Skip to main content
Social Sci LibreTexts

9.3: Anchoring

  • Page ID
  • \( \newcommand{\vecs}[1]{\overset { \scriptstyle \rightharpoonup} {\mathbf{#1}} } \)

    \( \newcommand{\vecd}[1]{\overset{-\!-\!\rightharpoonup}{\vphantom{a}\smash {#1}}} \)

    \( \newcommand{\id}{\mathrm{id}}\) \( \newcommand{\Span}{\mathrm{span}}\)

    ( \newcommand{\kernel}{\mathrm{null}\,}\) \( \newcommand{\range}{\mathrm{range}\,}\)

    \( \newcommand{\RealPart}{\mathrm{Re}}\) \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\)

    \( \newcommand{\Argument}{\mathrm{Arg}}\) \( \newcommand{\norm}[1]{\| #1 \|}\)

    \( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\)

    \( \newcommand{\Span}{\mathrm{span}}\)

    \( \newcommand{\id}{\mathrm{id}}\)

    \( \newcommand{\Span}{\mathrm{span}}\)

    \( \newcommand{\kernel}{\mathrm{null}\,}\)

    \( \newcommand{\range}{\mathrm{range}\,}\)

    \( \newcommand{\RealPart}{\mathrm{Re}}\)

    \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\)

    \( \newcommand{\Argument}{\mathrm{Arg}}\)

    \( \newcommand{\norm}[1]{\| #1 \|}\)

    \( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\)

    \( \newcommand{\Span}{\mathrm{span}}\) \( \newcommand{\AA}{\unicode[.8,0]{x212B}}\)

    \( \newcommand{\vectorA}[1]{\vec{#1}}      % arrow\)

    \( \newcommand{\vectorAt}[1]{\vec{\text{#1}}}      % arrow\)

    \( \newcommand{\vectorB}[1]{\overset { \scriptstyle \rightharpoonup} {\mathbf{#1}} } \)

    \( \newcommand{\vectorC}[1]{\textbf{#1}} \)

    \( \newcommand{\vectorD}[1]{\overrightarrow{#1}} \)

    \( \newcommand{\vectorDt}[1]{\overrightarrow{\text{#1}}} \)

    \( \newcommand{\vectE}[1]{\overset{-\!-\!\rightharpoonup}{\vphantom{a}\smash{\mathbf {#1}}}} \)

    \( \newcommand{\vecs}[1]{\overset { \scriptstyle \rightharpoonup} {\mathbf{#1}} } \)

    \( \newcommand{\vecd}[1]{\overset{-\!-\!\rightharpoonup}{\vphantom{a}\smash {#1}}} \)

    \(\newcommand{\avec}{\mathbf a}\) \(\newcommand{\bvec}{\mathbf b}\) \(\newcommand{\cvec}{\mathbf c}\) \(\newcommand{\dvec}{\mathbf d}\) \(\newcommand{\dtil}{\widetilde{\mathbf d}}\) \(\newcommand{\evec}{\mathbf e}\) \(\newcommand{\fvec}{\mathbf f}\) \(\newcommand{\nvec}{\mathbf n}\) \(\newcommand{\pvec}{\mathbf p}\) \(\newcommand{\qvec}{\mathbf q}\) \(\newcommand{\svec}{\mathbf s}\) \(\newcommand{\tvec}{\mathbf t}\) \(\newcommand{\uvec}{\mathbf u}\) \(\newcommand{\vvec}{\mathbf v}\) \(\newcommand{\wvec}{\mathbf w}\) \(\newcommand{\xvec}{\mathbf x}\) \(\newcommand{\yvec}{\mathbf y}\) \(\newcommand{\zvec}{\mathbf z}\) \(\newcommand{\rvec}{\mathbf r}\) \(\newcommand{\mvec}{\mathbf m}\) \(\newcommand{\zerovec}{\mathbf 0}\) \(\newcommand{\onevec}{\mathbf 1}\) \(\newcommand{\real}{\mathbb R}\) \(\newcommand{\twovec}[2]{\left[\begin{array}{r}#1 \\ #2 \end{array}\right]}\) \(\newcommand{\ctwovec}[2]{\left[\begin{array}{c}#1 \\ #2 \end{array}\right]}\) \(\newcommand{\threevec}[3]{\left[\begin{array}{r}#1 \\ #2 \\ #3 \end{array}\right]}\) \(\newcommand{\cthreevec}[3]{\left[\begin{array}{c}#1 \\ #2 \\ #3 \end{array}\right]}\) \(\newcommand{\fourvec}[4]{\left[\begin{array}{r}#1 \\ #2 \\ #3 \\ #4 \end{array}\right]}\) \(\newcommand{\cfourvec}[4]{\left[\begin{array}{c}#1 \\ #2 \\ #3 \\ #4 \end{array}\right]}\) \(\newcommand{\fivevec}[5]{\left[\begin{array}{r}#1 \\ #2 \\ #3 \\ #4 \\ #5 \\ \end{array}\right]}\) \(\newcommand{\cfivevec}[5]{\left[\begin{array}{c}#1 \\ #2 \\ #3 \\ #4 \\ #5 \\ \end{array}\right]}\) \(\newcommand{\mattwo}[4]{\left[\begin{array}{rr}#1 \amp #2 \\ #3 \amp #4 \\ \end{array}\right]}\) \(\newcommand{\laspan}[1]{\text{Span}\{#1\}}\) \(\newcommand{\bcal}{\cal B}\) \(\newcommand{\ccal}{\cal C}\) \(\newcommand{\scal}{\cal S}\) \(\newcommand{\wcal}{\cal W}\) \(\newcommand{\ecal}{\cal E}\) \(\newcommand{\coords}[2]{\left\{#1\right\}_{#2}}\) \(\newcommand{\gray}[1]{\color{gray}{#1}}\) \(\newcommand{\lgray}[1]{\color{lightgray}{#1}}\) \(\newcommand{\rank}{\operatorname{rank}}\) \(\newcommand{\row}{\text{Row}}\) \(\newcommand{\col}{\text{Col}}\) \(\renewcommand{\row}{\text{Row}}\) \(\newcommand{\nul}{\text{Nul}}\) \(\newcommand{\var}{\text{Var}}\) \(\newcommand{\corr}{\text{corr}}\) \(\newcommand{\len}[1]{\left|#1\right|}\) \(\newcommand{\bbar}{\overline{\bvec}}\) \(\newcommand{\bhat}{\widehat{\bvec}}\) \(\newcommand{\bperp}{\bvec^\perp}\) \(\newcommand{\xhat}{\widehat{\xvec}}\) \(\newcommand{\vhat}{\widehat{\vvec}}\) \(\newcommand{\uhat}{\widehat{\uvec}}\) \(\newcommand{\what}{\widehat{\wvec}}\) \(\newcommand{\Sighat}{\widehat{\Sigma}}\) \(\newcommand{\lt}{<}\) \(\newcommand{\gt}{>}\) \(\newcommand{\amp}{&}\) \(\definecolor{fillinmathshade}{gray}{0.9}\)

    How tall do you think Mt. Everest is? (Don’t Google it—that kind of defeats the purpose) You probably don’t know the exact number, but do you think it’s taller or shorter than 150 feet? Assuming you said “taller”[1], make a guess. How tall do you think it is?

    Figure 2. Mt. Everest

    Mt. Everest is roughly 29,000 ft. in height.

    How’d you do? If I were to guess, based on the psychology I’m about to share with you, you probably undershot it. Even if you didn’t, most people would.[3] The reason is what’s called the anchoring heuristic.

    Back in the 1970s, Amos Tversky and Daniel Kahneman identified a few reliable mental shortcuts people use when they have to make judgments. Oh, you thought people were completely rational every time they make a decision? It’s nice to think, but it’s not always what happens. We’re busy! We have lives! I can’t sit around and do the math anytime I want to know how far away a stop sign is, so I make estimates based on pretty reliable rules of thumb. The tricks we use to do that are called heuristics.

    The Basics of the Anchoring Heuristic

    The basic idea of anchoring is that when we’re making a numerical estimate, we’re often biased by the number we start at. In the case of the Mt. Everest estimate, I gave you the starting point of 150 feet. You though “Well, it’s taller than that,” so you likely adjusted the estimate from 150 feet to something taller than that. The tricky thing, though, is that we don’t often adjust far enough away from the anchor.

    Let’s jump into an alternate timeline and think about how things could have gone differently. Instead of starting you at 150 feet, this time I ask you whether Mt. Everest is taller or shorter than 300,000 feet. This time you’d probably end up at a final estimate that’s bigger than the correct answer. The reason is you’d start at 300,000 and start adjusting down, but you’d probably stop before you got all the way down to the right answer.

    Coming Up With Your Own Anchors

    In general, this is a strategy that tends to work for people. After all, when we don’t know an exact number, how are we supposed to figure it out? It seems pretty reasonable to start with a concrete anchor and go from there.

    In fact, some research has shown that this is how people make these estimates when left to their own devices. Rather than work from an anchor that’s given to them (like in the Mt. Everest example), people will make their own anchor—a “self-generated anchor.”

    For example, if you ask someone how many days it takes Mercury to orbit the sun, she’ll likely to start at 365 (the number of days it takes Earth to do so) and then adjust downward. But of course, people usually don’t adjust far enough.

    Biased By Completely Arbitrary Starting Points

    This paints an interesting picture of how we strive to be reasonable by adopting a pretty decent strategy for coming up with numerical estimates. When you think about, even though we’re biased by the starting point, it sounds like a decent strategy. After all, you’ve got to start somewhere!

    But what if the starting point is totally arbitrary? Sure, the “150 feet” anchor from before probably seems pretty arbitrary, but at the time you might have thought “Why would he have started me at 150 feet? It must be a meaningful starting point.”

    Figure 3. A game of Roulette

    The truth is that these anchors bias judgments even when everyone realizes how arbitrary they are. To test this idea, one study asked people to guess the percentage of African countries in the United Nations. To generate a starting point, though, the researchers spun a “Wheel-of- Fortune” type of wheel with numbers between 0 – 100.

    For whichever number the wheel landed on, people said whether they thought the real answer was more or less than that number. Even these random anchors ended up biasing people’s estimates. If the wheel had landed on 10, people tended to say about 25% of countries in the UN are African, but if the wheel had landed on 65, they tended to say about 45% of countries in the UN are African. That’s a pretty big difference in estimates, and it comes from a random change in a completely arbitrary value.

    People’s judgments can even be biased by anchors based on their own social security numbers.

    Biased By Numbers in the Air

    Through all of these examples, the anchor has been a key part of the judgment process. That is, someone says “Is it higher or lower than this anchor?” and then you make a judgment. But what if the starting point is in the periphery?

    Even when some irrelevant number is just hanging out in the environment somewhere, it can still bias your judgments! These have been termed “incidental anchors.”

    Figure 4. Restaurant seating

    For example, participants in one study were given a description of a restaurant and asked to report how much money they would be willing to spend there. Two groups of people made this judgment, and the only difference between them is that for one group, the restaurant’s name was “Studio 17” and for the other group, the restaurant’s name was “Studio 97.” When the restaurant was “Studio 97,” people said they’d spend more money (an average of about $32) than when the restaurant was “Studio 17” (where they reported a willingness to spend about $24).

    Other research has shown that people were willing to pay more money for a CD when a totally separate vendor was selling $80 sweatshirts, compared to when that other vendor was selling $10 sweatshirts.

    In both of these examples, the anchor was completely irrelevant to the number judgments, and people weren’t even necessarily focused on the anchor. Even still, just having a number in the environment could bias people’s final judgments.

    Raising the Anchor and Saying “Ahoy”

    Across all of these studies, a consistent pattern emerges: even arbitrary starting points end up biasing numerical judgments. Whether we’re judging prices, heights, ages, or percentages, the number we start at keeps us from reaching the most accurate final answer.

    Figure 5. Lifting our mental anchors

    This has turned out to be a well-studied phenomenon as psychologists have explored the limits of its effects. Some results have shown that anchoring effects depend on your personality, and others have shown that they depend on your mood.

    In fact, there’s still some debate over how anchoring works. Whereas some evidence argues for the original conception that people adjust their estimates from a starting point, others argue for a “selective accessibility” model in which people entertain a variety of specific hypotheses before settling on an answer. Still others have provided evidence suggesting that anchoring works similarly to persuasion.

    Overall, however, the anchoring effect appears robust, and when you’re in the throes of numerical estimates, think about whether your answer could have been biased by other numbers floating around.

    Problem 2 (adapted from Joyce & Biddle, 1981):

    We know that executive fraud occurs and that it has been associated with many recent financial scandals. And, we know that many cases of management fraud go undetected even when annual audits are performed. Do you think that the incidence of significant executive-level management fraud is more than 10 in 1,000 firms (that is, 1 percent) audited by Big Four accounting firms?

    a. Yes, more than 10 in 1,000 Big Four clients have significant executive-level management fraud.

    b. No, fewer than 10 in 1,000 Big Four clients have significant executive-level management fraud.

    What is your estimate of the number of Big Four clients per 1,000 that have significant executive-level management fraud? (Fill in the blank below with the appropriate number.)

    in 1,000 Big Four clients have significant executive-level management fraud.

    Regarding the second problem, people vary a great deal in their final assessment of the level of executive-level management fraud, but most think that 10 out of 1,000 is too low. When I run this exercise in class, half of the students respond to the question that I asked you to answer.

    The other half receive a similar problem, but instead are asked whether the correct answer is higher or lower than 200 rather than 10. Most people think that 200 is high. But, again, most people claim that this “anchor” does not affect their final estimate. Yet, on average, people who are presented with the question that focuses on the number 10 (out of 1,000) give answers that are about one-half the size of the estimates of those facing questions that use an anchor of 200. When we are making decisions, any initial anchor that we face is likely to influence our judgments, even if the anchor is arbitrary. That is, we insufficiently adjust our judgments away from the anchor.

    This page titled 9.3: Anchoring is shared under a CC BY license and was authored, remixed, and/or curated by Mehgan Andrade and Neil Walker.

    • Was this article helpful?