The Confirmation Bias: Seeing What We Expect to See

Anecdotes, the topic of the post at this link, are often used as evidence for claims, even though they have weaknesses that limit severely their value as evidence. Anecdotes are important for another topic we’re discussing in my PSY 101 course: the confirmation bias, which is a strong tendency to readily accept evidence that seems to support beliefs we already have (i.e., our preconceptions) and to examine closely evidence that seems to contradict our preconceptions in order to find problems with the disconfirming evidence that allow us to discount it.

Figure 1. Discounting evidence that disconfirms one's preconceptions can lead to the development of bizarre beliefs that people hold fervently.
Figure 1. Discounting evidence that disconfirms one’s preconceptions can lead to the development of bizarre beliefs that people hold fervently.

In everyday life, however, we often do not even become aware of disconfirming evidence. Sometimes, this is because we can easily avoid it (e.g., we might avoid watching certain news programs that we know will make claims that contradict our beliefs). Other times, our cognitive limitations make it difficult for us to notice the disconfirming evidence (i.e., it doesn’t reach the conscious level). One example of the latter involves the belief that washing our cars causes it to rain. Many of us are able to point to occasions (anecdotes) on which it rained after we washed their cars, which seems to be compelling evidence of the truth of this belief. We often fail, however, to notice (and, therefore, don’t remember) occasions on which these events didn’t occur together. In order to get better evidence for the accuracy of the belief, we would need to make the kinds of observations indicated in the following table.

Table 1. Each cell of this table indicates the observations that would need to be made to determine if washing one's car causes it to rain.
Table 1. Each cell of this table indicates the observations that would need to be made to determine if washing one’s car causes it to rain.

In making observations that would allow us to fill in the cells of the table, we force ourselves to pay attention not only to evidence that supports our belief, but also to evidence that might disconfirm that belief:

  • The cell labelled A shows the number of times we washed the car and it rained.
  • The cell labelled B shows the number of times we washed the car and it didn’t rain.
  • The cell labelled C shows the number of times we didn’t wash the car and it rained.
  • The cell labelled D shows the number of times we didn’t wash the car and it didn’t rain.

Let’s say that we make the relevant observations for one year and get the following results.

Table 2. The cells of the table show, for a one-year period, the proportion (percentage) of days we either washed or did not wash the car and it either rained or did not rain.
Table 2. The cells of the table show, for a one-year period, the proportion (percentage) of days we either washed or did not wash the car and it either rained or did not rain.

Was it more likely to rain on the days we washed the car?

  • Cell A shows that, on 20% of the days, we washed the car and  it rained.
  • Cell B shows that, on 80% of the days, we washed the car and it didn’t rain.
  • Cell C shows that, on 20% of the days, we didn’t wash the car and  it rained.
  • Cell D shows that, on 80% of the days, we didn’t wash the car and it didn’t rain.

Thus, regardless of whether we had just washed our car or not, it rained on 20% of the days that year (and it didn’t rain on 80% of the days). In other words, washing our car was not associated with whether or not it rained.

The confirmation bias is caused, in part, by our unconscious tendency to ignore, avoid, or distort information that would show a preconception to be wrong. In the present example, people tend to pay attention only to the first cell of the table and to ignore the rest. This is because, in general, we are much more likely to notice when something happens than when something doesn’t happen. By forcing ourselves to pay attention to all relevant information in such situations, we are more likely to realize when our preconceptions are inaccurate.


To summarize: Cognitive researchers have found that we have an automatic (unconscious) tendency to seek out and readily accept information that agrees with (confirms) our preconceptions, and to ignore, distort, or discount information that contradicts (disconfirms) them. This confirmation bias serves to maintain and strengthen our preconceptions: we are much more likely to perceive and remember experiences that confirm our prior beliefs, and to discount or reinterpret those that disconfirm them. Thus, over time, the confirmation bias results these beliefs becoming so well established in our minds that eventually we consider them to be common sense (i.e., obviously true). If we wish to minimize the effects of the confirmation bias, we must force ourselves to look for and examine closely both confirming and disconfirming evidence.


Anecdotes & Testimonials: Good Evidence for Claims?

In my PSY 101 class, we recently discussed some problems with using anecdotes and testimonials as evidence for claims. In this post, I want to begin to explore this issue in more depth. In future posts, I will discuss further some of the issues touched upon here.


One afternoon, Eileen Lipsker was sitting in her family room watching Jessica, her red-headed five-year-old daughter, play with her friends. Eileen later reported that she felt “spaced out” and was “thinking of nothing.” Lenore Terr (1994), a psychiatrist who spoke with her on many occasions, described what Eileen said happened next:

Jessica twisted her head to look at her mother. To ask something? Her chin pointed up in inquiry. She looked up and over her shoulder. Her eyes brightened. How odd! The young girl’s body remained stationary, while her head pivoted around and up…. And at exactly that moment Eileen Lipsker remembered something. She remembered it as a picture. She could see her redheaded friend Susan Nason looking up, twisting her head, and trying to catch her eye.

Eileen, eight years old, stood outdoors, on a spot a little above the place where her best friend was sitting. It was 1969, twenty years earlier. The sun was beaming directly into Susan’s eyes. And Eileen could see that Susan was afraid…. [Eileen] looked away from those arresting eyes and saw the silhouette of her father. Both of George Franklin’s hands were raised above his head. He was gripping a rock. He steadied himself to bring it down. His target was Susan. (pp. 2-3)

Eileen told Terr that this is how she first recovered her repressed memory of Susan Nason’s murder by her father. Eileen’s recounting of the memory recovery is an example of an anecdote: a brief story told by an individual about a personal experience. No matter how interesting or compelling an anecdote may be, it doesn’t provide good evidence for a claim because it is based on interpretations and memories of personal experiences. In other words, an anecdote is inadequate evidence for a claim because it does not control for factors that affect how a personal experience is (a) initially perceived and interpreted, and (b) eventually remembered.

Figure 1. An example of an anecdote used to support the claim that extraterrestrials visit earth
Figure 1. An example of an anecdote used to support the claim that extraterrestrials visit earth


Autistic Disorder is a severe mental disorder that develops in children before the age of three years. It has three main symptoms: a severe impairment in social interaction, a severe impairment in the ability to communicate, and a severely restricted range of interests, activities, and behaviors. On occasion, new treatments for autism are announced that seem to offer hope for either a cure or, at least, a dramatic reduction of symptoms. One such well-publicized treatment used injections of secretin–a hormone that assists in the digestion of food. Some have claimed that secretin improves the social and language skills of autistic individuals by affecting specific behaviors such as the amount of eye contact made, the level of awareness of one’s surroundings, the degree of sociability, and the amount of speech. One proponent of secretin therapy provided the following evidence for this claim:

The good news is that confirmatory evidence of the power of secretin keeps coming. A national newspaper told of Florida pediatrician Jeff Bradstreet’s own four-year-old son, Matthew, shocking his parents by holding his first normal conversation with them the day after his first secretin infusion. And Virginia pediatrician Lawrence Leichtman told me of his “miracle case”: a five-year-old who had previously said only two words amazed all in the office by saying, 15 minutes after his infusion, “I am hungry. I want to eat.” Most cases are much less dramatic, but the autism world is excited, and for good reason. (Rimland, 1998, p. 3)

Is this good evidence for the effectiveness of secretin in the treatment of autism? The evidence consists of two testimonials. A testimonial is an anecdote that describes the supposed merits of a product or service. Testimonials are not good evidence for a claim because they are anecdotes and, as stated above, anecdotes don’t control for factors that might distort our observations and interpretations of a personal experience, as well as how we remember it later on. For example, we may misremember exactly what happened during the event, or may have misinterpreted what we observed during the event.

Figure 2. A testimonial from a celebrity about a brand of cigarettes
Figure 2. A testimonial from a celebrity about a brand of cigarettes (circa 1951)

In testimonials about therapeutic treatments, one very important factor that causes distortions of perceptions, interpretations, and memories is people’s expectations for the treatment. These expectations may cause them to conclude that their symptoms improved or disappeared even when they haven’t. This generally happens in one of two ways:

  1. The expectations may cause observers (e,g, patients, family members, doctors) to misperceive or misinterpret the behavior of those receiving a treatment, thereby concluding that the behavior has changed when it really hasn’t.
  2. The expectations may cause observers to experience improvement that has nothing to do with the nature of the treatment itself (e.g., the placebo effect or a self-fulfilling prophecy; see later posts in this series).

Specific examples of some of the problems mentioned here will be described in future posts.

You may contact me at


Pendergrast, M. (1996). Victims of memory: Sex abuse accusations and shattered lives (2nd ed.). Hinesburg, VT: Upper Access.

Rimland, B. (1998). The use of secretin in autism: Some preliminary answers. Autism Research Review International, 12(4), 3. Retrieved January 23, 2013, from

Terr, L. T. (1994). Unchained memories: True stories of traumatic memories, lost and found. New York: BasicBooks.

The Evolution of Domestic Dogs — Part 1

The dog was the first domesticated animal. Domestication is an “evolutionary process [that] has been influenced by humans to meet their needs” (Secretariat, 1992, p. 3). In other words, domestication of a species causes biological changes over generations through selection by humans for favorable traits (i.e., traits that are useful, valuable, aesthetically pleasing, etc.).

Domestication led to extraordinarily large changes in the behavioral characteristics of domestic dogs, as well as in their physical characteristics (which is obvious when one compares the many breeds of dogs). Scientific research on the evolution of dog behavior began in the mid-1800s, most notably in the work of Charles Darwin (Darwin, 1872). In the middle of the twentieth century, a deeper understanding of the evolution of dog behavior was gained by combining behavioral analyses with classical genetic analyses of dog breeds (for a review, see Scott & Fuller, 1965).

Figure 1. An extreme example of the effects of domestication on the physical characteristics of a subspecies of the gray wolf.

Over the past 20 years, archaeological discoveries in combination with the results of highly sophisticated genetic analyses have shed a great deal of light on the evolution of domestic dogs (for a review, see Larson, Karlsson, Perri, et al., 2012). For example, there now is little doubt that domestic dogs evolved from the gray wolf, which is found in many parts of Europe and Asia (Honeycutt, 2010; Wayne & Ostrander, 2007). In fact, domestic dogs (Canis lupus familiaris) are considered to be a subspecies of the gray wolf (Canis lupus lupus). This means that, although dogs and wolves have physical features that often are very different, they can mate and produce fertile offspring.

Nevertheless, there still is much controversy about when and where domestic dogs originated. These disagreements are focused on the answers to two questions: when did domestic dogs “split” from gray wolves and where did this happen? These questions have proved difficult to answer because the results of genetic and archaeological research are complex and, hence, very difficult to interpret.

Genetic research on differences in DNA sequences have led to a wide range of estimates about when dogs and wolves first diverged: sometime between 20,000 to 100,000 years ago. One reason for the wide variation in these estimates is that dogs and wolves probably continued to interbreed, not only over long periods of time but also in many locations (Vilà, Savolainen, Maldonado, et al., 1997).

Archaeological researchers find no clear evidence for the existence of domestic dogs until about 15,000-30,000 years ago (Germonpré, Sablin, Stevens, 2009; Ovodov, Crockford, Kuzmin, et al., 2011). A major difficulty with interpreting archaeological findings, however, is that physical characteristics typically used to distinguish domestic dogs from wolves (e.g., the size and position of the teeth, the size and shape of the skull, etc.) probably varied much more in ancient dog populations than they do today (Larson, Karlsson, Perri, et al., 2012). In addition, nothing is known about the variation of these traits in populations of ancient wolves. In other words, the physical characteristics used to distinguish modern dogs and wolves probably overlapped to a relatively large extent in ancient dogs and wolves, thereby making it very difficult for archaeologists to know if they are looking at the bones and teeth of a wolf or a dog.

Figure 2. The 33,000-year-old skull of a wolf-like animal reputed to be a dog (Ovodov, Crockford, Kuzmin, et al., 2011)
Figure 2. The 33,000-year-old skull of a wolf-like animal reputed to be a dog (Ovodov, Crockford, Kuzmin, et al., 2011)

As of now, both the archeological and genetic evidence allow us to conclude with certainty that domestic dogs existed at least 15,000 years ago (Larson, Karlsson, Perri, et al., 2012). It is still an open question, however, if ancient dog populations leading to modern domestic dogs first diverged from wolves earlier than that.

In the next post, I’ll review what recent research seems to tell us about the genetic differences linked to the many physical and behavioral differences between (a) modern domestic dogs and their ancestral species, and (b) the various breeds of modern domestic dogs.

You may contact me at


Darwin, C. (1872). The expression of the emotions in man and animals. London: John Murray. Retrieved January 21, 2013, at

Germonpré M, Sablin MV, Stevens RE, Hedges REM, Hofreier M, Stiller M, et al. (2009). Fossil dogs and wolves from Palaeolithic sites in Belgium, the Ukraine and Russia: Osteometry, ancient DNA and stable isotopes. Journal of Archaeological Science, 36, 473-490. doi: 10.1016/j.jas.2008.09.033

Honeycutt, R. L. (2010). Unraveling the mysteries of dog evolution. BMC Biology, 8(20). doi:10.1186/1741-7007-8-20

Larson, G., Karlsson, E. K., Perri, A., Webster, M. T., Ho, S. Y. W., Peters, J., et al. (2012). Rethinking dog domestication by integrating genetics, archeology, and biogeography. Proceedings of the National Academy of Sciences, 109, 8878–8883. doi: 10.1073/pnas.1203005109

Ovodov, N. D., Crockford, S. J., Kuzmin, Y. V., Higham, T. F. G., Hodgins, G. W. L., & van der Plicht, J. (2011). A 33,000-year-old incipient dog from the Altai Mountains of Siberia: Evidence of the earliest domestication disrupted by the Last Glacial Maximum. PLoS ONE 6(7): e22821. doi:10.1371/journal.pone.0022821

Scott, J. P., & Fuller, J. L. (1965). Genetics and the social behavior of the dog. Chicago: University of Chicago Press.

Secretariat, C. B. D. (1992). The Convention on Biological Diversity. Retrieved December 30, 2012, from

Vilà, C., Savolainen, P., Maldonado, J. E., Amoim, I. R., Rice, J. E., Honeycutt, R.L., et al. (1997). Multiple and ancient origins of the domestic dog. Science, 276, 1687-1689.

Wayne, R. K., & Ostrander, E. A. (2007). Lessons learned from the dog genome. Trends in Genetics, 23, 557–567. doi: 10.1016/j.tig.2007.08.013

Natural Selection and Genes

When particular expressions of a characteristic are naturally selected and those expressions are associated with particular gene variants, those gene variants will be more likely than others to be passed on to future generations. For example, let’s say that there exists a gene with two variants, T and t, and that these variants are associated with difference in the average height of plants from a particular (fictional) species, Pirasus arizonensis. From a baseline height of five inches, T increases the average height by 1/2 inch, whereas t decreases the average height by 1/2 inch.

If P. arizonensis seeds are transported by birds into an environment in which the fully grown plant is surrounded by plants from another (fictional) species, Torensi mojavensis, that has an average height of six inches, the taller species will limit the smaller species’ access to sunlight. This, of course, would be detrimental to the survival and reproductive success of P. arizonensis. Thus, any P. arizonensis plant that grows taller than the average height of five inches will tend to survive longer and reproduce more.

Let’s say that the following 300 P. arizonensis plants have grown in this new environment:

100 TT plants, which will have an average height of 6 inches;
100 Tt plants, which will have an average height of five inches;
100 tt plants, which will have an average height of four inches.

The plants that are smaller than the surrounding plants will be much more likely to die before reproducing. And of course, other environmental factors also will affect plant survival, but in a more random fashion. Let’s say that 80 of the tt plants die before producing seeds, 50 of the Tt plants die before producing seeds, and 10 TT plants die before producing seeds. The result: 72% of the seeds in the next generation will contain T but only 28% will include t. Thus, if 300 plants grow in the next generation, their numbers will be as follows:

156 TT plants;
120 Tt plants;
24 tt plants.

As you can see, there are 276 plants with at least one copy of the T variant, which is much larger than the number of plants with at least one t variant (144 plants).

The third requirement of evolution by natural selection — the increased reproductive success of individuals with particular expressions of a characteristic —must remain stable over generations. This means that the “selective pressure” on P. arizonensis plants with respect to their heights must not change. The six-inch tall (on average) T. mojavensis plants must continue to limit the amount of sunlight obtained by smaller P. arizonensis plants. In this case, the following change in the frequencies of the T and t gene variants in this species should occur:

Figure 1. Changes in the Frequencies of T and t Over Five Generations.

As you can see, the frequency of T becomes almost 100% within five generations, which means that the population in this new environment now consists almost entirely of plants that are about six inches tall. Thus, over a very short period of time, natural selection can lead to a large change in the average expression of a characteristic in a population when individual differences in that characteristic are strongly associated with genetic differences.

Now, let’s return to the example from the previous post: the founding population of fruit flies on a tiny and isolated island buffeted by strong winds. Differences in the size of fruit-fly wings are strongly linked to differences in genes (Robertson & Reeve, 1952), to a degree similar to that described above for height differences in the fictional plant species. Thus, if the windy environment naturally select flies with smaller wings, gene variants correlated with smaller wings will increase in frequency over generations. This means that the population will evolve a smaller average wing size. This is shown in Figure 2 (the black bars represent the founding population and the white bars represent the population after a number of generations of natural selection for smaller wing sizes).

Figure 2. Evolution of a Population of Fruit Flies Undergoing Natural Selection for Smaller Wing Size

Although these two examples are fictional ones, there are many examples of natural selection and artificial selection (selection in which humans breed organisms that express particular characteristics). For introductions to and histories of the concepts of evolution and natural selection, see Colby (1996-1997), Endler (1986), and Zimmer (2001).

You may contact me at


Colby, C (1996-1997). Introduction to evolutionary biology (Version 2). Retrieved November 14, 2012, from

Endler, J. A. (1986). Natural selection in the wild. Princeton, NJ: Princeton University Press.

Robertson, F. W., & Reeve, E. C. R. (1952). Studies in quantitative inheritance. I. The effects of selection for wing and thorax length in Drosophila melanogaster. Journal of Genetics, 50, 416-448.

Zimmer, C. (2001). Evolution: The triumph of an idea. New York: HarperCollins.

The Evolutionary Approach in Psychology

In its most general sense, biological evolution refers to changes over generations in a population—changes in features of the body, mind, or behavior.

The evolutionary approach attempts to explain mind and behavior in terms of biological structures and processes that have evolved over hundreds to thousands of generations. This approach assumes that species have evolved ways of responding (cognitively, emotionally, and behaviorally) to environmental events because these responses led to greater survival and reproductive success in ancestral populations.

To take one example, the human spinal cord develops in such a way that it can rapidly process sensory information related to the temperature of objects. When we touch an object that is very hot, the spinal cord immediately activates a reflexive response that rapidly pulls the finger away from the object. Because this response occurs automatically, we can’t explain it as the result of conscious choice. In fact, the hand typically is jerked away before the information reaches the cerebral cortex (activity in the cortex is necessary for the conscious perception of pain). The existence of this spinal-cord reflex may be explained as the product of evolution: individuals that quickly pulled a body part away from painful stimuli were more likely to survive and, hence, reproduce because this proto-reflex prevented severe bodily damage. This explanation asserts that evolutionary changes in spinal-cord reflexes were caused by natural selection.

Evolution By Natural Selection

Evolution refers specifically to changes in the frequencies of variants of a characteristic (biological, psychological, or behavioral) over generations. A characteristic is a feature of an individual, such as eye color, that can be distinguished from other features, such as hair color. Characteristics often have variants that involve observable individual differences. For example, eye color has many variants, such as shades of brown, green, gray, and blue. Hair color also has many variants, such as shades of black, brown, red, and blonde. We will refer to such variants as expressions of the characteristic. Evolution, therefore, is a change over generations in the frequencies of expressions of a characteristic within a population of organisms. An analogous way of saying this is evolution is a change over generations in the average expression of a characteristic within a population of organisms. For example, a population consisting of 99% brown-eyed individuals and 1% blue-eyed individuals may evolve over generations into a population consisting of 1% brown-eyed individuals and 99% blue-eyed individuals. The average expression of eye color in this population evolved from brown to blue.

What causes evolution to occur in populations? For two decades beginning in 1836, Charles Darwin developed a credible naturalistic theory able to explain evolutionary changes — a theory that he began to develop when trying to interpret observations he had made during his five-year voyage on the H.M.S. Beagle (Darwin, 1839), as well as in research that he and others performed during the 23 years after Darwin returned from that voyage. This was the theory of evolution by natural selection. He published a detailed description of the theory in the first edition of the book, On the Origin of Species (Darwin, 1859; the sixth edition generally is considered to represent Darwin’s mature views on evolution and its causes.) No one before Darwin had so masterfully marshaled such an enormous amount of supporting evidence for the evolution of organisms. In addition, no one before Darwin had outlined such a compelling explanation of evolution: natural selection. Natural selection may be defined as the increased reproductive success of individuals with particular expressions of physical, mental, and/or behavioral characteristics. To put it most simply, Darwin argued that natural selection occurs when a subset of individuals in a population produce a greater number of offspring, on average, than others because they express a physical, mental, or behavioral variant that allows them to adapt better to their environments.

Let’s consider, for example, a fictional species of fruit fly that has just arrived on a windy and tiny island hundreds of miles from any other land. And let’s say that, in this founding population, there exists a a broad range of individual differences in wing size, as shown in the following graph.

Figure 1. The Percentage of Individuals in a Founding Population of Fruit Flies With Wings of Various Lengths

As can be seen in the graph, some individuals have large wings, which are advantageous for flying speed and for the ability to stay airborne, whereas others have small wings, which result in slower flying speeds and greater difficulties with staying airborne. On this small and windy island, larger-winged flies probably would be more likely to get blown out to sea, whereas the smaller-winged flies would be less likely to suffer that fate. Thus, smaller-winged flies would be more likely to survive long enough to reproduce than larger-winged flies.

This fictional example illustrates well the simple idea behind natural selection: individuals differ in their reproductive success because they have variants of characteristics associated with the ability to adapt to local environmental conditions. Because individuals with particular variants adapt better relative to individuals with other variants, the former survive longer, on average, and, hence, have more opportunities to reproduce. In other words, the local environmental conditions consist of factors that impose biological, psychological, and behavioral demands on organisms. These factors “naturally select” those organisms best able to deal with the environmental demands: they survive longer and reproduce more than others in their local population.

Given the obvious fact that natural selection occurs, how does it produce evolutionary changes in populations of organisms? There are three requirements that must be met in order for evolution in the average expression of a characteristic to occur through natural selection:

  1. There must be individual differences in the expression of the characteristic.
  2. These individual differences must be associated with genetic differences.
  3. The increased reproductive success of individuals with particular expressions of the characteristic must remain stable over generations.

The first requirement has already been discussed (see Figure 1). The second requirement involves the existence of genetic variants that affect the development of characteristics. A gene is the basic unit of biological heredity. Genes consist of sequences of chemical units (sections of DNA molecules) that are contained in chromosomes carried by the sperm of males and the ova (eggs) of females. In human reproductive cells (sperm and ova), there are 23 chromosomes, which together contain about 22,000 genes (Pertea & Salzberg, 2010). This means that, on average, each human chromosome contains about 1000 genes.

What Do Genes Do?

Genes influence the production of proteins and their use in developing and maintaining the body (for a history of the concept of the gene, see Rheinberger & Müller-Wille, 2010). For example, there are probably at least 16 genes that affect the development of eye color in humans (White & Rabago-Smith, 2011). But it seems that only two or three have major effects on individual differences in eye color. So, for purposes of explanation, let’s assume that there are only three genes that influence the development of eye color, which, for the sake of simplicity, we’ll refer to as Gene A, Gene B, and Gene C. As can be seen in the following table, babies receive one copy of each gene from their biological fathers (labelled as 1) and one copy of each gene from their mothers (labelled as 2):

Gene A has two variants: a brown variant and a nonbrown variant. If at least one brown variant is inherited from either parent, then, regardless of what is inherited at Gene B and Gene C, the person will develop brown eyes:

If, on the other hand, the nonbrown variant is inherited from each parent, then eye color is determined by what is inherited at Gene B and Gene C. Gene B has two variants: a brown variant and a blue variant. If at least one brown variant of Gene B is inherited from either parent, the person will develop brown eyes, regardless of what is inherited at Gene C:

If, on the other hand, the blue variant is inherited from each parent, the person will develop blue eyes depending on what is inherited at Gene C (which we will ignore for the moment):

Gene C has two variants: a green variant and a blue variant. If the blue variant of Gene B is inherited from each parent, then, if at least one green variant of Gene C is inherited from either parent, the person will develop green eyes:

If, on the other hand, the blue variant of Gene C is inherited from each parent, the person will develop blue eyes:

Thus, in our simplified example, eye color is determined by interactions among variants of three genes. The actual situation is much more complex: there are other genes as well as environmental factors that produce the many shades of eye color we see in real life.

Our example shows that gene variants, and interactions among them, contribute to the development of the physical characteristics of the body. In fact, you see evidence for this claim all around you: biological relatives often bear a strong resemblance to each other, as do conspecifics (members of the same species). Members of two closely related species typically don’t mate, and if they do, the mating typically doesn’t produce offspring. When interspecific matings are successful, however, the offspring generally express physical characteristics intermediate between the two species. For example, matings between male donkeys and females horses produce mules; and matings between male horses and female donkeys produce hinnies. Mules and hinnies have physical and behavioral characteristics that are intermediate between those of horses and donkeys. We’ll come back to this when we talk about matings between dogs and species that are closely related to them

The next post will look more closely at natural selection at the level of genes.

You may contact me at


Darwin, C. (1839). The voyage of the Beagle. Retrieved November 12, 2012, from

Darwin, C. (1859). On the origin of species by means of natural selection, or the preservation of favoured races in the struggle for life (6th ed.). Retrieved November 12, 2012, from

Pertea, M. & Salzberg, S. L. (2010). Between a chicken and a grape: Estimating the number of human genes. Genome Biology, 11, 206. doi:10.1186/gb-2010-11-5-206

Rheinberger, H-J, & Müller-Wille, S. (2010). Gene. The Stanford Encyclopedia of Philosophy. E. N. Zalta (Ed.). Retrieved November 12, 2010, from

White, D., & Rabago-Smith, M. (2011). Genotype–phenotype associations and human eye color. Journal of Human Genetics, 56, 5-7. doi:10.1038/jhg.2010.126
White & Rabago-Smith, 2011

Skepticism and Empiricism

One of the major goals of this website is to help people learn about scientific research in psychology and how it can help all of us to better understand why we do what we do in our everyday lives. This goal requires that we learn about some of the attitudes and assumptions indispensable to doing good research in the behavioral sciences. In this post, I will argue that two essential attitudes for scientific researchers are skepticism and empiricism. In fact, these attitudes form the foundation of the scientific approach to understanding ourselves and the world around us.

Evaluating Claims About Mind and Behavior

Angel Therapy works on the belief that everyone has guardian angels, and these angels perform God’s will of peace for us all. When we open ourselves to hear our angels’ messages, every aspect of our lives become more peaceful…. You can connect with your angels and guides. According to the therapy, everyone has at least 2 guardian angels, and a variety of spirit guides, souls who have agreed to work with you throughout your life. These angels and guides are loving entities, and are here to help you in every aspect of your life. They are believed to be the source of intuition and inspiration, and there to support you during times of need. (Quoted from The Body Guide website)

Three main claims are made in this passage. (1) We all have at least two guardian angels as well as countless other angels and spirit guides that we can “connect with.” (2) These supernatural beings want to help us in every aspect of our lives. (3) This help can be therapeutic: it can reduce or eliminate psychological problems and even provide “intuition and inspiration.”

But what is the evidence that these supernatural beings actually exist and that, if they do, that they want to help us? Susan Stevenson, a therapist, has claimed that the evidence is all around us, but that we need to pay close attention to see it:

My life seems to be teeming with angelic connections, and the momentum is building. Have you noticed this in your own life? Angelic reminders that they are with us- ‘whispers’ in our ear, ‘taps’ on the shoulder, brushes of air across your skin or changes in air pressure, ‘flutters’ from deep inside, glints of light and color- all these gentle hints to pay closer attention to their presence. Think back- have you been paying attention, listening, responding? (Carroll, 2012)

When one makes a claim, one is stating that something is a fact. In other words, a claim is a statement that is thought by at least one person to be true; but of course, it may turn out to be false. Claims often involve interpretations of experiences. For example, you may interpret a “brush of air” across your skin as an angel who has just passed by, or you may interpret it as a breeze that has wafted through the room from an open window. A glint of light may be the sign that an angel is nearby, or it may be the sign that sunlight just reflected off a passing car. In other words, two people may interpret the same experience in different ways. In deciding which interpretation is more likely to reflect reality, we need to evaluate the different interpretations. We do this by examining relevant evidence.

In your everyday life, you probably often have heard claims made about psychological problems and psychological therapies; and you probably think that you already know quite a bit about psychology. In order to get a sense of what you might know, please take the following brief quiz.

Which of the following claims are true?

  • dream images are known to have particular meanings that involve unconscious desires and conflicts
  • eating sugar causes children with attention-deficit hyperactivity disorder to become even more hyperactive
  • a person who commits suicide must have been clinically depressed
  • a 40-year-old man who has sex with a 15-year-old girl would be diagnosed with pedophilia
  • there are more admissions to mental hospitals during full moons than at other times
  • unconscious memories of traumatic events can be remembered in detail with hypnosis
  • a person who exhibits two or more personalities is diagnosed with schizophrenia
  • low self-esteem is known to cause most self-destructive behaviors
  • most mental disorders can be cured by remembering and mentally reliving distressing past experiences

You may be surprised to learn that none of these claims is known to be true. In fact, all but a few are known to be false, and the remaining ones are controversial at best. In order to avoid basing important decisions on false claims, clinicians (professionals who study and treat psychological problems), or those who aspire to be clinicians (perhaps you), need to learn to think critically about claims made about psychological problems and their treatments. Of fundamental importance to this goal is the development of skeptical and empirical attitudes regarding claims.


In some religions, a shaman is said to be a mediator between the visible natural world and an invisible supernatural world. The shaman claims to be able to journey to the supernatural world in order to help heal the ill, foretell the future, and control natural events. Some mental-health workers use shamanic journeying to help those suffering from psychological problems. Sharon Van Raalte (1998) gave an example of her shamanic work with a client:

Through image and symbol, the shamanic journeys revealed levels of knowing that were often beyond what could be perceived or expressed by the clients or the psychiatrist. For example, Luke was dying from a brain tumor. An early journey suggested that I teach his wife, Suzanne, to work with him. Learning to journey to find her power animal proved to be helpful when it came time for her husband to die. At another point, I was journeying on a question for myself, when the focus abruptly changed. I found myself sitting with [Luke and Suzanne] in a boat that began moving to a farther shore. On the other side, Luke got out of the boat and went toward a group of people waiting to greet him. I had the sensation that the pain they had caused him in his life was washed away as they surrounded him with love. The classic shamanic experience (known as conducting the souls of the dead) had come unbidden. (p. 164)

In other words, Van Raalte claimed that she and Suzanne had accompanied Luke to the “other side” as he was dying, and then saw him being reunited with others who had died before him. An apparent confirmation of this interpretation came later:

Only after I had reported this journey to the psychiatrist did I learn what had literally happened. In his delirium as he was dying, Luke had called out the name of his dead sister, with whom he had had a painful relationship. Drawing from the experience of her single [shamanic] journey, Suzanne knew what he was seeing and urged him to run to his sister. (p. 164)

In these passages, Van Raalte is making a number of claims: (1) She is able to journey to a spirit world. (2) She saw Luke being reunited with his dead sister. (c) Shamanic journeying is an effective treatment for at least some psychological problems. When hearing claims such as these, scientific psychologists are trained to be skeptical–to doubt the claim unless it is supported by adequate evidence. These particular claims may be true, but we need to see good supporting evidence before we accept them. As an ideal, we should be skeptical of any claim that may have an important impact in our lives, even a claim that seems on its surface to be convincing. It probably is impossible to reach this ideal, but we should strive to develop our skepticism as much as we can in order to improve our decision-making and problem-solving abilities. And it should be incumbent upon people who work in mental-health fields, especially those offering therapeutic services, to develop their skepticism as fully as they can since their beliefs and actions have important consequences for those with psychological problems.

When confronted with a claim, a skeptical thinker needs to do two things. First, because a claim is based on a particular interpretation of an experience, a skeptical thinker always needs to consider other possible interpretations of that experience. For example, a shamanic therapist who claims to be journeying to a supernatural realm may actually be doing so. On the other hand, she may only be vividly imagining that she is doing so, or she may be experiencing hallucinations. By considering other interpretations, a skeptical thinker is less likely to automatically accept the claimant’s interpretation and more likely to examine carefully the various alternatives.

Second, a skeptical thinker needs to determine if there is any evidence that contradicts the claim. For example, Van Raalte stated that she saw Luke being greeted by a “group of people,” all of whom had caused him pain during his life. However, the psychiatrist with whom Van Raalte worked stated that, at the time of his death, Luke mentioned only his sister’s name and was urged to run to his sister. This evidence seems to contradict the claim made by Van Raalte that she had seen Luke with a group of people at the time of his death. Without further clarification and more evidence, it is difficult to know whether to accept or reject her claim.


Evidence consists of observations that allow us to evaluate whether a claim is likely to be true or false. Let’s consider a very simple claim that probably everyone believes is true: “The sun will rise tomorrow morning.” For me, this claim is based on the following evidence:

  • As far back as I can remember, I have seen the sun rise each and every morning of my life.
  • No mention has ever been made in any historical document that the sun has ever failed to rise. It seems likely that something as significant as the sun not rising would have been recorded and reported.
  • Scientists and other experts tell us that the sun rises each morning because the Earth rotates on its axis, and that it should continue to do so for billions of years.

From all this evidence, it seems reasonable to infer that the sun will rise tomorrow morning. If someone claimed that he knew that the sun was not going to rise tomorrow morning, you would immediately ask him why he believed this claim (this is equivalent to asking him for his evidence). If he stated that he dreamed that this would happen and that his dreams often come true, most of us would be skeptical: the supporting evidence (his dream) does not seem adequate to accept his claim.

What is the best kind of evidence for supporting a claim? Should we rely upon what an expert tells us? Should we accept a person’s intuition? Are the statements of a channeled spirit guide acceptable evidence for a claim? Regarding the nature of evidence, scientific psychologists are trained to be empirical–to make direct observations of events in the natural world that are relevant to evaluating the claim. Empiricists do not consider statements made by authorities, armchair speculations, dream interpretations, or messages supposedly obtained from supernatural beings, to be adequate evidence for a claim. Instead, empiricists must see for themselves whether a claim is likely to be true or false. For example, in testing the claim that shamanic journeying is an effective treatment for at least some psychological problems, an empiricist would want to measure directly the severity of clients’ symptoms both before and after being told what was discovered about them during a shamanic journey. If their symptoms improved relative to those of a second group of clients who were told things about themselves that were not discovered during a shamanic journey, then this would be evidence that shamanic journeying (for whatever reason) is an effective therapeutic technique.

You may contact me at


Carroll, R. T. (2012). Angel Therapy. The Skeptic’s Dictionary. Retrieved November 9, 2012, from

The Body Guide. (2012). What is Angel Therapy? November 9, 2012, from

Van Raalte, S. (1998). Direct knowing. In W. Braud & R. Anderson (Eds.), Transpersonal research methods for the social sciences (pp. 163-166). Thousand Oaks, CA: Sage Publications.

Classical Conditioning of Drug Tolerance

In the case of psychoactive drugs, drug tolerance is a condition in which repeated use of a drug leads to reductions in its psychological effects, thereby requiring progressively larger doses in order for similar effects to occur. Many factors affect the development of drug tolerance, and learning is one of them. A classic study by Siegel, Hinson, Krank, and McCully (1982) was one of the first to demonstrate this. The researchers found that rats can be classically conditioned to develop tolerance to heroin.

Siegel, et al. (1982) wanted to understand why some addicts died after taking a dose of the drug that they had taken many times before. The official cause of death in these cases usually was said to be a “drug overdose” because the dose was high enough to kill most people not addicted to the drug. These addicts, however, had developed a high level of tolerance to the lethal effects of the drug, so ascribing their deaths to an overdose couldn’t be correct.

The researchers hypothesized that the addicts’ tolerance was due, in part, to classical conditioning. That is, the situation is which they typically took the drug had become a conditioned stimulus (CS) that elicited a conditioned response (CR) involving a change in biological processes that prepared their bodies to counteract the lethal effects of the drug. This explanation may be easier to understand by looking at a case study described in Siegel (2001). The study was of a man suffering from severe pain because of pancreatic cancer. In order to reduce the pain, he received four injections of morphine per day. Over time, the dosage had been increased to a high level because of the tolerance he had developed to morphine’s pain-reducing effects.

The patient stayed in his bedroom (which was dimly lit and contained apparatus necessary for his care), and received injections in this environment. For some reason, after staying in this bedroom for about a month, the patient left his bed and went to the living room (which was brightly lit and different in many ways from the bedroom/sickroom). He was in considerable pain in the living room, and, as it was time for his next scheduled morphine administration, he was administered his usual dose of the drug. The patient quickly displayed signs of opiate overdose (constricted pupils, shallow breathing), and died a few hours later. (p. 510).

The explanation given by Siegel and his colleagues involved the classical conditioning of tolerance to the lethal effects of morphine. The explanation is outlined in Figure 1:

Figure 1. Classical conditioning of tolerance to morphine

As shown in Figure 1, the stimuli in the bedroom comprise a CS that produces a CR consisting of physiological changes that counteract the effects of morphine (i.e., are opposite to the physiological changes caused by morphine). This CR is part of the change in biological processes that causes tolerance to drugs, thereby requiring people to use higher doses in order to experience the same psychological and physical effects of drugs. The CR develops because the CS is followed by the drug, which can be thought of as a UCS — in this case, the injection of morphine. The UCS causes a UCR consisting of physiological changes associated with the drug. By counteracting the physiological effects of the drug, the CR reduces its lethality (i.e., the probability that the drug will result in death).

When an addict takes a high dose of a drug in a different environment — in the case of the man taking morphine, the different environment was his living room — the CR does not occur and, therefore, the drug now has its full effect on the body, which makes it more likely that the person will die.

Siegel, et al. (1982) designed an experimental study to test the classical-conditioning theory of drug tolerance. They gave rats injections of heroin every other day for 30 days (a total of 15 injections). They increased the dose gradually over time so that the rats eventually could tolerate relatively high doses. On the days the rats didn’t get heroin,they were injected with a sugar solution (see Table 1 below).

Siegel, et al. (1982) tested two heroin-injected groups:

  1. Group 1. These rats received their heroin injections in Room 1, which was the room that housed all the rats (the “colony room”). They received their sugar injections in Room 2, which was a room that differed from Room 1 in two ways: (a) no rats were housed there;(b) a machine generated constant “white noise.”
  2. Group 2. These rats received their heroin injections in Room 2, and their sugar injections in Room 1.

The researchers also included a placebo-control group (Group 3) that received injections of the sugar solution in both Room 1 and Room 2 on the same schedule as the other rats. Table 1 shows the schedule for each group on the first four days of the 30-day experiment.

Table 1. The schedule of heroin and sugar injections over the first four days of the experiment

At the end of the 30-day period, rats in all three groups were given a very large dose of heroin (almost twice as much as those in Group 1 and Group 2 had received before). Figure 2 shows the results.

Figure 2. Mortality in heroin-tolerant and control rats after receiving a very high dose of heroin

The label Different in Figure 2 refers rats in Groups 1 and 2 that received the very large dose of heroin in the room in which they had been injected with the sugar solution during the first part of the experiment. The label Same refers to in Groups 1 and 2 that received the very large dose of heroin in the room in which they had been injected with heroin during the first part of the experiment. The label Control refers to rats that were injected only with the sugar solution during the first part of the experiment.

As can be seen in Figure 2, almost all the rats in the Control Group died after being injected with the large dose of heroin because they had no tolerance for its lethal effects. About 64% of rats in the Different condition died, whereas only 32% of the rats in the Same condition died. Siegel, et al. (1982) interpreted these results as supporting their theory: the environmental stimuli in which drug addicts usually take the drug serve as a CS that produces a CR that increases tolerance for the drug’s effects.

The researchers concluded that the sensory stimuli from the room in which the rats were injected with heroin made up a CS. The rats developed a CR to that room–a CR consisting of changes in biological changes that counteracted the effects of the drug they were about to receive. When they received the final injection in a different room, the CR didn’t occur, which increased their chances of dying from the overdose.

Another point I want to make is that, although in most experimental studies of classical conditioning, the CR and UCR are the same or very similar, the responses also may be very different. In fact, in the study by Siegel, et al. (1982), the CR and UCR were completely opposite from each other.


Siegel, S. (2001). Pavlovian conditioning and drug overdose: When tolerance fails. Addiction Research & Theory, 9, 503-513. doi: 10.3109/16066350109141767
Retrieved October 18, 2011, from

Siegel, S., Hinson, R. E., Krank, M. D., McCully, J. (1982). Heroin “overdose” death: Contribution of drug-associated environmental cues. Science, 23, 436-437.  doi: 10.1126/science.7200260

Experience and the Other-Race Effect

The other-race effect is the reduced ability to recognize strangers’ faces of another race relative to strangers’ faces of one’s own race. Many studies have demonstrated the reliability of this effect  (see Meissner & Brigham, 2001, for a review).  And a number of studies suggest that experience beginning in infancy is important for the development of the other-race effect (e.g., Bar-Haim, Ziv, Lamy, & Hodes, 2006). In short, beginning in the first year of life, children gradually become better at recognizing faces of those belonging to the race (or races) they most frequently interact with and gradually lose the ability to recognize faces of those belonging to other races.

Some researchers have tested the claim that the other-race effect depends on early experience by exposing young children to other-race faces. For example, Heron-Delaney, Anzures, Herbert, et al. (2011) noted that Caucasian children typically begin to develop the other-race effect between the ages of 6 and 9 months. Thus, they had parents periodically show their infants pictures of Chinese faces during this time. The researchers found that, even though the total amount of exposure to Chinese faces over the 3-month period was only about 70 minutes, these infants were able to recognize Chine faces as well as Caucasian faces, whereas infants not exposed to Chinese faces during that time showed the other-race-effect.

Some research suggests that early-childhood experiences may not have permanent effects on other-race facial recognition. One study found that adults of Korean origin who, between the ages of 3 to 9 years, moved to France, Switzerland, or Belgium after being adopted into Caucasian families were better at recognizing Caucasian faces than Asian faces (Sangrigoli, Pallier, Argenti, Ventureyra, & de Schonen, 2005). The researchers concluded that, because these individuals experienced primarily Caucasian faces in later childhood, not only was the other-race effect eliminated for them, it was reversed.

Another study, however, found that Chinese and Vietnamese children (6 to 14 years of age) adopted into Caucasian families in Belgium between the ages of 2 and 26 months recognized Caucasian and Asian faces equally well (De Heering, De Liedekerke, Deboni, & Rossion, 2009). In other words, experience with Caucasian faces eliminated the other-race effect but it did not cause a reversal of the effect.


Bar-Haim, Y., Ziv, T., Lamy, D., & Hodes, R. M. (2006). Nature and nurture in own-race face processing. Psychological Science, 17, 159-163. doi:10.1111/j.1467-9280.2006.01679.x

De Heering, A., De Liedekerke, C., Deboni, M., Rossion, B. (2009). The role of experience during childhood in shaping the other-race effect. Developmental Science, 13, 181–187. doi:10.1111/j.1467-7687.2009.00876.x

Heron-Delaney, M., Anzures, G., Herbert, J. S., Quinn, P. C., Slater, A. M., Tanaka, J. W., et al. (2011). Perceptual training prevents the emergence of the other race effect during infancy. PLoS ONE 6(5), e19858. doi:10.1371/journal.pone.0019858

Meissner, C.A., & Brigham, J.C. (2001). Thirty years of investigating the own-race bias in memory for faces: A meta-analytic review. Psychology, Public Policy and Law, 7, 3–35.

Sangrigoli, S., Pallier, C., Argenti, A.-M., Ventureyra, V. A. G., & de Schonen, S. (2005). Reversibility of the other-race effect in face recognition during childhood. Psychological Science, 16, 440-444. doi:10.1111/j.0956-7976.2005.01554.x