KONONENKO, Svetlana

004180-0114

Psychology Internal Assessment

An experiment into the effect of low anchor in numerical estimations of low
temperature
Psychology Higher Level
Name: Svetlana Kononenko
Candidate Number: 004180-0114
School: United World Colleges of South East Asia
Date of submission: October 2016

Word count: 1947

INDEX
Abstract .3

KONONENKO, Svetlana

004180-0114

Introduction4
Method...6
Design....6
Participants....6
Materials....6
Procedure...7
Results....8
Descriptive.....8
Inferential...9
Discussion....10
References....12
Appendices...13

Abstract
The aim of the study was to to investigate the effect of low anchor on judgement. Experiment
was based on Strack and Mussweilers study (1999a). It was predicted that participants
estimations in a anchored condition would appear lower in comparison to non-anchored
group. The IV was the low value of temperature (low anchor) and the DV was participants
estimations of the temperature. The only difference between two groups was the
question with anchor (Is the mean temperature in Antarctica higher or
lower than 56 C? following by What is the mean temperature in
2

KONONENKO, Svetlana

004180-0114

Antarctica?). The independent measure design was chosen with an

opportunity sample of 28 students of an international school. Participants
completed questionnaires following standardized instructions. Performed ttest shows that t value of t(26)=3.51 exceeded the critical value and turned to be highly
significant at 0.01 level of significance that supports the initial hypothesis that a low anchor
leads to lower estimations.
Word count: 157

Introduction
Cognitive psychology explores different mental processes. One of them is our perception.
Researchers say that our perception is a cognitive process concerning how sensory
information is understood (J. Willerton, J.M. Lawton, 2013). Every day people receive a
tremendous amount of information from the external environment through the sensory
system. However, people usually process and interpret the data according to their previous
experience, cultural influence and emotions. These all are said to be sources of peoples
biases when making-decisions. In fact, humans tend to rely even on an non significant piece
of information when trying to make an independent decision. These bits of information are
known as anchors.Thus, the anchoring effect - a biased estimate toward an arbitrary value
considered by judges before making a numerical estimate (Jacowitz & Kahneman, 1995).

KONONENKO, Svetlana

004180-0114

There is empirical evidence for the anchoring effect demonstrated by Wilson et al.
Researchers gave a quiz show to 116 undergraduate Virginia university students studying
different disciplines and asked to choose a random number from a bag containing "a large
range of numbers" and judge whether this number was less than, equal to, or greater than the
answer to a general knowledge question. Independent variable whether participants picked
the number or skipped the round. In fact, there was only one number in the bag (1930) that
was an anchor value. In this case the way to control conditions was the option skip a round
in the game which allowed researchers to compare the effect of the anchor on estimations.
Dependent variable was the participants estimates on the number of countries. Researchers
asked to compare the anchor number to the question they would answer next or to an
irrelevant question (Wilson et al., 1996). In the experimental conditions participants were
asked to deduce the number of United Nations countries according the number they got (less,
equal, more). The analysis of variance demonstrated a significant main effect of the
anchoring manipulation. People who received the anchor value guessed that there were more
countries in the United Nations than did people in the control condition according to Wilson
et al (1996).
Other studies attempted to demonstrate the anchoring effect in general in addition to the
effect of certain numerical values. Simmons et al. found the higher anchors have caused
higher estimates because of the main anchoring effect. Researchers manipulated anchor
values so that different questionnaires consisted either higher or lower anchors on a specific
question. It was stated that higher numbers indicate a higher anchoring effect (Simmons et al.,
2010).
Strack and Mussweiler designed a study that supported the notion that anchoring is a special
case of semantic priming. Researchers aimed to examine generated information as a crucial
determinant of the anchoring effect. Researchers recruited 20 participants and used 4 pairs of
difficult general-knowledge questions, constructed consisting of questions on comparative
judgments (higher or lower the given value) and another one on absolute judgments. The
anchors are typically set at one standard above and below the mean estimates of a calibration
group that answered absolute questions (Mussweiler & Strack, 1999a).
The findings showed that the absolute estimations were reliably assimilated towards the
provided anchor values . The effect was extremely robust even if participants tried to answer
against the given anchor, their estimates are typically assimilated towards the anchor values.
Moreover, the study provided evidence that a high anchor value led to higher estimates than
did a low anchor value.
The experiment is a modification of the Strack and Mussweiler, 1999a4 study. The aim of the
experiment is to investigate the effect of low anchor on judgement. The hypothesis is onetailed as the participants who are given a low anchor will potentially shift answer values to
lower estimations.

KONONENKO, Svetlana

004180-0114

Research Hypothesis (H1): Participants estimation of temperature will be lower when

given an low anchor value in comparison to participants with no anchor.

Null Hypothesis (H0): A low anchor value will have no effect on participants estimation
of temperature.

Method
Design:
An independent measure design was used. Participants were separated into two groups:
control (non-anchored) and experimental (anchored) groups. This type of design prevents
order effects and potential demand characteristics if participants try to work out the aim. The
experiment preserved similar environment in both groups (classroom settings, school
uniform) to control external variables. Participants signed an informed consent form to
participate in the experiment. They were explained the right to withdraw as well as
confidentiality of data and told that the study did not include any mental or physical harm.
There was no deception and groups were debriefed after the experiment procedure.
Independent variable: the low anchor value of -56 degrees Celsius in comparison to a
question without a given value
Dependent variable: participants estimations of the mean temperature in critical question

Participants:
28 participants were evenly split into two conditions randomly (by giving number 1 (control)
or 2 (experimental) to each person in the list). All participants are the same age group (16-17

KONONENKO, Svetlana

004180-0114

years old); the target population is students from various multicultural backgrounds studying
different sets of International Baccalaureate Diploma in an international school. Moreover,
opportunity sample was used. Participants available at that time were asked to participate. It
reduced the time spent on the experiment.

Materials:
-

Informed consent form (Appendix 1)

Briefing letter (Appendix 2)
Anchored questionnaire (Appendix 3)
Non-anchored questionnaire (Appendix 4)
Debriefing letter (Appendix 5)

Procedure: 1 / 2
Participants were informed that the questionnaire did not intend to test their knowledge on a
certain topic or IQ levels. Participants signed up an informed consent form. Both control and
experimental conditions groups were tested at the same settings.
After gathering the informed consent form the experiment in control group was started and
control condition questionnaires were distributed. Participants were given time of 2 minutes
to read instructions carefully and ask questions if any. 10 minutes were allocated to complete
the task which was a questionnaire consisted of 4 questions: 3 of them were filling questions
and 1 was critical. If the task was completed earlier students raised their hands and researcher
collected the paper. The rest of questionnaires were collected after 10 minutes. Researcher
debriefed participants with standardized debriefing letter. Students were thanked for
cooperation and told not to talk about the experiment to avoid revealing of the aim.

KONONENKO, Svetlana

004180-0114

Results
Descriptive
Experiment provided interval raw data (Appendix 6). The mean and Standard Deviation of
both groups were calculated and can be seen in Table 1. Although the data was dispersed in a
wide range with SD 19,57 for control condition and 22,02 for experimental condition, it can
be seen from the calculation of both groups mean temperature that anchor value led to
lower estimations than in control group (Table 1).
Table 1 - Processed data that represents non-anchored and anchored groups estimations of
temperature (Appendix 3).
Mean (C)

Standard Deviation

Control conditions (nonanchored group)

-25,67

19,57

Experimental conditions
(anchored group)

-51,33

22,02

Figure 1 - Difference in mean number of temperature from questionnaire answers in nonanchored (control) and anchored (experimental) conditions.

KONONENKO, Svetlana

004180-0114

Inferential Statistics
The hypothesis proposed was one tailed. An independent measures design was used, and data
collected from the experiment was interval. Despite there were not all of three parametric
criteria (interval data, normal distribution of data, and similar variance for both conditions)
met, unrelated t-test was used because it was robust enough to give significant results.
Calculations showed critical value at the 0.01 level of significance for a one-tailed test with
degrees of freedom at 26 = 2.48 (Appendix 8). The observed t value of t(26)=3.51 exceeded
the critical value. Therefore the null hypothesis can be rejected. There is a less than 1%
probability that the result occurred by chance (Appendix 6).

Discussion
This experiment showed that the low anchor value causes lower temperature estimations than
it appeared to be in non-anchored group (values lower that -56C) since the mean temperature
in experimental group was -51,33C compared to -25,67 in control conditions. Performed ttest showed that the results t(26) =3.51 that are highly significant at p=0.01 and support the
experimental hypothesis. Being consistent with Strack and Mussweiler (1995) experiment on
high anchors (Strack and Mussweiler, 1995). It was reported the anchoring effect taking place
and making participants tend to give higher values according to high anchor.

KONONENKO, Svetlana

004180-0114

Wilson et al. revealed that by manipulating the anchor researchers made a highly significant
effect on participants estimations (Wilson, Houston, Etling, & Brekke, 1996). Results are
concordant with the experiment showing a high significance exceeding the critical value t(26)
= 2.16 at p < 0.05. Nevertheless, statistical analysis of the current study showed even higher
significance exceeding critical value at 0.01. The study provides evidence for the effect of
anchors that supports the hypothesis of the current study in terms of the anchoring effect
influence. Data obtained also showed similar results to Simmons study that supported the
idea of higher/lower anchor provides a higher/lower estimated value respectively. Findings
gathered depicted the significance at probability lower than 0.01 which rejects probability of
answers by chance (Simmons, LeBoeuf, & Nelson, 2010).
To conduct the study the independent measure design was used. This type is chosen to avoid
order effects. The main variable in this case is the differences in participants, for example,
variations in gender (groups are not equal in proportion of males and females) and social
background (regardless of the fact that there is a strong effect of the values and mission on
the student's body, the school is an international school and each student has such a unique
and individual background) that can not be unheeded. For example, students from countries
that dont have winter as a season might not make an accurate estimations of negative
temperatures due to lack of experience. Solution of the limitation might be a matched pairs
design where one pair must be assigned to the experimental group and the other to the control
group. Students can be matched based on countries of their origin, language they speak or a
set of subject that they take.
Control group consisted a class of psychology students while experimental group task was
performed by non-psychology students who had the higher probability of not knowing the
study. Hence experimental condition was less likely to show demand characteristics.
However, control group was familiar with the researcher a while ago therefore control group
potentially could be biased or intentionally give results to support hypothesis due to liking
and friendship. Improvement to overcome this issue is the use of purposive sample method
that helps to exclude psychology students to obtain unbiased data and increase representation
of target population.
Another limitation that has to be considered is the intelligibility of the standardized
instructions. Concern is that the question for the independent variable was purposely chosen
to be unusual. The task in the experimental condition required the close reading when asking
to estimate the temperature higher or lower than -56C (Appendix 1). It might have happened
that participants interpreted the scale differently as instructions might have not been clear in
terms of difference between absolute and negative temperature values. This could mislead
participants in their estimation. Hence the nature of the question that was not considered in
the performed t-test could pull participants to one or the other side of the scale that does not
represent their actual values. Therefore there is a chance of type 1 error as the statistical
results rejected the null hypothesis. Indeed, a pilot study can be done to improve standardized
instructions as it usually shows weaknesses of given guidelines.

KONONENKO, Svetlana

004180-0114

In conclusion, the results align with the anchoring effect specifying significance of low
anchor values that is consistent with previous research on high anchors. Hence, the study
gives an empirical support to the idea that the low anchor will provide even lower values on
an estimation. However, the study did not consider the role of participant variables in the
effect of low anchor on judgements. This could be a theme for further research which should
adopt a repeated measures design to address this.

References 1 / 2
1. Lawton, J.-M., Willerton, J., Green, S., & Gammon, J. (2013). Psychology for
the IB diploma, Hodder Education.
2. Mussweiler, T., Englich, B., & Strack, F. 0 anchoring effect. Retrieved from
https://www.hf.uni-koeln.de/data/dppsenglich/File/PDFSStudien/PsychPr_04.pdf
3. Simmons, J. P., LeBoeuf, R. A., & Nelson, L. D. (2010). The effect of
accuracy motivation on anchoring and adjustment: Do people adjust from provided
anchors? Journal of Personality and Social Psychology,99(6), 917932.
doi:10.1037/a0021540
4. Strack, F., Mussweiler, T., Fiedler, K., Ftrster, J., Hannover, B., Higgins, T.,
Wippich, W. Explaining the enigmatic anchoring effect: Mechanisms of selective
accessibility. Retrieved from http://soco.uni-koeln.de/files/jpsp73.pdf
5. Tversky, A., & Kahneman, D. (1985). Behavioral Decision Making. In (pp.
2541). Retrieved from
http://www.stat.columbia.edu/~gelman/surveys.course/TverskyKahneman1981.pdf
10

KONONENKO, Svetlana

004180-0114

6. Lowry, R. VassarStats: Statistical computation web site.

7. Wilson, T. D., Houston, C. E., Etling, K. M., & Brekke, N. (1996). A new look
at anchoring effects: Basic anchoring and its antecedents. Journal of Experimental
Psychology: General, 125(4), 387402. doi:10.1037/0096-3445.125.4.387

Appendices
Appendix 1

Informed consent form

11

KONONENKO, Svetlana

004180-0114

Appendix 2
Briefing letter
Welcome everyone!
Thank you for allowing time to participate in this experiment. That task that you are going to
be asked to perform does not test your intellectual (IQ) abilities.
In a minute I will distribute an informed consent form that I will ask you to signal your
agreement. If at anytime you should change your mind do know that you are entitled to
withdraw from this experiment. Please read all instructions carefully and do not talk to any of
the other participants. A task has been distributed for you to carry out. You will have 10
minutes to complete the task. You can put your hand up as long as you are done and I will
collect the paper. If you have any questions you are more than welcome to ask me.

12

KONONENKO, Svetlana

004180-0114

Appendix 3
Instructions
Please, read questions carefully and write our answers in the space given. You can use words
where appropriate. For other questions specify only one value for your answer which has to
be in numerical form. Use the space given to write your answer. Important! This is an
individual task. Please, do not discuss your answers with a person next to you.

1. What is the year of Leonardo da Vinci birth?

______________
2. What year did Albert Einsteins first visit to the US?
______________
3. Is the mean temperature in Antarctica higher or lower than 56 C?
______________
-What is the mean temperature in Antarctica?
______________
4. How old was Mahatma Gandhi when he died?
______________
Thank you!

13

KONONENKO, Svetlana

004180-0114

Appendix 4
Instructions
Please, read questions carefully and write your answers in the space given. You can give only
one value (in numerical form) for a question as your answer. Important! This is an individual
task. Please, do not discuss your answers with a person next to you.
1. What is the year of Leonardo da Vinci birth?
2. What year did Albert Einsteins first visit to the US?
3. What is the mean temperature in Antarctica?
4. How old was Mahatma Gandhi when he died?

______________
______________
______________
______________

Thank you!

14

KONONENKO, Svetlana

004180-0114

Appendix 5
Debriefing letter
This study was an investigation on the anchoring effect. The anchoring
effect is a cognitive bias that influences you to rely too heavily on the first
piece of information you receive. Stores use it all the time to convince you
to buy. For example, Manufacturer's Suggested Retail Price for a new
Lexus is $39,465. You negotiated a price for $35,250. You feel terrific. You
believe you got a great deal. The anchoring effect has worked! This is an
example of a real-life anchor. So in this experiment I measured how an
anchor value shaped even such a small judgment as answers in a
questionnaire. This was achieved by giving 2 different questionnaires with
4 questions of the similar content. There was only one simple
manipulation, one set didnt have alternative questions (i.e. How old was
Mahatma Gandhi when he died?), and the other had both types:
alternative and non-alternative (i.e. Is the mean temperature in Antarctica
higher or lower than 56 C?). The main focus of the experiment was the
question about the mean temperature that consisted a low anchor
(temperature in Antarctica). Experiment aimed to see whether or not a
low anchor plays a role in judgments. You can find out the results of the
experiment contacting me any time. Please contact me at the following email address if you have any questions regarding this study:
lana.kononenko97@bk.ru
THANK YOU AGAIN FOR YOUR CO-OPERATION!

15

KONONENKO, Svetlana

004180-0114

Appendix 6
Table 1 - Raw data table
Participants

Control conditions/ no
anchor present (C)

Experimental conditions
using -56 C anchor value
(C)

P1

-27

-30

P2

-50

-48

P3

-21

-20

P4

-30

-64

P5

-5

-72

P6

-5

-40

P7

-20

-60

P8

-40

-18

P9

-20

-80

P10

-70

-80

P11

-10

-50

P12

-10

-40

P13

-34

-75

P14

-18

-48

16

KONONENKO, Svetlana

004180-0114

Appendix 7
The t table for critical values
1 tails =

0.1

0.05

0.025

0.01

0.005

2 tails =

0.2

0.1

0.05

0.02

0.01

df =1

3.078

6.314

12.706

31.821

63.656

1.886

2.920

4.303

6.965

9.925

1.638

2.353

3.182

4.541

5.841

1.533

2.132

2.776

3.747

4.604

1.476

2.015

2.571

3.365

4.032

1.440

1.943

2.447

3.143

3.707

1.415

1.895

2.365

2.998

3.499

1.397

1.860

2.306

2.896

3.355

1.383

1.833

2.262

2.821

3.250

10

1.372

1.812

2.228

2.764

3.169

11

1.363

1.796

2.201

2.718

3.106

12

1.356

1.782

2.179

2.681

3.055

13

1.350

1.771

2.160

2.650

3.012

14

1.345

1.761

2.145

2.624

2.977

15

1.341

1.753

2.131

2.602

2.947

16

1.337

1.746

2.120

2.583

2.921

17

1.333

1.740

2.110

2.567

2.898

18

1.330

1.734

2.101

2.552

2.878

19

1.328

1.729

2.093

2.539

2.861

20

1.325

1.725

2.086

2.528

2.845

21

1.323

1.721

2.080

2.518

2.831

22

1.321

1.717

2.074

2.508

2.819

17

KONONENKO, Svetlana

004180-0114

23

1.319

1.714

2.069

2.500

2.807

24

1.318

1.711

2.064

2.492

2.797

25

1.316

1.708

2.060

2.485

2.787

26

1.315

1.706

2.056

2.479

2.779

27

1.314

1.703

2.052

2.473

2.771

28

1.313

1.701

2.048

2.467

2.763

29

1.311

1.699

2.045

2.462

2.756

30

1.310

1.697

2.042

2.457

2.750

60

1.296

1.671

2.000

2.390

2.660

120

1.289

1.658

1.980

2.358

2.617

1.282

1.645

1.960

2.326

2.576

Appendix 8

18

KONONENKO, Svetlana

004180-0114

The calculator was used to find the mean and the standard deviation:
Data Summary
A

Total

14

14

28

-360

-725

-1085

13600

43237

56837

SS

4342.8571

5692.3571

14793.25

mean

-25.667

-51.333

-38.75

X2

Retrieved from VassarStats: Statistical computation web site.

Set A represents control conditions
Set B represents experimental conditions
Where mean= x/N
When:
= total of values
x = each value
n=number of value in the set
x = mean
= standard deviation

19

KONONENKO, Svetlana

004180-0114

20