Design
This is a mixed study that was conducted in accordance with the Creswell method in three phases (1- qualitative phase by content analysis, 2- development of item pools, 3- quantitative phase to determine the psychometric properties of the tool) from May 2016 to January 2018(10).
In the first phase, in order to identify the key concept (women’s QOL), we used a qualitative approach. Purposive sampling was used to select the population. Participant criteria for inclusion in the study were all women and girls of reproductive age (15 to 60 years) with the ability to read and speak Persian, without any disabilities and other physical and psychological ailments, excellent ability of communication and expression, willingness to do an interview )45 to 60 minutes( and talking about their life situation, and their worries and concerns in their daily life (totally 40 women when we reach data saturation).
In the second phase, from all items that were obtained inductively through interviews and terms of the participants, the main themes and pool of items were formed. Then, with the research team's opinion, the concepts that were conceptually similar were removed or merged. In this way, the initial tool for evaluating women's quality of life was designed based on perceived definitions in the form of a methodological process.
In the third phase, in the third phase of the study, the psychometric properties of the primary tool were done in 5 steps.
First, Face validity
Qualitative face validity was determined by asking the points of view of 10 women in the target group. Furthermore, a quantitative assessment of face validity was conducted with the calculation of impact scores (Impact score = frequency(%)🞨 importance) to reduce and remove inappropriate items. The question that received a score of above 1.5, was identified as an important item(11).
Second, content validity
In terms of qualitative content validity, 10 experts in the field of health, quality of life, nursing, and questionnaire designers were asked to write their editorial opinions in writing after carefully studying the instrument (about grammar, Wording, Scaling, and Item Allocation). To assess quantitative content analysis, content validity ratio (CVR).was used as well as the content validity index (CVI). To determine CVR, 10 experts were asked to score each item based on the three-part spectrum (essential, useful but not necessary, and not necessary). According to the Lawshe table, the minimum value was determined as 0/62 (12).Then based on Waltz and Bausell's criteria, CVI (Item content validity index & scale content validity index was calculated. To assess I -CVI, 10 experts scored the relevancy, of each statement in WQOLI using a 4-point rating scale (1 = not relevant, 2 = somewhat relevant, 3 = quite relevant, and 4 = extremely relevant), Then, the score for each item was calculated by dividing the number of specialists agreeing with the phrase having grades 3 and 4 on the total number of specialists. A rate of 0/78 and above is appropriate(11, 13). The Scale Content Validity Index Average method (S-CVI/Ave), was used to determine the average value of I-CVI for the whole scale. A reference value of more than 0.90 is desirable(11). Based on the results of the pilot study and experts’ opinions, necessary changes were made and the WQOLI was modified accordingly and the final instrument was obtained.
Third, Construct validity
We administered an exploratory factor analysis (EFAs) by a principal component extraction method for determining the construct validity of the WQOLI. The number of respondents required for factor analysis was 5 for each statement(14, 15). In this step, 590 women who had the criteria for entering the research were selected through randomized cluster sampling and completed the questionnaires from November 2017 to January 2018.
For indicating the degree of susceptibility of data for factorial analysis and sampling adequacy, the Kaiser-Meyer-Olkin (KMO) and Bartlett Sphericity tests were used. The recommended KMO value was 0.6. To determine the number of questionnaire constructs, eigenvalues greater than 1.0 was considered by the principal component extraction method. The varimax method was used to orthogonally rotation of the acquired constructs. For all 118 items factor-loading > 0.4 were considered(11).
Forth, Criterion validity
In the validity of the criterion, the relationship between the current instrument with the other criterion is tested(11). In this study, a standard instrument (SF36) was used to measure criterion-dependent validity. For this purpose, every 590 participants in this study were asked to complete the WQOLI in addition to SF36 as a completion criterion. Then, using the Pearson correlation test, the scores of the two instruments were compared.
Fifth, reliability
We used internal consistency and test—retest reliability methods with 14-day intervals to assess the reliability of WQOLI. Internal consistency was evaluated using Cronbach's alpha coefficient) 0.70 or more((16). In this research, Cronbach’s alpha of WQOLI was calculated in a sample of 30 women using a convenience sampling method. It was measured for each factor and also for the whole of the questionnaire. Values higher than 0.70 were considered desirable
The stability of a tool shows the reliability of obtained scores following a test-retest administration. In this step, the stability of WQOLI was assessed using the test-retest reliability measurement method. 30 women filled out the questionnaire on two different occasions, at 14-day intervals, and the Persian correlation coefficient was calculated. Values higher than 0.70 were considered desirable(11). Finally, the collected data were imported into version 22.0 for further analysis.