The data collection process

¶ … data collection process is an integral part of performing research studies -- particularly for the field of nursing. There are a number of different options of data collection methodology that researchers must consider before they begin this process. On a fundamental level, the methodology employed by a particular researcher will largely depend on the type of study that he or she is conducting, the specific research question, hypothesis, and other factors that are relevant for such a project. However, there is still a basic paradigm that virtually all researchers should follow when attempting to collect data. The first step in this paradigm is to devise a data collection plan, which largely involves analyzing the particular needs of the researcher's study and prioritizing those needs in such a way that they are addressed as expediently and prudently as possible. Next, individuals must prepare the particular instruments they will utilize during their study, which primarily depends on the needs that were assessed. Finally, there are a host of data collection methods that are available to researchers. Generally, however, these methods are distinguished between those that are structured or unstructured, and are in no small part determined by whether one's research is quantitative or qualitative.

Although there are plenty of opportunities for researchers to engage in qualitative research no matter what particular field of study they are involved in, there is a definite tendency of nursing research to be qualitative in scope and in focus in order to elucidate the most common type of issues studied within this profession. In support of this fact is the reality that "most nursing research studies involve the use of self-reports" and that "self-reports are the most frequently used type of data collection method in qualitative studies" ( Polit & Beck, 2008, p. 390). If self-reports, which are principally different varieties of interviews and questionnaires are the primary method for used for data collection in qualitative nursing research studies, the secondary method is observation -- of which there are also several varieties.

However, in order to successfully determine the type of self-report and method of observation that will yield the most benefit to a research study, researchers must first assess and hierarchize the needs they have for their data. In consideration of these needs, it is important for researchers to consider the type of variables that they will be testing, and also attempt to descry any potential biases that have the potential to compromise the integrity of their data. An integral component of the analysis necessary for a data collection plan is the consideration of administrative needs and pragmatic concerns such as costs and ability to analyze and actually use data in a timely fashion.

Once a team of researchers has determined just what exactly it needs from its data, it must then ascertain how best to get that data. Specifically, this step calls for the implementation and possible creation of instruments that will be used in the collection process. Since most nursing studies will inevitably utilize some form of self-reporting method or -- and possibly in conjunction with -- observation, creating specific tools for use in the data collection process may very well be feasible. However, it is essential for researchers to thoroughly test these instruments before employing them in the actual data collection process in order to minimize the threat of compromised data and to ensure they actually measure the construct that the study is based upon (Kimberlin & Winterstein, 2008, p. 2280). Additionally, specific members of the research team that will collect data should become acquainted with all instruments and procedures before the collection process actually begins.

One of the principle reasons that self-reports are so valuable to research in the field of nursing is because they allow for a free exchange of information from sample participants to the researcher that the latter may choose to filter in any way he or she feels is most effective. In terms of their actual implementation, however, researchers have the option of utilizing either unstructured or structured methods that can regulate how much, and what sort of information is gained via these reports. Self-report data collection that is either loosely structured or unstructured can produce a copious amount of information from members of a sample population, although the different types of reports associated with this method can require inordinate amounts of time both to implement and analyze. This is largely due to the fact that the data from these reports is determined almost solely by the restrictions of the individual subjects of the sample and can be as voluble or as terse as such a subject himself or herself is. One of the principle unstructured self-report methodologies is that of interviews, which can take the form of individual or group interviews, encompass aspects of their life history or include detailed entries in journals (Polit & Beck, 2008, p. 411). These interviews may also be in the form of participation-observation research, in which researchers "have firsthand involvement in relevant activities" (Levine et al., 1980, p. 38) of the lives of their subjects

In order to perform these self-reports effectively researchers and data collectors must be well-versed in the subtleties required to get subjects to provide adequate data, which requires a skill-set that usually develops over time. Interviews, however, also have certain inherent interview biases that can potentially compromise the integrity of the data gained from them. Researchers can easily transform unstructured self-report methods to structured ones by utilizing a number of instruments that can funnel the amount and type of information into a way that is most efficacious for a particular study. Examples of structured self-report methods include interviews, questionnaires and composite psychological scales, all of which can be tailored to meet specific research objectives by posing the questions for them in an open-ended fashion. The myriad types of stratifications for open ended questions include those that utilize a matrix questions, forced choice questions, cafeteria questions, dichotomous questions, and rank-order questions (Polit & Beck, p. 2008, 445). All of these questions limit the structure of a questionnaire or interview so that respondent must only give information that researchers specifically ask for. This methodology advantages certain types of studies, whereas unstructured self-reports can provide valuable information that researchers were not aware of that may influence their study. However, since researchers should adequately prepare their data collection methods to meet the needs of a study that is ideally founded upon a research review and probes into a particular area of research, structured methods of data collection should benefit most nursing studies. Still, it may not be possible to grasp the entire scope of a study using structured measures (Chowdhurry et al., 1998, p. 73).

As previously denoted, observation is frequently used in conjunction with self-reports in nursing studies. The methodologies for employing observation to attain data can also be categorized into both structured and unstructured methods. Regardless of which type of observation is employed, researchers should be cognizant of observation biases that largely stem from the effect of their presence (or the knowledge of subjects that researchers are observing them) and how that can compromise the authenticity of data. Observation is typically recorded via the usage of logs or field notes, particularly for unstructured observation methods (Polit & Beck, 2008, p. 412) which encompass elements of comparisons between subjects (and between subjects and controls) as well as observations of specific occurrences within the data collection process. Observations are extremely complementary for certain self-report methods such as interviews, and can provide a breadth of data that otherwise may not be possible.