Study design and sampling
A cross-sectional online survey of 445 students was performed in Lebanon between March and May 2022 using the non-random snowball sampling technique to collect data from university students. The questionnaire was developed, then created on Google Forms (https://forms.gle/7ea86dWBBuneq8B6A) and distributed on social media (WhatsApp, LinkedIn, and Facebook). The students approached were from the first to final academic years, and postgraduate students. They were enrolled in different faculties of private and public universities. The inclusion criteria were: being a university (undergraduate or graduate) student, age above 18 years, with internet access. Participation in the study was voluntary, and participants received no compensation in exchange for their participation.
It is noteworthy that university students in Lebanon are enrolled in public and private universities [27], although no official figures are available regarding the current distribution of students. In the academic year 2019–2020, students were estimated to be distributed over the Lebanese University (the only public university, tuition-free; around 79,000 students), the Lebanese International University (the largest private university, relatively inexpensive; about 20,000 students), and other private and more expensive universities, including the American University of Beirut (AUB), the Saint Joseph University of Beirut (USJ), the Lebanese American University (LAU), the University of Balamand (UOB), Beirut Arab University (BAU), and the Modern University for Business and Science (MUBS), which have 120,000 students enrolled.
Questionnaire
The study focused on the perception of students of the teaching and research skills of instructors. The self-report anonymous questionnaire was available in English (Appendix) and consisted of two sections.
The first section included sociodemographic and student characteristics such as age, gender, area of residence, marital status, current academic year, current university, GPA level, the highest level of education, the study major, employment status, and monthly income. The latter was divided into no income, low (< 1.500.000 LL), moderate (1.500.000–3.000.000 LL), and high (> 3.000.000 LL). A question was also added to inquire if the student regularly checks if the instructors carry out research activities or have published articles for adjustment purposes.
The second section included the following assessment scales:
The Student Perception of Research Integration Questionnaire (SPRIQ)
This questionnaire was designed to identify how students perceive research integration into their courses [28]. It consists of 40 items graded on a 5-point Likert scale from 1 (very rarely) to 5 (very frequently). The SPRIQ is divided into three constructs, i.e., research integration, quality of the course, and beliefs about research integration [28]. All items were summed, and a total score was created, with a higher score indicating a better perception of research integration.
The Adapted-Teachers’ Quality Assessment Questionnaire (A-TQAQ)
The TQAQ was designed to measure the teacher’s academic qualification, professional qualification, and years of experience [29]. In this study, the questionnaire was adapted by including four additional items deemed suitable and necessary to create the A-TQAQ, i.e., “Research activity of an instructor is dependent on one’s academic qualification”, “Excellent mastering of one’s subject as an instructor is dependent on one’s research activity”, “Students taught by more experienced researchers perform academically better”, and “A researcher is a role model for students”. This scale is graded on a 5-point Likert scale ranging from 1 (most unlikely) to 5 (most likely) [29]. All items were summed to yield a total score, with higher scores indicating a better perception of students of their instructors’ qualifications.
The Student Evaluation of Teaching—Short Form (SET37-QS)
This tool measures students’ subjective perception of learning and overall satisfaction with a course [30]. It helps obtain student feedback on internal practices and processes that monitor and enhance the quality of higher education instruction [30]. Derived from the SET37 [31], the SET37-QS consists of nine items rated on a 5-point Likert scale from 1 (strongly disagree) to 5 (strongly agree) [30]. The total score was calculated by summing all nine items, with higher scores indicating a better evaluation of teaching quality. In this study, the scale was used twice; in the first instance, students were asked to answer when thinking about the instructor with the highest research activity and then when thinking about the instructor with the lowest research activity in their institution. The variation between the highest and lowest level of the SET37-QS was calculated, and a new variable was created, i.e., the difference between the student evaluation of teaching quality (highest vs. lowest); this variation estimated the difference in perception of teaching quality between instructors according to their research activity (high-caliber researchers versus non-researchers).
The knowledge and attitudes towards health research questionnaire
In this scale, ten multiple-choice questions assess knowledge [32], and the proportion of correct answers is determined for each student as a measure of knowledge score [32]; six questions examine their attitudes toward research, with each response being rated on a scale from 0 (unfavorable attitude) to 1 (favorable attitude) [32]. Since the study was intended for all students from health and non-health specialties, the term “health” was removed. These scales were validated in a separate paper and showed to have appropriate structure validity and reliability (submitted article).
Statistical analysis
Data were analyzed using SPSS software version 25. The principal component analysis was used to assess the construct validity of the used scales, and Cronbach’s alpha was calculated to assess their reliability (internal consistency). A confirmatory factor analysis was carried out to assess the structure of the scales used. Several goodness of fit indicators were re-ported: the Relative Chi-square (χ2/df) that serves as goodness of fit index (cut-off values: < 2–5), the Root Mean Square Error of Approximation (RMSEA) that tests the fit of the model to the covariance matrix (close and acceptable fit are considered for values < 0.05 and < 0.11, respectively), the Goodness of Fit Index (GFI), the Adjusted Goodness of Fit Index (AGFI), comparative fit Index (CFI) and Tucker Lewis Index (TLI) (acceptable values are ≥ 0.90) [33].
A descriptive analysis was done using counts and percentages for categorical variables and means and standard deviations for continuous measures. The sample was normally distributed, as checked by visual inspection of the histogram, and skewness and kurtosis were below |1.96| [34]. In addition, the normality of scales was verified by the normality line of the regression plot and scatter plot of the residual. After checking the normality of both variables (perception of teaching score and assessment of the quality of teaching score), the independent-sample t-test was used to compare the means between two groups, the dependent-sample t-test to compare dependent groups, and the ANOVA test was applied to compare three or more non-dependent means. A p-value < 0.05 was considered significant.
All variables that showed a p-value < 0.2 in the bivariate analysis were included in all models to avoid potential confounders. A multivariate analysis of covariance (MANCOVA) was carried out, taking the perception of teaching scale and the assessment of the quality of teaching scale as the dependent variables. Moreover, the variable “difference in the evaluation of teaching quality between the highest and the lowest research-active teachers” was considered a dependent variable in the multivariable linear regression analysis.