Available online at www.sciencedirect.

com

Systems
Engineering
Procedia

SystemsSystems
Engineering
Procedia:
Editor Desheng
WU 00 (2011) 000000
Engineering
Procedia
1 (2011) Dash
414421

www.elsevier.com/locate/procedia

2011 International Conference on Risk and Engineering Management (REM)

A study on preference reversal on WTA, WTP and Choice

under the third generation prospect theory
Yongmei Liu a, Tiantian Li a *, Junhua Hu a, Chen Fan a
a

Department of Management Science and Infromation Management, Central South University,

Changsha, Hunan,410083, China

Abstract
The value of subjective attitude parameters in and abroad were introduced in this paper on the basis of the third generation
prospect theory, aiming to make a further study on the preference reversal between willingness to accept, willingness to pay and
choice. Results suggest that people's significant reversal activities are relatively constant, and there are standard preference
reversal in the group of WTA/WTP and the group of WTA/Choice, and non-standard preference reversal in the group of
WTP/Choice. Besides, Chinese are more risk-seeking in gain and more risk-aversion in loss than Americans while Americans are
more likely to present preference reversal compared to Chinese.

by by
Elsevier
B.V.Ltd.
Selection and/or peer-review under responsibility of the Organising Committee of The International Conference
2011
2010Published
Published
Elsevier
of Risk and Engineering Management.

Key words
prospect theory; preference reversal; loss-aversion; dynamic reference point

The study of preference reversal originated from the experimental research by Lichtenstein and Slovic in 1968. It
recorded an experiment on lotteries with similar expectation value, in which subjects are supposed to choose
between a lottery with a high probability of winning a low amount (P-bet) and a lottery with a low probability of
winning a high amount ($-bet) and price them. Results showed that subjects tended to choose P-bet over $-bet and
they stated a higher selling price for $-bet than P-bet. Thus, if choice and selling prices are viewed as two forms of
preference expression, individual preferences seem to depend on which one is used, implying that preferences may
be reversed when shifting from one form to another. It is called preference reversal phenomenon (PRP).
Since then, studies on this phenomenon are done by a lot of researchers, who found that this irrational behavior
exists when WTA (willingness to accept),WTP(willingness to pay) and choice are matched by pairs. Some of
empirical documents average WTA and WTP as pricing, and then study the preference reversal between reservation
price and choice. Ganzach(1996) found preference reversal between WTP and WTA. Casey(1991) discovered the
interaction between income value and preference reversal. Schmeltzer(2004) found that when information is
presented in a fixed pattern, the PR in choice and pricing disappears. Grether and Plott(1979) put doubt into the

* Corresponding author. Tel.: 15974161827;

E-mail address: doublesweetli@gmail.com.(Li Tiantian)

2211-3819 2011 Published by Elsevier B.V. Selection and/or peer-review under responsibility of the Organising Committee of The International
Conference of Risk and Engineering Management.
doi:10.1016/j.sepro.2011.08.062

Yongmei
Liu et al.FanChen/
/ SystemsSystems
Engineering
Procedia
1 (2011)00
414421
Liu Yongmei,
Li Tiantian,
Engineering
Procedia
(2011) 000000

415

existence of PR and proposed that it may be caused by incentive, experiments errors or income effect. However in
their modified experiments PR still exists systematically and robustly.Besides, Loomes etc(2010) controled factors
neglected in previous studies and found subjects PR did not decay in repeated market.
There are also lots of other explanations to this irrational behavior, such as Regret Theory, Contingent Weight
Theory(Tversky,1988), Expression Theory(Goldstein and Einhorn,1987), Anchoring and Insufficient Adjustment
Theory(Slovic and Lichtenstein,1983) etc. It is believed in evaluation theory(Hsee,1999) that in separate evaluation
mode, easily-evaluated attributes plays a comparative critical role in decision; In joint evaluation mode, difficultlyevaluated attributes can become more easily-evaluated through comparing with each other and import more
influence on decision. When this change is significant to subjects' preference, preference reversal will occur. In
choosing between the two lotteries, pricing is mainly affected by simple attribute result- income, and choice is
mainly impressed by difficult attribute-probabilitySlovic and Lichtenstein1968; Tversky, 1990; Irwin,1994).
Above all, there are already large amount of documents on preference reversal and its formation. Among the
documents, the explanation based on the first generation prospect theory (Kahneman and Tversky,1979) and
Cumulative Prospect Theory(Tversky and Kahneman, 1992) are the mainstreams. According to Prospect theory, we
hold the opinion that people are generally risk-aversion in gain, but when winning probability of $ bet decreases to
almost zero and the possible income increases to a high extent, people will amplify the difference between perceived
possible incomes" and assign $ bet greater weight, which results in a preference to $ bet. On the contrary, people
are generally risk-seeking in loss, but with the decrease of winning probability of $ bet and the increase of loss in
$ bet, people still tend to be risk-averse even if the loss of P bet is fixed. All in all, the differences exist mainly
because of the altering of choice preference with the different probability portfolio, whereas in seperate evaluation,
the behavior tendency toward WTA and WTP will remain the same.
In 2008, Schmidt introduced dynamic reference points based on the first and second generation prospect theory
and put forward the third prospect theory PT3. Compared with the former prospect theories, PT3 raises the function
in behavioral modeling and theoretic explanation to risk and uncertainty transactions. Schimidt analyzed the
preference reversal between WTA and choice by introducing status quo as dynamic reference point, and stated that it
is mostly determined by the coactions of risk attitude parameter , weight parameter and loss aversion parameter .
The influence of status quo bias and loss aversion on risk choices have been proven as early as Harlesss experiment
in 1989. It fully demonstrates the capacity of the third generation prospect theory in explaining preference reversal
in lotteries. But how these subjective parameters influence preference reversal specifically. What about their
relationships. Can they explain the practice and guide it. These questions remain unanswered and call for further
studies.
Prospect theory has been developed for more than 30 years and scholars have accumulated certain researches on
risk attitude parameter , weight parameter and loss aversion parameter . This study intends to investigate the
correlativity and preference reversal between WTA, WTP and choice on the basis of these three subjective attitude
parameters. The paper will be of certain theoretical and practical significance in analyzing the theory's reliability of
the explanations to risk transactions and decision-making behaviors.
1 Dynamic modeling based on PT3
PT3 is modeled by Schmidt as follows:
Considering a finite state space S = {s | i = 1, n} , each state si corresponds a probability pi , pi = 1 . X is the
result of State S on the premise of P . A is a set of all acts. Specific acts h, f A . They are functions from S to x ,
so h( s ) X f ( s ) X . is risk attitude toward objective income, and is loss aversion parameter in loss
situations. Let h be the dynamic reference point, then power function of PT3 is:
i

if f h
[ f ( s ) h( s )]
v ( f , h) =
[h( s ) f ( s )] if f < h

The definition of rank-dependent decision weight follows the RDSEU (reference-dependent subjective expected
utility theory). It should be determined considering probability and ordered position of act f relative to reference
point h. For example, we have j > i if and only if v[ f ( s ), h( s )] > v[ f ( s ), h( s )] ,.When h is a constant act, the third
j

416

Yongmei Liu
et al. / Systems
414421
Liu Yongmei, LiTiantian,
FanChen/
SystemsEngineering
EngineeringProcedia
Procedia1 (2011)
00 (2011)
000000

generation theory equals to CPT (cumulative prospect theory).

Compare f and h under state s , if f ( s i ) < h ( s i ) then there is a strong loss. Define m as the number of states with
strong loss. If f ( s ) h ( s ) , there is a weak gain. Define m = n m as the number of states with weak gain. Then the
weight function of PT3 is:

w + ( pi ) if i = n
+
+
w ( ( j i ) p j ) w ( ( j > i ) p j )
W ( si ; f , h ) =

w ( ( j i ) p j ) w ( ( j < i ) p j )
w ( p ) if i = 1
i

if m + 1 i < n

if 1 < i m

Where w( p) = p /( p + [1 p] )1/
Let F be the set of states si such that
utility function of PT3 is:

3
f ( si ) h( si ) and

let G be the set of states s j such that

f ( s j ) < h( s j )

, then the

4
The function refers to the utility of act f relative to dynamic reference point h,. When V(f, h) 0 ,
V(f, h) = 0 , V(f, h) < 0 , then f h , f ~ h or f h , they present taking h as the reference point, f is weakly preferred to h, f
is indifferent to h, or f is strongly preferred to f.

V ( f , h) = iF ( f [ si ] h[ si ]) W ( si ; f , h) jG (h[ s j ] f [ s j ]) W ( s j ; f , h)

2 Introduction of present parameters

Smith has been mentioned in his paper that since PT3 develops mainly based on the first and second generation
prospect theory, the present parameters can also be applied to PT3. Grounded on the study of Booij, we analyzed the
value of prospect theory parameters in present documents and elicited only two papers fully considered all three of ,
and ( see Tversky and Kahneman(1992); Harrison and Rutstr m(2009)). To fully study the difference in Chinese
and western, we also refer to Chinese studies and finally get four groups of prospect theory parameters.
As to power function and weight decision function, PT3 does not consider the difference under gain and loss. We
elicit these parameters by averaging gain and loss situation. So the groups in the end show as follows.
Table 1 selected prospect theory values
Sample source

Tversky and Kahneman(1992)

America

0.88

0.65

1.38/2.25

Harrison and Rutstrom(2009)

America

0.72

0.91

1.38/2.25

Zeng (2005)

China

1.12

0.52

1.38/2.25

Peng (2008)

China

1.15

0.57

1.38/2.25

3 Study on preference reversal in WTA, WTP and CE

3.1 Modeling deduction of boundaries in WTA, WTP and choice
Given h lottery: an increment of wealth x ( or an decrement of wealth x) with probability p, and nothing
otherwise. Act A refers to selling lottery in WTA. Act P refers to buying lottery in WTP. Act C refers to the certainty
equivalent of the lottery h. As is shown in table 3:
Table 2 derivation of WTA, WTP and CE
x>0

x(-x)

1-p

WTA

WTP

CE

WTA

WTP

CE

Yongmei Liu et al. / Systems Engineering Procedia 1 (2011) 414421

Liu Yongmei, Li Tiantian, FanChen/ Systems Engineering Procedia 00 (2011) 000000

Let

V ( A, h) = 0 , V (h, P) = 0 , V (h,0) = V (C,0)

417

and insert in formula(4). Then formulate WTA,WTP and CE.

Table 3 expressions of WTA, WTP and CE in loss or gain

x>0

x gain

-x loss

x
1 p /
(
) (1 / ) 1 / + 1
p

x
1 p / 1/
(
) +1
p

WTP

x
1 p / 1/
(
) +1
p

x
1 p /
(
) 1 / 1 / + 1
p

CE

xp /
[ p + (1 p) ]1 /

xp /
[ p + (1 p) ]1 /

WTA

According to table 3, WTA in gain is opposite to WTP in loss, while WTP in gain is opposite to WTA in loss.
Based on PT3, WTA and WTP differ with the reference point altering. So WTA in gain is the compensation of WTP
in loss. Since there is no disparity in gain and loss situation no matter when it is or , CE are opposite each other
in gain and loss.
Set P bet: an increment of wealth x=1/p (or an decrement of wealth x=1/p) with probability p, and nothing
otherwise; Set $ bet: an increment of wealth y=r/q or an decrement of wealth y=r/q) with probability q, and nothing
otherwise. It requires p>q, rp>q, which means P bet is high probability but low reward while $ bet is low probability
but high reward. So the standard and non-standard preference reversal can expressed as follows:
Table 4 expressions of preference reversal
preference reversal

WTA/WTP

WTA/

WTP/

WTA

WTP

WTA

WTP

standard PR

non-standard PR

Insert the rewards and probability of P bet and $ in prospect theory model( formula3) get the expressions of WTA,
WTP and choice as AP A$ PP P$ C P C$ . Then the method Schimidt used was introduced to achieve
the boundaries of WTA, WTP and choice that let AP = A$ , PP = P$ and C P = C$ . On the boundaries, P bet and
$ are both preferred equally.
Table 5 expressions of boundaries
boundaries
1/ p

gain
WTA
AP = A$
loss

gain
WTP
PP = P$

1 p /
(
) (1 / )1 / + 1
p

r/q
1 q /
(
) (1 / )1 / + 1
q

1/ p
r/q
=
1 p / 1/
1 q / 1/
(
) +1 (
) +1
q
p
(

1/ p
r/q
=
1 p / 1/
1 q / 1/
) +1 (
) +1
q
p

loss

1/ p
r/q
=
1 p /
1 q /
1/
(
) (1 / ) + 1 (
) (1 / )1 / + 1
q
p

gain

p / 1
rq / 1
=
[ p + (1 p ) ]1/ [q + (1 q ) ]1/

loss

p / 1
rq / 1
=
[ p + (1 p ) ]1/ [q + (1 q ) ]1/

choice
C P = C$

418

Yongmei Liu et al. / Systems Engineering Procedia 1 (2011) 414421

Liu Yongmei, LiTiantian, FanChen/ Systems Engineering Procedia 00 (2011) 000000

3.2 preference reversal analysis in WTA, WTP and choice

the impact of loss aversion parameter and lottery relative value r

Table 6 code
codes
r=1.2
r=1
r=0.6
=1.38,=0.88,=0.65
r=1.2
r=1
r=0.6

0.9
0.8

WTA
A1
A2
A3

=2.25,=0.88,=0.65
r=1.2
r=1
r=0.6

0.9
0.8

C1 P1(A1)

0.6

q=0.6*p

0.4

C2

0.6

P2(A2)

q=p

0.5

P1(A1) C1

0.7

C2

A2(P2)

A2(P2)

0.7

q=p

0.5

q=0.6*p

P2(A2)

0.4

0.3

0.3

A3(P3)

0.2

0.1

0.2

0.3

0.4

0.5
p

0.6

0.7

C3

0.1

P3(A3)
0

A3(P3)

0.2

C3

0.1
0

choice
C1
C2
C3

WTP
P1
P2
P3

P3(A3)
0.8

0.9

fig.1. when=1.38,preference with different r

0.1

0.2

0.3

0.4

0.5
p

0.6

0.7

0.8

0.9

fig.2 .when =2.25, preference with different r

In the figure, it is loss situation in parentheses, else is gain. The choice boundary C is equivalent in gain and loss.
Besides, because P>q, rp>q are the basic conditions for the lottery, r = 1 only if the right area of q = p was
established; else if r = 0.6, it will be only the right area of q = 0.6 p.
In gain above each boundary, P bet is preferred in WTA, WTP and choice while $ bet is preferred below it. When
r=1.2, P bet is preferred on the right of boundary, while $ bet is preferred to the left. In loss, it is opposite.
Property 1 When r is a fixed value, under the different probability combinations of lottery P and $, people will
behave different preference in WTA, WTP and choice. The mostly decision behavior will act as follows:
Table 7 preference under different probability combination
WTA/WTP

WTA/choice

WTP/choice

PR

PR

PR

NO

NO

NO

standard PR

standard PR

NO

standard PR

NO

non-standard PR

NO

NO

NO

area/preference

WTA

WTP

choice

Through the table above, there are only standard PRs in WTA/WTP or WTA/choice, while only non-standard
preference reversal between WTP and choice. Lichtenstein and SlovicKnez and Smith, Casey, etc found that the
standard PRs are significantly less than non-standard PRs in WTA and choice, and sometimes even does not exist. It
seems that WTP in pricing tends to reduce standard PRs and increase non-standard ones. Nevertheless, in many
empirical researches although it is not remarkable, but there still exist non-standard PRs in WTA/choice and
WTA/WTP as well as standard PRs in WTP/choice. We can infer that the small part of findings which are different
from this paper is because of their different subjective parameters with the present Chinese and American situations.
Property 2 As increases, the possibility increases, that $ bet has a higher selling price and P bet has a higher
buying price. There is no change in choice.
As is shown in chart 1, chart2, as increases, in gain WTA boundary move upwards, WTP boundary move
downwards. In loss, it is opposite.

419

Yongmei
Liu et al.FanChen/
/ SystemsSystems
Engineering
Procedia
1 (2011)00
414421
Liu Yongmei,
Li Tiantian,
Engineering
Procedia
(2011) 000000

According to Prospect theory, in respect that $ bet yields higher than P bet, sellers take selling bet as loss, and
$ bet will loss more. As a result, the larger sellers have, the higher selling price $ bet will be endowed. Meanwhile,
as buyers, money is considered loss to them, and they must cover the loss by the expected earnings from the lottery.
So, the larger the buyers have, the more possibly they will choose P bet.
Property 3 When other parameters are fixed, as r increases, people prefer $ bet in gain, as well as P bet in loss.
As is shown above, with the increase of r, all boundaries go to upper right, expanding the regions in which the
$ bet is favored in gain and disliked in loss. The intuition for this is that, as the value of r increases, the attraction of
$ bet is amplified with its larger rewards, which is reverse in loss situation.
Property 4 Preference reversal is more possible in large probability portfolios than small ones.
According to figure 1 and 2, area 1 and 2 are expanded with the increase of probability p and q. Liu(2008)
demonstrates that there is significant discrepancy in preference scores from large probability portfolios to small ones.
When items are presented with large probabilities, preference reversal is higher.

Sino-us situation comparison

0.5

0.5

=2.25r= 0.6
=0.72,=0.91
=0.88,=0.65

0.45
0.4

0.4

A(P)

0.25

A(P)

q=0.6.p

0.3

0.3
0.25

P(A)

0.2

0.2

0.15

0.15
0.1

0.1
0.05

0.05

P(A)
0

0.1

0.2

0.3

0.4

0.7

0.6

0.5
p

0.8

0.9

0.2

0.1

fig.3.

0.4

0.5
p

0.6

0.7

0.8

0.9

=2.25r= 1.2
=0.72,=0.91
=0.88,=0.65

0.8

=2.25r= 1.2
=1.12,=0.52
=1.15,=0.57

0.9
0.8

0.7

C
P(A)

0.7

q=p

P(A)

0.4

0.4

P(A)

0.3

0.5

0.5

q=p

0.6

0.6

0.3

0.2

0.2

0.1
0

0.3

fig.4.

1
0.9

A(P)

0.35

q=0.6.p

0.35

=2.25r= 0.6
=1.12,=0.52
=1.15,=0.57

0.45

0.1

A(P)
0

0.1

0.2

0.3

0.4

0.5
p
p

0.6

0.7

0.8

0.9

0.1

0.2

0.3

0.4

0.5
p

0.6

0.7

0.8

0.9

fig.5.
fig.6.
In gain, in the left area of each boundary, $ bet is preferred in WTA, WTP and choice while P bet is preferred in
the right area. If and only if =0.72=0.91, P bet is preferred in the area below the boundary, while $ bet is
preferred in the area above the boundary. In loss, it is opposite.
Property 5 Considering the preference of Sino-us, Chinese are more risk-seeking than Americans in gain, and more
risk-averse in loss.
The charts illustrate that no matter whether r>1 or r<1, Chinese boundaries are higher than Americans. It means
that compared with Americans, in WTA, WTP and choice in gain, Chinese are more likely to choose $ bet than
Americans and show a larger risk-seeking inclination. However, in loss Chinese prefer P bet more likely than
Americans, which shows more risk-aversion in this act. The studies of Bontempo (1997), Weber and Hsee(1998),
Yates (1998), Fan and Xiaoet al(2006) indicate that Asians has stronger risk-seeking tendency and higher level of

420

Yongmei Liu
et al. / Systems
414421
Liu Yongmei, LiTiantian,
FanChen/
SystemsEngineering
EngineeringProcedia
Procedia1 (2011)
00 (2011)
000000

overconfidence. Weber analyzes the impact of risk propensity coefficient and probability perception to Sino-us. The
results note that the former factor is not significant; instead, it is mainly because Chinese undertakes the probability
risk.
Property 6 Americans presents higher-leveled preference reversal than Chinese.
As is shown in figure 3-6, Americans own larger preference reversal areas in regional 1 and 2. So, preference
reversal is more probable in Americans than Chinese. Joyce (2003) found that different from that in risk-neutral state,
preference reversal reflected in strong risk-seeking or risk-averse subjects is much weaker. Corresponding with
Property 5 preference reversal is more possible to Americans due to Americans lower risk attitude in gain and loss.
Property 7 If >, the area of preference reversal where r<1 will be larger than that where r>1. Else if <, it is the
opposite.
According to figure 3 and 5, if >, compared with those where r=1.2, preference reversal area 2 and 3 are larger
where r=0.6. Else if <, there are larger preference reversal areas where r=1.2.
Where >, leads to the increase in the irrationality to lottery's probability so that the $ bet relatively become
more attractive. If r<1, it will weaken this irrationality and enlarge P bet's effect. Therefore, it will have larger
preference reversal areas. Where <, people will reveal stronger bias in risk attitude and perform more serious riskaversion in gain as well as risk-seeking in loss. It also means they will prefer P bet in gain and $ bet in loss. At this
moment, if r>1, it can weaken the effect of this phenomenon and expand preference reversal areas.
4 Conclusions
In this paper we investigated the preference reversal in WTA, WTP and choice among Americans and Chinese
based on the values of three existed prospect theory subjective attitude parameters. We hold the opion that people
have different preference to different lotteries in WTA, WTP and CE due to distinct probability portfolios as well as
the variety of their risk attitude parameters, weight decision parameters and loss aversion parameters.. Several
conclusions are as following:
Firstly, the results suggest that people's significant reversal activities are relatively constant, and there exist
standard preference reversal in the group of WTA/WTP and the group of WTA/Choice, and non-standard preference
reversal in the group of WTP/Choice. Secondl y, Chinese are more risk-seeking in gain and more risk-aversion in
loss than Americans while Americans are more likely to present preference reversal compared to Chinese.
Although in practice, preference reversal behavior in WTA, WTP and choice are not as significant as that in
laboratories, but as early as in 1993, Bohm and Lind, etc found it still exists doubtlessly under a certain proportion.
Therefore, it can be used to risk transactions and decision-making engineering problems in reality, based on the
analysis of its reliability under the third generation prospect theory and scientific testing. For example, in marketing
field, through predicting people's choosing or buying tendency, it can lend insight to making pricing-decision to risk
commodities such as stock, lottery and insurance. From the perspective of policy management such as tax planning,
it can help to refer a extent to tax evasion punishment which can drive people pay taxes regularly, etc.
Restricted to the functions of third generation prospect theory, we did not analyze the situations with both gain
and loss. Thereby, we cannot judge and predict these kinds of cases in actual problems. Besides, this article analyzes
WTA, WTP and choice in pairs and reckon without the absolute value of the commodity, i.e. we eliminate the
influence of framing effect. However in many documents, significant influence of framing effect on the preference
reversal in WTA, WTP and choice has been validly proved and it can be referred for our further study. Above all,
part of inferences in this article has already been empirically tested, but there still are problems calling for further
empirical study. And it is needed to be tested in evidence whether the conclusions in risky commodities can be
applied to non-market goods as environment and health.
Acknowledgements
This work was supported by the National Natural Science Foundation of China under Grant No.71071164 and No.
70971139; the Innovative Research Group National Science Foundation of China under Grant No.70921001.

Yongmei Liu et al. / Systems Engineering Procedia 1 (2011) 414421

Liu Yongmei, Li Tiantian, FanChen/ Systems Engineering Procedia 00 (2011) 000000

421

References
1. P.Bohm and H. Lind Preference reversal, real-world lotteries, and lottery-interested subjects. Journal of Economic Behavior &
Organization, 22(3)(1993) 327-348.
2. R. Bontempo,W. Bottom and E. Weber Cross-Cultural Differences in Risk Perception: A Model-Based Approach. Risk Analysis,
17(4)(1997) 479-488.
3. A. Booij,B. Van Praag and G. Van de Kuilen A parametric analysis of prospect theorys functionals for the general population. Theory and
Decision, 68(1)(2010) 115-148.
4. J. T. Casey Reversal of the preference reversal phenomenon. Organizational Behavior and Human Decision Processes, (1991).
5. J. Fan and J. Xiao Cross-cultural differences in risk tolerance: a comparison between Chinese and Americans. Journal of Personal Finance,
5(3)(2006) 54-75.
6. Y. Ganzach Preference reversals in equal-probability gambles: A case for anchoring and adjustment. Journal of Behavioral Decision
Making, 9(2)(1996) 95-109.
7. W. Goldstein and H. Einhorn Expression theory and the preference reversal phenomena. Psychological Review, 94(2)(1987) 236.
8. D. Grether and C. Plott Economic theory of choice and the preference reversal phenomenon. The American Economic Review, 69(4)(1979)
623-638.
9. D. Harless More laboratory evidence on the disparity between willingness to pay and compensation demanded. Journal of Economic
Behavior & Organization, 11(3)(1989) 359-379.
10. G. Harrison and E. Rutstr m Expected utility theory and prospect theory: One wedding and a decent funeral. Experimental Economics,
12(2)(2009) 133-158.
11. C. HseeG. LoewensteinM. Bazerman, et al. Preference reversals between joint and separate evaluations of options: A review and
theoretical analysis. Psychological Bulletin, 125(5)(1999).
12. J. Irwin Buying/selling price preference reversals: Preference for environmental changes in buying versus selling modes. Organizational
Behavior and Human Decision Processes, 60(1994) 431-457.
13. E. B. Joyce,W. D. John and A. R. Thomas Preference reversals and induced risk preferences:evidence for noisy maximization. Journal of
Risk and Uncertainty, 27(2)(2003) 139.
14. D. Kahneman and A. Tversky Prospect Theory: An Analysis of Decision under Risk. Econometrica, 47(2)(1979) 263-292.
15. M. Knez and V. Smith Hypothetical valuations and preference reversals in the context of asset trading. Laboratory Experimentation in
Economics: Six Points of View, (1987) 131-154.
16. S. Lichtenstein and P. Slovic Reversals of preference between bids and choices in gambling decisions. Journal of experimental psychology,
89(1)(1971) 46-55.
17. W. Liu, the influence of personality and probability to preference reversal, Northease normal university.(2008)
18. G. Loomes,C. Starmer and R. Sugden Preference reversals and disparities between willingness to pay and willingness to accept in repeated
markets. Journal of Economic Psychology, 31(3)(2010) 374-387.
19. M. Peng, newsvendor problem study with preference and mental accounts of decision maker, Central South University (2008)
20. C. Schmeltzer,J. Caverni and M. Warglien How does preference reversal appear and disappear? Effects of the evaluation mode. Journal of
Behavioral Decision Making, 17(5)(2004) 395-408.
21. U. Schmidt,C. Starmer and R. Sugden Third-generation prospect theory. Journal of Risk and Uncertainty, 36(3)(2008) 203-223.
22. P. Slovic and S. Lichtenstein Relative importance of probabilities and payoffs in risk taking. Journal of experimental psychology,
78(3)(1968) 1-18.
23. P. Slovic and S. Lichtenstein Preference reversals: A broader perspective. The American Economic Review, 73(4)(1983) 596-605.
24. A. Tversky and D. Kahneman Advances in prospect theory: Cumulative representation of uncertainty. Journal of Risk and Uncertainty,
5(4)(1992) 297-323.
25. A. Tversky,S. Sattath and P. Slovic Contingent weighting in judgment and choice. Psychological Review, 95(3)(1988) 371.
26. A. Tversky,P. Slovic and D. Kahneman The causes of preference reversal. The American Economic Review, 80(1)(1990) 204-217.
27. E. Weber and C. Hsee What folklore tells us about risk and risk taking: Cross-cultural comparisons of American, German, and Chinese
proverbs. Organizational Behavior and Human Decision Processes, 75(2)(1998) 170-186.
28. J. YatesJ. LeeH. Shinotsuka, et al. Cross-cultural variations in probability judgment accuracy: Beyond general knowledge overconfidence?
Organizational Behavior and Human Decision Processes, 74(2)(1998) 89-117.
29. J. Zeng Experiments test cumulative prospect theory. Journal of jinan university(natural science edition), 123(01)(2007) 44-47.