How Do Psychologists Collect Research Data?
Psychology Research Methods
The basic research methods used in Psychology are usually classified into three types - descriptive, correlational and experimental. Choosing which research method to use depends upon the goals of the research study and the possible limitations that one can encounter during the data collection process. The list of common research methods enumerated below is not exhaustive of all the methods practiced in the field, but it can provide you a general idea of how psychologists collect research data.
Descriptive Research in Psychology
Research problems, particularly psychological research, come from questioning an existing theory, or having an interest on a particular psychological phenomenon. Descriptive research in Psychology is usually sparked by an interest to describe the nature of certain phenomena in Psychology. There are many ways in which this can be done, among them are the following:
Observation is a research method widely practiced in almost any kind of research, be it in other social or physical sciences. However, unknown to many, it is also one of the most difficult to do; it requires much skill and practice. Psychologists usually follow a structure or a system of observing. Psychologists identify beforehand how they will go about the observation process, who and what they will observe, when and where they will observe, and in what form the observation will be recorded. psychologists may choose to conduct observation in two kinds of location - the laboratory or the natural environment. The setting in which the observation will take place depends upon the degree of control a researcher needs to impose on the study. Laboratory observation is usually done to isolate factors psychologists want to observe from unwanted factors found in a natural environment. For instance, a psychologist may want to conduct a research study on the effects of different levels of noise on reading comprehension. In this study, other factors aside from noise levels, such as lighting and temperature, must be made constant. There are many drawbacks in conducting a laboratory observation. One is that participants are aware that they are being observed and may act unnaturally, thereby reducing the reliability of the research results. Another drawback was identified by cross-cultural psychologists. They noted that participants for a laboratory observation are oftentimes representatives of a major racial type because marginalized people are usually unwilling to participate in the study, or that Psychologists are unable to reach them. Naturalistic observation, on the other hand, lacks the kind of isolation present in a laboratory setting. psychologists who use this setting make the effort to conduct observation anonymously. Jade Goodall, for example, spent almost 30 years observing chimpanzees in their natural habitat. Choosing between conducting a laboratory or a naturalistic observation requires an understanding of both their strengths and weaknesses. Again, the method to be used in a psychological research must align with the research goals of the study.
Surveys and Interviews use questionnaires to gather quick and massive data across a large population. They may be administered in paper or in person, by means of the telephone or the internet. Questions should be clear and unbiased, requiring unambiguous answers from participants. They may be structured or unstructured (open-ended), that is, the participants may choose only among the available answers the researcher has provided, or the participants may answer the questions freely and without limitations and restrictions from the researcher. The problem with conducting surveys and interviews is that participants may exhibit Social Desirability. This psychological phenomenon is defined as people's tendency to project a more likable image to the researcher. There are many ways psychologists do to reduce the likelihood that participants answer questions in a socially desirable manner, but that topic will be discussed in a separate article.
Standardized Tests are used to put individuals in context by tallying and comparing scores. Standardized tests may be administered orally or by writing. They use percentiles to describe individuals by indicating the percentage of people who scored below you. For example, you belong to the 92nd percentile in the Stanford-Binet Intelligence Test. This means that 92% of those who took the same test scored lower than you did. The use of standardized tests, like all other research methods, has its own set of problems to deal with. Because tests are standardized, they fail to account for variations in human behavior, so that results are not always predictive in non-test situations. For instance, intelligence tests oftentimes serve as major criteria for accepting university applicants because they generally predict college success; however, not all applicants who passed the test make it through their graduation. Clearly, performance in standardized tests does not always show in real-life setting. Another problem is that most tests are based from the data gathered in western countries, particularly in the United States and in Europe. Although these tests are increasingly being adapted in local areas across the world, the primary data context in which local test-takers are compared remains to be from western participants. For example, almost all countries use the Stanford-Binet Intelligence Test (adapted in the local language) to assess intelligence.
Case Studies or Case Histories are oftentimes used by clinical psychologists to obtain an in-depth look at a single individual. Case studies try to uncover the range of possibilities in a human behavior by focusing on special cases like Sybil, a woman reported to possess 16 personalities, and the famous Developmental psychologist Erik Erikson's (1969) study of Mahatma Gandhi. Unique life experiences merit a more individualized understanding of these people. The problem lies, however, in the generalizability of this kind of research. Because conclusions are drawn from a single to few studies, it is unlikely that results are applicable to all people who have multiple personalities, or those who have influential spiritual advocacies. Furthermore, the process used for arriving at the conclusions is made only by the researcher conducting the study and is usually of unproven judgments of reliability. Hence, this method is the most widely criticized among scientists, but the insights gathered from this kind of research remains to be valuable in the academic community, particularly in the area of Clinical Psychology.
Correlational Research in Psychology
Some researchers wish to go beyond describing psychological phenomenon and hope to uncover relationship between psychological factors. For example, a psychologist may want to know the relationship between high school grades and college grades. The relationship can be summed up by the Correlation Coefficient. A positive correlation means that as one factor increases, the other factor also increases; while a negative correlation means that as one factor increases, the other factor decreases. The strength of the relationship is indicated by the number of the correlation coefficient. A 1.00 correlation coefficient indicates perfect relationship, 0.76-0.99 indicates a very strong relationship, 0.51-0.75 indicates a strong relationship, 0.26-0.50 indicates a moderate relationship, 0.01-0.25 indicates weak relationship, and 0.00 indicates no existing relationship between the two variables. Going back to the high school-college grade example, a +0.81 correlation coefficient between the two variables means that high school grades and college grades have a very strong relationship, and are highly predictive of each other. Thus, an A+ high school student will most likely have high grades in college, or a college summa cum laude most likely had good high school standing. Notice that the relationship between high school and college grades in this example is not perfect. This means that not all A+ high school students will get good grades in college, although the likelihood that they will is very high.
A common misconception in a Correlational Research is that relationship equals causation. An uncritical eye may interpret the above example on high school and college grades as such: "High school grades cause college grades." Of course, high school grades don't cause college grades. It could well be that A+ students practice good study habits that carry on through their college years. Thus, attempting to establish causation in a Correlational Research poses the problem of a reverse causal possibility in a relationship (i.e., "College grades cause high school grades."), and the presence of an unidentified third variable (e.g., good study habits). Because Correlational Research cannot establish causation between variables, the method is oftentimes used in post-hoc psychological phenomena (instances that happened in the past and cannot be repeated, like the September 11 attack on the World Trade Center), and studies where establishing causes through experimentation proves to be unethical (like when a student is forced to get low grades in high school to see its effects in his or her performance in college).
Experimental Research in Psychology
Unlike Correlational Research, Experimental Research is done to identify causes of behavior and mental processes. This can be done by distinguishing the independent from the dependent variable. The Independent Variable is the potential cause of a psychological phenomenon, and is manipulated to produce effects called Dependent Variable. The independent variable is rightly called "independent" because it can be manipulated and changed without affecting other factors, except the dependent variable, for which it was rightly called "dependent." For example, in a research study that seeks to identify whether a particular reading strategy can result to better reading comprehension, the independent variable in this example is the reading strategy, while the dependent variable is performance in a reading comprehension test. Participants are randomly assigned either in the experimental or the control group. The Experimental Group is composed of participants who will experience the independent variable, while the Control Group serves as the baseline for comparison because the independent variable is withheld from the participants in this group. If the dependent variable from the experimental and control group are significantly different from each other, then it is safe to say that the independent variable caused the difference between the two. Put simply, if the participants in the experimental group significantly outperformed those in the control group in the reading comprehension test, then the researcher can conclude that reading strategy can result to better reading comprehension.
Manipulation in doing Experimental Research requires a degree of control that can only be achieved in the laboratory. Therefore, as with laboratory observation (as discussed in the early parts of Descriptive Research), Experimental Research faces similar problems on participant awareness and lack of representatives from marginalized societies. Furthermore, bias contaminates most experiments. Experimenter Bias is the researcher's expectations of the outcome of the study. A study by Robert Rosenthal (1944) found that students who were told that their rats are maze-bright report that their rats outperformed those rats belonging to students who were told that their rats are maze-dull (even if there is no actual difference between the two groups of rats). Because there is actually no difference between the rats, it was therefore the students' expectation that caused the "maze-bright" rats to do better than "maze-dull" rats. Participant Bias, on the other hand, is the expectations of the participants on the outcome of the study. For example, a no-effect drug (called a placebo) is usually injected to medical patients who were told that they were given the prescribed drug. If the patients showed positive changes in health, the result is therefore caused by the expectation that they were given the appropriate treatment. Thus, the prescribed drug may not be as effective as it is supposed to be. Participant bias is usually prevented by a Single-Blind Experiment, in which the participants do not know whether they belong to the experimental or the control group. However, in order to rule out the possibility that the research results are contaminated both by experimenter and participant bias, the researcher may opt to design a Double-Blind Experiment, in which both the experimenter and the participants are unaware which of the participants receive the independent variable or not.
On Choosing the Best Research Method in Psychology
Most students of Psychology do not know which method to use when conducting their research. A good way is to follow the scientific method, step by step. Below is a guide that can help you identify which method to use in your research.
- Would you want to explain or pinpoint the cause of a psychological phenomenon? If yes, go to question 4. If no, proceed to question 2.
- Would you want (instead) to identify the relationship between two or more variables? If yes, go to question 5. If no, proceed to question 3.
- Would you want (instead) to describe and explore the nature of a psychological phenomenon? If yes, go to question 6. If still no, please consult with your professor.
- Can you identify the independent and dependent variables in your research study? If yes, go to question 7. If no, proceed to question 2.
- Can you identify the variables you wish to study? If yes, go to question 8. If no, proceed to question 3.
- Would you want to observe the phenomenon? If yes, go to question 9. If no, proceed to question 12.
- Is it ethical to use the independent variable to your participants? If yes, go to question 10. If no, proceed to question 2.
- Can you measure the variables you wish to study? If yes, go to question 11. If no, proceed to question 3.
- Can you identify who, what, when, where and in what form of recording you're planning to conduct the observation? If yes, it is recommended that you do a Descriptive Research, particularly observation. If no, proceed to question 12.
- Is it ethical to measure the dependent variable from your participants? If yes, it is recommended that you do an Experimental Research. If no, proceed to question 2.
- Is it ethical to measure those variables? If yes, it is recommended that you do a Correlational Research. If no, proceed to question 3.
- Would you want (instead) to conduct a survey or an interview? If yes, go to question 15. If no, proceed to question 13.
- Would you want (instead) to administer a standardized test? If yes, go to question 16. If no, proceed to question 14.
- Would you want (instead) to conduct a case study? If yes, go to question 17. If still no, please consult with your professor.
- Can you provide clear and unbiased questions? If yes, go to question 18. If no, proceed to question 13.
- Can you identify and obtain which standardized test to use? If yes, it is recommended that you do a Descriptive Research, particularly standard testing. If no, proceed to question 14.
- Can you identify what makes this person special? If yes, go to question 19. If no, please consult with your professor.
- Will the questions yield unambiguous answers? If yes, it is recommended that you do a Descriptive Research, particularly survey or interview. If no, proceed to question 13.
- Do you have access to this person? If yes, it is recommended that you do a Descriptive Research, particularly a case study. If no, please consult with your professor.
Don't forget to consult with your professor after taking the survey. Remember that some research studies can be done using multiple methods, so it's best to consult your professor on ways to enhance the research design.