The Handy Psychology Answer Book. Lisa J. Cohen

John Watson’s treatment of Little Albert provides another example.

      Starting in the 1940s, a series of national and international laws were instituted to protect the rights of human subjects in research studies. In 1947 the Nuremburg code laid down an international code of ethics regarding human experiments. In the 1960s, a series of laws was passed in the United States further developing these protections. The establishment of independent review boards to oversee the safety and ethics of human research in all American research institutions dates from this period.

      Currently, Institutional Review Boards (IRBs) or Human Subjects Review Committees must approve all studies conducted with human subjects. Most academic journals require IRB approval of any study submitted for publication.

      Do all psychological studies use numbers?

      Most psychological studies are quantitative and rely on the translation of psychological traits and behaviors into variables that can be analyzed statistically. Qualitative research, however, also has a place in psychological research. In qualitative research a smaller number of subjects are observed or interviewed intensively. The observations are recorded not in numbers but in a long, detailed narrative. From these narratives, the researcher identifies themes that can be explored with greater precision in later quantitative research. Thus, qualitative research is hypothesis generating vs. hypothesis testing. It is more broad-based and open-ended than quantitative research but less precise and reproducible. It is best understood as a preliminary type of research.

      Prisoners at the Buchenwald concentration camp in Germany during World War II line up for roll call. In response to the Nazis’ horrifically brutal experiments on their prisoners, the Nuremberg Code, a set of ethical requirements regarding human research, was developed in 1947.

      How do the methods of the social sciences differ from those of the hard sciences?

      In general, the role of math is different in the social sciences than in the hard sciences. In the hard sciences, especially physics, mathematics is used to identify fixed laws of nature. Once a mathematical equation is identified to explain the behavior of an object, the equation can be used to predict the behavior of the object with extraordinary precision. Consider the equations that send rockets into space. Thus, mathematics in the hard sciences is predictive and deterministic. All of the object’s behavior can be predicted by the equation.

      Aspect of modern physics, such as quantum mechanics and Heisenberg’s Uncertainty Principle, do contradict this certainty, however, although only in the realm of the extremely small (e.g. the sub-atomic particles) or the extremely large. In psychology, the topics of study are so complex that it is not possible to predict all human behavior with mathematical equations. Whether that will ever be possible is debatable, but it certainly has not been done yet. What does that mean for the role of mathematics in psychological science? In psychology, mathematics is probabilistic. We estimate the likelihood that certain statements are true or not. Further, these estimates are based on the aggregate, on the behavior of groups. Therefore, while we may be able to say a lot about the likely behavior of groups, we are unable to predict the behavior of any given individual with certainty. For example, based on samples showing increased beer drinking among male college students compared to females, we can predict that, in general, male college students will drink more beer than females but we cannot predict the behavior of any given college student.

      Why is sample selection important in psychological research?

      In psychological research, we try to draw conclusions about a larger population from observations of a small sample. We cannot study all male college students or all people with schizophrenia so we study a sample of the population of interest and then try to apply our findings to the larger population. For this reason it is critical to make sure the sample is similar to the larger population. There are many ways the sample can vary from the larger population. The way we recruit our study subjects may bias the sample right from the start. For example, if you want to study illegal behavior, you are likely to find your sample in the judicial system. Right off the bat, your sample is biased toward people who have been arrested, leaving out the people who never got caught. If you want to study people with depression, you are likely to study people in the mental health system and your sample will be biased toward people who seek treatment. Because it is virtually impossible to remove all problems from sample selection, researchers must carefully describe their samples so that the applicability to a larger population, or the study’s generalizability, can be assessed.

      What are statistics and how do they work?

      Millions of psychology majors grit their teeth and roll their eyes at the very thought of statistics. Nonetheless, statistics are a fundamental part of psychological research. Statistics provide a mathematical technique to measure the relationships between two or more variables (traits of interest such as intelligence, aggression, or severity of depression). Statistics can show how these variables relate. It can show the strength of their relationship and the probability that the relationships found in a given study are likely to be true findings, rather than a statistical fluke, i.e., due to chance. The most common statistics are measures of central tendencies, specifically: the mean, median, and the mode; measures of group differences such as the t-test, analysis of variance (ANOVA) and multivariate analysis of variance (MANOVA); and measures of covariation, such as correlation, factor analyses, and regression analyses. Measures of covariation assess the degree to which two or more variables change in relation to each other. For example, height and weight covary (or are correlated) while age and ethnicity do not. In general tall people weigh more than shorter people while ethnicity does not change with age.

      This is a graph showing how 69 people scored on a test measuring traits of a personality disorder. The vertical axis represents the count, or the number of people who attained each score. The horizontal axis represents the scale score. In this graph, the vast majority of the subjects scored on the low end of the scale while a few people had much higher scores. When the distribution of scores is concentrated toward one end of the scale, we say there is a skewed distribution. In a normal distribution the majority of scores are in the middle with a few scores moving out toward the far ends of either side. When the distribution is skewed like this, the mean, median, and mode separate from each other.

      What are measures of central tendency?

      These are ways to characterize a population or a sample. The mean is the average score. It is calculated by dividing the sum of scores by the number of scores. For example, the average of the series {4, 7, 8, 9, 9} equals 7.4, (4 + 7 + 8 + 9 + 9 divided by 5). The median refers to the number that falls in the middle of the sample; half of the scores lie above it and half lie below. In this case the median is 8. The mode refers to the most common score. In this case the mode is 9. Each measure of central tendency has different advantages and disadvantages.

      What is the difference between the median and the mean and why does it matter?

      The mean is very sensitive to extreme values, also known as outliers, and so can give a distorted view of a population when some values are much higher than the rest. The median is not affected by outliers and thus can be a more stable measure of central tendencies. For example, the mean of the series {8, 8, 9, 12, 13, 102} is equal to 26.4 but the median is equal to 10.5. This distinction is very important when describing characteristics such as national income. Due to a small percentage of people with very large incomes, the average or mean income in the United States is higher than the median income. Because of this the U.S. Census only reports median income. The mean, on the other hand, is more useful in statistical analyses.

      What is the standard deviation?

      In order to characterize a sample, it is not only necessary