Title: Anchoring and Adjustment

Introduction:

Heuristics-
Invention of Heuristics as a concept:

During the 1970s, psychologists Amos Tversky and Daniel Kahneman

presented their research on cognitive biases. They proposed that these biases
influence how people think and the judgments people make. As a result of these
limitations, we are forced to rely on mental shortcuts to help us make sense of
the world. Simon's research demonstrated that humans were limited in their
ability to make rational decisions, but it was Tversky and Kahneman's work that
introduced the study of heuristics and the specific ways of thinking that people
rely on to simplify the decision-making process.

● Cognitive Heuristic-

A heuristic is a mental shortcut that allows an individual to make a decision,

pass judgment, or solve a problem quickly and with minimal mental effort.
While heuristics can reduce the burden of decision-making and free up limited
cognitive resources, they can also be costly when they lead individuals to miss
critical information or act on unjust biases. It is any approach to problem-
solving that uses a practical method or various shortcuts in order to produce
solutions that may not be optimal but are sufficient given a limited timeframe or
deadline. Heuristics methods are intended to be flexible and are used for quick
decisions, especially when finding an optimal solution is either impossible or
impractical and when working with complex data.

● Anchoring and Adjustment Heuristic:

It is a cognitive heuristic where a person starts off with an initial idea and
adjusts their beliefs based on this starting point. Anchoring and adjustment have
been shown to produce erroneous results when the initial anchor deviates from
the true value. Awareness of anchoring, monetary incentives, giving careful
consideration to a range of possible ideas, expertise, experience, personality,
and mood can all modify the effects of anchoring. It can be used to advantage in
sales and price negotiations where setting an initial anchor can influence
subsequent negotiations in your favour.

The anchoring and adjustment heuristic is of great interest to psychologists

because it helps to explain a wide variety of different psychological phenomena.
For example, people’s estimates of what other people are thinking are often
egocentrically biased (i.e., people assume that others think more similarly to
how they themselves think than is actually the case) because they tend to start
with their own thoughts and then adjust (insufficiently) for another person’s
perspective. People suffer from a hindsight bias, thinking that past outcomes
were more predictable at the time than they really were, because they anchor on
current knowledge and then adjust (insufficiently) for the fact that certain things
that are known now were not known back then. Also, people tend to assume that
they will do better than others on easy tasks because they start with an
assumption that they will do well themselves and then adjust (insufficiently) for
the fact that other people are also likely to do well on such easy tasks.

● Other Heuristic Methods:

In the paper “Judgment Under Uncertainty: Heuristics and Biases” (1974),

Daniel Kahneman and Amos Tversky identified three different kinds of
heuristics: availability, representativeness, and anchoring and adjustment.

a. Availability Heuristic:
The availability heuristic, also known as availability
bias, is a mental shortcut that relies on immediate
examples that come to a given person's mind when
evaluating a specific topic, concept, method or
decision.
b. Representativeness Heuristic:
The representativeness heuristic is used when making
judgments about the probability of an event under
uncertainty.The representativeness heuristic is
associated with prototype theory.6 This prominent
theory in cognitive science provides an explanation
for object and identity recognition.
 ● Past Research:

Anchoring effects are remarkably robust and easy to demonstrate

experimentally. In the most common experimental paradigm, participants are
first asked to make a comparative judgment about whether a target value is
more or less than a given anchor value, and then asked to make an absolute
estimate about the actual value of the target (Tversky & Kahneman, 1974). The
typical result in this two-stage paradigm is that the absolute estimate is biased in
the direction of the initial anchor, even if care is taken to explain that the anchor
value was generated randomly and is therefore not a useful clue to the absolute
estimate. In one study, for instance, people asked to estimate the height of
Mount Everest provided a median estimate of 8000 feet after first considering
whether Everest was taller or shorter than 2000 feet, but provided a median
estimate of 42,500 after considering whether it was taller or shorter than 45,500
feet .These results and others like them were originally interpreted as the
product of insufficient adjustment from the initial anchor—resulting from the
use of what Tversky & Kahneman (1974) called the anchoring and adjustment
heuristic. According to this account, people start by rejecting the anchor value
as a plausible estimate and then adjust in a deliberate fashion until they reach a
satisfactory answer (‘‘I don’t need 18 Snickers bars, but 14 would be nice, and
12 would be just right’’). Because adjustment requires mental effort that people
may be either unwilling or unable to expend, adjustments tend to be insufficient
and final estimates are biased in the direction of the initial anchor value.

Problem: To study the effect of Anchor and Adjustment heuristic , a

mental shortcut on participant`s response style.

Hypothesis:
● Null: There will be no difference in the response pattern found for high or
low anchor conditions.
● Alternative: There will be a significant difference in the response pattern
found for high and low anchor condition participants
a. Participants in high anchor condition will respond with
larger percentages for different questions.(51 to 100)
b. Participants in low anchor condition will respond with
smaller percentages for different questions.(1 to 50)

Design: Random group design or Between subject design

Operational definition of variables:

● IV: Anchoring condition (high and low) 

The high anchor condition participants were randomly exposed to 10
digits from 51 to 100 for all 10 questions.
The low anchor condition participants were randomly exposed to 10
digits from 1 to 50 for all 10 questions.
● DV: Participants estimated answer after having been exposed to the high
or low anchor condition
● CV: Socioeconomic status, age, education, language 

Method : 

Participants: 
Individual Data 

Name (Optional)  Age  Class  Emotional State

Participant 1 Vedant Mankar 21 TYBA Stable
Participant 2 Ritik Dhanore 21 TYBA Stable

Group Data 
College Students: N= 23 participants 

Material:

Wooden screen, paper, pencil, chits 1 to 100, question sheet, Individual data
table

Question sheet:

1. On an average, what percentage of typing is done with the left hand? 

2. What percentage of the earth’s land is covered by glaciers? 
3. What percentage of plant life is found in the ocean? 
4. What percentage of bones in the body is located in the feet? 
 5. What percentage of the world’s spices come from India? 
6. What percentage of households in India are single parent families? 
7. What percentage of police personnel in India are women? 
8. What percentage of the world population consumes rice as a staple
food? 
9. What percentage of their lives do cats spend sleeping? 
10. What percentage of the world population lives in Asia?

Procedure:

The participants were given brief instructions about conduction of the

experiment. They were asked to download a random number generator app with
the link provided to them. They were presented with the questions given in the
manual. Based on the questions they were asked to generate a number on the
random number generator. Then they were asked to tell if their actual estimate
number was higher, lower or equal to the random number generated. Actual
estimate number was asked to them. Based on the data individual graph and
table was drawn. In the end conclusion was noted.

Instructions:

This is a simple experiment. I will ask you 10 questions one at a time. Your task
is to write the answer to each question as accurately as possible. Before you
answer each question, you have to pick out 1 chit, on which a number is written.
You have to write whether your answer to the question, which will be a number,
will be higher, lower or the same as the number on the chit. Then you have to
write your actual estimate. 

Post Task Questions:

Participant 1:

1. What do you think this experiment is about?

A. I think this experiment is maybe about testing something

2. How did they feel being the part of the experiment?

A. Found the experiment quite interesting
Participant 2:

1. What do you think this experiment is about?

A. I think the experiment is about mental health

2. How did they feel being the part of the experiment?

A. They enjoyed being the part of it.

Experimenter’s Observation: Non-verbal cues

Participant 1: It has been observed that the participant was quite confident and
seemed stable during the course of the experiment. A firm body posture and
good speech was also observed.

Participant 2: The participant seemed relaxed and was enjoying the experiment.
He gave an active participation during the experiment.

Debriefing:

This experiment was conducted to test the anchor and adjustment heuristic
which is basically a cognitive bias. It says that when people are given an anchor
for example- a number or fact, it tends to impact the outcome, even if the
anchor is completely absurd or unrelated to the actual estimate. Holding an
anchor is a subconscious phenomenon. Whereas adjustment is a conscious
action.
This experiment will check whether you employed the anchor and adjust
heuristic. Thank you for your active participation in this experiment.

Results and Discussion:

 Individual Data Discussion

From the data it can be seen that both of the participants answered
opposite to the numbers generated. With this it is evident that they used
anchor and adjustments heuristics. A distinct gap was seen in the answers
given in question 5 with the answers as 30 and 61, where the anchors
were 43 and 81 respectively. A wide gap was seen in the answer of
participant 1 for question number 9 with 4 as the answers, and anchor
being 45. With comparing both the participants data we can see that both
the participants stayed around the anchor but the second participant
stayed much closer as compared to the first participant.
Individual Data Table:

A. Low Anchor Condition

Q.No. Anchor Estimate Actual

High/ Low Estimate No.
1 21 Equal 21
2 28 Higher 35

3 3 Higher 15
4 9 Higher 21
5 43 Lower 30

6 12 Equal 12
7 15 Higher 30
8 39 Higher 50
9 45 Lower 4

10 46 Higher 49

Total: 267 Mean: 26.7

 Low Anchor Condition
60

50

40

30

20

10

0
0 Q.1 Q.2 Q.3 Q.4 Q.5 Q.6 Q.7 Q.8 Q.9 Q.10 0

Column1

High Anchor Condition:

Q.No. Anchor Estimate Actual

High/ Low Estimate No.
1 63 Higher 70
2 95 Lower 52

3 59 Lower 55
4 64 Lower 56
5 85 Lower 61

6 61 Higher 80
7 66 Equal 66
8 93 Lower 81
9 98 Lower 90

10 62 Lower 54

Total: 657 Mean: 65.7

 High Anchor Condition
100
90
80
70
60
50
40
30
20
10
0
0 Q.1 Q.2 Q.3 Q.4 Q.5 Q.6 Q.7 Q.8 Q.9 Q.10 0

Column1 Series 3

Group Data:

Mean estimated percentages

P’s  High Low anchor

Anchor 

1 65.7 26.7

2 45.5 26.3

3 36.6 36.3

4 42 21.5

5 66 55

6 55.5 33.4

7 68.9 24.4

8 51.0 44.5

9 69.3 49.5

10 29.3 31.5
11 27.2 31.6

12 42 44

13 61.7 45.6

14 55.5 39.1

15 53.6 31.3

16 39.7 32.2

17 59.5 51.7

18 68.7 23.1

19 70.5 29.2

20 44.82 31.7 

21 67.2 29.8

22 46.5 35.4

23 25.8 36.4

High Anchor Condition-

Total: 1192.52
Mean: 52

Low Anchor Condition-

Total: 810
Mean: 35.22
 Group data graph
80

70

60

50

40

30

20

10

0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 -

High anchor  Low Anchor

Calculations:

Sum of squares and standard deviation:

Calculation of sum of squares(high) Standard deviation (high) Calculation of sum of squares (low) Standard deviation (low)
Square root of (X-M)2/N) Square root of (X-M)2/N)
High Low
anchor  X-Mean (X-M)^2 14.03171 anchor X-Mean (X-M)^2 9.073708
65.7 13.7 187.69 26.7 -8.3 68.89
45.5 -6.5 42.25 26.3 -8.7 75.69
36.6 -15.4 237.16 36.3 1.3 1.69
42 -10 100 21.5 -13.5 182.25
66 14 196 55 20 400
55.5 3.5 12.25 33.4 -1.6 2.56
68.9 16.9 285.61 24.4 -10.6 112.36
51 -1 1 44.5 9.5 90.25
69.3 17.3 299.29 49.5 14.5 210.25
29.3 -22.7 515.29 31.5 -3.5 12.25
27.2 -24.8 615.04 31.6 -3.4 11.56
42 -10 100 44 9 81
61.7 9.7 94.09 45.6 10.6 112.36
55.5 3.5 12.25 39.1 4.1 16.81
53.6 1.6 2.56 31.3 -3.7 13.69
39.7 -12.3 151.29 32.2 -2.8 7.84
59.5 7.5 56.25 51.7 16.7 278.89
68.7 16.7 278.89 23.1 -11.9 141.61
70.5 18.5 342.25 29.2 -5.8 33.64
44.82 -7.18 51.5524 31.7 -3.3 10.89
67.2 15.2 231.04 29.8 -5.2 27.04
46.5 -5.5 30.25 35.4 0.4 0.16
25.8 -26.2 686.44 36.4 1.4 1.96
4528.442 Total 1893.64
Calculation of degree of freedom

Degree of freedom df
df (high) 23-1=22
df (low) 23-1=22
df = 22+22=44

Calculation of pooled variance

Pooled variance
SP2 = ss1 + ss2/df1 + df2
4528.44+1893.64 / 22 + 22
6422.08/ 44
145.95

Calculation of standard error of mean

Standard error of the mean

S(M1-M2)
√145.95/23+ 145.95/23
√6.345+6.345
√12.691
square root =3.56

Calculation of t test and t critical

t= M1-M2/S(M1-M2)
t = 52-35/ 3.56
t = 17/3.56
t = 4.77
t calculated = 4.77
t critical= 1.697
calculated t is greater than the critical t
the t value is significant at 0.05 level

The calculated t is greater than the critical t, the t value is significant at 0.05
level. Participants have given greater responses for high anchor conditioning
(M=52, SD = 14.03) than low anchor conditioning (M = 35, SD = 9.07)
T = 4.77, p < 0.05
Group Data Discussion:

The mean percentage of 23 students has been shown through the group data
table and graph. The mean score for high anchor condition is 52 and the mean
for the lower condition is 35. In this test the independent variable being the high
anchor and low anchor in this case. The t value has been observed as 4.77. By
the table of critical values of t, the tail test of df44, has to have the t value as
1.697. Thus, the t value was found to be significant at 0.05 level. Based on this
information we can reject the null hypothesis. Participants in high anchor
condition will respond with larger percentages for different questions. (51 to
100); Participants in low anchor condition will respond with smaller
percentages for different questions. (1 to 50)’ is therefore considered as correct
and validation occurs.

Conclusion:

With the information available from the individual data and group data we can
conclude that there is a impact on the responses of participants in both high and
low anchor condition. People do use heuristics to seek out answers to questions.

References:

A. Putting Adjustment Back in the Anchoring and

Adjustment Heuristic: Differential Processing of Self-
Generated and Experimenter-Provided Anchors
https://www.researchgate.net/publication/11792422_P
utting_Adjustment_Back_in_the_Anchoring_and_Adj
ustment_Heuristic_Differential_Processing_of_Self-
Generated_and_Experimenter-Provided_Anchors

B. Why do we take mental shortcuts?

https://thedecisionlab.com/biases/heuristics/