Descriptive statistics and inferential statistics
Descriptive statistics refers to the branch of statistics of describing, as the name suggests, the basic statistical features of a given set of data. It involves ways of obtaining certain measures on a set of data and representing the information to provide some kind of visual impression. In addition, it forms the basis for quantitative analysis.
The measures of central tendency like mean, mode, and median and the measures of dispersion like standard deviations are used in descriptive statistics (Shaughnessy, Zechmeister & Zechmeister, 2009, p239). Generally, measures of central tendency give the value around which a set of data is concentrated, while measures of dispersal describe how data is spread around the mean. The descriptive statistics are used when we are only interested in shedding light on the state of the data.
For instance, the total number of patients that visited a psychiatrist due to psychological disorders in one week was recorded to be 56. In the research report, it would be reported that the mean number of patients with the psychological disorders is 8. This implies that on the average, eight patients are admitted but of course with some deviations.
On the other hand, inferential statistics involves the analyses of the data that has already been provided to derive some conclusion that are not within the scope of the measures given (Shaughnessy, Zechmeister & Zechmeister, 2009, p255). Inferential statistics is used when we want to obtain or seek to inquire information that is not provided by the descriptive statistics.
A researcher would be seeking to know if individuals performing stressful duties are more vulnerable to physical illnesses. After performing the relevant tests for association and correlation, it would be reported in the research paper that there is an observed degree of correlation between the kind of stressful tasks performed and the cases of physical illness recorded.
The inferential statistics relies on the basic information that is obtained from the descriptive statistics. Most of the analyses involved in inferential statistics will need measures like the mean and the standard deviations of a given set of data (Shaughnessy, Zechmeister & Zechmeister, 2009, p240).
In essence, the inferential statistics attempts to make conclusions concerning a given set of data to a more general condition. On the other hand, descriptive statistics will be handy again in interpreting the results from inferential statistics (Shaughnessy, Zechmeister & Zechmeister, 2009, p300).
Case studies and single –subject experimental research designs
A case study involves a research carried out on an individual. It entails the description and analysis of the individual with respect to the variables under study (Shaughnessy, Zechmeister & Zechmeister, 2009, p.324). The data that is used in the case study research method can be obtained primarily through an interview, observation or mailed questionnaires or it may be obtained from secondary sources. The case study provides a psychologist with a special criterion for studying and understanding the behavior of individuals.
It differs from the experimental methods particularly due to its goals and the methods that are used (Shaughnessy, Zechmeister & Zechmeister, 2009, p.324). An example of a case study can be seen if a researcher investigates the early life of an individual who is currently an alcohol addict.
The parental care the individual received the group with which he associated, and the current family status of the individual. The case study method gives an exploratory approach to studying the behavior of individuals. In this respect, the case study method is used when the researcher seeks to develop a hypothesis for further study in the behaviors of individuals.
There are certain concepts learned from the case study approach. The approach is observed to be interrelated to the other research methods as it helps develop research hypotheses (Shaughnessy, Zechmeister & Zechmeister, 2009, p.326). The studies provide opportunities for developing new ideas about some character of an individual and sheds light on the possible causes of a given behavior. They provide some kind of evidence to support theories.
On the other hand, single-subject experiment involves the study of changes in behavior witnessed in an individual or a small group of individuals due to the different treatments that are given (Shaughnessy, Zechmeister & Zechmeister, 2009, p.335). In this case, the researcher is able to maneuver the variable under study and monitor the effects of such manipulations on the behavior of individual. This method is also exploratory and can be used instead of the case study due to the control that the researcher has on the treatment variables.
A researcher can decide to investigate the number of cases of irregularities recorded for some college student throughout the semester following the previous counseling programs that have been conducted. Due to the influence that the researcher has, the method is used to develop a unique solution to the problem affecting a particular individual. An important point to note with this approach is that the researcher is able to use the baseline details to predict the future treatment effects on some behavior.
Experiments
Experiments are research techniques involving the tests of hypotheses, are aimed at examining the causes of some phenomena, and can provide an explanation to the associated theories. The experimental approach seeks to put in place a cause-and-effect relationship between some variables under study. They seek to examine the effects of different treatments given to the subjects under study (Shaughnessy, Zechmeister & Zechmeister, 2009, p222).
A typical experiment is characterized by an intervention or treatment and the kind of control that the researcher has. A measure of comparison is also necessary. The researcher is able to manipulate the independent variable in order to develop an insight of the kind of relationship that occurs between the variables.
The internal validity of an experiment is characterized by certain conditions. These include covariation, time-order relationship and the absence of other possible causal measures. An experiment needs to test the hypothesis that there is a causal relationship between an independent variable manipulated by the researcher and certain dependent variable.
This is then followed by manipulation of the independent variable, while at the same time observing the corresponding changes in the dependent variable in order to establish a time-order relationship. Besides, an experiment will be effective if there are no other causative factors to explain the observed behaviors. This is ensured through proper control of the experiment.
Generally, ensuring that the experimental conditions are kept constants followed by balancing and manipulations are proper ways of ensuring that the threats to internal validity are eliminated (Shaughnessy, Zechmeister & Zechmeister, 2009, p223). A good experiment would be an investigation into the effectiveness of corporal punishment in correcting the behavior of a child. One group is subjected to corporal punishment while the other is left as a control and the effects noted.
The major difference that is observed between experimental and non-experimental design is that the former allows the researcher to manipulate the treatments and study their effects, whereas in the latter case, the researcher has little control on the treatments. Consequently, the non-experimental designs have the weakness that they are not able to explain the causes of associations that are observed.
Quasi-experimental designs
Quasi-experimental designs are similar experimental designs but lack the randomization that is performed in true experimental design. These experimental designs are important due to their simplicity and ease with which they can be performed. The main difference that is observed between quasi-experimental designs and the true experimental designs is this absence of randomization in the former (Shaughnessy, Zechmeister & Zechmeister, 2009, p364).
The true experiments are effective in studying the relationship of some variables. However, performing the true experiments in the natural setting (outside the laboratory) has some obstacles that often render it ineffective. Firstly, the researcher is faced with the problem in seeking permission to carry out the experiment in a natural setting (Shaughnessy, Zechmeister & Zechmeister, 2009, p354).
Secondly, getting access to the participants and explaining the purpose of the study is often a nightmare (Shaughnessy, Zechmeister & Zechmeister, 2009, p354). The quasi-experimental designs are used when the true experimental methods cannot be applicable or where they are very costly.
An example of quasi-experimental design is where a psychologist seeks to observe the relationship between the kinds of video shows that some group of children have watched in a period of two months and their behavior in relating to others in the society. The study records the number of romantic video shows watched by the individual students and a corresponding record also given for cases of immoral sexual behaviors observed for the respective individuals.
Reference
Shaughnessy, J., Zechmeister, E. and Zechmeister, J. (2009). Research Methods in Psychology. Eighth edition. New York: McGraw-Hill.