TCA Lectures What is experiment and how you will conduct experimental research, what will be the threats to internal and external validity to minimize these threats ? BEd

If you want to view other related topics. Click Here.

Answer : 

Non-Experimental & Experimental Research

Alright! It's time to learn something using research by … performing a non-experimental study?

Wait, wait, and wait! Is it possible to have a non-experimental study? Is that sort of like sugar free candy? Is it something that you're supposed to have that is replaced by something that makes you scratch your head? Before we discuss research designs, though, you need a brief walkthrough of some of the terms I am going to throw at you.

A predictor variable is the portion of the experiment that is being manipulated to see if it has an effect on the dependent variable. For example, do people eat more Gouda or cheddar cheese? The predictor variable in this is the type of cheese. Now, every time you eat cheese, you'll think about predictor variables. When I say subjects, I just mean the people in the experiment or the people being studied.

Experimental research is when a researcher is able to manipulate the predictor variable and subjects to identify a cause-and-effect relationship. This typically requires the research to be conducted in a lab, with one group being placed in an experimental group, or the ones being manipulated, while the other is placed in a placebo group, or inert condition or non-manipulated group. A laboratory-based experiment gives a high level of control and reliability.

Non-experimental research is the label given to a study when a researcher cannot control, manipulate or alter the predictor variable or subjects, but instead, relies on interpretation, observation or interactions to come to a conclusion. Typically, this means the non-experimental researcher must rely on correlations, surveys or case studies, and cannot demonstrate a true cause-and-effect relationship. Non-experimental research tends to have a high level of external validity, meaning it can be generalized to a larger population Experimental and non-experimental research design

By The U.S. Food and Drug Administration [Public domain], via Wikimedia Commons

Experimental research design Experimental research designs are based on a clear hypothesis, the purpose of the research is to confirm or refute the validity of the hypothesis. Experimental research designs have an independent variable, a dependent variable, and a control group. Most of the experiments are conducted in a laboratory in a controlled environment. The purpose of the experiment is to find out the causation and experimental studies are causal studies. These studies answer what, why and even how questions in the research. The experimenter can manipulate the variables and he has a control group and a placebo. The control group receives the treatment that

the experimenter wants to test and the placebo group is tested without any treatment. The change in the results of both groups are compared. The experimenter repeats the test in the same environment more than one time to get most valid results. There are basically three different types of experiments: controlled experiments, quasi experiments, and field experiments.

Methodology

The experimenter determines two groups to perform an experiment: one group is called the experimental group and the other group is called a placebo group. The experimental group gets the treatment and the placebo does not get any treatment. The experimental group is similar to the control group except that the control group does not get any treatment while the experimental group gets the treatment. The experimenter repeats the test twice or thrice to increase the validity of the results.

Non-experimental research design

Non-experimental research designs are carried out in natural settings, it does not involve manipulation of the situation, event, circumstances or people. Survey, case studies, correlational studies comparative studies and descriptive studies are some of the examples of non-experimental research design. Longitudinal studies are also non-experimental research and the purpose of these studies is to study a situation, people or phenomenon over a period of time to observe the change.

A correlation can be established using non-experimental research design but causation cannot be established. To establish causation, the researcher should be able to say that the result is the outcome of the observed variable and not something else. They do not have a control group and the research design is highly flexible. Due to the absence of the control group the researcher cannot ascertain that the final results are the direct effect of the variable that has been studied. The non-experimental research design study the phenomenon, people or situation in a natural setting without manipulating it, therefore, the findings can be applied to a wide audience.

Methodology

In a non-experimental research, the researcher does not manipulate the variables to be tested and therefore he cannot ascertain the effect of the independent variable on the dependent variable. A causation is established in some of the non experimental studies but not in all of them. The causation is established by determining that a certain variable has an impact on another variable. It can also compare the results of two or more groups or people on one or more variables

Steps in Experimental Research

The basic steps involved in conducting experimental research will be presented in this module.

Learning Objectives:

• List the steps that are used in conducting experimental research.
• Describe each step involved in the process.

The following YouTube vide, Experimental Design, offers a visual representation of the process of conducting experimental research and serves as introduction to the steps discussed in this module.

The following list of steps explains the process of conducting experimental research in more detail. Researchers should follow these steps in order to ensure the integrity of the process.

1. Select a topic. This involves simply identifying an area of interest or general subject.
2. Identify the research problem. Given the topic or subject, the researcher must now identify specific problems or questions that relate to the subject. The researcher may be familiar with subject and may already know the problem they want to research. If the researcher is new to the topic, it may be helpful to examine literature and previous studies, as well as talk to other researchers. The problem selected should be important to the field and be of significance to others in the discipline.
3. Conduct a literature search. Once the research problem is identified, a literature search should be conducted before proceeding to design the experiment. It is helpful to know what studies have been performed, the designs, the instruments used, the procedures and the findings. This information will guide the researcher and help them create a project that extends or compliments existing research.
4. Construct a hypothesis. In this step, the researcher states the research question as a hypothesis.  This provides the basis for all other decisions in the process and therefore, it is a critical step.
5. Determine the design of the research. The researcher should review the hypothesis and verify that an experimental design is the appropriate research design needed to answer the question. Additional information regarding different types of experimental research design will be covered in the next module.
6. Determine the research methods. In this step, the researcher will identify and plan the details necessary to conduct the research. This includes identifying the test subjects, materials, data collection instruments and methods, and the procedures for the conducting the experiment.
7. Conduct the research and test the hypothesis. The experimental procedures will be carried out in this phase.
8. Analyze the data. Experimental research data lends itself to a variety of potential statistical analyses. The appropriate analysis is determined by the research question and the type of data.
9. Formulate conclusions. Review the data and determine if it confirms or disproves the hypothesis.

Method Advantages Disadvantages

Survey

Advantages: Many people can be included. If given to a random sample of the population, a survey’s results can be generalized to the population.

Disadvantages:Large surveys are expensive and time consuming. Although much information is gathered, this information is relatively superficial.

Experiments

Advantages: If random assignment is used, experiments provide fairly convincing

data on cause and effect. 

Disadvantages:Because experiments do not involve random samples of the population and most often involve college students, their results cannot readily be generalized to the population.

Observation (field research)

Advantages: Observational studies may provide rich, detailed information about the people who are observed.

Disadvantages:Because observation studies do not involve random samples of the population, their results cannot readily be generalized to the population. 

Existing data

Advantages: Because existing data have already been gathered, the researcher does not

have to spend the time and money to gather data.

Disadvantages:The data set that is being analyzed may not contain data on all the variables in which a sociologist is interested or may contain data on variables that are not measured in ways the sociologist prefers.

Surveys

The survey is the most common method by which sociologists gather their data. The Gallup poll is perhaps the most well-known example of a survey and, like all surveys, gathers its data with the help of a questionnaire that is given to a group of respondents. The Gallup poll is an example of a survey conducted by a private organization, but sociologists do their own surveys, as does the government and many organizations in addition to Gallup. Many surveys are administered to respondents who are randomly chosen and thus constitute a random sample. In a random sample, everyone in the population (whether it be the whole US population or just the population of a state or city, all the college students in a state or city or all the students at just one college, etc.) has the same chance of being included in the survey. The beauty of a random sample is that it allows us to generalize the results of the sample to the population from which the sample comes. This means that we can be fairly sure of the behavior and attitudes of the whole US population by knowing the behavior and

attitudes of just four hundred people randomly chosen from that population.

Some surveys are face-to-face surveys, in which interviewers meet with respondents to ask them questions. This type of survey can yield much information, because interviewers typically will spend at least an hour asking their questions, and a high response rate (the percentage of all people in the sample who agree to be interviewed), which is important to be able to generalize the survey’s results to the entire population. On the downside, this type of survey can be very expensive and time consuming to conduct.

Because of these drawbacks, sociologists and other researchers have turned to telephone surveys. Most Gallup polls are conducted over the telephone. Computers do random-digit dialing, which results in a random sample of all telephone numbers being selected. Although the response rate and the number of questions asked are both lower than in face-to-face surveys (people can just hang up the phone at the outset or let their answering machine take the call), the ease and low expense of telephone surveys are making them increasingly popular. Surveys done over the Internet are also becoming more popular, as they can reach many people at very low expense. A major problem with web surveys is that their results cannot necessarily be generalized to the entire population because not everyone has access to the Internet.

Surveys are used in the study of social problems to gather information about the behavior and attitudes of people regarding one or more problems. For example, many surveys ask people about their use of alcohol, tobacco, and other drugs or about their experiences of being unemployed or in poor health. Many of the chapters in this book will present evidence gathered by surveys carried out by sociologists and other social scientists, various governmental agencies, and private research and public interest firms. 

Experiments

Experiments are the primary form of research in the natural and physical sciences, but in the social sciences they are for the most part found only in psychology. Some sociologists still use experiments, however, and they remain a powerful tool of social research.

The major advantage of experiments, whether they are done in the natural and physical sciences or in the social sciences, is that the researcher can be fairly sure of a cause-and-effect relationship because of the way the experiment is set up. Although many different experimental designs exist, the typical experiment consists of an experimental group and a control group, with subjects randomly assigned to either group. The researcher does something to the experimental group that is not done to the control group. If the two groups differ later in some variable, then it is safe to say that the condition to which the experimental group was subjected was responsible

for the difference that resulted.

Most experiments take place in the laboratory, which for psychologists may be a room with a one-way mirror, but some experiments occur in the field, or in a natural setting (field experiments). In Minneapolis, Minnesota, in the early 1980s, sociologists were involved in a much-discussed field experiment sponsored by the federal government. The researchers wanted to see whether arresting men for domestic violence made it less likely that they would commit such violence again. To test this hypothesis, the researchers had police do one of the following after arriving at the scene of a domestic dispute: They either arrested the suspect, separated him from his wife or partner for several hours, or warned him to stop but did not arrest or separate him. The researchers then determined the percentage of men in each group who committed repeated domestic violence during the next six months and found that those who were arrested had the lowest rate of recidivism, or repeat offending

(Sherman & Berk, 1984). This finding led many jurisdictions across the United States to adopt a policy of mandatory arrest for domestic violence suspects. However, replications of the Minneapolis experiment in other cities found that arrest sometimes reduced recidivism for domestic violence but also sometimes increased it, depending on which city was being studied and on certain characteristics of the suspects, including whether they were employed at the time of their arrest (Sherman, 1992).

As the Minneapolis study suggests, perhaps the most important problem with experiments is that their results are not generalizable beyond the specific subjects studied. The subjects in most psychology experiments, for example, are college students, who obviously are not typical of average Americans: They are younger, more educated, and more likely to be middle class. Despite this problem, experiments in psychology and other social sciences have given us very valuable insights into the sources of attitudes and behavior. Scholars of social problems are increasingly using field experiments to study the effectiveness of various policies and programs aimed at addressing social problems. We will examine the results of several such experiments in the chapters ahead.

Observational Studies

Observational research, also called field research, is a staple of sociology. Sociologists have long gone into the field to observe people and social settings, and the result has been many rich descriptions and analyses of behavior in juvenile gangs, bars, urban street corners, and even whole communities.

Observational studies consist of both participant observation and nonparticipant observation. Their names describe how they differ. In participant observation, the researcher is part of the group that she or he is studying, spends time with the group, and might even live with people in the group. Several classical social problems studies of this type exist, many of them involving people in urban neighborhoods (Liebow, 1967; Liebow, 1993; Whyte, 1943). In nonparticipant observation, the researcher observes a group of people but does not otherwise interact with them. If you went to your local shopping mall to observe, say, whether people walking with children looked happier than people without children, you would be engaging in nonparticipant observation.

Similar to experiments, observational studies cannot automatically be generalized to other settings or members of the population. But in many ways they provide a richer account of people’s lives than surveys do, and they remain an important method of research on social problems.

Existing Data

Sometimes sociologists do not gather their own data but instead analyze existing data that someone else has gathered. The US Census Bureau, for example, gathers data on all kinds of areas relevant to the lives of Americans, and many sociologists analyze census data on such social problems as poverty, unemployment, and illness. Sociologists interested in crime and the criminal justice system may analyze data from court records, while medical sociologists often analyze data from patient records at hospitals. Analysis of existing data such as these is called secondary data analysis. Its advantage to sociologists is that someone else has already spent the time and money to gather the data. A disadvantage is that the data set being analyzed may not contain data on all the topics in which a sociologist may be interested or may contain data on topics that are not measured in ways the sociologist might prefer.

The Scientific Method and Objectivity

This section began by stressing the need for sound research in the study of social problems. But what are the elements of sound research? At a minimum, such research should follow the rules of the scientific method. As you probably learned in high school and/or college science classes, these rules—formulating hypotheses, gathering and testing data, drawing conclusions, and so forth—help guarantee that research yields the most accurate and reliable conclusions possible.

An overriding principle of the scientific method is that research should be conducted as objectivelyas possible. Researchers are often passionate about their work, but they must take care not to let the findings they expect and even hope to uncover affect how they do their research. This in turn means that they must not conduct their research in a manner that helps achieve the results they expect to find. Such bias can happen unconsciously, and the scientific method helps reduce the potential for this bias as much as possible.

This potential is arguably greater in the social sciences than in the natural and physical sciences. The political views of chemists and physicists typically do not affect how an experiment is performed and how the outcome of the experiment is interpreted. In contrast, researchers in the social sciences, and perhaps particularly insociology, often have strong feelings about the topics they are studying. Their social and political beliefs may thus influence how they perform their research on these topics and how they interpret the results of this research. Following the scientific method helps reduce this possible influence.

*********************************************************************************