Identifying the challenges of online education from the perspective of University of Medical Sciences Students in the COVID-19 pandemic: a Q-methodology-based

This cross-sectional study was performed using the Q methodology during the following six steps using Barry and Proops method [19].

Stage 1 and 2: defining the concourse

At this stage, a concourse space was formed with the identification of the subject or idea of the study. The presented views on the issue raised for the concourse can be formed from a review of texts and experts in this field [19].

In this study, the topic and idea for the concourse were the challenges of online education during the COVID-19 pandemic. The concourse included a collection of diverse materials related to the research topic that was discussed among students. The students (P-set) who also had contributed earlier to the development of the initial set of statements. Thirty-one students participated in semi-structured interviews, and we tried to identify their subjectivity about the research topic using the Q method [20].

In this study, the concourse (sample of people) included students of the University of Medical Sciences (paramedical students) who had enough information about online education during the COVID-19 pandemic.

Step 3: screening and selection of statements (Q-sample)

During the semi-structured interviews with 31 students, 70 statements were extracted about the perceived challenges of online education. The Q items were selected very carefully so that items did not overlap, and at the same time, no perspective should be missing. Therefore, the selection process takes the most time and effort of all the steps of the Q methodology. Therefore, research team removed similar unrelated, and ambiguous statements from the Q set. Finally, 50 statements were selected.

Stage 4: selected P-set

Students who participated in the concourse (interviews) were selected as a sample of individuals to participate in sorting in the Q study (P-set). In the present study, students were selected by purposive sampling to include students who had an educational, professional, experimental relationship or previous knowledge about the subject of study. This selection of samples made the participants with more diverse mentalities enter the study. It is recommended that in Q studies, the number of participants to sort statements should be less than the number of statements around the study subject [21]. In the present study, the number of participants who ranked the challenges of online education programs was 31 (Table 1).

Table 1 The Q-set statements and factor arrays in the study of challenges online education amongst students

Stage 5: Q-sort

At this stage, the normal distribution table in the form of a Likert scale from − 5 to + 5 was designed offline. Tips on distributing the expressions on the normal distribution table were provided. In the first stage, the purpose of the study is the number of statements selected through the interview. In the second stage, place the statements in three columns: “I agree”, “I have no opinion,” and “I disagree. In the third stage, the statements (mandatory) are distributed in the normal Likert distribution diagram (− 5 to 5+), explaining the reason for choosing the two ends of the Likert scale from their point of view and finally entering the demographic information. Thus, in Q, the sorting process is subjective [19]. In other words, sorting items in the normal distribution allow each participant to present their internal perspective through sorting.

Stage 6: analysis and interpretation of factors

Students’ data obtained from Q sorting were entered into PQ-Method software version 2.35. The process of analysis and interpretation was performed in three stages: (a) identification of factors, (b) conversion of factors into factor arrays (c) interpretation of factors using factor arrays.

  1. A)

    Factor Identification

The extraction of factors in PQ-Method software was done through the following sequential steps: (a) principal component analysis, (b) identification of latent factors, (c) varimax rotation and evaluation of loading factors for specific values above 1.00, d) estimation of the percentage of variance explained by the identified factors and (e) differentiation of interpretable factors with at least two correlated Q types [22].

  1. B)

    Convert factor to factor arrays

The correlation between each Q sort and one identified factor indicates the degree of interaction between the Q sorts and the identified factors [19, 23]. The manual flagging in PQ-Method software was applied for this study. The correlation coefficients of at least 0.364 were considered as the cut-off point (the absolute value of the factor load is greater than (\(\frac{2.58}{\sqrt{N}}\)). That factor load was 99% significant, respectively, and the value of N was equal to the number of Q statements (N = 50). Sorted for identified factors [24]. Specifications specified on a factor are used to create a factor array for that factor. The factor array represents the sorting of that factor (point of view) using z-scores. The factor array for each factor determined the degree to which each expression was in the spectrum, so a more accurate interpretation of each factor (subjectivity) was obtained according to the position of each expression. (P-value< 0.05 vs. 0.01) is also determined from the Z score to distinguish expressions [25].

  1. III)

    Factor interpretation using factor arrays

Distinct Q expressions were identified, and factors were interpreted textually. The defining expressions for a factor were those that had a rating value of “+ 5”, “+ 4”, “4-,” 5- “in factor arrays that had different scores (P < 0.05) in a given factor Compared to their scores on other factors, the post-P-set interview was conducted at the end of Q sorting to confirm the diagnosis and interpretation of item subgroups among the identified factors.