The Greatest Goal of Social Researchers - Essay Example

Chapter Seven Summary The greatest goal of social researchers is to observe people and events that are seldom observed and therefore not understood.It is because of this lack of observation that researchers are unable to make sense of problems or concerns, or else do not have the information they could use to help them in another area of research. By observing and investigating that which is usually ignored or misunderstood, unique insights can be made in regard to these unknown areas. Part of this involves sampling, or a selection of people to observe. Since sampling is an important part of observation, it is vital that there is an understanding of the logic behind sampling. Though probability sampling using precise statistical techniques are better to use, there are times when non-probability techniques are more appropriate. One of the reasons that non-probability tends to be less accurate is because it is usually based on the available subjects, which may not be an accurate representative of the population being sampled. In researching and observing, especially when it involves trying to understand a certain group of people, it is important for the researcher to ensure that they sample group they have is an honest representative of the population in general, or otherwise what is observed has the possibility of being false. Probability sampling techniques, however, make it easier to select a proper sample that represents large populations. This is done through random selection, which means that there is no conscious bias when selecting the sample. This improves the chance of getting a more accurate sample regardless of how small the sample is. It must be noted that there is no such thing as a perfect representation of the population; there will always be, to some degree, sampling error. Probability sampling just decreases these chances more than non-probability sampling. If the researcher is able to predict the distribution of samples with respect to the target parameter, it becomes possible to estimate the amount of sample error in any given sample. Since error is expected, it is used to express confidence levels and confidence intervals. There are many types of sampling designs that are available to researchers. The most basic of random sampling tends to be the must fundamental technique in seldom, even though it is seldom used in research and observation. The most practical method of sampling is systematic sampling, which involves selecting every kth member from a sampling frame. This ensures that there is no bias in the selection of the sample. Stratification is the sampling technique that involves grouping together members of a population into homogeneous strata before sampling, which helps to improve the representativeness of a sample by lessening the possibilities of sampling error. A multistage cluster sampling is a complex sampling technique, but is used quite frequently when the study does not come with a list of all of the members of a population and such a list is unable to be obtained. Due to not having this list, researchers need to balance the number of clusters and the size of each cluster in an attempt to achieve a certain sample size. Stratification can be used during this time to reduce sampling errors that may arise due to the lack of knowing the members of a population. Probability proportionate to size is an efficient technique that researchers can use for multistage cluster sampling, which further helps to ensure accuracy and unbiased sampling. As with any part of research and observation, even if it is just in the writing process of the research article, ethics must be observed and ensured to those in the experiment and those that are reading about it. In regard to those that are reading the research articles, the researcher must make it known that there is always the probability for errors that might make the results misleading. Science is not always an exact art, since other researchers and scientists are able to work off of previous studies to simply further the researcher. As such, a lot of this further research stems from inaccuracies or gaps in information in previous studies. Furthermore, when non-probability sampling methods are being used to obtain the breadth of variations in a population, the researcher needs to make it clear that the readers are not misled in regard to the variations and what is typical in the population. Sampling is perhaps one of the more important aspects of a research study. The sample acts as the population as much as possible, either comparing behaviors or beliefs, or seeing how often something is undergone, all depending on the purpose of the study. Without the sample population, researchers cannot even begin to fathom the results of their research questions. Therefore, it becomes important that researchers understand the various types of sampling techniques that are available to them, and they need to be aware that they will always face errors in their research and that there are methods that can be undergone to decrease how many. Chapter Eight Summary Experimentation is the heart and soul of any research or observation project. By conducting an experiment, researchers are able to focus on the important aspects of their research question, wether the variables are dependent or independent. In experiments, the researcher gathers a selected group of subjects, one that is based on the purpose or intended result of the experiment, does stuff to them (the experimental part) that can either provoke a wanted response or observe natural behaviors without outside influence, and then observe the effect of what took place. Many topics can use experiments to understand old concepts, discover new information, or test information that is already available. Experiments provide a wonderful chance to implement controlled testing of casual processes. Though there are different methods of experiments and more are being designed as science and research evolves, the classic experiment is the one that is used most, being the tried-and-true method. The classical experiment involves the testing of an independent variable, or the experimental stimulus, on a dependent variable through various levels of testing of experimental and control groups. The experimental and control groups must be similar to each other, something that is of great importance to the experiment. However, it is not as important for the experimental subjects to be representative of the larger population. Double-blind experiments, a special technique, disables experimenter bias, which causes the experimenter and the subject to be unaware of which subjects are in the control and experimental groups. Campbell and Stanley are the most vocal about experiments and their various forms. They describe three different forms of pre-experiments, and they are as follows: the one-shot case study, the one-group pretest-posttest design, and the static-group comparison. Within these, Campbell and Stanley also point out eight important sources of internal invalidity in experimental design, and they are as follows: maturation, history, instrumentation, testing, statistical regression, experimental mortality, demoralization, and selection biases. Campbell and Stanley go on to discuss some of the problems that researchers may have with experiments, with the largest one being external invalidity, which is when the experimental findings do not entirely or truthfully reflect real life. By implementing the Solomon four-group design or other versions of the classical experiment, the researcher can safeguard against invalidity. Furthermore, Campbell and Stanley agree that as long as there is proper randomization in the assignment of the subjects to the experimental and control groups then there is no purpose for pretesting in experiments. Experiments originally started off as natural experiments, or experiments that often occur in the course of the social life in the real world, and that social researchers can use in similar ways that they would design and conduct laboratory experiments. Due to the amazing changes in technology, the Internet has become a common source for performing social experiments. It is easier to access more accurate sample groups and subjects, as well as using the Internet itself to research further information. Also, with how common and implemented the Internet is in the daily lives of millions of people, it provides researchers with even more concepts and curiosities to study on, such as the use of the Internet. Part of experimentation is deceiving the subjects. To get the best results, most of the subjects are either unaware that an experiment is taking place or, if they are aware that they are part of an experiment, they are not confided in with what roll they are playing. A lot of measures have to be taken to ensure the most accurate results and these often include keeping the subjects in the dark. However, ethics need to be taken into consideration to make sure that no rights of the subjects are violated. Experiments allow the possibility of causing damage to the subject. As such, researchers need to guarantee that the experiments that are undergone will not harm the subject, and if there is a chance that it could, the subjects need to be told and given the chance to opt out of the experiment. In the cases that subjects are aware of being in experiments, they have to be given the chance to not participate or to participate anonymously. It is important that the subjects remain safe from every and any dangerous aspect of the experiment. Experiments are excellent ways - and often the only ways - to fully observe and understand something, especially in regard to people and behaviors. Experiments often involve a few different variables - the independent variable, the dependent variable, and the control group. They all compliment each other in the experiment and move the research forward. There are a variety of different experiments, all of which are dependent on the type of research being undergone. Since most experiments are done out of deception to the subjects, it is vital that their rights are taken into consideration. Read More

A multistage cluster sampling is a complex sampling technique, but is used quite frequently when the study does not come with a list of all of the members of a population and such a list is unable to be obtained. Due to not having this list, researchers need to balance the number of clusters and the size of each cluster in an attempt to achieve a certain sample size. Stratification can be used during this time to reduce sampling errors that may arise due to the lack of knowing the members of a population.

Probability proportionate to size is an efficient technique that researchers can use for multistage cluster sampling, which further helps to ensure accuracy and unbiased sampling. As with any part of research and observation, even if it is just in the writing process of the research article, ethics must be observed and ensured to those in the experiment and those that are reading about it. In regard to those that are reading the research articles, the researcher must make it known that there is always the probability for errors that might make the results misleading.

Science is not always an exact art, since other researchers and scientists are able to work off of previous studies to simply further the researcher. As such, a lot of this further research stems from inaccuracies or gaps in information in previous studies. Furthermore, when non-probability sampling methods are being used to obtain the breadth of variations in a population, the researcher needs to make it clear that the readers are not misled in regard to the variations and what is typical in the population.

Sampling is perhaps one of the more important aspects of a research study. The sample acts as the population as much as possible, either comparing behaviors or beliefs, or seeing how often something is undergone, all depending on the purpose of the study. Without the sample population, researchers cannot even begin to fathom the results of their research questions. Therefore, it becomes important that researchers understand the various types of sampling techniques that are available to them, and they need to be aware that they will always face errors in their research and that there are methods that can be undergone to decrease how many.

Chapter Eight Summary Experimentation is the heart and soul of any research or observation project. By conducting an experiment, researchers are able to focus on the important aspects of their research question, wether the variables are dependent or independent. In experiments, the researcher gathers a selected group of subjects, one that is based on the purpose or intended result of the experiment, does stuff to them (the experimental part) that can either provoke a wanted response or observe natural behaviors without outside influence, and then observe the effect of what took place.

Many topics can use experiments to understand old concepts, discover new information, or test information that is already available. Experiments provide a wonderful chance to implement controlled testing of casual processes. Though there are different methods of experiments and more are being designed as science and research evolves, the classic experiment is the one that is used most, being the tried-and-true method. The classical experiment involves the testing of an independent variable, or the experimental stimulus, on a dependent variable through various levels of testing of experimental and control groups.

The experimental and control groups must be similar to each other, something that is of great importance to the experiment. However, it is not as important for the experimental subjects to be representative of the larger population. Double-blind experiments, a special technique, disables experimenter bias, which causes the experimenter and the subject to be unaware of which subjects are in the control and experimental groups. Campbell and Stanley are the most vocal about experiments and their various forms.

They describe three different forms of pre-experiments, and they are as follows: the one-shot case study, the one-group pretest-posttest design, and the static-group comparison.

Read More