Result of phase 1 (The Translation and adaptation Process of CWBS):
Stage I: forward translation
This stage included the translation and cultural adaptation of the original CWBS from English to Persian to provide more equivalency and highlight ambiguous meanings in the original questionnaire; the CWBS translated by two bilingual translators who had different background or profiles (one PhD nurses, a professional translator with no medical or clinical background). Each translator, whose mother tongue was the Persian language, worked independently and they were blinded to the work of the other translator. They produced two translations. During the forward translation, some problematic words and phrases were identified. In the ADL dimension, there was difficulty translating item 5 (Participating in events at church and/or in the community), of the original version’s Activities domain, “church”, was translated into Persian as “amaken mazhabi [religious places]” by one translator and the other translator used “masjed [mosque]”. However, the phrase “amaken mazhabi [religious places]” was retained by reached consensus between the two translators' reports, because the majority of people in Iran are Muslims and attend in events at mosque and other religious places.
Stage II: Synthesis
The two Persian versions were compared together. Ambiguities or poor wording in the original version was negotiated by the translators. The results of the translations were synthesized to produce the Persian version of the questionnaire by reached consensus between the two translators' reports.
Stage III: Back Translation
Direct translation of an outcome measure developed for one language or culture to another language may not result in a valid instrument. The questionnaire was translated back into the original language to make sure that the translation reflects the same item content as the original to semantic equivalence achievement. Backward translation was done independently by another two competent translators who were blind to the original English version, with no medical background and their first language was English. This step magnified unclear wording and imperfections in the translations. This was a process of validity checking to make sure that the translated version is equivalent in meaning for a consistent translation. Comparing two back-translations with each other and the original, showed no significant differences in meaning and the translations were recognized to be equal. The translated CWBS was then provided to the expert panel.
Stage IV: Reconciliation Experts Committee
The fourth step of the translation process was a final reconciliation with committee of experts to consolidate. This committee, was composed of researchers, a professional translator, a methodologist, and four translators reviewed all versions and reached a consensus on any discrepancy and verified the cross-cultural equivalence of the instrument. This stage was conducted to consolidate all the information from the previous translation stages to achieve cross-cultural equivalence between the source and target version in Semantic, idiomatic, experiential, conceptual. They discussed and reached a consensus on discrepancies and verified the cross-cultural equivalence of the instrument which led to the pre-final Persian version of the CWBS created.
Stage V: Test of the Pre-Final Translated Version
The pre-final version of CWBS was tested through cognitive face to face interview with a convenience sample of 30 participants who was representative of population age, sex and education level. During this step, the interviewer read each item. After which, participants answered to that item, they were asked about the comprehensibility, content, difficulty and ambiguousness with the item. They were given the opportunity to state their viewpoints and suggestions on items because they were familiar with the construct through deep personal experience. This stage resulted in modifications to two items of the original version. In ADL scale: Item 3 (‘‘Attending to medical needs’’ was changed to the ‘‘Attending to medical needs (preventive and curative)); Item 8 (‘‘Making plans for your financial future’’ was changed to “Making plans for your financial future (having job and/or income)‘‘. This stage provided insight into how a person interprets the items on the questionnaire ensured the linguistic and conceptual equivalence of the translation.
Result of phase2
The translated instrument tested for its psychometric characteristics.
Participant characteristics
Descriptive statistics including mean, frequency, and standard deviation (SD) were determined for all variables and were expressed as mean ± SD for normally distributed variables. We invited 150 schizophrenia caregivers to participate in the study (six refused). Of the 144 participants included in the validation process, 75 % were female. Participant ages ranged from 25 to 75 years with a mean of 52 (SD =12.4). The majority of the sample had some college or higher education (91% of the participants were educated. 31% were educated to diploma level. 77% were married, and they were most likely to be parents (40.3% were mothers) and had a mean of 3 children. The caregivers had provided care for a mean of almost 10 years.
content validation (qualitative and quantitative content validity)
This step aimed to assess content validity of the Persian version of the CWBS.
Qualitative content validity
Paying attention to content validity is critical step in assessing an instrument’s validity because If the content of an instrument is an adequate reflection of a construct, then the instrument has a better chance in reaching its measurement objectives.(26) Content validity indicates how well the instrument has the appropriate items that comprise all important aspects of the construct being measured. Content validity is often measured by relying on the knowledge of relevant experts(26). To assess the qualitative content validity, a panel composed of 15 experts in the fields of mental health, who were familiar with the psychometric process, were asked to provide their views on the accuracy of the items’ content in written form regarding its ease of use, wording, grammar, item allocation, and scaling. Some items from the original version were subjected to more clarification and modifications during qualitative content validity steps. In ADL scale: Item7 (‘‘Treating or rewarding yourself’’ was changed to “Treating or rewarding yourself (doing favorite activities, relaxing, exercising, having a hobby, laughing, etc.)”). In BN scale: Item5 (‘‘Feeling good about yourself’’ was changed to “Feeling good about yourself (feeling valuable, having purpose in life) ‘‘). Since the experts couldn’t agree on clarifying item 4 “Having adequate shelter”, the author was asked about it and item was changed to “Having adequate shelter (having a safe place to live in)”.
Quantitative content validity
Content validity ratio (CVR) and content validity index (CVI) were used to achieve the quantification content validity of this scale.
CVR
The content validity ratio (CVR) by Lawshe’s method was used to evaluate whether each item was essential to determine the rejection or retention of items. Eleven experts were each asked to rate each item into one of three categories: (Essential = 3; useful, but not essential = 2, and not essential = 1). Then, the CVR value was computed for each item in the following way: CVR= (Ne - N/2)/(N/2); Where, ne= the number of total experts in panel, N = the total number of experts judging an item ‘essential’ (2 and 3). Based on the Lawshe table(34), a CVR of at least 0.59 was considered acceptable. In this study result of CVR was 0.6 to 1.0. Further details are presented in Table 1.
Table 1.Result of CVR
CWBS
|
Items
|
CVR
|
DL
|
1. Buying food
|
0.6
|
2. Taking care of personal daily activities (meals, hygiene, laundry)
|
1
|
3.Attending to medical needs (preventive and curative)
|
0.6
|
4.Keeping up with home maintenance activities (lawn, cleaning, house repairs, etc.)
|
1
|
5- Participating in events at religious places and/or in the community
|
0.73
|
6- Taking time to have fun with friends and/or family
|
1
|
7- Treating or rewarding yourself (doing favorite activities, relaxing, exercising, having a hobby and laughing, etc.)
|
1
|
8- Making plans for your financial future (having job and /or income)
|
0.73
|
BN
|
1-Eating a well-balanced diet
|
1
|
2- Getting enough sleep
|
1
|
3- Receiving appropriate health care
|
0.88
|
4- Having adequate shelter(having a safe place to live in)
|
0.73
|
5- Feeling secure about your financial future
|
0.88
|
6- Feeling good about yourself (feeling valuable, having purpose in life)
|
1
|
After inclusion of items based on their CVR values in the final scale form, the relevancy of all items was computed by CVI. The CVI is the mean of the CVR values of the retained items. The scale was given to the experts’ panel in order to determine the CVI. They separately assessed the items in the Persian version using a 4‐point Likert scale for relevancy and rated each item from 1 to 4 for relevancy. 2 types of CVI can be computed using: 1) the individual-content validity index (I-CVI), and 2) the scale content validity index (S-CVI). To determine the relevancy of each item I-CVI calculation was conducted using the proportion of experts who rated the items as highly relevant and quite relevant, (a rating of 3 or 4). I- CVI values ranged from 0 to 1 where above 0.79, the items were adequately relevant and were regarded acceptable. If an item scored between 0.70 and 0.79, then it is somewhat relevant, its relevancy is questionable and the item needs to be revised and if the value is below 0.70, the item is not relevant and is deleted(35). In this study all items had ICVI = 1.00 thus, they were considered relevant.
CVI
S-CVI is calculated using the number of items in a scale that has achieved a rating of “highly or quite relevant” by experts. There are two different approaches to determine S-CVI, (the S-CVI/Ave being a less conservative approach). Presenting both indices has been recommended(36) used to calculate S-CVI: 1- (S-CVI/UA) are universal agreements by experts that calculated the proportion of the items in the scale that the experts valued as relevant, while 2- (S-CVI/Ave) is average agreement by experts. In the average approach, the sum of I-CVI in scale is divided by the total number of items. Researchers recommend that the acceptable standard for the S-CVI/UA and the S-CVI-Ave is above 0.8(27, 36) and a value exceeding 0.90 is considered excellent(27). It is recommended that the researchers should mention the approach used for CVI computing because the values obtained from both approaches may be different (37) However, in this study S-CVI was conducted by both indices for all items obtained 1. Thus, the values were considered excellent. Due to the possibility of a chance of agreement in the Content Validity Index, Polit et al., recommended propose Cohen’s modified Kappa coefficient should be calculated. This coefficient is an index of appraiser agreement on the relevance of each item and provides the degree of agreement among experts beyond chance. Kapa coefficient above 0.75 is considered excellent, between 0.6-0.74 good and less than 0.59 weak. Therefore, Kappa coefficient was computed based on following formula: K = (I-CVI – Pc)/ (1- Pc), where Pc = [N! /A! (N-A)!] * 0.5N (26, 35). The kappa statistic values also obtained 1 because all items had received high relevancy scores by the experts.
Constructs validity
Convert and divergent validity
Pearson’s correlation coefficient was used to evaluate the Inter-scale correlations(38). Convert and divergent validities were supported if correlation between BN and ADL with their hypothesized scale was higher than their correlation with the other scales. We considered levels classified as follows: less than 0.30, weak correlation; between 0.30 and 0.50, moderate correlation; and above 0.50, strong correlation(39). As it was shown in table2., both subscales of CWBS showed a significant strong positive correlation with total CWBS (BN: r = 0.81 and ADL: r= 0.88), moderate positive correlation with SOC (BN: r = 0.42 and ADL: r = 0.46), and moderate negative correlation with CBI (BN: r = -0.38 and ADL: r= -0.47), as expected. The correlation relationship between the BN and ADL and with its hypothesis scale (CWBS), was higher than their relationship with other scales.
Table 2.The inter scale correlation between ADL and BN with CWBS, CBI and SOC
|
Convergent
validity
|
Divergent
validity
|
scales
|
Total
CWBS
|
CBI
|
SOC
|
ADL
|
0.88
|
-0.47
|
0.46
|
BN
|
0.81
|
-0.38
|
0.42
|
Exploratory factor analysis
The exploratory factor analysis (EFA) was conducted to determine the construct validity of CWBS. An item would be retained if it was unique and sufficient by loading >0.40 or better (26, 39). Prior to exploratory factor analysis, The Kaiser‐Meyer‐Olkin (KMO) test of sampling adequacy was conducted and Bartlett test of sphericity was analyzed (26). The result of the KMO test of sampling adequacy was 0.781, above the recommended valued of 0.6 and the result of the Bartlett test of sphericity was (X2 = 675.646, and df = 91, P < 0.001), indicating the factor ability of the correlation matrix suitability for factor analysis.
The items to total correlations and item loadings resulted from the factor analysis with Varimax rotation to test the construct validity of the subscales are presented in Table 3.
In the un-rotated solution, most items were loaded on a single factor. As a result, a three factor solution was done. In ADL, all items were loaded on to the same factors as scale developer report (2013) with accounting for 67% and 77% of the variance for ADL and BN, respectively. In ADL all items loaded onto three factors with loadings higher than 0.40, and the lowest and highest loadings were 0.557 (item 3) and 0.833 (item2), respectively. In BN all items loaded onto three factors had factor loadings higher than 0.40, and the lowest and highest loadings were 0.757 (item 1) and 0.951 (item5), respectively. The items of ‘Feeling good about yourself’ (item5), in BN scale, itself was a separate factor from other factors.
Table 3.Result of Exploratory factor analysis.
|
Item
|
Item to total correlation
|
Factor loading
|
Factor1
|
Factor2
|
Factor3
|
Self-Care
|
Buying food
|
0.598
|
0.779
|
|
|
Taking care of personal daily activities (meals, hygiene, laundry)
|
0.577
|
0.833
|
|
|
Attending to medical needs (preventive and curative)
|
0.671
|
0.557
|
|
|
Keeping up with home maintenance activities (lawn, cleaning, house repairs, etc.)
|
0.695
|
0.712
|
|
|
Connectedness
|
Participating in events at religious places and/or in the community
|
0.650
|
|
0.824
|
|
Taking time to have fun with friends and/or family
|
0.641
|
|
0.804
|
|
Time for Self
|
Treating or rewarding yourself (doing favorite activities, relaxing, exercising, having a hobby and laughing, etc.)
|
0.648
|
|
|
0.735
|
Making plans for your financial future (having job and/or income)
|
0.662
|
|
|
0.828
|
Physical Needs
|
Eating a well-balanced diet
|
0.724
|
0.757
|
|
|
Getting enough sleep
|
0.661
|
0.886
|
|
|
Receiving appropriate health care
|
0.720
|
0.787
|
|
|
Self-Security
|
Having adequate shelter(having a safe place to live in)
|
0.710
|
|
0.784
|
|
Feeling secure about your financial future
|
0.689
|
|
0.854
|
|
Emotional Needs
|
Feeling good about yourself (feeling valuable, having purpose in life, etc.)
|
0.625
|
|
|
0.951
|
Reliability
Internal consistency was obtained using Cronbach's alpha. Internal consistency reliabilities for CWBS and its subscales were acceptable. The Cronbach’s alpha was 0.79 and 0.77 for the ADL and BN dimension respectively and a total Cronbach’s alpha of 0.842 for the items. To determine consistency and repeatability of the CWBS, test-retest reliability was computed for the stable group by using Intraclass Correlation Coefficient (ICC) and Pearson correlation coefficient. No information about the previous CWBS scores were provided to the study participants. Reliability coefficients over 0.70 were considered acceptable(40). In the test–retest analysis, the ICC was 0.794 (95% confidence interval [CI], 0.725 to 0.847) and 0.815 (95% confidence interval [CI], 0.751 to 0.863) for the ADL and BN dimension respectively and for the total scale it was 0.872 (95% confidence interval [CI], 0.827 to 0.906). Mean time interval between the initial and second test was 2.5 weeks (SD = 0.69). These results demonstrated that there was not much variability between the test‐retest of the CWBS. Although participants were not informed of their initial test scores, their responses in the retest indicated a high correlation, consistency and reliability.