Biological Bases

This section covers the following topics

Requirements for a good experiment/study
The types of research
Statistical methods for combining research
Section Summary

Research forms the foundation of psychology; much of what we know about psychology today is because of research. There are specific requirements for research. A study or experiment must be both reliable and valid, and the procedure should be ethical according to the American Psychology Association’s standards. Within psychology, there are multiple types of research: case studies, naturalistic observation, experiments, surveys, and more. Once research is completed, descriptive and inferential statistics help researchers to make sense of all of the data.

Introduction

The Importance of Research

Psychology is based on research in its different forms and methods, and the way that a psychologist conducts research is important because it determines how trustworthy the conclusions of the research are.

There are two categories of research.

Applied research: Research that is done for a clear, practical purpose. For example, a study to determine whether breast-feeding is beneficial to children’s later intelligence is an example of applied research because it has an immediate practical use.
Basic research: This builds psychology’s general knowledge base but has no immediate applications. For example, a study determining whether identical twins’ personalities are nature- or nurture-oriented is an example of basic research.
Terminology

A theory (in psychology) is an idea that explains some aspect of human behavior and psychology. Theories give psychologists material for new, testable hypotheses. A theory may be supported or disproved through many research studies based on that theory.

A hypothesis is an idea that explains the relationship between two specific variables. A variable is an element that can vary, or change, among different people.

The independent variable is the variable that is being tested. A good way to remember this is that the “I”ndependent variable is the variable that “I” change.
The dependent variable is the variable that the experimenter has no control over. In a study on whether or not eating breakfast in the morning makes one live longer, the independent variable is the act of eating breakfast and the dependent variable is the length of life.
Confounding variables are differences between the experimental group and the control group that could affect the dependent variable. For example, a confounding variable in the breakfast experiment could be the exercise activity that each of the subjects engages in.
Types of Research

Guidelines for Research

A good study must be both reliable and valid. Reliability means that the results are consistent. A reliable study can be repeated over and over again with the same results. On the other hand, a valid study measures what it was intended to measure.

One of the most important aspects of a study is the sampling method. The people who are involved in the experiment are the participants, and the process of choosing these participants is sampling. A sample should be representative of the population, or the larger group of people, from which they are chosen. The sample should be chosen through random selection, which means that each person has an equal chance of being selected. One type of sampling method is stratified sampling, in which the sample represents the population according to some criteria.

The American Psychological Association (APA) has established guidelines for human and animal research. Academic research and experiments must conform to the guidelines of the APA, and psychologists must first propose their study to the ethics board or institutional review board (IRB) of the APA. The guidelines are different for animal and human research. For human research, the guidelines are:

No coercion—people must volunteer for the experiment/be willing to be involved.
Informed consent—the subjects in the study should have an idea of what they are doing.
Anonymity or confidentiality—the subjects’ privacy must be protected, and the scientists cannot match subjects with their names during the experiment; if anonymity cannot be achieved, confidentiality is required.
Risk—participants should not be put in harm’s way during the experiment, and the researchers should ensure their safety and well-being, both physical and emotional.
Debriefing—participants should be told the complete purpose of the study following the experiments, especially in experiments involving minor deception.
Experimental Method

The experimental method is the only type of research that can show causation. The experimental method is used for field experiments—where the experiment takes place in the real world—and laboratory experiments—where the situation is highly controlled.

The advantage of laboratory experiments is that the psychologist can prevent confounding variables. Ideally, each student should take a test with the exact same level of difficulty. Situation-relevant confounding variables occur when the situation (setting) for the experimental and control groups is not the same. Examples of situation-relevant confounding variables include the time of day, weather, and presence of other people in the room during the experiment.

Another type of confounding variables is participant-relevant confounding variables. For example, in the experiment measuring whether or not more sleep means better test scores, if the students could choose to be in the experimental group (a limited period of sleep) or the control group (a normal level of sleep) the composition of the groups would probably be different. Because of participant-relevant confounding variables, the experimenters must assign participants to the experimental or control group. Random assignment means that each participant has an equal chance of being chosen for the experimental or control group. Random assignment is necessary for a reliable experiment. One method of random assignment that incorporates some criterion to the selection of the groups is group matching. For example, if an experimenter wanted to see the differences between young and older people, the experimenter could divide the participants by age and then randomly assign the two groups into the experimental and control groups.

The way in which an experiment is conducted is important for an experiment to be reliable and valid. Experimenter bias, in which the experimenter unconsciously incorporates his or her own beliefs and expectations into the experiment, can affect the dependent variable. To eliminate experimenter bias, the experimenter can use a double-blind experiment. In a double-blind experiment, neither the experimenter nor the participants know who is in the experimental group and who is in the control group. On the side of the participants, this eliminates multiple problems:

The placebo effect, in which a participant feels a change simply because he or she received a treatment.
The Hawthorne effect, in which participants act differently because they know that they are part of an experiment.
Depending on the experiment, a psychologist could use counterbalancing to use the participants as their own control group. For example, in an experiment that tests the impact of positive thinking before an IQ test, the psychologist could have participants take an IQ test without positive thinking, then instruct the participants to think positively, and take the IQ test again. To counterbalance, the psychologist should have another group think positively first, take the IQ test, then take the IQ test without instruction to think positively. This should eliminate the possible effects of taking the test two times in a row, which are known as order effects.

Case Study

The case study is a type of research that focuses on one or more participants. A case study cannot prove causation, but it can provide rich insight into the psychology and development of a particular participant. Sigmund Freud and Jean Piaget both used case studies to develop their theories.

Case studies are limited, however. Because they focus on such a small group of participants, the results of case studies cannot be generalized to the larger population. Just because one person suffers from a condition in a certain way does not mean that the rest of the population would respond in the same way. Another problem with case studies is the potential for researcher bias. A researcher may project his or her ideas or expectations onto the case study.

Survey

Survey research involves asking a large sample of participants questions through an interview or through a written questionnaire. The advantage of survey research is that it is a relatively quick way to gain large amounts of data from a sample. This data is easy to process and quantify.

However, a researcher has very little control over the people who take the survey, and cannot eliminate many of the variables that could influence the results of the survey. For example, in conducting a survey on personality, there could be a particular personality type that would be predisposed to filling out the survey and sending it back. This personality type would then be overrepresented in the results of the survey. Social desirability bias refers to the fact that people often give answers that are politically correct even if they do not personally agree with the answer.

Correlation

A correlational study is used to establish the relationship between two variables. A correlational study cannot prove causation, but it can provide insight into the relationship between the two variables. A correlational study can show positive correlation, which means that if one variable is present, the other variable is likely to be present. On the other hand, a negative correlation means that the presence of one variable is likely to predict the absence of the other variable.

Example of positive correlation: A study of a supermarket shows that when there is a sale, the number of people who enter the store increases. The sale and the number of people who enter the store are positively correlated.
Example of negative correlation: A study of a supermarket shows that when there are two managers, the number of store complaints decreases. The two managers and the number of store complaints are negatively correlated (an inverse relationship).
Naturalistic Observation

Naturalistic Observation is when a researcher simply observes a phenomenon in its natural state. The benefits of naturalistic observation include a realistic view of participants’ behavior outside of the artificial environment of the laboratory. However, the presence of an observer can affect the behavior of the participants; this is known as observer effects. Additionally, naturalistic observation cannot establish causation and can be limited by researcher bias.

Evaluating the Research

One of the most important steps following research is the evaluation of the research. This will help the experimenter to reach conclusions regarding the research.

Descriptive Statistics

Descriptive statistics just describe information, and can be organized into frequency distributions, frequency polygons or histograms.

A frequency distribution simply groups the data by how frequently a particular variable occurs. A frequency polygon is a line graph based on a frequency distribution, and a histogram is a bar graph.

Measures of central tendency intend to show the center of a group of data. The mean is the average, and it can be distorted by extreme scores, or outliers. For example, a study of average incomes in a low-income part of the town with one extremely wealthy person on the edge could distort the mean. The median is the center score in the distribution, and if the number of scores is even, the median is the average of the two middle numbers. The mode is the score that appears the most in a distribution. A distribution can be bimodal when two different numbers appear with the same frequency.

If a distribution is not symmetrical, it is skewed, either positively (with extremely high scores) or negatively (with extremely low scores). A positively skewed graph has more low scores with some unusually high ones; the mean is higher than the median. A negatively skewed graph has more high scores with unusually low ones; the median is higher than the mean.

Measures of variability include range, variance, and standard deviation. The purpose of these measures is to show the diversity of the distribution. The range is the difference between the highest and lowest scores. The standard deviation is the square root of the variance, relating the average distance of any score in the distribution from the mean. Essentially, the standard deviation and the variance show how widely distributed the scores are. If the standard deviation and variance are high, the scores are very spread apart.

Scores can be converted from the different distributions into z scores, which measure the distance of a score from the mean in units of standard deviation. For example, if the standard deviation is 5, Jordan scored a 90, Jackson scored an 80, and the mean was 90, then Jordan’s z score=0 and Jackson’s z score=-2.

Percentiles show the distance of a score from 0. For example, someone scoring in the 99th percentile scored better than 99% of the people taking the test.

Correlations show the relationship between two variables. For example, studying hard is positively correlated with good grades. There is a relationship between the two variables, for when one is present, the other is also present. Correlations can be strong or weak, shown by a correlation coefficient, ranging from -1 to +1, with -1 showing a perfectly negative correlation and +1 showing a perfectly positive correlation. A scatter plot can show a correlation, and the closer the points come to forming a line, the stronger the relationship. A line of best fit/regression line shows how close the points come to forming a line.

Inferential Statistics

The purpose of inferential statistics is to determine whether or not findings can be applied to a greater population than the original population from which the sample was selected. The difference between the sample and the population is the sampling error. Inferential statistics help psychologists decide whether their findings can be applied to the larger population. All of the tests to find these inferential statistics take into account the magnitude of the difference found and the size of the sample. All of the tests will produce a p value.

The smaller the p value, the more significant (applicable to the general population) the results are.
The higher the p value, the less likely the results are statistically significant.
A p value of .05 is the cutoff for statistically significant results. For example, a p value of .06 would be too high for the results to be applicable to the rest of the population.