¶ … psychologists to study behaviors, are unique in that the studies use small groups of individuals, rather than large samples. Through the use of an in-depth, longitudinal examination of a single instance or event, this method can lead to a deep understanding of why a certain event occurred. Further, the researcher can examine what possible issues he or she may need to examine extensively in future, larger studies (Miles, et al. 1984). For example, a researcher could use case study examinations to determine possible reasons for increased anger behavior in elderly women recently admitted to hospitals. The process would begin with a careful selection of a small group of individuals who fit the above criteria. Generally, there small groups would include ten or less subjects (Miles, et al., 1984).

The researcher would then fully review the patients' histories. This would include conversations with hospital staff, interviews with family, written history from previous physicians or mental health representatives, and interviews with the patients. Additionally, the researcher would examine all previous medication schedules, dietary plans, living conditions, and health conditions.

Following this in-depth analysis of all applicable data, the researcher would then attempt to find similarities between the cases that could represent reasons for increased anger behaviors. If, for example, the research found that the medication needs of many of the patients were not followed by hospital staff, this may lead to a possible correlation between medication changes and anger behavior. Conversely, if the researcher noted changes in dietary behavior which altered the patients' routine, he or she may infer that these changes correlate to the changes in behavior. By examining a small group of subjects in-depth, the researcher will have a better understanding of the full range of possible factors relating to these behavioral changes.

Traditionally, researchers have used a literature review to discuss the finding of other studies on the topic they are choosing to study. These literature reviews begin with the gathering of all available studies pertaining to the topic. Once all viable information has been gathered, the researcher then chooses which other studies to use within their own work. These studies are then reviewed, summarized in narrative style, and discussed in terms of their key points which relate to the topic at hand (Light, et al., 1984).

The problem, however, is that these reviews are very subjective in nature. While the strengths and weaknesses of some previous research on the topic are clearly identified, this identification is limited to the researcher writing the literature review. Further, since not all research on a topic can be used within a given literature review, the writer of the review also must use their own subjectivity in choosing which studies to include. Without clear and precise procedures for minimizing this level of subjectivity, the literature review can lose merit (Light, et al., 1984).

The meta-analysis, on the other hand, provides a method of using "statistical analysis of a large collection of results from individual studies for the purpose of integrating the findings" (Glass, 1976, p. 5). Since the results from various studies investigate different dependant variables, measure on different scales, the meta-analysis compares the results of each study using a standard measure of effect size. This can be done with any study that describes findings using correlation coefficients. Since standard literature reviews do not take sample issues into effect, the meta-analysis often measures how much difference exists in any study between the trial group and the control group. This analysis is done on each study to be used in the review. In this way, researchers are able to effectively use only those studies which truly show a correlation (Glass, 1976).

A prime example of an effective meta-analysis is that of the 2003 meta-analysis by Norman Staunton, "A Meta-Analysis of Adventure Therapy Outcomes." The goal of this analysis was to effectively determine whether adventure therapy was effective, and if so, which types of adventure therapy were most effective. Staunton searched electronic databases, bibliographies, and adventure therapy websites, and used common research, listserv requests, and existing data for his meta-analysis. Only those studies which were empirically based and used descriptive statistics were used. Additionally, only studies examining adventure therapy with diagnosed populations were included. Using these criteria, the selection was limited to 17 studies (Staunton, 2003).

Staunton then calculated the effect sizes for each study, and correlated them based on the relationship between effect size and primary and secondary variable. He used regression analysis to ensure all returned significant correlates....

Additionally, the study used the file drawer effect to estimate the number of non-significant effects that would negate findings (Staunton, 2003).
The end result of the mea-analysis included 17 studies, using calculations of 95 effect sizes, with no identification of negative effects. Significant effects were found with dissertations, medium and high quality, ABC programs, continuous programs, intermittent programs, residential programs, single sex and co-ed groups, adolescent and adult groups, and self construct groups. Only outpatient programs were found to have non-significant effect sizes. The highest significance was found with ABC programs and mixed diagnosis programs (Staunton, 2003). Staunton thus concluded, using meta-analysis, that the above adventure therapy programs were effective.

While questionnaires are certainly a useful tool in psychological research, there are certainly potential problems with the method. The primary challenge is to avoid common mistakes when creating the questionnaire items. If flaws within the items exist, the questionnaire becomes a no-viable tool (Whitney, 1972).

One major flaw can be found if questions are worded which allow inference as to a specific response being more favorable than another. Since questionnaires rely on the subject's inferences when testing, it is vital to avoid any wording of questions which can lead to an inference of 'goodness' with a specific answer. To avoid this problem, or to verify results, researchers can use like-question validity checks. By asking the same question more than once, researchers can determine whether a question's wording is appropriate (Whitney, 1972).

Another flaw of questionnaire items is the ambiguous question. Again, since the questionnaire's validity is directly related to the ability of the subject to respond to the questions, providing clear and concise items is vital. If the subject does not comprehend the question, the questionnaire loses merit. To avoid this issue, performing a pilot test using colleagues or other professionals prior to actual distribution to subjects can be beneficial. By having someone else examine the questions, researchers can discover potentially ambiguous questions prior to the actual study, and can correct them (Whitney, 1972).

Still another flaw with questionnaire items can be the open-ended answer. By leaving a questionnaire item open, rather than including a specific set of multiple choice answers, the researcher can cause problems with the coding of responses. Additionally, since the subject is allowed to write their own answer, the answer given may not fit into the coding strategy. To avoid this, it is advisable to a specific set of choices for the respondent to choose from (Whitney, 1972.)

If we are interested in the effects of Drug A on learning, we must design an experiment to test the effects of that drug. In order to test this, we use a one-factor experiment. For this experiment, we will need two groups of subjects, those of the control group, and the experimental group. The control group will receive a placebo, while the experimental group will receive Drug A. The purpose of the control group is to test for a placebo effect. The dependant variable is the test scores of subjects, and the independent variable is the drug.

We would begin by identifying the null hypothesis. For this experiment, the null hypothesis is that there is no difference in the mean test scores between groups. The experimental hypothesis is that there is a difference in the mean test scores between groups.

Next, we would need to choose subjects. In order to control for the effects of age and gender, we choose subjects who were male, age 23 to 25. Additionally, to control for sociological effects, all subjects are chosen from a similar pool, such as college students. We require all subjects to go through a complete physical to ensure the drug will not cause potential, known issues. Once selected, we assign 100 subjects at random to either the control group or the experimental group. Those in the control group act as a comparison against the experimental group. Randomization is used to ensure validity of the findings.

Each group of subjects is brought to a testing lab, where they are given a list of 10 words, and their corresponding definitions. The words are chosen at random from The Dictionary of Uncommon Words (Urdang, et al., 1991). Subjects are given one-half hour to learn the definitions of each word. Following the learning period, each subject is given a written examination, consisting of seven of the words from the previous list. Beside each word are four definition choices. Subjects are told to choose which definition is correct.

Following the pre-test, subjects in the control group are given a placebo for a period of five weeks. The subjects in the experimental group are given…