Descriptive and Inferential Statistics - Essay Example

? Research Methods and Statistics: Short Answers Prepare written responses to the following questions. The response to each question should be at least 250 words in length.   a)     What are the similarities between descriptive and inferential statistics? What are the differences? When should you use descriptive and inferential statistics? Descriptive statistics are a set of techniques that help in describing the data at hand in the most effective and detailed manner as possible. Inferential statistics on the other hand are techniques that help the researcher to explore the trends that exist in the population with some confidence on the basis of the collected sample (Howitt & Cramer, 2008). Both sets of techniques are similar in that they describe the data collected, and help in evaluating the similarities and differences between the groups being studied. Both inferential and descriptive statistics help the researcher identifying and exploring the trends observed and to make sense of the relationships that are shared by the variables being studied. The main difference between descriptive statistical techniques and inferential ones is that while the descriptive statistics provide information that is limited to the data available, inferential techniques allow the researcher to verify if the observations made from the sample collected may be considered representative of the population from which the sample is drawn (Howitt & Cramer, 2008). While descriptive statistics provide an exact description of the data used, inferential statistics provide an estimate about a larger group. Descriptive statistics are used when it is important to describe the similarities and differences in different groups, to explain the trends observed in the data for a particular group of subjects, and to provide information that may be used to develop hypotheses about the population (Howitt & Cramer, 2008). When all relevant subjects are included in the study or when the group being studied is very specific, descriptive statistics help in understanding the group better than other techniques. On the other hand, when the population being studied is very large, and it is not possible to study all members for reasons of practicality (like necessary time, finances and value of the study) it is preferable that the researcher carefully selects a small sample and uses inferential statistics to make estimates about the population (Howitt & Cramer, 2008). b)    What are the similarities between case studies and small-N research designs? What are the differences? When should you use case studies and small-N research designs? Case studies are usually conducted using very few participants, and could also be focused on a single participant (McBurney & White, 2009). These studies describe the experiences of the individuals included in natural, clinical or experimental conditions in great detail, and use techniques of describing data that are both qualitative and quantitative (Christensen, Johnson & Turner, 2010). Case studies are used to describe each individual studied so that the particular aspects of the variables being studied may be discussed. Typical examples of case studies would describe an individual’s pertinent background, their experiences and the way they respond to the chosen stimuli within a setting. While case studies are considered as a type of small-N research design, not all the Small-N studies describe the particular individuals as much as the case study. Some small-N designs are extremely quantitative and although the data collected is from a very small sample, the research study follows a very experimental design (McBurney & White, 2009). Small-N designs are useful when the researcher is trying to evaluate a rarely occurring condition or when a new treatment procedure is being tried out. Case studies, small-N studies and single participant research have been found to provide valuable insights in the fields of psychophysics, clinical research and cognitive psychology (McBurney & White, 2009). Although the main criticism about these studies is that they lack both internal and external validity (McBurney & White, 2009), results from such studies have allowed the researcher to develop hypotheses about a larger population as well as develop theories of behavior and methods of treatment for clinical participants. c)     What are true experiments? How are threats to internal validity controlled by true experiments? How are they different from experimental designs? True experiments are a kind of research design that are most favored by researcher as they provide the most trustable data among all the types of research designs (McBurney & White, 2009). It is important to differentiate between true experiments and experimental designs. An experimental design is the manner in which the researcher attempts the study the research problem at hand. A good research design allows the researcher to use a control group as well as the ability to randomly assign participants to groups (Christensen, Johnson & Turner, 2010). While a true experiment showcases the two characteristics that are important to a good experimental design, other research designs may be developed keeping in mind the needs of the research question. These designs may allow the researcher to either assign subjects randomly or to use a control or baseline group. True experiments are research studies that are completely controlled by the experimenter (McBurney & White, 2009). In a true experiment, the researcher is able to assign participants randomly to different research groups, use a control group to verify the baseline responses and manipulate the independent variable so create as many levels or conditions as necessary (Christensen, Johnson & Turner, 2010). A true experiment also allows the researcher to control the extent to which known extraneous variables affects the study results. Although it is not possible to control all external factors completely, it allows the researcher to gauge the extent to which these factors may play a role and to control them or to measure them systematically. Internal validity is the extent to which study results truly reflect the relationship shared by the variables (McBurney & White, 2009). By allowing the researcher to assign subjects randomly and by comparing the results with a baseline group, a true experiment reduced the chances of systematic errors affecting the data (Christensen, Johnson & Turner, 2010). Thus, a true experiment is most likely to reflect the nature of the relationship between the independent and dependent variables. d)    What are quasi-experimental designs? Why are they important? How are they different from experimental designs? Quasi – experimental designs are research designs that allow the researcher to control some aspects of the experiment being conducted, but do not allow for random assignment of subjects (Christensen, Johnson & Turner, 2010). It is possible for the researcher to use a control or baseline group and to control some extraneous variables. This may be done either by setting limits during sample collection, so that only participants who fulfill certain criteria are included in the study, or by statistically controlling the effects of a known confounding variable. Although these designs seem weaker than true experiments, they provide valuable information that cannot be studied using true experimental designs (McBurney & White, 2009). Typical conditions where quasi experimental designs are used are ones where the effect of gender, age or other such variables needs to be verified. Quasi-experiments allow the researcher to study naturally occurring groups that share some significant distinction, and also provide results that are likely to have higher external validity as compared to true experiments (Christensen, Johnson & Turner, 2010). A well designed quasi-experiment resembles a true experiment in most details except for the aspect which is not under the researcher’s control. On the other hand, quasi-experiments have some significant weaknesses, and thus should be avoided if true experiments are possible (McBurney & White, 2009). Results of a quasi-experiment are more susceptible to the effects of confounding factors since random assignment of participants to experimental conditions is not possible (Christensen, Johnson & Turner, 2010). These studies typically have lower internal validity as compared to true experiments. This is because it is difficult to ensure that no variable associated with the non-manipulated variable has affected the results obtained by limiting or exaggerating the effects of the independent variables. References Christensen, L. B., Johnson, R. B. & Turner, L. A. (2010). Research methods, design, and analysis (11th ed.). Boston: Allyn and Bacon. Howitt, D. & Cramer, D. (2008). Introduction to statistics in psychology. (4th ed.) New York: Pearson. McBurney, D.H. & White, T.L. (2009). Research Methods, (8th ed.). Belmont, CA: Wadsworth/Thomson Learning. Read More