analysis of
only a portion of that group. The method is valid since
"sufficient similarity exists among large numbers in any
population to permit the selection of a few as representa-
tive of the entire group," 2 and "although some sample items
will underestimate the true value of their groups, other
sample items will overestimate their respective true
values.3
There are many ways in which cluster sampling may be
applied. This study makes use of the method technically
known as "cluster sampling with subsampling: unequal-sized
2..bid.,p. 65. .3.bid, p. 66.
11
clusters." When using cluster sampling, each element of
the population is not considered as a sampling unit, but
instead, groups of elements are considered. The most
widely known method of cluster sampling is the area sample
where mall compact areas such as city blocks are used as
the sampling units. Either the area as a whole may be used,
or one may choose individual elements in the area to pro-
duce the random sample. No matter what combinations of
sampling procedure are used, each time a sample is taken
it must be done by random methods. When this procedure is
followed, each sampling unit and each subsample has the
same chance of being chosen.
To accomplish the objectives of this study, it was
necessary to measure the purchases and classify the re-
spondents according to households. The interest was not in
data concerning individuals but instead the actions of the
households. The city of Denton was used as the total uni-
verse, and the city blocks were considered as areas for the
sampling units. Within these city blocks, an equal pro-
portion of potential respondents was chosen to be repre-
sented in the sample. Through the use of this design, every
household in Denton had an eqal chance of being chosen.
J. H. Lorie and H. V. Roberts, Basic Methods of Market.ig Research, first edition, p. 128.
12
When following the design previously outlined, the
author found much of the required information at the Denton
City Hall. The first step was to obtain a map of all the
residential blocks inside the city limits. This map was
furnished at the City Hall with the blocks numbered sepa-
rately and the city divided into four major sections. A
list of all these blocks was compiled giving them the same
number that appeared on the map, and then a random sample
was taken. Of the total 426 blocks in the city, 107 were
chosen to be used as clusters by beginning with the second
block and taking every fourth block thereafter. The 426
blocks in the universe did not include those known to be
made up of only business establishments.
The next step was to compile a list of every residence
in the 107 clusters taken from the city map. Here again,
the City Hall furnished the necessary information with a
tax file giving the addresses of every residence and the
vacant lot numbers in each block of the city. One diffi-
culty, that of no distinction between the addresses of
residences and those of business establishments, was en-
countered. These addresses found in the City Hall were
imprinted on Address-o-Graph plates, and the blocks were
numbered exactly as they were on the city map. For example,
if a block was numbered 384 on the city map, one would look
under block 384 in the Address-o-Graph file and find all the
13
addresses of the residences in this particular block im-
printed on the plates. All the plates for each of the
blocks previously chosen to be used as clusters were then
removed, and the address of each building in that block was
stamped consecutively, using separate sheets of paper for
each block. This funished a list of all the addresses in
the chosen 107 clusters, with the list totaling 1,294
addresses. The Address-o-Graph file had recently been
brought up to date so that it was ideal for this use.
The next step was to select those addresses that would
make up the sample. It had been previously decided that the
sample should be composed of approximately 250 households.
This number would provide a fairly reliable sample after
subtracting vacant residences and absent residents from
the list. Actually, 258 addresses were taken by beginning
with the third address on the list and taking every fifth
address thereafter.
Method of collecting data.--It would have been im-
possible for the author to make all the interviews nec-
essary for this study. To overcome this difficulty, the
students in the Business Administration 470 Marketing
Research class at North Texas State College were used as
interviewers. This was satisfactory since they had had
some previous experience in interviewing an d were in the
midst of a course directly related to the methods used.
The 258 addresses representing the sample were written
on index cards and distributed among the twenty-six mem-
bers of the class. The students were given maps of Denton
on which they plotted the approximate locations of the
potential respondents. Then the questionnaires were dis-
tributed among the students, and a class discussion was
held on the contents. The next step was to distribute
instruction sheets describing in detail the information
the interviewers were to obtain. The instruction sheets
were discussed thoroughly with the class, thus avoiding
misunderstandings. These instruction sheets are placed in
the Appendix of this thesis for further reference. Also,
discussion was held and instruction given on methods to
be used in approaching the respondents, the psychology
underlying the ways in which questions should be asked,
plus other pertinent techniques of interviewing. The stu-
dents were allowed two weeks to complete their interviews
and bring the results back to class. They made contact
with each household personally and made call backs when
necessary, which resulted in the completion of 215 question-
naires. - The remainder could not be found, or the addresses
were those of commercial establishments.
The elimination of bias is one of the most important
elements in the collection of data. It results from faulty
interviewing as well as faulty responses. These errors
14
15
should be recognized and eliminated as far as possible
before the collection of data begins. The purpose of the
discussions in class on the techniques of interviewing was
to overcome, as much as possible, any bias induced by the
interviewers. The bias could not be controlled as well
when received through responses; however, the interviewers
were instructed in ways they could get the desired informa-
tion with the least likelihood of biased answers. They
were also instructed to discount any responses that were
obviously biased. For instance, if the respondent stated
he owned a television set, yet no antenna was on the house,
the response was discounted since it is known that televi-
sion sets in this area generally require antennae to be
effective. The techniques of searching for all the types
of information sought for in this study are too extensive
to be discussed here; however, the interviewers had been
subjected to the study of these techniques in the scope of
their marketing research course and were familiar with them.
The estionnaire and its. development.--The question-
naire was the sole method of collecting data, so it was
necessary to give much time and consideration to its de-
velopment. The following paragraphs discuss the procedure
used in concluding mhat should be asked. Also, the four
sections of the questionnaire are discussed, giving the
objectives and value of each question.
16
In deciding what information should be collected to
determine how much shopping goods trade is leaving Denton
and why it is leaving, the first step was to list the in-
formation considered essential. Next, questions were de-
vised that should furnish this essential information.
Questions the author considered relevant to the studra were
brought before the Marketing Research class and fully
analyzed as to their potential value. Then others were
suggested by the class and analyzed in the same manner.
Again the class was helpful by aiding in the determination
of the final questions. After several periods were spent
in discussing what the questionnaire should include, the
final arrangement of each question was written on the black-
board and tested for ambiguities and other flaws. Next
the questions were copied in logical sequence, then mime-
ographed on 8*-by-l34-inch paper. The sequence of the ques-
tions ranged from the general to the specific so as to
focus properly the attention of the respondent on the sub-
ject. A copy of the final questionnaire has been placed
in the Appendix of this thesis.
The first section of the questionnaire was concerned
principally with classification data. The "a" part asked
for the race of the respondent, which was to be answered
through observation by the interviewer. Only two races
were considered, white and colored. The "b" part was
17
concerned with the approximate income level of the re-
spondents, and here again the interviewer supplied the
answer through observation. The remaining parts inquired
as to the size of the household, the composition of the
household including the ages of the daughters and sons,
the family position of the respondent, the occupation of
each person employed, and the city where each worked. All
these data were necessary to determine characteristics of
different groups and to test the influence certain factors
might have on where the respondents made their purchases.
Section II of the questionnaire had only one portion
which asked what newspaper or newspapers were taken by the
respondent. In the original design of the questionnaire,
it was thought that there might be a close relationship
between the newspapers taken and the cities in which the
respondents made their purchases.
The "a" portion of Section III stated as a fact that
many Denton residents do much of their shopping in Dallas
and Fort Worth, and then asked the respondents why they
supposed these people went so far to shop. No definite
value can be placed on this question, but it was hoped
that indications would be made as to why people shopped
in Dallas and Fort Worth rather than Denton. Even though
the respondent did very little of his or her shopping out-
side of Denton, if his answer stated that goods were cheaper,
18
it would indicate his probable attitude. If it was found
that this attitude was prevalent, the Denton retail mer-
chants would be supplied with an indication of possible
reasons for trade losses. If the answers indicated that
people had no reason to go so far to shop, it would sug-
gest that the respondents were satisfied with Denton re-
tail merchants. No exact conclusions can be drawn from
this question, but if one or more answers were given a
number of time, it would likely indicate the attitude of
a large portion of the residents of the city.
The second question in Section III asked if the re-
spondent preferred to buy clothing in Dallas, Fort Worth,
or Denton. This question was specific and some indications
could be deducted from it. However, the possibility of
bias was recognized since there is a tendency to feel an
obligation toward trading in the town in which one lives,
and since some people would say they preferred to shop in
Dallas or Fort Worth even though they never did.
The next portion of the questionnaire was one of the
most important parts. It asked in what town the last pur-
chase of a number of articles was made. These articles
included men's suits, hats, and shoes; women's dresses,
hats, coats, and shoes; children's shoes, and clothing;
plus furniture and rugs. Valuable information was obtained
19
from this question since it probably referred to some re-
cent event, and it required the recollection of a concrete
experience. This was designed to show where the respondents
actually made their last purchases no matter what town they
stated as their preference. After applying the theory
that true values will sometimes be overstated and others
understated, which results in offsetting the incorrect
values, valid conclusions can be drawn from these answers.
Section IV of the questionnaire required the listing
of the appliances owned by the respondents. The appliances
listed were an electric mixer, radio, television set, gas
or electric range, sewing machine, vacuum cleaner, washer,
home freezing unit, gas or electric refrigerator, ice box,
and a hot water heater. The interviewer was to ask in
what town the purchase of each appliance was made if the
appliance was bought since the respondent had been living
in Denton. This also was an important part of the survey
since it solicited concrete information that could easily
be remembered by the respondents. Places of actual pur-
chases were shown, providing valuable information from
which conclusions could be drawn. The latter portion ofSection IV, and the last question on the survey, asked
whether or not the household owned a car or cars, the make,the model, and where bought if purchased since the respondent
had been living in Denton.
20
Processing the Data
The processing of data involves summarizing and pre-
paring the facts found in the survey for presentation.
Lorie and Roberts describe it as an operation "designed to
convert raw data into a form more suitable for analysis and
presentation."5 Editing the surveys for the purpose of
correcting obvious errors is the first stage of processing.
Next, the answers must be classified or coded and a tabu-
lation made of all the answers falling into the different
classes. Then the relationships between the primary classi-
fications must be found through cross-tabulations. The
last step in processing the data is to present an over-all
picture found from the facts in a logical and organized
manner. This should be done through the use of charts,
tables, illustrations, and the like so as to facilitate the
preparation of the final report.
In applying the above procedure to this study, the
surveys were edited as they were returned by the inter-
viewers, and a running tabulation of the results was kept
on large accounting worksheets. In making the primary
classifications, categories were set up for each question
and tabulations made accordingly.
The next major step in processing of data was to find
the relationships between the primary classifications
5 i., p. 22.
21
through cross-tabulations. Four separate sets of cross-
tabulations were considered necessary to aid in determining
the findings of the study. These four sets included tabu-
lations for the different income levels; for those who said
they preferred to shop in Dallas or Fort Worth; for those
who work in Dallas, Fort Worth, or their immediate areas;
and for those who take Dallas and Fort Worth newspapers.
In making all the cross-tabulations, the same categories
were used as those for the primary classifications; however,
some of the questions were not taken into account, depend-
ing on whether or not they had any value toward the ob-
jectives of the cross-tabulation. The method of cross-
tabulation was simply to remove the sets to be cross-tabulated
from the total group of questionnaires, and tabulate them
separately. For example, all the questionnaires that stated
the respondent took a Dallas or Fort Worth paper were re-
moved and tabulated individually from the total group.
Actually, the sets had to be broken down again so that
separate tabulations could be made depending on the char-
acteristics in question of the particular set. As in making
the cross-tabulation for those who took Dallas or Fort Worth
papers, it was necessary to tabulate those who took a
Dallas paper separately from those who took a Fort Worth
paper.
22
The final step in processing the data was to present
an over-all picture found fram the survey in a logical and
organized manner. The best method of doing this was to
compare the results percentage-wise and set the percentages
up in tabular fonm. In placing the data concerned with
primary classifications in final form, the results were
given percentage-wise for all the significant factors that
were found. Comparisons can be made from these percentages.
Placing the data in final form for the cross-tabulations
was done in a manner similar to that for the primary classi-
fications. The same categories were used, but the percent-
ages were broken down further depending on the objective
of the cross-tabulation. For example, in making the cross-
tabulation for the different income levels, the same cate-
gories were used as in the primary classifications, but
percentages were figured according to income levels.
The Margin of Error
In the previous discussions reference was made to the
statistical evaluation of the possible error or tolerance
of a sample, but there has been no explanation of its use.
The purpose of calculating tolerance is to determine sta-
tistically the number of interviews necessary in the sample
to produce reliable results. There must be enough interviews
23
so that the margin of error will not be large enough to
cause the findings of the survey to be invalid.
The following formula has been devised to show the
necessary number of cases required in the sample if it
is to be within 90 per cent of certainty:
N .2.71(pq)E2
Explanation of symbols:
N is the number of cases required in the sample
p is frequency of occurrence of phenomenonmeasured (successes)
q is 1-p (failures)
E is maximum allowable error either way in per-centage points (tolerance)
Two phenomena were being measured in this study; first,
the portion of shopping goods trade Denton retail merchants
were retaining; and, second, the portion of shopping goods
trade the merchants lose to other cities. It was necessary
to estimate percentages representing the frequency of oc-
currence of these two phenomena. This estimation was set
at 80 per cent and 20 per cent respectively. Also, it was
desired that the maximum allowable error either way be no
more than 5 per cent.
Lyndon 0. Brown, Marketing and Distribution Research,P. 475.
214
After deciding on the allowable tolerance and esti-
mating how frequently the phenomena being measured occur,
one can then substitute the figures into the formuLa to
find the number of interviews necessary to produce a sample
within 90 per cent of certainty. Substituting the figures
into the formula, the following was found:
N : Number of cases required in the sample
p z 80 per cent or the portion of trade retainedin Denton
q- 20 per cent or the portion of trade leavingDenton
E::5 per cent or the maximum allowable erroreither way
N 2.71(80x20)25
~o 2 5
No: 174
To have the sample within 90 per cent of certainty, it
was necessary to take at least 174 interviews. Actually,
215 interviews were taken and the respective phenomena oc-
curred 81 per cent and 19 per cent of the time. This caused
the sample to have a tolerance of 4.4 per cent rather than
5 per cent. It proves that nine times out of ten, the
sampling error will be 4.4 per cent or less. Seldom is it
necessary to have a sampling error smaller than this.
When using the previous formula, the same margin of
error computed for the primary classifications cannot be
applied to the cross-tabulations. In either case the in-
dividual cross-tabulations must stand alone. It was not
necessary to determine the sampling error for the cross-
tabulations in this study since the results were referred
to only as indications or influences and not as precise
facts.
CHAPTER III
FINDINGS OF THE STUDY
The results of this survey are presented in tabular
fom so that comparisons can easily be made and signifi-
cant facts recognized. Throughout the tabulations, cate-
gories were established to record the individual responses
which allowed each answer to fit into at least one cate-
gory. These categories can be seen in the tables present-
ing the findings of the study. The tables have been
analyzed pointing out the significant results and explain-
ing the possible meanings of the results.
The Primary Classifications
Section I of the questionnaire dealt only with classi-
fication data. Before taking this survey, it was impossible
to tell exactly what classification data would be necessary
to produce significant findings. Some of the questions in
Section I were not essential and were designed partly to
provide a conversational starting point between the inter-
viewer and the respondent and to focus the respondent's
attention on the subject. However, some of the data were
pertinent and were necessary to find the relationships
between the primary classifications. Only the data which
26
27
resulted in significant findings and the data that might be
of interest to Denton merchants or the Chamber of Commerce
are presented and discussed.
The tabulation f or the "a" part of Section I concern-
ing the race of the respondents shows that 181, or 89.8 per
cent, were white, and 22, or 10.2 per cent, were colored.
The exact proportions of colored and Yhite people in Denton
could not be found, but the cluster sample used for the
study should have picked up a reasonably proportionate
number of both races.
The categories used to tabulate the different income
levels of the households can be seen in Table 1. The inter-
viewers were directed to give close consideration in es-
timating the income level in which the households should
fall. The interviewers were to answer the question last
after evaluation of such factors as the principal sources
of income for the household, the size and condition of
the home, the appliances owned, the make and model of their
car, plus any other contributing factors. The households
classified in the upper level should have had an income
of approximately $20,00o or more a year. Those in the
upper-middle bracket should have had an income between
6,ooo and $10,000 a year. The middle level consisted
of households with a yearly income of about $14,000 to
$6,000. Those in the lower-middle bracket should have
28
had an income of approximately $2,000 to $4,000 yearly, and
any households that appeared to have an income smaller than
this should have been included in the lower bracket. The
number of households classified in each of these categories
can be seen in Table 1. Here again the exact proportions
of respondents that should have been included in the cate-
gories were not known. However, it appears that the sample
picked up a fairly good representation of various income
One can see from the table that there is very little
difference in where the two groups of respondents made
Placeof
Purchase
their purchases, since there is a margin of only 5 per cent.
With a margin this auall, it cannot be stated as a fact
that persons working in Dallas or Fort Worth do more of
their shopping in the two cities than other people. It
appears that most of the Denton residents working in
Dallas and Fort Worth are in the middle income group and
below. This would have a tendency to reduce the percent-
age of purchases in the two cities, since people in the
upper income level made the largest portion of purchases
outside of Denton.
The cross-tabulation concerning the respondents who
take Dallas and Fort Worth neK papers.--The purpose of
this cross-tabulation was to see if there was any relation-
ship between the newspapers taken and where purchases were
made. Percentages are given in Table 22 representing the
actions of those who did and those who did not take Dallas
or Fort Worth papers.
The percentages show there is a sizeable variation in
places of purchases between the two groups of respondents.
Those who took a Dallas or Fort Worth paper made approxi-
mately 10 per cent more of their purchases outside of Denton
than those who did not take either of the papers.
It was previously found that the higher the income
bracket the more likely it would be for the respondents to
make purchases outside of Denton. Also, it is logical
that the higher the income bracket the more likely the
respondent is to subscribe to a Dallas or Fort Worth paper.
These two factors go hand in hand and possibly cause the
pr oportion of purchases in cities other than Denton to be
TABLE 22
THE PER CENT OF PURCHASES BY MERCHANDISE CLASS AND THE TOTALPER CENT OF PURCHASES IN EACH LOCATION BY THE HOUSEHOLDS
THAT DO AND THOSE THAT DO NOT TAKE DALLAS AND/ORFORT WORTH NEWSPAPERS
Households That Do NotTake Dallas or Ft.Worth PapersPer Cent of Pur-chases by Merchandisc Class I
H
P4<4
- q~-
0
4-) 04
P4
A0H
0
44
TotalPerCent
Households That Do TakeDallas or Ft. WorthPaers-Per Cent of Pur-chases by Merchandise Class
H-0
F4
04P4
e
P4p.4
00
OHto0
45
rotalPerCent
Denton 84 83 91 71 85 72 70 81 69 75
Dallas 8 7 5 10 7 14 17 9 7 12
Ft. Worth 5 4 2 2 4 11 8 5 9 8
Others 3 6 2 17 4 3 5 5 15 5
Total 100 00 .100100O-100 100 100 100 100 100
higher for those who take outside papers. It cannot be
stated that taking a Dallas or Fort Worth paper definitely
Placeof
Purchases
causes any trade loss; however, this cross-tabulation in-
dicates that it likely contributes to influencing the
shopping habits of Denton consumers.
CHAPTER IV
SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS
Summary
The respondents in this survey suggested that se-
lectivity and lower prices were the principal reasons
residents of Denton went to Dallas and Fort Worth to shop.
They also suggested that people shopped in the larger
cities because of habits and social reasons, for better
merchandise, and for ease of shopping. Twenty-one per
cent of the respondents stated that they preferred to
buy apparel in Dallas and Fort Worth rather than Denton.
Probably these same respondents preferred to shop gen-
erally in the two cities. Also, the study indicated
that approximately 19 per cent of the residents hold jobs
in cities other than Denton.
A larger portion of automobiles were bought outside
of Denton by residents of the city than any other mer-
chandise class. Furniture was second, apparel third,
and appliances fourth. Combining the purchases in the
four merchandise classes, it was found that Denton re-
tained approximately 81 per cent of the trade, Dallas and
Fort Worth obtained 15 per cent, and other cities obtained
4 per cent.
The cross-tabulation concerning where purchases were
made by different income groups indicated that there was a
definite relationship between the income level of resi-
dents and where they did their shopping. People in the
upper income bracket did more of their shopping outside of
Denton than did any other income group. Combining the
purchases of all four merchandise classes, the people in
the upper level made only about 70 per cent of their pur-
chases in Denton, whereas the middle level made 81 per
cent, and the lower level made 90 per cent. It appears
that as the income level of residents increases, there is
more of a tendency to shop outside of Denton.
One significant result of the cross-tabulation con-
cerning respondents who preferred to buy apparel in Dallas
and Fort Worth was the finding that the proportion of
people who preferred to shop in the two cities increased
as the income level went from the lower to the upper-middle.
The cross-tabulation also pointed out that the people who
preferred to buy apparel in Dallas and Fort Worth actually
made more of their other purchases in these two cities.
The respondents who work in Dallas and Fort Worth
appear to do about 5 per cent more of their shopping outside
of Denton than those who do not work in the two cities.
58
Results of the study also showed that people taking Dallas
and Fort Worth newspapers made about 10 per cent more of
their purchases outside of Denton than those who did not
take papers from either of these cities. These two fac-
tors probably have some influence on where Denton con-
sumers make their purchases. It may also be assumed that
there are a number of things influencing where purchases
are made that were not found by this study.
Conclusions
The major finding of this study is the indication
that approximately 81 per cent of Denton's retail shopping
goods trade is actually done in Denton. Since there is no
standard against which this finding can be evaluated, it
is not possible to state categorically that it is good or
bad. However, it does appear that inevitable trade losses
do occur from any city, and that relatively large trade
losses might be expected in a small city which lies within
the retail trade orbits of two much larger cities.
To estimate whether or not Denton merchants are losing
any trade in excess of the amount that would be profitable
for them to retain and the amount that will inevitably be
lost, it will be necessary to make a thorough analysis of
the findings of this study and other factors that are known
to exist. The study shows that only a small portion of
Denton's trade is being lost to cities other than Dallas
and Fort Worth. Any substantial reduction of trade loss
will require that Denton merchants concentrate their
efforts on reducing the amount going to these two cities.
Therefore, the analysis is principally concerned with
the trade being lost to Dallas and Fort Worth.
First, we will consider some of the general problems
Denton retailers have in retaining trade when competing
with Dallas and Fort Worth merchants. If a resident of
Denton intended to buy an automobile and decided on a
Chevrolet, he would have only one dealer in Denton from
which to get an offering. When considering the purchase
of an automobile, many consumers want to get the offerings
of several dealers before deciding on the actual purchase.
To do this, they must go outside of Denton, thus increas-
ing the possibility of their buying elsewhere. When buy-
ing furniture, many consumers want a large variety from
which to choose. The size of Denton limits the number of
furniture stores it can support and the inventories they
can afford to carry. This, too, increases the possibility
of consumers going outside of Denton to find the desired
variety in selection and prices. Many consumers are dis-
criminating when considering the purchase of certain
articles of apparel. Large and numerous stores can offer
6o
the necessary variety to satisfy these consumers' wants,
where stores with a relatively small volume cannot.
The study shows that consumers are less likely to
buy appliances outside of Denton than any other merchandise
class. Still, some of the trade is lost and Denton mer-
chants are limited in ways to combat this loss. For in-
stance, more special sales may be offered in Dallas and
Fort Worth than Denton merchants could possibly afford,
thus sometimes causing appliances to be cheaper in the two
cities. These problems and limitations are difficulties
about which Denton merchants can do little, and because
of them it is inevitable that some trade will be lost to
Dallas and Fort Worth.
A variety of suggestions was made as possible reasons
why Denton residents went to Dallas and Fort Worth to
shop, but the two prevalent suggestions were that there
was a wider selection of merchandise and that goods were
cheaper. It is likely that these two factors are guiding
influences in determining where many consumers make their
purchases, and are responsible for some of the trade loss.
Therefore, the two influences must be at least partially
overcome if trade losses are to be reduced appreciably.
However, the question arises as to whether this would be
possible or profitable.
The volume of business done in Dallas and Fort Worth
enables the retailers to carry a wide selection of mer-
chandise. To minimize the influence this has on where
Denton residents shop, it would be necessary for the mer-
chants of this city to increase their inventories in order
to provide wider selection. If inventories were increased
enough to even approach the variety of merchandise found
in Dallas and Fort Worth, there probably would not be
enough demand for some of the products to make it profit-
able for them to be carried.
It is problematical as to whether shopping goods are
actually cheaper in Dallas and Fort Worth than in Denton.
However, the mere fact that many Denton residents appear
to believe that this is true makes it a serious problem.
If prices actually are higher in Denton, sane price re-
ductions might retain part of the trade which is now being
lost. If prices are not higher in Denton, local merchants
seem to need to get this idea across to the people who now
believe otherwise.
The study distinctly points out that consumers above
the middle income level are most prone to make purchases
outside of Denton. If merchants are going to reduce the
trade loss, this is the group on which their efforts should
be concentrated. The principal reasons found in the study
as to why consumers above the middle income level shopped
61
62
in Dallas and Fort Worth were the wider selection, habits,
and social reasons. Still, these answers do not cover
all the probable causes for their shopping in the larger
cities. Possibly, one reason just as important is con-
sumers above the middle income level shop in the two
cities because they have the time and money to go there.
Also, they are attracted to the two cities for a number
of reasons other than for shopping, but they buy things
while they are there.
Another study could be made to find that action on
the part of Denton merchants would be necessary to retain
part of this trade. However, it can be seen without the
study that retaining the trade would be quite expensive.
This study shows that approximately 30 per cent of the
purchases by consumers above the middle income level are
not made in Denton. It is inevitable that some of the
trade will leave Denton, and that some consumers would
shop in Dallas and Fort Worth no matter what action Denton
merchants took. Therefore, it is doubtful that the action
required to reduce the amount of outside shopping appreci-
ably would be justified in view of the limited amount of
trade that could possibly be retained. If Denton merchants
offered a wider selection and became more competitive in
63
prices with Dallas and Fort Worth, it is doubtful that in-
creased sales would more than offset increased expenses
and price reductions.
There are many factors over which Denton merchants
have little control that undoubtedly are responsible for
influencing the purchasing habits of some consumers. For
example, the study indicated that taking Dallas and Fort
Worth newspapers probably contributed to influencing where
purchases are made. The fact that some residents of Denton
work in the two cities possibly causes trade loss. The
attractions held by Dallas and Fort Worth such as athletic
events, professional skills, recreation, and specialized
medical care cause trade loss. These factors appear to
be inherent in the problem of a small city competing with
a large one.
From the findings of the study, it appears that Denton
retailers may be doing a good job of satisfying the demands
of consumers in view of the limitations with which they
are confronted. Also, it does not seem that merchants are
losing much trade in excess of the amount that would be
profitable for them to make efforts to retain. It appears
that the 19 per cent of the shopping goods trade being
lost by Denton merchants is not unduly large.
64
Recommendations
The author suggests that the possibility of retaining
trade that is being lost by Denton retailers be investigated
further. Additional investigation might reveal some areas
in which profitable action could be taken. For example,
the investigation might show what would be profitable for
Denton merchants to do in order to retain a larger por-
tion of the trade done by consumers above the middle income
group. It might be found that it would be profitable to
handle a wider selection of certain articles of merchandise
or become more competitive in prices with Dallas and Fort
Worth on some articles. However, until further study has
been made and it is known what action would produce profit-
able results, it would not be wise for Denton merchants to
begin blindly making efforts to reduce the trade losses
without regard to the costs.
APPENDIX
Instruction Sheet
The first thing you should do is make an alternate copyof the addresses to be surveyed by you. Do it now andput it some place where it will not be lost.
Address.,--Use the address of the residence inter-viewed, apartment, et cetera.
I. (a) Race.--We are using only white and colored. Ifnot colored, consider the respondent as white. Circlethe "W" for hite and "C" for colored, depending on thecategory in which the respondent should fall.
(b) Income status.--Give this close consideration. Fillit in last. Five categories have been provided so as tohave more distinction. Be careful not to show any biasthrough central tendencies of rating everyone too highor too low. Very few people should fall in the high in-come bracket. If they do, they should have incomes ofapproximately $20,000 or more a year. Even upper middleincome classes should be what the average person considersfairly well off, and there will not be very many of these.Probably most of the people here in Denton will fall inthe middle and lower middle brackets and probably morewill fall in the lower bracket than most people realize.Many interviewers have a tendency to place respondents ina category higher than they belong. Do not let this biasenter into your choice.
In determining the income status, the purpose ofdoing it last is to give the interviewer a chance tobring into consideration such important factors as thenumber of people in the household, the number employed,the type of jobs held, the size and condition of thehouse, the make and model of their car, and appliancesthey own. After considering these factors, simply circlethe "H" for high income, "UM" for upper middle, "M" formiddle, "LM" for lower middle, and "L"' for low , depend-ing on the category in which they should fall in youropinion.
66
(c) Year moved to Denton.--If they have been liv-ing here all their lives, put down the approximate numberof years. If they lived here, moved away, then moved back,put down when they moved back.
(d) Size of family.--This is not the number in thefamily but the number of people living in the residenceat this particular time. Visitors should not be includedbut instead more or less permanent parties. The totalnumber of this should equal the number in "C" or the com-position of the family.
(e) Composition.--This refers to the compositionof the household. Place a check mark in the blank if thefather and mother live in this residence. For thedaughters and sons presently living in the residence,place their age in the blanks. "Others" would be suchpersons as the grandfather or son-in-law who lives inthe residence and is considered as part of the family unit.You should write in the relationship that the "others"have to the family. This may require some questioningand interpretation on the part of the interviewer, butbe especially careful that you have it correct.
(f) F . .pos. of resp., _or family position ofrespondent.--A check mark should be placed in theseblanks next to the family position of whoever is inter-viewed. For instance, if the mother is interviewed acheck mark should be placed in the blank provided inline with the word "mother." In case there is more thanone daughter or son, and either one is the respondent,then place a check mark in the blank provided in linewith the word "daughters" or "sons" as the case may be,then circle the age of the daughter or son that was therespondent. Do not interview any person who you thingsis incapable of answering the questionnaire effectively.
Under "occupation" give the work of any person inthe household that holds a full time or part time job,whether it be the manager of a department store or carry-ing a paper route. However, be sure to identify the workif it is only part-time by placing "part-time" after thedescription.
Under "city" where employed, we want to find thenumber of co mmuters and the number who work here inDenton. Give the town of employment for everyone in thehousehold who holds a job.
67
II. What newspapersdo you take means at the present timeand only major city publications such as the Denton RecordChronical or the Dallas Morning News..
III. (a) This question should be carefully read by theinterviewer to the extent that he is positive the respon-dent understands what he wants. However, do this withoutsuggesting any possible answer that the respondent mightuse. It is very important that the respondent makes uphis own mind and answers this question individually. Ifthis is not done, the answers will be worthless in computa-tion.
(b) This question is self-explanatory, but theinterviewer should again caution that he does not giveany assistance in obtaining an answer. When reading thequestion do not weight the possible answers with changesin vocal tones but instead read it in one phrase, "Doyou prefer to shop in Dallas, Fort Worth, or Denton,"and then let the respondent determine his preference.Place a check mark in the blank of their choice.
(c) This question will probably require a greatdeal of explanation on the part of the interviewer. Itis one of the most important and should be clearly under-stood by the respondent. In asking this question, donot necessarily read it with the exact wording on thequestionnaire. One might put it in his own words byasking, "Do you know where your husband bought his lastsuit?" or "Can you tell me where you purchased your lastdress?" Make it clear that this means the town of thevery last purchase of the listed items, and take care notto influence the respondent in any way.
IV. In question IV simply check those appliances therespondent owns and the town in which they were boughtin the blank to the right. Put a check mark in the blankprovided for every appliance they own that is listed, butplace the town in which the appliance was bought onlyif purchased since the household has been living in Denton.If they own more than one of a j rticular appliance, placethe number in the blank to the left and the towns in whichthey were purchased to the right. If they do not knowwhere the appliance was bought, put "don't know" in theblank provided for the town.
68
In answering the question concerning ownership of anautomobile, do not consider any business car unless itis actually the method of transportation used by thehousehold. Also, trucks should be considered only ifthey are the means of transportation for the household.In the blanks to the right, place the town in which thecar or truck was bought if purchased since the householdhas been living in Denton.
Questionnaire
Address
I. (a) Race W C (c) Year moved to DentonHUm
(b) Income status MLML
(d) Size of family Fam. pos. Occupation Cityof resp. where
employed
Father
Mother(e) Comp. Daughters
(age)Sons
(age)
Others
II. What newspaper (papers) do you take?
III. (a) A great many people do a lot of their shoppingin Dallas or Fort Worth. Why do you supposethey go so far to shop?
(b) When you are buying clothing for yourself or yourfamily, do you prefer to shop in Dallas ?Fort Worth ? Denton ?
(c) In what town was the last purchase of the follow.ing articles made?1. Man's suit2. Man's hat3. Man's shoes
69
Woman's dressWoman's hatWoman's coatWoman's shoesChildren' s shoesChildren's clothingFurniture ( *hat)Rug
IV. Check appliances respondent owns:
Electric mixerRadioTelevisionGas or electric rangeSewing machineVacuum cleanerWasher (auto. or agitator type)Home freezing unit (deep freeze)Gas or electric refrigeratorIce boxHot water heater
If bought sinceliving in Den-ton, where? (town)
Own a family car (cars)?MakeMod
Interviewer
6.7.8.9.
10.11.
BIBLIOGRAPHY
American Marketing Association, The Techniques of MarketingResearch, A Report Prepared by the Cammittee on Mar-keting Research Technique of the American MarketingSociety, Ferdinand C. Wheeler, chairman, New York,McGraw-Hill Book Company, 1937.
Beckman, Theodore N., and Maynard, Harold H., Principlesof Marketing, fourth edition, New York, The RonaldPress Company, 1946.
Blankenship, Albert B., editor, How to Conduct Consumerand Opinion Research, New YorksHarper and Brothers,1946.
Blankenship, Albert B., and Heidingsfield, Myron S.,Market and Marketing Anallsis, New York, Henry Holtand Company, 1947.
Bradford, Ernest S., Marketing Research, New York, McGraw-Hill Book Company, Inc., 19TT
Brown, Lyndon 0., Marken and Distribution Research,New York, The Ronald Press Company, 1949.
Brown, Theodore H., The oUseof Statistical Techniques inCertain Problemsof Marketing Researc Boston,Harvard University, Graduate School of Business Ad-ministration, Bureau of Business Research, 1935.
Ferber, Robert, Statistical Techniques in Market Research,New York, The McGraw-Hill Book Company, Inc.,T1.
Jones, Fred M. A Survey of a Retail Trading Area, Bulle-tin No. 4, Urbana, University of Illinois, Bureauof Business Research, 1932.
Lorie, James H., and Roberts, Harry V., Basic Methods ofMarketing Research, New York, McGraw-Hill BookCompany, Inc., 1951.
Luck, David J., and Wales, Hugh G., Marketing Research,New York, Prentice-Hall, Inc., 192,