Description

As evidenced from the pages of nursing and midwifery education journals and other nursing and midwifery journals which publish educational research a range of methods is used to investigate nursing and midwifery education (Long and Johnson, 2002, Schneider et al., 2013, Morton, 2017). These include both qualitative and quantitative designs and within the quantitative designs a range of methods from correlational research through to experimental research. While there has been an increase over the years in the use of experimental designs in clinical nursing and midwifery research (Baldi et al., 2014), there is a dearth of rigorous experimental designs being used in nursing and midwifery education, particularly randomised controlled trials (RCTs; MRC 1948). For example, a cursory search of the relevant journal webpages shows that most issues of Journal of Advanced Nursing (JAN) will contain several RCTs and many issues of Journal of Clinical Nursing (JCN) likewise.

Within-subjects designs (or repeated measures) are attractive to educational researchers for several reasons. The main reason is convenience in the sense that a group of students who can be tested before and after an intervention is relatively easy to find. These designs are also statistically powerful as they do not suffer from individual variation; participants serve as their own controls. As discussed above, the main features of betweensubjects research (blinding; control; and randomisation) may not be as easily achieved in educational research as they are in clinical research.

Research in Nursing and Midwifery Education

Against a background of a paucity of rigorous experimental research in nursing and midwifery education, we consider above the reasons why this is the case. There are many valid reasons why research in nursing and midwifery education often cannot apply the rigorous and relatively simple experimental approaches used, for example, in randomised controlled trials. In fact, each of these journals has published hundreds of RCTs as evident if these are specifically searched for in each journal. However, in the two leading nursing and midwifery educational journals Nurse Education Today (NET) and Nurse Education in Practice (NEP), experimental studies are rare and RCTs, as a subset of experimental studies, are even rarer.

In the past 5 years 6.8 % of the content of NET has been represented by experimental work (58 studies) and RCTs represent 1.8 % of the content. The respective figures for NEP are 2.2 % (23 studies) and only 0.8 % of the content was RCTs. In the corresponding period JAN will have published several hundred. Clearly, it could be expected that a generic journal such as JAN and a clinical journal such as JCN would publish more RCTs. However, we wish to explore the paucity of RCT type studies in nursing and midwifery education and to compare designs.

Experimental designs have been scarcely adopted in nursing and midwifery research over time and quantitative multivariate analyses need also to be further implemented. Most quantitative designs is correlational, comparative or descriptive as stated by Yarcheski et al. in an overview of nursing and midwifery research from 1985 to 2010. When considering specifically nursing and midwifery education research, the 8 % of study designs are RCTs compared with the 18 % of clinical nursing and midwifery research. In the same vein, in nursing and midwifery education the 89 % of research moved beyond descriptive analysis against the 95 % of clinical nursing and midwifery research. It is not entirely clear why there is a paucity of the more rigorous designs in nursing and midwifery education research. Within-subjects studies are easier to conduct than between-subjects studies.

Within-subjects studies can conveniently be conducted in a single site and require smaller sample sizes than betweensubjects studies. As such they are cheaper to conduct which is possibly a driver as the funding allocated to nursing and midwifery research is orders of magnitude less than that allocated to, for example, medical research; according to the numbers presented above, it is clear that less of that is allocated to nursing and midwifery education research. One explanation could be that nursing and midwifery scientists express anxiety towards statistics generally and develop a fear of the advanced statistical methods required to design and analyse experimental studies.

The term ‘experiment’ is general and simply describes an investigation designed to test an hypothesis. Designs vary in their ability to distinguish between cause and effect and extent to which they are externally valid. External validity refers to the extent to which the outcome of a study may be generalised to other situations. The design which best distinguishes cause and effect is the true experiment where a group receiving an intervention is compared with one that does not receive the intervention (a control group). Ideally the allocation to the two groups is random and participants and data collectors are blind to the allocation. However, all aspects of a true experiment described above are not always possible and any compromise in control, randomisation or blinding creates a quasi-experiment. The relationship between rigour and external validity is reciprocal and it must be acknowledged that true experiments are low in external validity.

The relationship between rigour and external validity is reciprocal and it must be acknowledged that true experiments are low in external validity.