Research critique methodologies and applications

Nursing Research Report

The structure of a research report is simple. It is almost the same as the structure of the research itself: the problem, the methodology, the results, the conclusions, and the interpretations. The purpose of the research report is to inform readers about the problem investigated, the methods used to solve the problem, the results of the investigation, and the conclusions inferred from the results (Polit & Beck, 2004). The function of the research report is not, however, to convince the reader of the virtue of the research previously conducted. At all times the report is to be so written that the reader can reach their own conclusions as to the adequacy of the research and the validity of the reported results and conclusions. Research reports must strive to be succinct, objective, and crystalline. The ultimate test of an excellent research report is in its ability to be replicated by those who read it. If this criterion cannot be met then the report is inadequate in terms of knowledge transfer.

Within the field of nursing research is a vital part of providing optimal patient healthcare. With properly executed research, albeit in the area of critical care, ER medicine, pediatrics, or administrative practice the data collected from effective research provides the theoretical and applied sustenance for the improvement of nursing practice. Improved practice not only relates directly to the nurses themselves but also provides the patient, the ultimate benefactor, improved medical care through research expertly conducted. Although a critical evaluation of all possible research in every area of nursing would most definitely go beyond the scope of this assignment, one particular area is worthy of additional investigation; namely, potential changes in the perception of nursing work groups and nursing leaders visa via a selected professional practice model. In order to determine whether or not this area of investigation was fairly evaluated a research article entitled "The Effects of a Professional Practice Model on Staff Nurse Perception of Work Groups and Nurse Leaders" by Ingersoll, Schultz, Hoffart, and Ryan (1996) was put to the critical test. The research endeavor and resulting research report with be critically evaluated and scrutinized following the Polit and Beck (2004) evaluation formula. The Polit and Beck evaluation formula pays close attention to the following: a study's empirical development; supportive literature for the basis upon which the study was designed; the theoretical framework encompassing the research; the study's design, data collection, data quality, and data analysis procedures; maintenance of ethical standards; the representation of results; the strength for knowledge inference based on data interpretation; and finally, a completeness rating and a hierarchical placement position.

Research Problem(s), Research Question, and Hypothesis Research activities, whether clinical trial based, focus group oriented, or historical in design, must exhibit and command interest, enthusiasm, and passionate commitment. To the enthused researcher there must be debate, discussion and even argument if there is to exist intelligent conviction regarding the nature, design, analysis, and inference of the phenomenon under investigation. The introductory phase of the Ingersoll and Beck study fell completely short of giving the reader any indication of passion for what had taken place let alone adhere to the rigors of best-fit research by first stating a research problem and hypothesis. In fact the authors dove right into the inner workings of the study by relating to the reader the intended design and from where the sample was derived: "To determine the effect of and enhanced professional practice model (EPPM)... Two experimental units within the medical center...(p.2)." Before any attempt is made to state a design there must exist an expressed need for the study, a direct statement of the null hypothesis to be tested, and at what probability level of acceptance, and a full definition of the variables being investigated - including a definition of all necessary terms such as perception, work environment, EPPM, urban medical center, rural hospitals, and community hospitals.

The most significant error committed by the investigators in their opening remarks was to state "The study was also expected to clarify...increased staff nurse control over work environment... was expected to produce...(p.3)" totally denies the use of ethical research standards. At no time is a researcher to present biasing or directional statements anywhere within a research report. Doing so introduces investigator bias untold. The investigators should have turned these biasing statements into testable null hypotheses such as: There will exist no statistically significant difference, at the alpha level of 0.01, in the effects of EMMP on a nurses perception of the work environment (work group and ideal) as measured by the Perceived Group Attractiveness and Cohesive Scale and the Leadership Opinion Questionnaire. Form this main effect's null hypothesis the second, third and interaction level null hypothesis could have been formulated. By not fulfilling the aforementioned requirements the authors did little to enforce the rules applied to experimental research.

Notwithstanding the fact that the researchers did state three hypotheses, but positioned within the report wrongly (i.e., after the review of literature instead of before), they were not in null form, which is a requirement in experimental research situations. Additionally, the hypotheses presented (p.5) were not in null form and most decisively directional as can be seen in the following wording: "...is associated with...leaders influence...(p.5)." As a result of the omission of a null hypothesis, a biased direction of the hypothesis presented, and the inappropriate placement of the hypotheses stated, the reviewer was left with an initial feeling of incompleteness and bewilderment with respect from where the research idea even came and for what purpose.

Literature Review The purpose of a literature review is to give credence and support to the investigation that is or has taken place. Unfortunately Ingersoll, et al. had as their review a scant four short paragraphs that cited five supportive documents. The literature review included neither support for their investigation nor did it enlighten the reader. There should have existed ample review material to support the need to conduct the study as well as provide the reader a sense intrigue. In addition, the authors should have discussed previous studies that supported not only the selected independent variable (EPPM) but also the need to include the three types of hospital sample populations. A research investigator should always keep in mind that the research problem is stated first, followed by a research question, testable hypothesis, and definition of terms. After that a review of literature is presented which reinforces the need to research the selected topic.

Theoretical and Conceptual Framework Theoretical frameworks and conceptual designs are a direct result of information gathered through the review of literature in support of a formulated research question and null hypothesis. As there is no cited research problem or null hypothesis there is no theoretical framework upon which to interpret any resulting assessment data.

Even though the researchers have a section entitled "Conceptual Model" (p.4) there exists no reference to the model outlined visa via the review of literature or as an identifiable factor in a research question or hypothesis. Complicating matters further the investigators reference the theoretical model before stating their hypothesis - in effect working backwards.

In addition to formatting a theoretical and conceptual framework there must also be evidence of a direct relationship between the theoretical and the results obtained. Again, the omission of a null hypothesis and research question, plus the lack of historical support, has, for this article produced a situation wherein theoretical interpretations cannot be made. The closest the researchers came in an attempt to lay the groundwork to test a theoretical model in nursing was mainly in the "Description of Study" section wherein a general study purpose was stated. Again, the positioning of the study's purpose should have occurred during the introduction section; i.e., prior to a statement of the research question, null hypothesis, and review of literature (Potts & Beck, p.656).

Research Design and Sampling For all experimental investigative situations the research design section of a study outlines for the reader the basis for the most appropriate way in which to analyze the data. When a research endeavor is introduced as an experimental study (Ingersoll, et al., p.1) it signals to the reader that certain stringent research requirements have been met; namely the precise identification of the independent and dependent variables, the presence of possible nested variables, and the likelihood of orthogonal trends being present. Further, by labeling a study as experimental there must always exist the statement of a testable null hypothesis, the probability level against which the null hypothesis will be accepted or rejected and the type of data the assessment instrument(s) will produce; i.e., nominal, ordinal, interval, or ratio. None of the aforementioned concerns were even alluded to in the Ingersoll, et al. study.

Inherent within an experimental design is also the need to define the selected type of statistical data analysis technique in greater depth than simply stating that an Analysis of Variance was used. ANOVA techniques come in many different forms from one-way classifications, to two-way classifications, to covariance models, to orthogonal designs, to correlation methods - all requiring a different set of assumptions for use.

Other considerations that a researcher must account for within the design section of the study are directly related to the manner by which a sample has been selected, the fashion by which the assessment instrument is administered, possible limitations and delimitations surrounding the investigation, and the applied control procedures. Should these factors not be taken into account there exists a strong possibility that the results, even though statistically analyzed, will be contaminated with error (Polit & Beck, p.8). The Ingersoll, et al. research endeavor fell short of adherence to the aforementioned requirements. Not only did the investigators fail to report on the sampling selection procedure but also the size of their sample simply was not presented or discussed in the "Procedure" section (p.7). Without a full explanation as to sampling procedure and sample size the reader is left with a vacant feeling as to the efficacy of the results (Polit & Beck, p.660). As best that can be determine by way of reviewing this particular article is that the sampling procedure was non-random, without a statement as to size, and without dependent variable identification. The Ingersoll group also committed one of the greatest biasing errors in sampling known to experimental research, namely offering an incentive (monetary)(p. 7) to groups of nurses who failed to complete the questionnaire visa via the initial mailing. The effects of the $100.00 incentive can, and will, affect the results of this investigation in uncounted for and uncontrollable ways.

Put into proper perspective the Ingersoll, et al. research group failed to adequately design their study with respect to sampling procedure, the setting of limitations, the accounting of extraneous and external variable influence, the setting of proper controls, and the choice of statistical methods used. In the end all that can be said is by way of mentioning that the study needs to be re-constructed in proper research methodology format. Some suggestions to the investigators include, but not limited to, randomly select the sample if possible, stratify the sample selection according to the number of dependent variables identified, omit the incentive bias, test for secondary and interaction main effects.

Data Collection and Data Quality One of the most important research tenets to keep in mind is that statistical data procedures produce numerical results that are only as good as the inputted data. For example, some statistical tests can only function properly where assessment data is ordinal, while others exist properly when the data is interval in nature. Unfortunately, however, Ingersoll, et al. did not identify the type of measurement data returned from the two assessment instruments. Not knowing the type of data makes it impossible to determine whether or not the ANOVA technique is best suited. Should the instrument data, for example, be that of the nominal and ranked type then it is possible that the statistic of choice is a Kruskal-Wallis ANOVA rather that the chosen Fisher ANOVA. Again, however, not knowing the type and classification of data garnered from the two administered assessment instruments it is impossible to determine which is more suited.

With reference to the procedure followed in administering the two questionnaires the reviewer has, again, some serious concerns. As stated earlier the investigators should have foregone the incentive practice, as there is not way in which to account for the effect this has on the interpretive results. Had the investigators known beforehand as to the possibility of there being a shortage in returned responses, and that an incentive would be used, then the incentive should have been built into the study as an additional independent variable.

An additional area of concern is in reference to the time span that governed the return of results. According to information contained within the "Procedure" section (p. 7) respondents were assessed on a longitudinal basis as to the effects of the a professional practice model on nurses' perception of work groups and nurse leaders. This is the first time the authors mention the use of a longitudinal research procedure. Such a statement clearly belongs in the design section and not the procedure section. Further, statistical procedures are highly sensitive to measurement data gathered visa via longitudinal studies data by way of compounding or accumulated error, sampling attrition, and personal changes that might have taken place with regard to the respondent. As a result a study must be formatted to take into account these extraneous variables.

Other infractions pertaining to data collection and data quality correctness evolve around the areas of questionnaire description, rationale for instrument selection, supportive literature as to the efficacy of instrument choice, and type of data the assessment instrument produces. When choosing an appropriate measurement instrument the investigators must always inform the reader as to the validity and reliability of the measurement technique as well as whether or not the instrument is standardized or non-standardized. As is true for sampling, the appropriateness of the selected statistical tool is dependent on the type of measurement instrument chosen.

Knowing that the Ingersoll, et al. set forth by way of an "experimental" study to test the effects of an independent variable (EPPM), a control situation is necessary. Mailing questionnaires to respondents wherein geographical boundaries are not know, observations cannot be made, and instructions have no follow-up or immediate clarification, greatly biases the results. Control is most important in an experimental study and these particular investigators were not able to exercise the control needed. With reference to the type of scaling being reported by use of this particular measurement instrument there exist no information. The authors clearly needed to have given a great deal more consideration as to the type of measurement instrument chosen and how it was administered Polit & Beck, p. 663).