1.3 The Research Process: How We Find Things Out

All science ultimately begins with observations. Although these observations may begin as impressions or anecdotes, to be the first step in science they must move beyond that—they must lead to systematic observations. Scientists want to know the facts, as free as possible from interpretations of their implications. Facts are established by collecting data Careful descriptions or numerical measurements of a phenomenon., which are careful, objective descriptions or numerical measurements of a phenomenon. Data are “objective” because researchers try to eliminate distortions based on their personal beliefs, emotions, or interpretations; while data are collected, their interpretation must be set aside and saved for later. In addition, if data are systematically collected (for example, by being sure to collect data from different parts of the country when comparing changes over time in national attitudes about free speech), the investigator or somebody else should be able to collect comparable data by repeating the data collection procedure; when a study is repeated so that the data can be compared to those collected originally, the study is called a replication Repeating the method of a study and collecting comparable data as were found in the original study. .

Scientists often prefer quantitative data (numerical measurements), such as how many men versus women are diagnosed with depression, in part because such data are relatively easy to summarize, analyze, and report. However, many scientists—especially in the relatively early phases of an investigation—rely on systematic qualitative observations (descriptions), which may simply document that a certain event occurs or may describe specific characteristics of the event, such as which toys children play with while at their preschool.

Objectively and systematically observing events is just the first step in the scientific method because events typically can be interpreted in more than one way. For example, a smile could mean happiness, agreement, politeness, or amusement. To sort out the proper interpretation of data, researchers need to follow up observations with the additional steps described in the sections to follow.

Most of science is about answering questions that are raised by observations. For example, a scientist might note that even as a child, Tiger Woods’s abilities made him stand out from his peers. Why? Was there something different about his brain or the way he was being raised? Questions can come from many sources: For example, in addition to direct observation (which occurs when the scientist notices something that piques his or her interest), they can arise from a “secondhand” observation, reported by someone else (and often published in a journal article), or from reflection (which may even occur when the scientist is puzzling about why a certain event does not occur—such as toddlers learning to read at exactly the same time that they learn to talk). 

To try to answer a question, scientists make an educated guess about the answer, called a hypothesis A tentative idea that might explain a set of observations.. A hypothesis is a tentative idea that might explain a set of observations. Hypotheses typically specify a relationship between two or more variables (such as between the amount of time a child sleeps and his or her activity level during the day); hypotheses often propose that one variable (or set of variables) cause another. By the term variable An aspect of a situation that can vary, or change; specifically, a characteristic of a substance, quantity, or entity that is measurable., researchers mean an aspect of a situation that can vary; more precisely, a variable is a characteristic of a substance, quantity, or entity that is measurable. Examples of variables that could be of interest to psychologists include the amount of time someone needs to choose a particular reward, whether or not a person was raised in a single-family household, or what part of the country a person lives in.

For example, say you decide to take golf lessons. On the golf course, you notice a particularly good player and ask her for tips. She confesses that when she’s off the course, she often practices her swings mentally, visualizing herself whacking the ball straight down the fairway or out of a sand trap; she assures you that the more time she has spent doing such mental practice, the better her game has become. You can translate this idea into a hypothesis, speculating that there’s a connection between two variables—the time people spend visualizing swings and their golf score. This idea appeals to you, in part because it means that you can practice at night or whenever you are bored.

But at this point all you have is a hypothesis. Before you go to the trouble of imagining yourself swinging a club, over and over and over, you want to test the hypothesis to find out whether it’s correct.

To test the hypothesis, you need to collect new data. But before you can do that, you must make the key concepts of the hypothesis specific enough to test. An operational definition A definition of a concept that specifies how it is measured or manipulated.  defines a concept by indicating how it is measured or manipulated. For example, you could define “improvement in playing golf” in terms of the number of putts sunk or the distance the ball is driven.

In fact, much research has been conducted on mental practice, and many studies have documented that it actually does improve performance, as it is measured—operationally defined—in various ways. In a typical study, people are divided into two groups, and the performance of both groups is assessed. One group then uses mental practice for a specified period, while the other is asked to take an equal amount of time to visualize familiar golf courses (but does not imagine practicing). Then the golf performance of both groups is assessed again. Usually, the people who engage in mental practice show greater improvement (Driskell et al., 1994; White & Hardy, 1995; Yagueez et al., 1998).

To test hypotheses, scientists make two types of observations: (1) those that directly address the object of study, such as how many times in an hour a mother speaks to her infant or how far a golf ball traveled after being hit, and (2) those that indirectly address the object of study, which might be the nature of thoughts, motivations, or emotions. Many sorts of behavior can be used to draw inferences about mental events. For example, when people smile without really meaning it (as they often do when posing for photographs), the muscles they use are not the same ones they use when they produce a sincere smile (Ekman, 1985). By studying what’s observable (the particular muscles being used), researchers can learn about the unobservable (the mental state of the smiler).

Now we consider the aspect of the scientific method that uses “such evidence to formulate a theory.” A theory Concepts or principles that explain a set of research findings.  consists of concepts or principles that explain a set of research findings. Whereas a hypothesis focuses on possible relationships among variables, a theory focuses on the underlying reasons why certain relationships may exist in data. In our example, the idea that mental practice leads to better performance is a hypothesis, not a theory. A theory might explain that mental practice works because the brain structures that are used to perform an action (playing golf) are also engaged when you mentally practice that action; thus, after practicing it mentally, the relevant brain mechanisms operate more efficiently when you later perform the actual behavior.

Finally, what do we mean by “testing the theory”? Researchers evaluate a theory by testing its predictions. It’s not enough that a theory can, after the fact, explain previous observations; to be worth its salt, it has to be able to predict new ones. As illustrated in Figure 1.1, theories produce predictions A hypothesis that follows from a theory, which should be confirmed if the theory is correct., which are new hypotheses that should be confirmed if the theory is correct. For example, the theory of mental practice predicts that the parts of the brain used to produce a behavior—in our example, swinging a golf club—are activated by merely imagining the behavior. This prediction has been tested by using brain-scanning techniques to observe what happens in the brain when an individual imagines making certain movements. And, in fact, parts of the brain that are used to control actual movements are activated when the movements are only imagined (Jeannerod, 1994, 1995; Kosslyn et al., 1998; Parsons, 1987, 1994; Parsons & Fox, 1998).

Each time a theory makes a correct prediction, the theory is supported, and each time it fails to make a correct prediction, the theory is weakened. If enough of its predictions fail, the theory must be rejected and the data explained in some other way. The theorist is no doubt disappointed when his or her theory is shot down, but this is not a bad thing for science. In fact, a good theory is falsifiable; that is, it makes predictions it cannot “squirm out of.” A falsifiable theory can be rejected if the predictions are not confirmed.

Now that you understand all the steps of the scientific method, you could use it to investigate an enormous range of questions. For instance, say you’ve heard that putting a crystal under your bed will improve your athletic ability. Should you believe this? The scientific method is just what you need to decide whether to take this claim seriously! Let’s walk through the steps of the scientific method, applied to the role that crystals under people’s beds might play in improving their athletic performance:

Any individual researcher need not go through all of the steps of the scientific method in order to do science. In fact, he or she can focus solely on the first step, observing events systematically; some research is devoted simply to describing “things as they are.” Although—as we’ve just seen—there is more to the scientific method than observation, it’s no accident that “observing events” is a key part of the scientific method: Theorizing without facts is a little like cooking without ingredients. Researchers carry out such observations in several ways: naturalistic observations, case studies, and surveys.

Researchers can use another method to study the relationships among variables—a method that relies on the idea of correlation.  A correlation is a relationship in which two variables are measured for each person, group, or entity, and the variations in measurements of one variable are compared to the variations in measurements of the other variable. For example, height is correlated with weight: Taller people tend to be heavier than smaller people. A correlation coefficient is a number between -1 and +1 that indicates how closely related two measured variables are; the higher the coefficient (in either the positive or negative direction), the better the value of one measurement can predict the value of the other. The correlation coefficient A number that ranges from -1.0 to 1.0 that indicates how closely interrelated two sets of measured variables are; the higher the coefficient (in either the positive or negative direction), the better the value of one measurement can predict the value of the other. Also simply called a correlation.  is often simply referred to as a correlation because the coefficient is the numerical summary of the relationship between two variables.

Figure 1.2 illustrates three predicted correlations between variables Graph 1 shows a positive correlation, a relationship in which increases in one variable (height) are accompanied by increases in another (weight); a positive correlation is indicated by a correlation value that falls between 0 and 1.0. Graph 2 shows a negative correlation, a relationship in which increases in one variable (age) are accompanied by decreases in another (health); a negative correlation is indicated by a correlation value that is between -1.0 and 0. The closer the correlation is to 1.0 or to -1.0, the stronger the relationship; visually, the more tightly the data points cluster around a slanted line, the higher the correlation.  Finally, Graph 3 shows a zero correlation, which indicates no relationship between the two variables (height and aggressiveness); they do not vary together.

Correlations always compare two sets of measurements at a time. Sometimes several pairs of variables are compared, but each pair of measurements requires a separate correlation. Thus, correlational research has two steps:

The main advantage of correlational research is that it allows researchers to compare variables that cannot be manipulated directly (which is what happens in experiments, which we discuss in the following section). The main disadvantage is that correlations indicate only that the values of two variables tend to vary together, not that values of one cause the values of the other. For example, evidence suggests a small correlation between poor eyesight and intelligence (Belkin & Rosner, 1987; Miller, 1992; Williams et al., 1988), but poor eyesight doesn’t cause someone to be smarter! Remember: Correlation does not imply causation.

Much psychological research relies on conducting experiments, controlled situations in which the investigator observes the effects of altering variables. Experiments provide the strongest way to test a hypothesis because they can provide evidence that changes in one variable cause changes in another.

Like an experimental design, a quasi-experimental design includes independent and dependent variables and assesses the effects of different values of the independent variable on the values of the dependent variable. However, unlike true experiments, in quasi-experiments, participants are not randomly assigned to conditions and the conditions typically are selected from naturally occurring situations—that is, they are not created by the investigator’s manipulating the independent variable. For example, if you wanted to study the effects of having been in an earthquake on how well people sleep at night, you would need to select and compare participants who had or had not been in an earthquake. You could not randomly assign participants to the different groups.

Quasi-experimental designs are used because it is not always possible or desirable to assign people to different groups randomly. For instance, let’s say that you want to discover whether the effects of mental practice are different for people of different ages, and so you decide to test four groups of people: teenagers, college students, middle-aged people, and elderly people. Obviously, you cannot assign people to the different age groups randomly. Rather, you select groups from what nature has provided. When composing the groups, you should control for as many variables—such as people’s health and education levels—as you can in order to make the groups as similar as possible. By “control for,” we mean that you should ensure that values of variables (such as education level or health) don’t differ systematically in the different groups. (In fact, in the example we just gave of dividing participants into four age groups, a variety of variables are not controlled for and therefore create confounds. For instance, the middle-aged and elderly participants will be more likely to have finished college than the teens and college students.) Similarly, if you want to track changes over time, it is not possible to assign people randomly to the groups because you are taking measurements only from people you have measured before.

Unfortunately, because groups can never be perfectly equated on all characteristics, you can never be certain exactly which of the differences among groups produce the observed results. And thus, you cannot draw conclusions from quasi-experiments with as much confidence as you can from genuine experiments.

Table 1.3 summarizes the key research methods used in psychology, along with their relative strengths and weaknesses. When thinking about the various methods, keep in mind that even though experiments are the most rigorous, they cannot always be performed—particularly  if you are interested in studying large groups or if it is difficult (or unethical) to manipulate the independent variables. In addition, we must stress that different combinations of the methods are often used. For example, observational methods are used as part of correlational research, and observational, correlational, or experimental research can be conducted with single individuals (case studies).

Meta-analysis A statistical technique that allows researchers to combine results from different studies on the same topic in order to discover whether there is a relationship among variables. is a statistical technique that allows researchers to combine results from different studies on the same topic in order to discover whether there is a relationship among variables (Cooper, 2010). A meta-analysis is particularly useful when the results from many studies are not entirely consistent, with some showing an effect and some not. Thus, the technique is not a way to collect new data by observing phenomena—it is a way to analyze data that have already been collected.

A meta-analysis can sometimes reveal results that are not evident in all the individual studies that go into them. This is possible because studies almost always involve observing or testing a sample from a population; the sample A group that is drawn from a larger population and that is measured or observed. is the group that is measured or observed, and the population The entire set of relevant people or animals. is the entire set of relevant people or animals (perhaps defined in terms of age or gender). The crucial fact is that there is always variation in the population; just as people vary in height and weight, they also vary in their behavioral tendencies, cognitive abilities, emotional reactivity, and personality characteristics. Thus, samples drawn from a population will not be the same in every respect; and if samples drawn from different populations are relatively small, the luck of the draw could obscure an overall difference that actually exists between the populations. For example, if you stopped the first two males and first two females you saw on the street and measured their heights, the females might actually be taller than the males. In this example, the two populations would be all males and all females, and the samples would be the two people drawn from each population. And these samples would—by the luck of the draw—not accurately reflect the general characteristics of each of the populations.

The problem of variation in samples is particularly severe when the difference of interest—the effect—is not great and the sample sizes are small. In our example of male and female heights, if men averaged 8 feet tall and women 4 feet tall, small samples would not be a problem; you would quickly figure out the usual height difference between men and women. But if men averaged 5 feet 10 inches and women averaged 5 feet 9 inches (and the heights in both populations varied by an average of plus-or-minus 6 inches), you would need to measure many men and women before you were assured of finding the difference. By combining the samples from many studies, a meta-analysis allows you to detect even subtle differences or relations among variables (Rosenthal, 1991).

Some data should be taken more seriously than other data because data differ in their quality. One way to evaluate the quality of the data is in terms of reliability. Reliability Consistency; data are reliable if the same values are obtained when the measurements are repeated. means consistency. A reliable car is one you can count on to behave consistently, starting even on cold mornings. Data are reliable if you obtain the same findings each time the variable is measured. For example, an IQ score is reliable if you get the same score each time you take the test.

Something is said to be valid if it is what it claims to be; a valid driver’s license, for example, is one that confers the right to drive because it was, in fact, issued by the state and has not expired. In research, validityA method does in fact measure what it is supposed to measure.  means that a method does in fact measure what it is supposed to measure. A study can be reliable without being valid, but to be valid it must be reliable.

The outcome of a study can also be affected by what the participants or the researchers believe or expect to happen or by their habitual ways of responding to or conceiving of situations. Bias When conscious or unconscious beliefs, expectations, or habits alter how participants in a study respond or affect how a researcher sets up or conducts a study, thereby influencing its outcome. occurs when (conscious or unconscious) beliefs, expectations, or habits alter how participants in a study respond or affect how a researcher sets up or conducts a study, thereby influencing its outcome. Bias can take many forms. For example, response bias A tendency to respond in a particular way regardless of respondents’ actual knowledge or beliefs. occurs when research participants tend to respond in a particular way regardless of their actual knowledge or beliefs. For instance, many people tend to say “yes” more than “no,” particularly in some Asian cultures (such as that of Japan). This sort of bias toward responding in “acceptable” ways is a devilish problem for survey research. For example, even though residents of a community in Japan might not want a local, publicly funded golf course, the politician pushing the idea could take advantage of this cultural response bias by having a pollster ask residents “Do you support using public funds to build golf courses, which will increase recreational options?” Respondents would be more likely to say “yes” simply because they are more likely to answer “yes” to any question.

Another form of bias is sampling bias A bias that occurs when the participants are not chosen at random but instead are chosen so that one attribute is over- or underrepresented., one form of which occurs when researchers do not choose the participants at random but instead select them so that an attribute is over- or underrepresented. For example, say you wanted to know the average heights of male and females, and you went to shopping malls to measure people. You would fall victim to sampling bias if you measured males outside a toy store (because you would likely  be measuring little boys) but measured females outside a fashion outlet for tall people (because you would likely be measuring especially tall women). (As you’ve probably noticed, such bias leads to a confound—in this case between where the samples were chosen and the type of sample.)

Sampling bias isn’t just something that sometimes spoils otherwise good research studies. Have you read about the U.S. presidential election of 2016? Based on surveys, most pundits incorrectly predicted that Hillary Clinton would be the winner. What led them astray? In part, sampling bias. The news organizations that conducted the surveys did not ask a representative sample of the relevant population—namely people in specific districts. In the United States, presidents are elected not by popular vote, but by the electoral college—and hence the winner is determined by who wins enough delegates. To be an accurate predictor, the surveys needed to sample among all likely voters in the various key districts, determining who would win each district and therefore each key state. By focusing on the overall popular vote, or even overall state votes, the pundits were misled about who would be the likely winner.

Another set of problems that can plague psychological research goes by the name of experimenter expectancy effects, even though these effects can occur in all types of psychological investigations (not just experiments). Experimenter expectancy effects Effects that occur when an investigator’s expectations lead him or her (consciously or unconsciously) to treat participants in a way that encourages them to produce the expected results.  occur when an investigator’s expectations lead him or her (consciously or unconsciously) to treat  participants in a way that encourages them to produce the expected results. Here’s a famous example, with an unusual participant: Clever Hans was a horse that lived in Germany in the early 1890s. He apparently could add (Rosenthal, 1976). When a questioner (one of several) called out two numbers to add (for example, “6 plus 4”), Hans would tap out the correct answer with his hoof. Was Hans a genius horse? Was he psychic? No. Despite appearances, Hans wasn’t really adding. He seemed to be able to add, but he responded only when his questioner stood in his line of sight and knew the answer. It turned out that each questioner, who expected Hans to begin tapping, always looked at Hans’s hoof right after asking the question—thereby cuing Hans to start tapping. When Hans had tapped out the right number, the questioner always looked up—cuing Hans to stop tapping. Although, in fact, Hans could not add, he was a pretty bright horse: He was not trained to do this; he “figured it out” on his own.

The story of Hans nicely illustrates the power of experimenter expectancy effects. The cues offered by Hans’s questioners were completely unintentional; they had no wish to mislead (and, in fact, some of them were probably doubters). But such cues led Hans to respond in specific way. In psychological research, the cues can be as subtle as when a researcher makes eye contact with or smiles at a participant. For instance, if you were a researcher interviewing participants about their attitudes about alcohol, your own views could affect how you behave, which in turn affects what the participants say: If you smile whenever they mention an attitude that is similar to yours and frown when they mention an attitude that is different from yours. In turn, the participants may try to please you by saying what they think (perhaps unconsciously) you want to hear.

At least for experiments, you can guarantee that experimenter expectancy effects won’t occur by using a particular type of experimental arrangement called a double-blind design. In a double-blind design A study in which the participant is “blind” to (unaware of) the predictions and therefore cannot consciously or unconsciously produce the predicted results; the investigator is also “blind” to the group to which the participant has been assigned or to the condition that the participant is receiving and therefore cannot produce the predicted results., the participant is “blind” to (unaware of) the predictions of the study—and hence unable consciously or unconsciously to serve up the expected result— and the experimenter is also “blind” to the group to which the participant has been assigned to or the particular condition the participant is receiving, and thus is unable to induce the expected results. Clever Hans failed when the questioner did not know the answer to the question. In an experiment, participants are no different than Clever Hans; they also can’t respond to cues that aren’t provided to them.

The field of psychology—rooted in science—must be distinguished from pseudopsychology Theories or statements that at first glance look like psychology but are in fact superstition or unsupported opinion, not based in science. : theories or statements that at first glance look like psychology but in fact are superstition or unsupported opinions. For example, astrology—along with palm reading and tea-leaf reading and all their relatives—is not a branch of psychology but is pseudopsychology. Pseudopsychology is not psychology at all. It may look and sound like psychology, but it is not science.

Unfortunately, some people are so interested in “knowing more about themselves” or in getting help for their problems that they turn to pseudopsychology. So popular is astrology, for example, that you can get your daily horoscope delivered daily to you via the Internet. And advice found in some (but not all) self-help books falls into the category of pseudoscience. For instance, at one point in history, some self-help gurus told people that screaming was good because it would “let it all out” (the “it” being anger and hurt)—but there was absolutely no evidence that such screams did any more than annoy the neighbors. In fact, research suggests that venting anger in this way sometimes may end up increasing your angry feelings, not diminishing them (Bushman, 2002; Lohr et al., 2007). (It’s a good idea to check whether the advice dispensed in a self-help book is supported by research before you shell out your hard-earned cash and possibly waste your time reading  it.)

It’s not always obvious what’s psychology and what’s pseudopsychology. One key to distinguishing the two is the method; we’ve already seen what goes into the scientific method, which characterizes all science—regardless of the topic. For example, is extrasensory perception (ESP) pseudopsychology? ESP refers to a collection of mental abilities that do not rely on the ordinary senses or abilities. Telepathy, for instance, is the ability to read minds. This sounds not only wonderful but magical. No wonder people are fascinated by the possibility that they, too, may have latent, untapped, extraordinary abilities. But the mere fact that the topic may seem “on the fringe” does not mean that it is pseudopsychology. Similarly, the mere fact that many experiments on ESP have come up empty, failing to produce solid evidence for the abilities, does not mean that the experiments themselves are bad or “unscientific.” One can conduct a perfectly good experiment—taking care to guard against confounds, bias, and expectancy effects, for instance—even on ESP (Moulton & Kosslyn, 2008).

For example, let’s say that you want to study telepathy.  You might arrange to test pairs of participants, with one member of each pair acting as “sender” and the other as “receiver.” Both the sender and receiver would each look at different sets of the same four playing cards (say an ace, a two, a three, and a four). The sender would focus on one of the cards (say, the ace) and would “send” the receiver a mental image of the chosen card. The receiver’s job would be to try to receive this image and report which card the sender is focusing on. By chance alone, with only four cards to choose from, the receiver would be right about 25% of the time. So the question is, can the receiver do better than mere guesswork? In this study, you would measure the percentage of times the receiver picks the right card and compare this to what you would expect from guessing alone.

But wait! What if the sender, like the questioners of Clever Hans, provided visible cues (accidentally or on purpose) that have nothing to do with ESP, perhaps smiling when “sending” an ace, and grimacing when “sending” a two. A better experiment would have sender and receiver in different rooms, thus controlling for such possible confounds. Furthermore, what if people have an unconscious bias to prefer red over black cards, which leads both the sender and the receiver to select red cards more often than would occur by chance? This difficulty can be countered by including a control condition, in which a receiver guesses cards when the sender is not actually sending. Such guesses will reveal response biases (such as a preference for red cards), which exist independently of messages sent via ESP.

Just following the scientific method will not eliminate the possibility of pseudopsychology. For instance, if solid studies conclusively show that there is nothing to ESP, then people who claim to have it or to understand it will be trying to sell a bill of goods—and will be engaging in pseudopsychology. If proper studies are conducted, we must accept their conclusions. To persist in beliefs in the face of contradictory results would be to engage in pseudopsychology.

Psychologists have developed a set of rules to ensure that researchers follow sound ethical standards, especially when participants’ rights may conflict with a research method or clinical treatment. Certain methods are obviously unethical:  No psychologist would cause people to become addicted to drugs to see how easily they can overcome the addiction or would abuse people to help them overcome a psychological problem. But many research situations are not so clear-cut. 

New therapies are developed continually, but only qualified therapists who have been trained appropriately or are learning the therapy under supervision can ethically provide it. For instance, imagine that Dr. Singh has developed a new type of therapy that she claims is particularly effective for patients who are afraid of some social situations, such as public speaking or meeting strangers. You are a therapist who has a patient struggling with such difficulties and is not responding to conventional therapy. You haven’t been trained in Singh’s therapy, but you want to help your patient. According to the American Psychological Association guidelines (see Table 1.4), you can try this therapy only if you’ve received training or will be supervised by someone trained in this method.

Some ethical decisions, such as not administering treatments you haven’t been trained to provide, are relatively straightforward, but the ethical code can also come into conflict with the state laws under which psychologists and other psychotherapists practice. For instance, generally speaking, psychologists should obtain specific permission from a patient before they communicate about the patient with people other than professionals who are treating him or her. However, the law makes certain exceptions to the need to maintain confidentiality, such as when a life or (in some states) property is at stake.

Indeed, difficult cases sometimes cause new laws to be written. In 1969,  a patient at the University of California told a psychologist at the student health center that he wanted to kill someone and named the person. The campus police were told; they interviewed the patient and let him go. The patient later killed his targeted victim. The dead woman’s parents sued the psychologist and the university for “failure to warn.” The case eventually wound its way to California’s highest court. One issue the court debated was whether the therapist had the right to divulge confidential information from therapy sessions, even when someone’s safety is at risk. The court ruled that a therapist is obligated to use reasonable care to protect a potential victim. More specifically, in California (and in most other states now), if a patient has told his or her mental health clinician that he or she plans to harm a specific other person, and the clinician has reason to believe the patient can and will follow through with that plan, the clinician must take steps to protect the targeted person from harm—even though doing so may violate the patient’s confidentiality. Similar guidelines apply to cases of potential suicide.

Further, a therapist should not engage in sexual relations with a patient or mistreat a patient physically or emotionally. The American Psychological Association has developed many detailed ethical guidelines based on the principles listed in Table 1.4.

As research in psychology continues to progress at an increasing pace, new issues have emerged that would have been in the realm of science fiction a few years ago. To address one set of these issues, a new branch of ethics, called neuroethics, is focusing on the possible dangers and benefits of research on the brain, including the use of neuroscience to predict and control individual behavior. Although a relatively young field, this field has its own journal, Neuroethics, and addresses questions that are hotbeds of debate.  So far, neuroethicists have more questions than answers. For example, is it ethical to scan people’s brains to discover whether they are telling the truth? To require young children to take medication that helps them pay attention better in school? To encourage adults with no cognitive problems to take “cognitive enhancing” pills or medical and surgical procedures (Darby & Pascual-Leone, 2017)? Should computer games that can enhance brain function be regulated—either by physicians or the government? These are the kinds of questions that neuroethicists try to resolve.

Consider the field of neuroethics applied to convicted and future criminals. Suppose that the brains of murderers could be shown conclusively to have a distinctive characteristic (perhaps one region that is much smaller than normal). Should we then scan people’s brains and watch them carefully if they have this characteristic? Should this characteristic be used as a criterion for parole for prisoners? Would it be ethical to alter prisoners’ brain function if it makes them less likely to reoffend when released (Craig, 2016)? Such questions have no easy answers.

One organization devoted to neuroethical questions, The Center for Cognitive Liberty and Ethics, asserts that two fundamental principles should form the core of neuroethics: First, individuals should never be forced to use technologies or drugs that alter how their brains function. Second, individuals should not be prohibited from using such technologies or drugs if they so desire, provided that such use would not lead the individuals to harm others. However, not everyone agrees with these principles.