How much are the IMRaD structures of diploma theses and original research articles similar? A cross-sectional analysis of diploma theses from the Dental Medicine study in Split

correlation with the overall length of IMRaD structure in diploma theses was found for Introduction (R=0.82; R 2 =0.67; P <0.001). In original research articles, the Discussion section was most highly correlated with the over-all length of IMRaD structure (R=0.62; R 2 =0.38; P <0.001). Conclusion: The IMRaD structure of diploma theses defended in graduate study of Dental Medicine in Split differs from the structure of original research articles published in dental scientific journals. Diploma theses have a comparative - ly longer introduction and shorter discussion section than original research articles.


Introduction
Before graduation, students of the graduate Study of Dental Medicine at University of Split School of Medicine must prepare and defend their diploma theses. Diploma thesis is a written report of scientific research conducted by each student under the supervision of a selected mentor. The design of diploma theses is based on the standard methodological postulates for writing original research articles derived from the International Committee of Medical Journal Editors (ICMJE) guidelines for writing original research articles [1,2].
According to these guidelines, original research article reports results of a study based on a clear research question and a defined hypothesis. As such, original research articles must adhere to the IMRaD (Introduction, Methods, Results and Discussion) structure, i.e., they must include four main sections -Introduction, Methods, Results, and Discussion [1]. Each IMRaD section has a specific purpose [1,[3][4][5]. The Introduction serves to present the background, objective, and main hypothesis of the research. The Methods section should describe the data collection and analytical procedures in sufficient detail to allow other researchers to replicate the study. The Results section contains the list of findings obtained through analytical procedures. In the Discussion section, the relevance of findings (in the context of findings from similar studies and the methodological strengths and limitations) and conclusion are presented. Depending on the complexity of the research topic and study design, the length of the IMRaD sections in original research article may vary. However, it is strongly recommended that the introductory section must be brief and concise [3].
Currently, there are two guidance documents for the preparation of diploma theses at University of Split School of Medicine. These documents -"Guidelines for the preparation of diploma thesis" and "Instructions for writing/formatting of diploma thesis" -are available on the official website of the "Department of Diploma Thesis" [6][7][8]. There are no specific instructions for writing the Introduction and Discussion sections in either document, except for the definition of the Introduction as a "brief and concise statement of the research background" in the "Guidelines for the preparation of diploma thesis" [7,9].
We analyzed diploma theses defended at the graduate Study of Dental Medicine at University of Split School of Medicine to investigate whether their IMRaD structure is quantitatively different from the IMRaD structure of original research articles. For this purpose, the length of the main sections within the IMRaD structure of diploma theses and original research articles was compared. Since the topics of the analyzed diploma theses are mostly related to clinical research, we specifically considered the guidelines stated in "Instructions for authors" of several scientific journals that publish original research articles mostly related to clinical research in dental medicine [10][11][12][13][14].

Data collection
Diploma theses defended in the graduate Study of Dental Medicine in Split were downloaded in PDF format from the digital academic archive and repository (DABAR) [15] during February and March 2022. Original research articles were downloaded in PDF format during September 2022 from the digital repositories of five dental scientific journals:

Standards for comparison of diploma theses and original research articles
The basis for comparing diploma theses and original research articles was partly derived from the "Instructions for authors" of dental scientific journals in which the analyzed original research articles were published. We also referred to the recommendations for writing original research articles from the relevant literature on the IMRaD structure. Dental scientific journals do not limit the length of individual sections in the IMRaD structure.
However, there are clearly defined requirements for the overall length of IMRaD structure in original research articles, which vary from 3000 words (JOP) to 3500 words (JOE, JCP) [10-14]. The requirements for the length of sections in the IMRaD structure are best defined for the Introduction. According to some authors, the Introduction section in original research articles should be between 400 and 600 words, or no more than 15% of the total length of the IMRaD structure [9]. The length of Methods, Results, and Discussion sections is not strictly limited, as it may vary depending on the complexity of study design [21][22][23].

Descriptive statistics
The length and proportions of each section in the IMRaD structure, as well as the overall length of IMRaD structure, were first analyzed with descriptive statistics and expressed as means with 99% confidence interval (CI) for both diploma theses and original research articles. Other parameters such as the total number of references and the percentage of references cited first in the Introduction section were also expressed as means with 99% CI.
Testing differences between IMRaD structure of diploma theses and original research articles To examine differences in IMRaD structure between diploma theses and original research articles, we used simple and multiple linear regression. To compare the magnitude and significance of possible variations between the observed patterns in the IMRaD structure of diploma theses and original research articles, we created an additional set of simple linear regression models. For each of these models, the proportion of the respective IMRaD section in the overall IMRaD structure was set as the outcome, whereas the predictor was a binary categorical variable for the type of research report (diploma thesis or original research article) coded as a single dummy variable with dummy coding.
The influence of several modifying factors (year of defense, type of research reported, and topic of research related to the branch of dental medicine) on the observed patterns in the IMRaD structure of diploma theses was also analyzed. Three linear regression models were created. The main outcome for each of those models was the length of the IMRaD section, which was found to be most strongly correlated with the overall length of IMRaD structure of diploma theses. The predictors were categorical variables for the modifying factors, which were coded as dummy variables using either dummy coding or effect coding.

Correlation of the length of IMRaD sections and IMRaD structure
The overall length of IMRaD structure was set as the main outcome (dependent variable), whereas the lengths of the sections of the IMRaD structure (Introduction, Methods, Results and Discussion) were set as predictors (independent variables). The regression analyses were performed under the assumption that (on average) the longest section of the IMRaD structure was the one most strongly correlated with the overall length of IMRaD structure itself.
Linear regression was chosen as the only statistical test for the analysis of data according to recommendations by Cohen [24]. The advantages of linear regression over standard approach to hypothesis testing are: (i) it does not assume a normal distribution of the input data; (ii) it is a parametric test, so there is no need to compromise statistical power by using nonparametric tests because of the nonnormality of the data distribution; (iii) it can also be used to compare two (such as t-test) or more group means (such as ANOVA), but (in the latter case) (iv) there is no need for post-hoc testing.

Composition of diploma theses and original research articles
A total of 150 diploma theses and 150 original research articles were analyzed. Diploma theses included topics from all branches of dental medicine, but the majority (105/150, 70%) was related to endodontics and restorative dentistry, oral medicine, prosthodontics, and pediatric dentistry. Most diploma theses reported results from clinical research
The difference in IMRaD structure between diploma theses and original research articles was found when analyzing the relative proportion of each IMRaD section in the overall length of IMRaD structure ( Table 1;

Correlation patterns of IMRaD sections and IMRaD structure in diploma theses and original research articles
The correlation between the length of IMRaD sections and the overall length of IMRaD structure was assessed separately for each IMRaD section of diploma theses and original research articles by simple linear regression (8 models) (Figure 2; Supplementary dataset -Spreadsheet 4). In diploma theses, the strongest correlation with the overall length

The influence of modifying factors on the observed patterns in the IMRaD structure of diploma theses
Considering that the length of the Introduction section was the most important predictor of the overall length of IMRaD structure of diploma theses, it was necessary to determine how established the style of writing diploma theses with lengthy introduction is and whether there is evidence of a change in trend during the observed period. It was also investigated whether there were differences depending on the type of research reported in diploma thesis (clinical/basic) and the topic of diploma thesis in relation to a particular branch of dental medicine. According to the models, modifying factors show no discernible influence on the variability of the length of Introduction section in diploma theses.
Therefore, writing diploma theses with lengthy introduction was widespread throughout the period studied, regardless of the type of research or the topic of the research reported in diploma thesis ( Table 2; Supplementary dataset -Spreadsheets 5-7).

Discussion
In diploma theses, the Introduction section occupied almost half of the overall length of IMRaD structure. In contrast, the Introduction section in original research articles was short, while the longest section was the Discussion section. According to findings, writing diploma theses with lengthy introduction was a widespread phenomenon throughout the period studied, regardless of the type of research or the topic of research reported in diploma thesis. One of the reasons may be the length requirements for diploma thesis on MEFST, which require the length of about 40-60 pages [6,8]. to 10 references, which is far less than the average of 29 references in the diploma theses analyzed in this study [3,9].
Without a qualitative analysis of the content, we cannot say to what extent the narrative sections of the diploma theses (e.g., the Introduction section) relate to findings from original scientific research as opposed to textbook knowledge. According to the principles for writing original research articles, the latter must be avoided [27][28][29]. Considering this, mentors, and evaluators of should emphasize the importance of the "target audience" during the evaluation of diploma theses before their final approval. Although diploma theses defended in graduate Study of Dental Medicine are stored in a public repository and are fully accessible to anyone, the potential readership will most likely consist of professionals in the field who are familiar with technical terms and textbook concepts. It would st-open.unist.hr be interesting to investigate whether the differences in the IMRaD structure of diploma theses and original research articles revealed in this study are related to students' and mentors' perceptions of who they think should/could read diploma theses.

Improving the IMRaD structure of diploma theses in Study of Dental Medicine in Split
-what can we learn from the guidelines for authors in dental scientific journals?
The observed features of the IMRaD structure of diploma theses defended in the graduate Study of Dental Medicine in Split can be partially related to the lack of clarity in the official guidelines for writing/formatting of diploma theses [2,[6][7][8]. is methodologically limited, is frequently done by inexperienced authors [27][28][29][30]. In the diploma theses defended in the graduate study of Dental Medicine in Split, the narrative section of choice is the Introduction section. Unfortunately, most mentors and members of review committees for diploma theses do not seem to discourage students from such practice. It would be interesting to investigate whether students, mentors, and evaluators consider proper IMRaD structuring to be a more important determinant of the quality of diploma thesis than the overall length of diploma thesis. st-open.unist.hr