Nonexperimental Research Methods Research Paper

This sample Nonexperimental Research Methods Research Paper is published for educational and informational purposes only. Free research papers are not written by our writers, they are contributed by users, so we are not responsible for the content of this free sample paper. If you want to buy a high quality research paper on any topic at affordable price please use custom research paper writing services.

Randolph Smith’s chapter on Experimental Designs (see Chapter 9) indicates that experimental research, in which researchers directly manipulate independent variables (IVs) about which they seek to gather information, record changes in dependent variables (DVs), and control extraneous variables that might pro-duce a flawed experiment, yields the strongest cause-and-effect results. However, not all research projects use the experimental method; in fact, researchers have developed a large number of nonexperimental techniques. Although they are classified as nonexperimental, these methods do not necessarily produce inferior results. The choice of a nonexperimental research strategy depends on the type of problem the researcher is interested in, where the research is conducted, and the type of data the researcher gathers. We will examine several types of nonexperimental techniques in this research-paper.

Observational Techniques

Naturalistic Observation

Researchers who use naturalistic observation seek answers to their research questions by observing behavior in the real world. For example, a researcher who is interested in people’s behavior at an ATM machine would locate an ATM machine and observe the behavior of people who use it.

The possibilities for conducting studies using naturalistic observation are limited only by our curiosity and insight into a potential research area. When researchers conduct a naturalistic observation study, they have two goals: (a) to describe behavior as it occurs in the natural setting without the artificiality of the laboratory and (b) to describe the variables that are present and the relations among them. In a naturalistic observation study it is important that the researcher not interfere with or intervene in the behavior being studied. For example, the observer at the ATM machine should be as inconspicuous and unobtrusive as possible; otherwise, the observer might influence or change the behavior of the people using the machine. The presence of an observer is not part of the natural setting for an ATM machine. In this context, researchers often mention the reactance or reactivity effect, which refers to the biasing of participants’ responses because they know they are being observed. For example, if the people using the ATM machine know they are being observed, then they may glance around more often and try to rush their transaction. Clearly, the knowledge that you are participating in an experiment and being observed may result in dramatic changes in your behavior.

Naturalistic observation’s main drawback is its inability to yield cause-and-effect results. If it does not lead to the development of cause-and-effect statements, why do researchers use this method? In many instances, the reason is quite obvious: Naturalistic observation may be the only option available if the researcher wants to conduct the project. Psychologists who are interested in reactions to natural disasters such as hurricanes, earthquakes, tornadoes, and fires, ethically cannot create such life-threatening situations just to study behavior; they must gather their data and make their observations under naturally occurring conditions.

Case Studies

When researchers conduct case studies, they observe and record the behavior of one, possibly two, participant(s) over an extended period of time. (Conducting research on the same participant[s] over an extended period of time is typically called a longitudinal design!) There are no established guidelines for conducting case studies, hence the procedures used, behaviors recorded, and final reports typically vary considerably. Researchers often use case studies in clinical settings to help formulate hypotheses and ideas for additional research. For example, Paul Broca (1824-1880), a French physician, reported the unusual case of “Tan.” His patient was nicknamed Tan because this word (in addition to obscenities when he was angry or frustrated) was the only word the patient uttered (Howard, 1997). Based on his observations, Broca hypothesized that the brain center that controlled speech production was located in the frontal lobe of the left hemisphere. An autopsy indicated that Broca’s hypothesis was correct; Tan’s left frontal lobe was damaged.

Although case studies can provide interesting and important results, you should keep in mind that their results may be applicable only to the patient whom the researcher observed. Researchers should not generalize beyond the patient they observe. However, if your research goal is to understand the behavior of that one patient, then the case study method certainly can achieve your goal.

Participant Observation

Participant observation is similar to naturalistic observation in that behavior is observed in the real world. Unlike naturalistic observation, however, researchers who use participant observation become part of the group they are studying; the researcher does not try to be unobtrusive. When the research goal is to learn about a specific culture or socioeconomic group, researchers frequently use this technique and typically assume the role of a member of the group in order to make the needed observations.

Glesne (1999) identified two types of participant observation that are related to the degree of participation. The observer as participant refers to a researcher who primarily observes a situation but also interacts with the others involved in the study. The participant as observer refers to a researcher who becomes a part of the culture by working and interacting extensively with the others. There is a cost-benefit relation with these two approaches. The more immersed you become in a culture, the more you can learn about that culture; however, greater immersion can result in a loss of objectivity.

Participant observation suffers from several additional drawbacks. First, it may take a long time before the participant observer is accepted as a member of the group that is under study. This extended period of time can increase the cost of the project substantially. Additionally, simply being part of the situation does not guarantee that the observer will be accepted. If the observer is not accepted, then the amount of information that the observer can acquire is limited. On the other hand, acceptance of the observer may result in a loss of the observer’s objectivity. Finally, as with other observation methods, participant observation lacks the ability to develop cause-and-effect statements.

Selecting Behaviors and Choosing Recording Techniques

The conduct of an observational study requires a great deal of planning and preparation before the researcher actually attempts to record behavior in the real world. First, researchers must decide which behaviors they will observe. Unless you plan to use video equipment, which may not be possible if you plan to be as unobtrusive as possible, you will not be able to record all behaviors. Second, you will need to decide when and where to record the behaviors in question.

These where and when questions are deceptively important. Consider the ramifications of observing in one location at one time. Even if the researchers make a large number of observations, their conclusions are limited to just that one situation and one time period. If the researchers are interested in behavior in only one situation occurring at only one specific time, then this limitation is not a problem. However, if the researchers want to generalize their findings to other situations, they will need to make observations in other, different situations. The process of making observations in order to increase the generalizability of the conclusions is called situation sampling.

Likewise, even if the researchers make a large number of observations during one time period, they cannot generalize their conclusions to other time periods. Should they wish to generalize their conclusions, the correction for this problem is called time sampling and is analogous to situation sampling; the researchers simply make observations during several different time periods. Of course, researchers can, and do, use situation sampling and time sampling simultaneously to determine which of the behaviors under observation are a function of context or time of observation.

Once researchers have decided whether to use time sampling and situation sampling, they must also decide whether they want to describe the behavior(s) of interest qualitatively or quantitatively. If they use the qualitative approach, their report likely will consist of a narrative description of the behavior(s) in question. (See Chapter 11 for an in-depth coverage of this methodology.) The researchers should take special care to make sure that the narrative reports use clear, precise language and avoid speculative language. The quantitative approach requires a thorough knowledge of the measurement techniques the researcher plans to use and the type(s) of analysis that this measurement system will require.

Interobserver Reliability: The use of More Than One Observer

Researchers using an observational method must also decide if they will use one or more observers. There are two main reasons for using two or more observers. Observers may disagree about what exactly was observed. Even the use of video recordings does not eliminate the potential need for multiple observers. Someone has to watch the video and evaluate the behaviors.

When researchers use two observers, they can mathematically determine the extent to which their observers agree. This agreement is called interobserver reliability. Low interobserver reliability indicates that the observers disagree about the behavior(s) they observed; high interobserver reliability indicates agreement.

What can influence the degree of agreement between the observers? Physical factors such as fatigue boredom, and emotional state, can exert an influence. Likewise, experience observing the behavior in question can influence interobserver reliability. Hence, it is the researchers’ task to make sure that both observers are well rested, interested in their task, in good physical and emotional health, and have received the same amount of training and experience with the research project and the behavior(s) under study. The trainer should give clear, precise definitions of the behavior(s) to be observed and should provide concrete examples of positive and negative instances of the behavior(s) in question.

How do researchers measure interobserver reliability? The simplest technique involves determining the number of times the two observers agree and the number of opportunities they had to agree. Once the researcher has determined these numbers, they are used in the following formula:

number of agreements/number of opportunities to agree x 100 = percent agreement

This simple formula yields the percent of agreement between the two observers, and many researchers use it as an indication of interobserver reliability. The more accepted measure of interobserver reliability is the square of the correlation coefficient between the two observers multiplied by 100 (i.e., r2 x 100). The resulting figure gives the percentage of variation that is due to observer agreement; the higher the percentage, the greater the agreement. Although there is not complete agreement, most researchers do agree that good interobserver reliability must involve at least 85 percent agreement.

So far we have seen that observational techniques are among the most widely used nonexperimental techniques available to researchers. In the next section we examine a nonexperimental technique that easily can appear to be an experimental procedure.

Ex Post Facto Studies

When researchers study the effects of independent variables (IVs) that they cannot or do not manipulate, they are conducting an ex post facto study. In Latin, ex post facto means after the fact; hence, this type of research involves using or examining an IV “after the fact”—it has already varied or changed before the researcher started the project. Thus, the ex post facto study clearly qualifies as a nonexperimental research technique. On the other hand, it does share some similarities with true experiments.

For example, just as in a true experiment, ex post facto research involves recording the behavior(s) of groups of participants and then performing appropriate statistical analyses on these data. For example, Licht (2000) evaluated occupational stress as a function of the sex of the participants and whether they worked for a nonprofit or for-profit organization. She found that (a) nonprofit employees perceived more occupational stress than did for-profit employees, and (b) men had more stress than women. Her research is classified as ex post facto because she had no control over the sex of the participants or the type of organization they worked for. For these reasons, she was unable to make cause-and-effect statements. For example, she did not randomly assign participants to jobs, hence she cannot say that working for the nonprofit organization caused the increased stress levels that she recorded.

Even though ex post facto research is widespread, the use of questionnaires, surveys, tests, and inventories may be even more widespread. We examine these popular techniques in the following sections.

Questionnaires And Surveys

Because researchers who use questionnaires or surveys to answer their research questions do not directly manipulate variables, these projects belong in the nonexperimental category. Without question, this type of research is extremely popular. One reason for this popularity is its deceptive simplicity; it seems so easy to conduct research using questionnaires and surveys. All the researcher has to do is secure a group of participants, have them complete the questionnaire or survey, and analyze the results. At that point, the project is completed. As we will see, much more is involved in this type of research than meets the eye.

Descriptive and Analytic Surveys

In order to determine participants’ opinions on a topic of interest, researchers may administer a descriptive survey or an analytic survey. Researchers use descriptive surveys when they are interested only in determining the participants’ characteristics, opinions, or beliefs. The polls that the media conduct on election night to predict the winners/losers are good examples of descriptive surveys; they only seek information (such as frequencies or percentages), not reasons or an understanding of the variables involved. Hence, the goal of the descriptive survey is to describe a specific behavior or characteristic as it exists in a participant sample. These sample results will hopefully be representative of the population from which the sample was drawn.

In contrast, researchers who use an analytic survey are interested in determining the reasons for a particular behavior, belief, or opinion—what are the relevant variables and how are they related? For example, the researcher who is interested in why voters preferred this or that candidate, in addition to determining the number of people who voted for each candidate, would use an analytic survey to delve into the “why” question.

Creating a Survey or Questionnaire

Because psychology is such a vast and diverse field, researchers may choose to conduct a project and find that a survey or questionnaire dealing with their topic does not exist. In this situation researchers may be tempted to create their own questionnaire. Researchers who find themselves in such a situation should exhaust all possible sources for such surveys and questionnaires before attempting to create their own instrument. It is not as simple as it may seem to create a good survey or questionnaire; several time-consuming but crucial steps are involved in developing such an instrument:

Preliminary (Pilot) Testing—The researchers will want to conduct preliminary tests with the new instrument. This preliminary testing is typically called pilot testing and may include (a) in-depth interviews with participants to determine the type(s) of questions that would be best for the final version of the instrument, and (b) testing small groups of participants to determine that the instrument is valid (measures what it is supposed to measure) and reliable (yields essentially the same score when administered again).
Determine How the Data Will Be Collected—Researchers need to decide if they will collect their data via mail questionnaires/surveys, face-to-face interviews, telephone interviews, or in a group setting. Who will administer the questionnaires/surveys? What sort of training will the assistants require? Clearly, these are important decisions that can impact the conduct and quality of the research.
Determine the Type of Questions—Researchers can choose among several types of questions: yes-no questions, forced alternative questions, multiple-choice questions, Likert-type questions (selecting a response alternative from a designated scale), and open-ended questions. Without a doubt, the type of questions researchers use directly influence they type of data they gather and the type of analysis they can conduct once they have gathered these data. For example, the yes-no format limits the researcher to counting frequencies and conducting nominal-scale analyses, such as chi square tests.
Writing the Questions—After deciding which type of questions to use, the researcher must actually write the items. Generally speaking, items should be clear, short, and specific; use familiar vocabulary; and be at the reading level of the intended participants. Researchers should be careful to avoid creating questions that would constrain or bias the participants’ answers in any way. For example, wording a question in a positive manner likely will result in a large number of positive answers. Also, the researcher should ask other experts in the research area to review the items for bias and unintended meanings. The researcher also will want to conduct pilot tests on the newly created items. During these pilot tests, the researcher will want to discuss the items with the participants; participant insights can be very valuable.
Determining Other Information to Obtain—Researchers also have to decide what other relevant information, typically called demographic data, they want to collect. Demographic data may include such items as sex, age, academic major, academic classification, size of home community, and annual income. Lest they forget an important bit of demographic information, researchers should review these items very carefully.
Determining Administration Procedures—Researchers must decide how they will administer the questionnaire or survey and what type of instructions they will provide for the participants. They also need to decide who will administer the informed consent forms and answer participants’ questions.

Clearly, if done correctly, the creation of a questionnaire or survey is very time-consuming and involved. Whether or not researchers create their own questionnaires, they must decide how they will administer their instruments. Four of the most popular approaches are mail surveys, personal interviews, telephone interviews, and group administration. We examine these approaches next.

Administering the Questionnaire or Survey

Mail Surveys—Researchers use this method to gather data on just about any topic you can imagine. This approach is popular because researchers can send their questionnaires and surveys to a very large sample and they do not have to be present when the participants complete them. Despite these advantages, mail surveys do have several drawbacks. Researchers never know who completed the surveys and in what order the respondent answered the questions. Also, low return rate is a problem with mail surveys (as few as 25 percent return surveys; researchers consider 50 percent a good return rate). A low return rate and not knowing who completed the surveys can result in a biased sample. Researchers can improve their return rates by (a) including a letter that clearly indicates the nature and importance of the research project, the process used to select the participants, and the researcher’s assurance that all responses are confidential, along with a prepaid return envelope in the initial mailing, and (b) sending additional mailings to the respondents.
Personal Interviews—If a trained interviewer administers a survey in a participant’s home, response rates go up appreciably; in fact, response rates as high as 90 percent are not uncommon. Compared to mail surveys, face-to-face interviews have several additional benefits: The researcher is sure who answered the questions, the order of answering the questions is controlled, and the interviewer is present to answer any questions. Despite these advantages, personal interviews have several drawbacks: They are costly in terms of both time and money, the potential for interviewer bias exists, and high crime rates are making it more dangerous to conduct such surveys in urban areas.
Telephone Interviews—Telephone interviews allow researchers to overcome many of the problems posed by mail surveys and personal interviews. For example, the interviewer actually talks to the respondent and, therefore, knows who completed the survey and the order in which the respondent answered the questions. Also, the interviewer does not have to enter dangerous, high-crime areas. Additionally, computer technology has made it possible to obtain a large, random sample with ease. In most situations the interviewer can enter the participant’s responses as they are made, hence the data are ready for analysis as soon as the interviewer makes the final call. Despite these attractions, telephone interviews have several drawbacks. It is relatively easy for potential respondents to screen calls and, thereby, refuse to participate. Also, telephone interviews preclude the use of visual, nonverbal cues, such as facial expression and body language, by the interviewer. Also, many interviewers and respondents find it difficult to establish rapport on the telephone.
Group Administration—Researchers who use college students as their participants frequently administer their questionnaires or surveys in a group setting. Group administration has several advantages: (a) because several participants are tested at the same time, there are savings of both time and money, (b) the researcher can answer questions, (c) the researcher can evaluate nonverbal cues, and (d) the researcher has some control over the order in which participants complete the questionnaires/surveys and the questions within each testing instrument. The main drawback to group testing appears to be the group setting itself; some participants may be tempted to look on their neighbor’s questionnaires/surveys and answer in a similar manner.

Inventories And Tests

As we have seen, researchers use questionnaires and surveys to determine and analyze participants’ opinions on an issue or topic. On the other hand, researchers use inventories and tests to evaluate a specific characteristic or ability. For example, educational psychologists designed the Scholastic Aptitude Test (SAT) to measure academic potential at the collegiate level. The most popular types of inventories and tests are achievement, aptitude, and personality instruments.

Characteristics of Good Inventories and Tests

Because most researchers use inventories and tests prepared by other professionals, they need to thoroughly examine the instrument before they use it. Some of the questions they will need to answer include the following: How long does it take to administer the inventory or test? Can it be administered to a group? Is the administration procedure complicated? Does a trained professional have to score and interpret the inventory or test? Additionally, researchers will be concerned with the validity and reliability of the inventory or test; these are the two most important characteristics of an inventory or test. When a test is valid, it measures what it was designed to measure. A test is reliable if repeated administrations yield comparable scores; in other words, the inventory or test is consistent.

Types of Inventories and Tests

Achievement Tests—Researchers use achievement tests when they want to evaluate a person’s level of accomplishment or mastery. For example, lawyers must pass the bar examination before they are allowed to practice law. Many achievement tests have a passing score that signifies the minimum level of achievement that a person taking the test must attain.
Aptitude Tests—Researchers use an aptitude test to determine a person’s ability or skill for a certain job or in a particular situation. For example, the student who plans to attend graduate school following completion of the baccalaureate degree most likely will be required to take the Graduate Record Examination (GRE). Graduate school admissions committees believe that GRE scores predict an applicant’s ability to successfully complete graduate-level courses.
Personality Tests or Inventories—Researchers use a personality inventory or test to measure a specific aspect of a person’s motivational state, interpersonal capability, or personality.

Summary

I hope that this research-paper has convinced you that psychological researchers do not have to directly manipulate IVs and record subsequent changes in DVs, while extraneous variables are controlled, in order to gather meaningful research data. Nonexperimental methods can produce excellent data; they simply cannot address cause-and-effect relations. Our examination of research methods has only scratched the surface of the available techniques; I encourage you to read further.

References:

Glesne, C. (1999). Becoming qualitative researchers: An introduction (2nd ed.). New York: Longman.
Howard, D. (1997). Language in the human brain. In M. D. Rugg (Ed.), Cognitive neuroscience (pp. 207-304). Cambridge, MA: MIT Press.
Licht, C. A. (2000). Occupational stress as a function of type of organization and sex of employee. Psi Chi Journal of Undergraduate Research, 5, 46-55.
Smith, R. A., & Davis, S. F. (2007). The psychologist as detective: An introduction to conducting research in psychology (4th ed.). Upper Saddle River, NJ: Prentice Hall.