This cross-sectional study conducted to evaluate the validity and reliability of the Iranian version of the MHLS with the participation of 1363 individuals of the general population in Gonabad, Iran in 2018.
Sample size
Based on the previous studies sample size of 1000 or more is estimated to be very good for factor analysis. In this study, the sample size of 1363 individuals was considered for assessment of the validity and reliability of the instrument[19-21].
Sampling method
The sampling in this study was conducted by multistage. Initially, the number of community health centers and the population of each center were determined. In the next step, each center was considered as one class and the sample size was determined according to the population of each class. Samples were then randomly selected from each center. It should be noted that the questionnaire for illiterate participants was completed by the interviewer. Inclusion criteria were age over 18 years old and informed consent of the individual to participate in the study.
Instruments
1) Demographic questionnaire: This questionnaire includes questions on gender, age, occupation, level of education, marital status, etc.
2) MHLS: This questionnaire was designed and evaluated by Connor et al. In 2015[22]. The questionnaire has 35 questions and 6 subscales of the Ability to recognize disorders, Knowledge of self-treatment, Knowledge of risk factors and causes, Knowledge of where to seek information, Knowledge of the professional help available, Attitudes that promote the recognition or appropriate help-seeking behavior.
- Ability to diagnose disorders: This construct has 8 questions that are assessed by a 4-point Likert scale (highly rare, rarely, likely, very likely).
- Knowledge of risk factors and causes: This construct is assessed with 2 questions and a four-option scale (highly rare, rarely, likely, very likely).
- Awareness of self-treatment: This construct is measured by two questions and a four-option scale (not very useful, not useful, useful, very useful).
- Awareness of the available professional assistance: This construct consists of 3 questions that are measured on a four-option scale (highly rare, rarely, likely, very likely).
- Awareness of the place for searching information: This construct consists of 4 questions that measure with a 5- option Likert scale (strongly disagree, disagree, no opinion, agree, strongly agree).
- Attitudes that promote cognition or appropriate behavior: This construct is with 16 questions and a 5-option Likert scale [(strongly disagree, disagree, no opinion, agree, strongly agree) or (definitely willing, probably willing, no opinion, probably unwilling, definitely unwilling).
In this questionnaire, the lowest score is 35 and the highest score is 160, and higher scores indicate better MHL. The validity and reliability of this questionnaire were evaluated in the Connor study. The internal consistency of this scale was measured by Cronbach’s alpha (Cronbach's alpha= 0.873)[22].
Validation: Face and content validities (qualitative and quantitative)
Firstly, the English version of the questionnaire was forward-translated into the Persian language by three faculty members and then reconciled. Then one skilled English expert who was not familiar with the specialized English text of psychology translated the text backward into English. The English text of the backward-translation was re-translated into the Persian language by three psychology professors fluent in the English language. Then, the final Persian version was prepared by adjusting the first Persian translated version of the questionnaire and making the necessary corrections. The validity and reliability of the questionnaire were then evaluated.
Qualitative formal validity
For this purpose, the Persian questionnaire was given to two faculty members fluent in English language and specialized vocabulary to evaluate the final Persian version of the questionnaire in terms of clarity (use of simple and comprehensible words), common language application (avoidance of the specialized and technical words). If required, changes were made to the used metrics to make it simpler and more comprehensible. Also, for knowing the comments of audiences, a face-to-face interview was conducted with some of the individuals in the target group to find out any difficulty in understanding of the words and phrases, the appropriateness and relevance of the items, the likelihood of ambiguity and misunderstandings, or any failure in conceptualization. In case of problems, their comments were applied to the questionnaire.
Qualitative content validity
To evaluate the validity, the questionnaire was provided to 20 experts and specialists to assess grammar, use of appropriate words, the importance of items, the correct placement of items, and the time for completion of the designed instrument. After collecting the expert evaluation results, necessary changes were made in consultation with the members of the research team. Given that the standard questionnaire has been used and translated, quantitative content validity and formal validity were not required to be measured[23].
Construct validity assessment (exploratory factor analysis)
Firstly, the Kaiser-Meyer-Olkin (KMO>0.8) and Bartlett’s test of sphericity (<0.05) were used to make sure of the reliability of sampling[24]. CFA was used to evaluate the Construct validity assessment. Before to CFA, data were analyzed using Mahalobis statistics for the outliers. The normality of data was also evaluated using skewness and kurtosis. CFA was performed using AMOS version 24 software. To obtain an acceptable model, questions with poor internal consistency were removed from the questionnaire.
The assessment of the model was conducted using the following fit indices: Chi-square ratio to degree of freedom (x2/df); root mean square residual (RMR); root mean square error of approximation (RMSEA); goodness of fit index (GFI); adjusted goodness of fit index (AGFI); parsimonious normed fit index (PNFI); parsimony comparative fit index (PCFI); incremental fit index (IFI); parsimony goodness-of-fit index (PGFI); comparative fit index (CFI); and parsimonious normed fit index (PNFI)[25-27]. The model was acceptable if the (x2/df) <5, RMSEA and RMR ≤ 0.08, PCFI, PNFI and PGFI>0.5, AGFI > 0.8, and other indices (IFI, GFI, CFI) > 0.9[25-28].
Reliability assessment
Finally, to establish the reliability, a pilot study was performed; the minimum sample size required for reliability assessment is between 25 and 30[29]. In this phase, the questionnaire was completed by 60 individuals who did not participate in the main study, and the internal consistency of the questions in each dimension was assessed by Cronbach’s alpha coefficient.
Ethical Considerations
In this study, written informed consent was obtained from all the participants before the study and they were assured that their information remained confidential with the research team.