Human beings spend significant amount of their time into sleeping. During their sleep, human beings often experience the ‘magical’ state, the dream. In the dream, people maybe part of a Wonderful Wizard of Oz, horror stories or even reality shows. This chimerical state has caused dream researchers to question the cause of dream or even the contents of dream. Despite the fact that the state of dreaming has gathered researchers’ attention for long time, researchers have only begun to conduct thorough investigation on dreaming. However, as Robert E. Haskell (1986) states, “…The study of dreams and dreaming is an area of research that seems either to stimulate almost boundless, yet imaginative, theories regarding the meaning and function of dreaming; or conversely, it stimulates an equally boundless and irrational skepticism, denying any meaning or function to dreaming whatsoever…” (p.1). As stated, the field of dream research is very elusive. Pioneers in dream research, such Jung and Freud, developed various notions and models to explain the cause and content of dream, however, these theories and models are controversial and cannot be solidified. One of the most important factors that make dream research perplexing is the method of obtaining data. For example, as discussed later, participant drop out rate, small sample size and accuracy of data report hinder researchers from developing accurate notions on dreaming. In order to establish a solid explanation for dream, it is imperative that the methodology used for obtaining the data be improved. One of the many problems in obtaining dream reports is the drop out rate of participants. For instance, a study was conducted by Winegar and Levin (1997) in order to study the adolescence maturity. The study initially began with 182 volunteers of ages between 15 to 18. Of these students, 115 had turned in at least two dreams of 35 words or more. From this experiment, Winegar and Levin concluded that girls showed more maturity in relationships than boys and the difference was greater in the later years (Domhoff, P.50). The problem with this experiment is the drop out rate. Also, the length of dream description by participants imposes a problem of small sample size, but problem on sample sizes will be discussed later. According Cote et al. (1996) and Lortie-Lussier, Schwab & de Koninck (1985) describes that it takes atleast several weeks to obtain a minimum of two dreams from a participant (Domhoff, P.50). Hence, large drop out rate becomes likely because of the length that it takes for a participant to clearly identify their dream. This may cause frustration to the participants; as a result, their participation may become oblivious to them. Although a clear solution is hard to obtain for this methodology, a study by Tonay (1990) provides a possible solution. Tonay’s study conducted among students in University of California, Berkely, indicates that most of students who had chosen not to volunteer for a dream diary for 2 week, 43% of them said that they would continue the dream diary if it were part of a course assignment (Domhoff, P.51). Instead of having dream diary as a volunteering, it should be assigned as a task that may affect the participants, such as the grade, or even provide substantial amount of reward to the participants. In order to develop conclusions for studies that suffer from drop out rate, researchers should develop a conclusion after removing participants who dropped out of the experiment. However, most of the times, the result of the experiment may then suffer from small sample size.
The issue of drop out rate consequently leads to another problem that dream reports suffer from, the small sample size. Beck (1967)’s study in the effect of divorce, such as depression, on dreaming indicates the problem of small sample sizes. Initially, the experiment started with 300 participants, half women and half men, who were either divorced or formerly divorced. Of the 300 participants, only 87 gave the report. Furthermore, to make the matters worse, the number of participants who were depressed dropped significantly. Originally, 47% of the men and 51% of the women were classed as depressed; however, during the follow up, only 7.5% of the men and 6.3% of the women remained depressed. Conclusion has yet been developed for this experiment for it is still in progress (Bootzin, P.184). As seen from this experiment, the result of this experiment had suffered heavily from the small sample size. Again, this has contributed by the length of the experiment. Not only did the length of time discourage people from remaining in the study, but the time was also long enough to convert the mood of people, and the mood was an essential factor of the experiment. Although the solution to preserving the mood of participants of experiment may not be evident, in order to encourage the participants to remain in the study, as stated before, experimenter should reward substantially to the participants. This would allow the sample size to remain similar to that of initial. In order to develop conclusions for studies that suffer small sample size, researchers should build conclusion and re-attempt the identical experiment with new set of participants and develop conclusion based on that group. In the end, they should develop the final conclusion after examining conclusions of the same experiment. Among the small sample size and drop out rate, another important issue is the accuracy of report. Bernstein & Belicki’s (1995) study demonstrates the distortion of data by self reporting. In order to examine the rank-order correlation between expressed opinion and content analysis of social interactions, four different samples of college students were asked about the frequency of the dream content, Friendly, Aggressive and Sexual interactions. The result of survey was as following: 1.Friendly interaction, 2.Sexual interaction, 3.Aggressive interaction. However, through other analysis, such as the Sleep Laboratory, the accurate survey result was as following: 1.Aggressive interaction, 2.Friendly interaction, 3.Sexual interaction. Hence the correlation turned out to be negative (Domhoff, P.41). It is always a challenge for dream researchers to confirm the validity of dream reports. Dream cannot be directly observed by third person observation unless the participant reports it. However, the solution to this problem has been the Sleep Laboratory until the government decided to reduce the funding for dream research. In the laboratory, observers may record the sleep talk, which accurately describes the dream of participant’s. In addition, researchers may interrupt participant’s sleep during REM (Rapid Eye Movement) cycle, when the dreams are seen and remembered most vividly. Unfortunately, due to the high cost expense and reduction of government funding, this solution has become impractical and the alternatives are under investigation by dream researchers (Domhoff P.47). For now, the results from the laboratory should be used to further explore the dream states for those results are more accurate. The field of dream research is still at rudimentary level. It is once said by Haskell (1986), “…dream data have not historically been incorporated into cognitive psychology because the canons of scientific research excluded the adequate control and manipulation of such data…”(P.3). It is dream researchers’ objective to build the model that can wholly illustrate dreams. In order to do so, like all the other scientific procedures require, an approved method, in which flaws are minimally presented, must be present. The dream researchers’ main goal for near future may have to be to establish procedures that would minimize most of the flaws, such as drop out rate, small sampling size and accuracy of data.