Research methods and applications in academic study

¶ … new reading program on a student's ability to learn to read. Because the program is designed specifically for helping new readers learn basic reading skills the experimenter chooses only beginning first-grade students as the population of interest. Thus, as the subject variable is one that cannot be the target of random assignment (you cannot randomly assigned subjects to the first-grade) a nonequivalent control group pre-test post-test quasi-experimental design is selected as the preferential design (Cozby & Bates, 2012). Moreover, it is often not feasible for programs like this to randomly assign different students within the same classroom to different breeding programs as there is typically only one teacher teaching the first graders and such a design would be cumbersome and most likely the study rejected by the school system. Thus, the incoming first-grade students are given a basic reading pretest at the beginning of the school year, and then one section or one of the first-grade teachers in a school uses the traditional reading instruction for first graders (control group), whereas a different teacher in a different classroom uses the new reading program (treatment group). At the end of the semester the two different classes are evaluated on their reading abilities in reading progress in order to determine the effectiveness of the new reading program.

The major threat to the internal validity of study include the fact that different teachers are using the different programs, thus there is the potential for some of the differences in the outcome to be due to the teacher's style of teaching, the particular classroom atmosphere, etc. rather than the reading program (Thompson, Diamond, McWilliam, Snyder, & Snyder, 2005). One way to control for this threat by using a true experiment would be to randomly assign students in each classroom to either the treatment or control condition and then have the teacher teach groups of students differently; however, as discussed above this would be very cumbersome.

Part 2. Not all research questions can be answered by true experiments. Typically researchers are interested in different types of subject variables and how these variables affect the behavior of the individuals that possess those (Cozby & Bates, 2012). For instance, in research attempting to assess how individuals with different political affiliations view different types of social issues survey research (correlational research) is necessary to help answer these questions. A researcher may be interested in how individuals with either a conservative or liberal political viewpoint feel about violence on television. The researcher could administer surveys to individuals that would include questions regarding their stance on certain political issues and their stances on certain levels of violence in the media.

Another classic example of correlational research is research on minimal groups and how individuals who belong to a particular group perceive other individuals who share the same group membership or different group membership (Cozby & Bates, 2012). While it is possible to manipulate group membership by creating pseudo -- groups that do not exist or that have no meaning (such as individuals who like certain styles of paintings) the inherent tendency to attribute favorable characteristics to people one believes are like them and less favorable characteristics to people one believes are not like them is an inherent quality that is not manipulated experimentally. What is being measured in research like this is the basis for the development of prejudice and in -- group favoritism.