xt7sxk84kq4z https://exploreuk.uky.edu/dips/xt7sxk84kq4z/data/mets.xml   Kentucky Agricultural Experiment Station. 1959 journals 083 English Lexington : Agricultural Experiment Station, University of Kentucky Contact the Special Collections Research Center for information regarding rights and use of this collection. Kentucky Agricultural Experiment Station Progress report (Kentucky Agricultural Experiment Station) n.83 text Progress report (Kentucky Agricultural Experiment Station) n.83 1959 2014 true xt7sxk84kq4z section xt7sxk84kq4z I Progress Report 83 November 1959
’ _ O
Farmers Use of Strateg1es
_ I I B - ·
1n Maclnnery Trades
F9 I Department of Agricultural Economics I
..`_..,.-»— `   . ~—-;· `  
   NQ   N as
B ~I  , ¤x   K    `   __T§“`w   .
r§'  A I    I-I /·’ ll      lr- I I h -‘ F I I rv ll
  I  »  °  { .r,/@&¥I§,}9 Ha N
r  ..   -. `     K ,2 I "   Y   " -·  I

By Albert N. Halter and John W. Hubbard
Department of Agricultural Economics
Farmers often find themselves in bargaining or trading positions with
other individuals. The trading or bargaining may be done with relatives,
neighbors, the corner grocery store, the feed man, or strangers. Usually
the persons involved in the trade do not possess the same information about -
the items to be traded, the personalities of the "opponent(s)" and other
I considerations that might enter into the bargaining. In this kind of situation
an individual may use certain devices, techniques, or expressions that we may
call "strategies" to enhance his position. In situations where strategies are
used we might consider such trades as "games. "
If we wish to extend our analogy we can classify trades according to the
various characteristics of the game involved. We could classify games
according to these characteristics: (1) number of players, (2) amount of pay·
1 off,"` (3) number of plays or moves in the game, and (4) amount of information
each player possesses. I
- Several examples will demonstrate the feasibility of classifying by the
above characteristics. (1) A trade between two neighbors could be called a
two-person game, whereas a deal between a farmer, a trucker, and the
J livestock buyer would be a three-person game. (2) The amount of payoff or
the net effect of the trade for the first player is $100 in the situation where
a farmer trades one of his cows worth $2.00 for two of his neighbors' sows _
worth $100. (3) While in chess or checkers a play of the game is easily
discerned as the moving of a "man, " it is not so easily recognized in real
life games. A move might include a whole sequence of actions, one spoken ’·
word, or even a gesture. (4) The amount of information available to the
players regarding past plays will vary with the individuals involved. Trades
between. strangers compared to trades between fathers and their sons will be
affected differently by the amount of information they possess about one another.
The most important feature of games or trades from the standpoint of
the participants is the outcome . The strategy or strategies that the players
use determine the outcome. By strategy we will mean the possible ways a
( player can play the game within the structure of the game given by the four
above characteristics. While a game or trade may be specified by the number
of players, their information on each other's position, the payoff, and the
number of moves, the final outcome of the game, (win, lose, or draw) is
determined by the combination of strategies that each player uses.

In a game such as football where the four characteristics are well _
specified, the set of strategies open to the quarterback to call are numerous
and, to a large extent, determine the outcome of the game, Extending this
analogy to real life situations is difficult, but obviously many strategies are
open to the players in trades and business deals that help to determine their
The idea we are trying to develop is that farmers use different strategies
in trades with other people and that it is the particular combination of strategies
that determines the outcome of the trade, Whether the set of strategies used by
an individual stems from ethical, economic, or prestige reasons is an interesting
question, but is not our concern here, Rather in the remainder of this report
we want to consider: (1) Do farmers use different sets of strategies in making
machinery trades? (2) If they do, will a particular set of;strategies determine
the outcome of the trade?
The general objective of this-pr0gress reporiiis to review the research ‘
now in process on the use of strategies by farmers. This research is only a
part of the research going on in decision-making studies in farm management.
This type of research had its beginning at the Kentucky Agricultural Experiment
Station in the publication of Bulletins 593 and 610., 1 The empirical results
reported here are from the Interstate Managerial Survey conducted in 1953 in
Kentucky and six other midwe stern states,
The specific objectives of this report are: p
l, Review the techniques used in obtaining answers to questions involving 3
farmers' use of strategies in a machinery trade,
2, Review the procedure and results of the analysis used to distinguish
sets of strategies used by farmers in a machinery trade,
3. Point out the kinds of research necessary to further the work in this
phase of the decision—making study,
The Interstate Managerial Survey concerned itself with the decision-
making functions of individual farm managers with gross incomes of $2., 500
or more in Kentucky, Iowa, Michigan, Indiana, South Dakota, Ohio, and Kansas,
G. I-., Johnson and C. B, I-Iaver, Decision Making Principles in Farm
Management, Kentucky Agricultural Experiment Station Bulletin 593
ilaexington. University of Kentucky 1953), G,   Joh.nson, Managerial
Concepts for Agriculturists, Kentucky Agricultural Experiment Station
Bulletin 619 (Lexingtonl University of Kentucky 1954),

 -3.. .
Although the results of this report are based on interviews from 362 farmers,
the total contacted was 1, 075. The main outline of the schedule used in the
survey can be summarized under the following two points:
1. Control questions asking for tenure status, size of farm, type of farm, l
background, education, income, liabilities, and other personal characteristics.
Z. Structured and unstructured questions concerned with:
a. Types of information used by farmers in organizing and operating
b. Use of managerial functions (observation, analysis, decision,
action, responsibility bearing) in solving analytical problems.
c. Sources and means of securing information used by farmers.
d. Expectations of prices, people, institutions, and events.
. e. Use of informal insurance practices.,
f. Utility of gains and losses of wealth.
( g. Recognition by farmers of knowledge situations.
{ h. Use of strategies. A
While the last item is the main concern of this report, answers to the
other questions are used in the analysis. »
  The specific questions that were asked with reference to strategies are: (
A We would like to ask you what you think should be done in the
· following situation. A farmer wants to trade his combine for a
i ` tractor. There are other farmers in the neighborhood who also
want to deal for a tractor. (RECORD ANY COMMENTS)
(a) While he's still looking around to see who has a tractor to trade for,
should he keep quiet about his intentions so as to keep people he
might want to trade with from having plenty of time to decide on
how much they would want to get'?

 -4- »
(b) When he finally decides who he’d like to trade with, is it a good idea
for him to act as though he's not sure whether he wants to trade so
that other farmers who might also be interested in a trade would think p
the tractor was not desirable?
(c) If he finds out that his neighbor is trying to make a trade for the same
tractor, should he improve his competitive position by trying to find
out what he neighbor is offering without letting his neighbor know what
his offer is ?
(d) If he meets so-nieone else who wants to trade for a tractor but doesn‘t
know about the one that he's interested in, is it better for him not to
mention that he knows about this tractor ?
Yes A
(e) Is it wise to for him to make the man he's dealing with think that a
combine is what he needs most, so that trades for other items won't
be given much consideration?
Yes T
(f) If he finds the tractor needs minor repairs the owner hasn't told
him about, is it better for him not to mention anything that might
be wrong with his combine so that he can make the trade successfully?
Yes _
The strategy questions are concerned with a hypothetical situation in that
they ask what the respondent thinks should be done in a trade of a combine for
a tractor. The situation is constructed so the respondent will place himself in

, the position of the farmer with the combine. It is believed the respondent will
1 answer the specific questions as though he were making the trade. Their
attitude toward use of strategies in this situation is believed to indicate the
extent to which farmers use strategies in their own affairs.
A set of questions using a direct approach was constructed and pre-tested
p prior to the main survey, and was found unsuccessful in obtaining answers.
( Apparently, farmers are reluctant to admit they use strategies. However, while
1 specific questions may have been unanswered, no farmer refused to answer the
entire set of indirect questions. It remains to be seen whether answers to l
hypothetical questions are meaningful with respect to other personal character-
istics and behavior of the farmers interviewed. The results of the analysis are
presented in a later section. Table 1 shows the distribution of answers to the six
T_&lE 1. - Distribution of Answers of 362 Farrners to the Six Strategy Questions
_ Question
Answer (a) (b) (c) (d) (e) (f)
“ Should or Yes 168 171 151 245 161 131
‘ °?‘Shou1dn‘t or No because bad strategy 18 13 7 2 5 17
*Shou1dn't or No because poor ethics 18 15 28 10 20 48
**Shouldn't or No, unspecified 139 152 161 95 162 156
1 No Answer 19 11 15 10 14 10
’7’Th.ese responses were interpreted from comments written by the interviewers on
A the schedules.
**Unspecified means there were no comments to interpret.
Answers to the six strategy questions indicate the opinion or attitude of
the respondent toward using certain strategies. It was believed that certain
patterns of responses not evident in Table 1 would be found in a closer __
examination of the answers. ln fact, it was believed that (l) if a numerical
weight could be given to each type of answer and (2) if these weights were
added into a score for all questions, then the response made by an individual
to a particular question could be predicted from knowledge of that individual's
score. The method of constructing the score is known as the Guttrnan technique
of scale analysis. 1
1The theory and techniques used in this report for scale construction are given
in chapters written by Guttman and Suchrnan in Measuremerit and Prediction,
by S. A. Stauffer, Louis Guttman, E. Suchman, P. F. Lazarsfeld, S. A. Star
and J. A. Clausen (Studies in Social Psychology in World War ll, Vol. 4;
Princeton, Princeton University Press, 1_950). ——
A shorter explanation is given by M. A. Hagood and D. O. Price in
Statistics for Sociologists (N. Y., Henry Holt and Company, 1952).

While we are not interested in giving the complete details of the technique (
here but rather results of analysis, we will summarize the construction of the i
scale in the following steps:
(1) The "most positive" reply, a "Yes" or "Should" in our case is
selected for each question. `
(2) The other responses are ranked according to their positiveness or 1
negativeness, in this case "No, Unspecified" is most negative.
(3) The responses are given weights from zero for the most negative
to 3 for the most positive.
(4) A total score is computed for each individual by adding up the
weights corresponding to his responses.
(5) All the individuals are then arranged in order of their total scores
from highest to lowest. `
(6) The questions are arranged in order from the one with the fewest
positive responses to the one with the most positive responses.
(7) Ideally, all the responses down to a certain point would be positive
and beyond this point all the responses would be negative. The
pattern might look like that of Table 2..
In the actual situation under consideration the ideal pattern did not appear
but only an approximation to it. In fact, the two responses "Yes" and "No, bad
strategy" were combined to make the most positive response and "No, poor
ethics" and "No, unspecified" were combined to make the most negative response.
The responses were re-weighted and new scores computed for each individual.
Again the individuals were arranged in order by their score and the question
ranked by their positiveness and negativeness.
The pattern of responses indicated that question numbers (a) and (c)
could not be used to form a scale. After eliminating these questions the four
remaining questions produced a scale with the patterns displayed in Table 3.
The table shows that question (d) was most positive, i.e. , more
individuals answered it either "Yes" or "No, bad strategy" than any other
question. Further, (f) was most negative in that more individuals answered it
"No, poor ethics" or "No, unspecified" than any other question. Individuals
denoted by scale type I were more positive about all the questions than the
other scale types. Thus, from knowledge of the scale type (or total score)
of an individual we can predict how he answered the four questions. Hence,
scale type V answered all four questions either "No, poor ethics" or "No,

 ` \
Table 2, - Ideal Pattern of Scores and Question Arrangement for Three Questions
______   and Individual Answer
_-._---.-i-L-m-- ___-GL-._-. ____.§L.-m. ‘
lndiv1diaLs__ Yes No No No Yes No No No Yes No No No
§p_n_1l·ie;£_€¥;ore. lB;S,}tP.l¢) (B.S.)(P.I-) (B.S.)(Ptl.)
l- 9 X   X X
2 7 X X X
_ 3 6 k X 1 X X
T 4 3 · X { X X °
__g__ 1 ll _ ____ X { X I X
Questiogmcore ___ 6 8 l2
Table 3. · Patterns of Responses to the Four Questions for the Guttnnian
Scale Types
  Extent of Scalability
Scale Type l”"`"`7§'§_'—`"fj_ lb) (d) Perfect Cases Error Cases
  Sl Response
I }   -i- -1- + 68 54
X li   . -i~ + + 28 is
ll} I - » + + 29 l-4
lv     .- .- + 49 32 A
V { ‘·“· ji  
__________W_§_>_________“________ 2 18 13 l
There were qi ;-ourse, deviations from the ideal situation given in Table Z -
with r s-j;·;tem, of errawre we in<,laca.¢ned by what is called a "coefficient of reproducibility
given by th  f<:·rni:-ala
C]cn·—3.‘.f1c1eii" c»€ raapro·rl·-;·;ibility =
l - _·__W_________» Number of errors
‘?·-»E¤_;nil;>-;-1· ol fiLl·2€llOT1S X number of individuals
For the   agx;es.i.12 Total
I 10.2 35.2 11.3 33.1 10.2 100
II 22.5 12.5 22.5 40.0 2.5 100
III 16.3 41.0 23.3 9.3 9.3 100
IV 3.3 36.0 19.7 32.8 A 8.2 100
V 10.1 29.1 24.1 30.4 6.3 100
aChi-square significant at 1 percent level. -
Table 5. — Scale Types and the Number of Re‘spondent‘s Dependentsa —
Scale Type Percent with Specified Number of Dependents
1 - 2 3 - 4 >4 Total
I 50.1 26.3 23.6 100
II 29.7 37.8 32.5 100
III 50.0 25.0 25.0 100
IV 41.7 28.3 30.0 100
V 44.6 12.2 43.2 100
3-Chi-square significant at 5 percent level.
bDependents include individuals who receive support from the respondent.
Table 6. - Scale Type and the Number of Children Living with the Respondenta I
Scale Type Percent with Specified Number of Children
None 1 - 2 > 3 Total
I 67. 2 25. 9 6. 9 100
II 47.5 37.5 15.0 100
III 51.2 34.8 14.0 100 A
IV 443 40.9 14.8 100
V 47.4 32.0 19.8 100
aChi-square significant at 10 percent level.
bChildren are 10 years or less.
ICharacteristics found to be statistically unrelated were 4-H membership and
variables related to farming experience.

Table 7. - Scale Types and the Stage in the Family Cyclea
Percent in Specified Stage
Couple with no
Children or Couple with Couple with I
Children less Children Children over
Scale Type than 5 years old 5-10 years old 10 years old Total
I 32.0 34.9 33.1 100
II 35.3 55.9 I 8.8 100
III 33.3 44.4 22.3 100
IV 41.3 34.5 24.2 100 I
V 44.0 30.0 26.0 100
aChi—square significant at 10 percent level.
Table 8. - Scale Types and the Structure of the Householda
Percent in Specified Category I
Married with
Respondent Married with Children or
Scale Type Unmarried I no Children Others Present Total
1 4. 8 29. 3 65. 9 100
II 12.5 12.5 75.0 100
Ill 4.8 28.6 66.7 100
IV 1.6 27.9 _ 70.5 100 ~
V 7. 5 20. 3 72. 2 100
8~Chi-square significant at 25 percent level.
These characteristics are all of a sociological nature as they are the
results of social institutions, or social interaction. This fact is interesting
in and of itself. It is also highly important when we consider their relationship
to the social act of bartering for some economic good. For example, consider
the specific relationship between education and the scale types. There is a
general indication that a group of similar farmers with no more than eight
years of school would be expected to be of a scale type with a larger index
number, i.e. , would be expected to say "No" to more of these questions than
those with more education. Evidence supporting this statement is increased
by the fact that more of those with 12 years of education or more could be
expected to be in a scale type of lower index and would be expected to use different
strategies than those with less education.
ln addition to education the relationship between the scale types and the
social institution of the family seems impressive. In relation to number of
dependents, scale type five has the largest percentage with one to two
dependents. This relationship is further generalized by Tables 6, 7, and 8
where the influence of children on the scale type is shown.
Table 9 shows the effect on scale type of having grown up on a farm.

Table 9. — Scale Type and Whether the Respondent had Childhood Experience
on a Farma
Percentage Having Specified Experience A
All of Part of None of
Scale Type Childhood Childhood Childhood
I 33. 8 50. 0 25. 0
. Il 12. 7 3. 8 0. 0
III 11.4 19.2 25.0
IV 17. 5 15. 4 37 . 5
V 24. 7 ll. 5 12. 5
Total 100 100 100
&Chi-square significant at the 30 percent level.
Although the statistical significance is relatively low generally those with
childhood farm experience are more likely to be scale types who answered "Yes"
to the strategy questions. This variable is even more interesting and perhaps
more information from the standpoint of bargaining in trades when one considers ‘
that age is not a statistically significant variable. Usually one thinks of our
older farmers as more likely to have had childhood farm experience, but since
age is unrelated to the scale types, the rural background may be more useful
in explaining differences in the strategies farmers use.
Some of the behavioral variables we found to be statistically related to
scale types were (1) geographical location, (2) debt position of respondent, and
(3) ranking of types of information.
The state of origin or location is a significant variable in understanding
why different farmers might be using different strategies. The debt position
of a farmer seemed a likely variable for explaining the differences in the way
farmers might actually make trades and hence in the set of strategies they
would choose. It may also indicate the outcomes of previous bargaining in
which strategies may have been used. ‘
Ranking various types of information as to difficulty in acquisition by
the individuals in the scale types may be the best indication that players in
any trading game have various kinds and levels of information; hence, ranking
aids in understanding the choice of strategy.
The relationship between these three variables and the scale types are ‘
shown in Tables 10, ll, and 12.
The state where the respondent lives appears to be a highly significant
variable from the statistical standpoint as shown in Table 10. The distributions
over the scale types for Kentucky, Iowa, and North Dakota are approximately
the same. However, theseidistributions are different from Ohio and Michigan
as a group whichalso differs from Indiana and Kansas as a group. As the type
of farming variable was not statistically related to the scale type, we cannot
conclude that it is due to type of farming differences. In addition, upon

 . \.
-13- .
Table 10. - Scale Types and Geographical Location by Statea
‘ State V
Scale Type Ky. Ohio Ind. Mich. N. D. Iowa Kansas
I 58.3 18.1 28.1 27.7 36.3 53.8 36.0
II 5.5 11.3 12.5 11.1 20.4 15.3 6.0 L
III 8.3 15.9 10.9 12.5 20.4 10.2 8.0
~IV 19.4 22.7 32.8 22.2 13.6 10.2 34.0
V 8.5 32.0 15.7 26.5 9.3 10.5 16.0
Total 100 100 100 100 100 100 100
3Chi-square significant at one percent level.
Table 11. - Scale Type Related to Debt Position of Respondent by Percentagea
Amount of Debt in Dollars
1,,000- 5,000- 10,000-·
Scale Type None (999 4, 999 9, 999 99, 999 Total
I 48.1 13.9 21.3 6.5 10.2 100
II 36.8 7.9 31.6 13.2 10.5 100
III 58.9 5.2 15.4 12.8 7.7 100
IV 35.7 7.2 34.2 10.0 12.9 100
V 38.5 7.7 ‘ 19.2 15.4 19.2 100
aChi—square significant at 20 percent.
re-grouping the farmers according to specialized or diversified farming and · '
relating this to scale type within the three state groupings, no significant
rel.ationship was found. »
An examination of Table ll reveals the interesting generalization that the
lower the index number on the scale type the more likely it is that the i.ndividuals
have some debts. In other words, if an individual has debts he i.s more likely
to say "No" to the strategy questions. Whether or not this indicates an avoidance
of further deal.s that might increase debts or a general conservatism on the part
of certain farmers in uncertain. A better way of measuring risk aversion than
answers to these strategy questions would seem desirable before this question
could be answered.
The outstanding feature in both sections of Table 12 is that Types 1 and
V agree that information about other humans is the most difficult to acquire.
A second feature is the differences between scale types. If the percentage is
taken as the ranking within each scale type and numbering the kind of information
from left to right across the top of the table, 1 then the scale types rank the type
of information thus:
I1 stands for price, 2 for production, 3 for new developments, 4 for human
and 5 for institutional information.

 -14- ·
Rank p
Scale Type lst 2nd 3rd 4th 5th
I 4 3 5 1 2
II 4 1 3 ~ 5 2
III 1 5 3 4 2
IV 2 5 3 1 4
V 4 3 1 5 2
Table 12. - Scale Types and Their Ranking of Information as to its Difficulty
in Acquiring
Percentage Who Ranked Number One
Kind of Informationa
(1) (Z) (3) (4) (5)
Scale New _
Type Price Production Developments Human Institutional Total
I 19.5 14.3 22.0 23.4 20.8 100
II 25.9 3.7 22.2 26.0 22.2 100
III 33.3 6.1 21.2 12.1 27.3 100
IV 18.2 29.1 18.2 14.5 20.0 100
V 20.6 5.9 23.5 35.3 14.7 100
Percentage Who Ranked Number Five
Kind of Informationb
I 39.5 21.1 6.6 11.8 21.0 100
II 36.0 28.0 4.0 28.0 4.0 100
III 46.2 19.2 7.7 15.4 11.5 100
IV 30.4 13.0 13.0 17.5 26.1 100
V 46.4 28.6 10.8 7.1 7.1 100
aChi-square significant at 10 percent level.
bCh1—square significant at 30 percent level.
Outside of Type I and II which are very similar there appears to be no
pattern. Lack of pattern could mean the other three types are not completely
defined by the scale technique. It could also be the first clear indication that
the four items used for the scale did not fit together as well as the scale indicated. _
Let us consider the four items that did scale and notice that (f) and (e) are
strategies used in a two-person game, i. e. , between the owner of the combine
and the owner of the tractor. However, the strategies (b) and (d) indicate at
least a three-person game, i. e. , between the owner of the combine and the
owner of the tractor and any one or more persons who might be interested in
the same trade. The fact that (f) and (e) fall on the most negative end and (b)
and (d) on the most positive end could further indicate the strategies are
appropriate for two different games, one with two persons and another with
more than two. The fact that our scale technique placed them in this order
adds strength to its usefulness in discerning sets of strategies. Had the
researchers who formulated the questions been more cognizant of the l

classification of games, the items could have been constructed in such a way
as to avoid asking about more than one game at a time. The chances of the _
respondents making errors would have been reduced., Actually, farmers are
probably participating in several games at one time; however, the stage to
which this kind of research has advanced dictates an over-simplified approach.,
Two of our objectives set forth at the beginning of this report were (1) to
review the technique used in obtaining answers to questions concerning use of
strategies, and (2) to review the procedure and results used in distinguishing
sets of strategies used by farmers in a machinery trade. We will set forth
certain conclusions from our efforts to accomplish these two objectives before
proceeding to the third objective, that of pointing out kinds of studies to further
the research in this area. These conclusions are given in the statements that
1, The indirect approach, where the respondent is asked to project
himself into a particular situation and th.en asked what he thinks
should be done, is more appropriate than the direct approach where
an individual is asked whether he uses a particular strategy. This
approach not only was e