Attitudes toward persons with ID

There is considerable research on communal attitudes toward persons with IDs [27,28,32–36]. As a psychological concept, attitude lies at the intersection of beliefs and the outcome of a given behaviour [37–39]. It is an interaction between the beliefs of an individual and their projection about the outcome of a given phenomenon [37]. Once individuals have a favorable belief and conceive the outcome of a behaviour as positive, their attitude toward the behaviour will also be positive [38]. However, attitude toward behaviour is fluid—not static—as evolves over time [23,33]. This has contributed to the development of scales to measure attitudes toward individuals with IDs in society. Some of the standard scales used are the Attitudes towards Intellectual Disability Scale (ATTIP; e.g., [28,33,40,41]) and the Community Living attitude Scale for ID (CLAS-ID; e.g., [27,32,34,36,41–43]).

The CLAS-ID comprises four sub-scales (empowerment, exclusion, sheltering, and similarity) which broader and directly captures some of the issues faced by parents who are raising children with IDs in sub-Saharan Africa. Empowerment refers to the self-belief and capacity of persons with IDs to take ownership and make decisions concerning their lives, while exclusion refers to the extent to which people think that individuals with IDs should be excluded in society [36,44]. Sheltering refers to the help needed by individuals with IDs to keep safe, while similarity has to do with the extent to which people think that persons with IDs are similar to others in society [36,44]. It is important to mention here that the nature of participants could sometimes influence their responses on some of the measure sub-scales such as the sheltering [44]. Where the study tends to focus on psychiatrists, they may be inclined to support the sheltering of persons with ID [44]. However, in the study environment, there are limited facilities to support the development of individuals with ID. This means that the participants may not have their own biases while completing the survey.

The barriers encountered by persons with disabilities are very broad which requires holistic approach to rectify the situation. For example, in sub-Saharan African countries such as Cameroon and Ghana, the exclusion of individuals with IDs with respect to the denial of their fundamental rights, societal rejection, and sub-human categorization [5,7,11,19] closely aligns with the concepts of the CLAS-ID. Also, the scale has yielded appropriate reliability scores in different contexts where they have been used to understand communal attitudes toward individuals with IDs [34–36,41–43]. This lends support for the use of the CLAS-ID in studying communal attitudes toward individuals with IDs in contexts where it has never been used. As part of efforts toward promoting the inclusion of persons with IDs across sub-Saharan Africa, the study reported here was conducted to provide baseline information and give directions to policymakers about whether attitudes toward persons with IDs are changing.

Students are the future leaders and are expected to play a central role in promoting the acceptance and full inclusion of individuals with IDs in society. Consequently, several studies have sought to understand the attitude of students toward individuals with IDs (e.g., [15,27,41,42,45–48]). Nevertheless, cross-national studies on student attitudes toward persons with IDs are very rare, with few exceptions being Benomir et al. [32] and Sheridan and Scior [42]. Benomir et al. [32] compared students’ attitudes toward persons with IDs in Libya and the UK, with the students in Libya recording less favorable attitudes toward persons with IDs than their counterparts in the UK. However, this comparison was conducted between different cultural contexts and countries characterized by other contextual differences. In particular, at the time of the data collection in Libya, there was a civil war in the country, which toppled the then government. Without a central authority and the breakdown of social support systems, the findings reported by the authors were expected. This arguably underscores the need for cross-national studies of countries with a similar culture, experiences, and systems and provides a more justifiable basis for comparison.

In another study, which was conducted by Sheridan and Scior [42] in the UK, the authors compared the attitudes of young White British and British Asians toward individuals with IDs. They reported that British Asians had lower scores on exclusion and higher similarity scores than their White British counterparts. Although different groups with diverse identities were recruited for this study, they shared the same characteristic with respect to living in the same context. In our view, to conduct a cross-national study, the groups being studied should reside in different contexts as this would enable a broader understanding of the findings. With commonalities between countries in sub-Saharan Africa such as Cameroon and Ghana, this study sought to develop a cross-national understanding of student attitudes toward persons with IDs.

There are also inconsistencies between the variables influencing people’s attitudes toward persons with IDs [33]. With the exception of contact and relation with a person with an ID, which have been found to be positively associated with attitudes toward persons with IDs (e.g., [18,23,40,42,49–51], inconsistencies have been noted among other background variables. An exception is a study in China that reported no differences between participants in terms of contact and relatives with IDs and their attitudes toward individuals with IDs [36]. With respect to gender, while some studies have reported that males were more positive than females toward persons with IDs (e.g., [33,45]), others have reported otherwise (e.g., [33–35,42,51]). Still, other studies have reported no differences between participants in terms of gender and attitude toward persons with IDs [32,36]. Other studies have reported further inconsistencies regarding general attitudes (positive [33,35,36] or negative [18,27,45] and demographic variables such as age (e.g., [education (e.g., [16], ethnicity [18,42], and religion [34,42]. The fluidity of communal attitudes toward persons with disabilities provides a solid justification for a contextual understanding that could inform national policies and planning. Unfortunately, sub-Saharan African countries’ share of research on the communal acceptance and influence of background variables on attitudes remains rare or unresearched.

The current study

While Cameroon is in central Africa with an estimated population of 26 million [52], Ghana is located in West Africa and has a population of around 30 million [53]. There is a shared understanding of ID in the two countries. For example, from a traditional perspective, the onset of IDs is linked to the punishment meted out to parents for sins committed, making them unwelcome or “unwanted” in most societies [6,54]. As IDs are linked to abnormality, people with IDs are seldom given any form of assistance by the state. Due to the national conception of ID, they become objects of ridicule, humiliation, and mockery [54,55]. In the context of national development, there is limited consideration of persons with IDs in planning and participation in productive activities [9,54,56–58]. Furthermore, in both countries, teachers in regular classrooms generally reject the idea of practicing inclusive education for students with IDs [9,54,56,58]. They believe that they were not trained to teach students with IDs in regular classrooms [24,25,56], thereby leaving students with IDs to access education in special schools [7,58]. However, there are limited special school facilities to accommodate the growing population of individuals with IDs in both countries [8,58]. It is not surprising that a critical mass of individuals with IDs are not in school and are unlikely to be employed in Cameroon [9,54,58] and Ghana [59,60]. Consequently, parents have to shoulder the burden of raising their children with IDs in hostile social contexts.

Despite the rejection and discrimination perpetuated against persons with IDs in both countries, they are both signatories to the CRPD, which encourages countries to formulate policies or intervention programs aimed at the inclusion of persons with IDs. While Cameroon signed the CRPD in 2008, Ghana ratified it in 2012. In Cameroon, the government passed decree No 2018/6233, which sought to encourage the acceptance of persons with IDs as well as promote their fundamental rights to essential services [61]. In Ghana, the government passed Disability Act 715, 2006, with the sole aim of promoting the fundamental rights and civil liberties of persons with disabilities, such as those living with IDs [62]. However, studies conducted in both countries point to the difficulties faced by persons with IDs and their families in their day-to-day living experiences (e.g., [9,55,57–59]). Research on communal attitudes toward individuals with IDs is very rare in both Cameroon and Ghana, and as students are important to shaping future policies, it is necessary to develop insights into their perspectives toward persons with IDs in both countries. Therefore, the purpose of this study was to develop deeper insight into students’ attitudes toward individuals with IDs using the CLAS-ID in the sub-Saharan African context, specifically Cameroon and Ghana, where the tool has never been implemented. The study reported here drew on students from diverse background who somehow mirrors the differences among the people in the society. Indeed, the students are the next frontiers who are expected to be change agents and leaders of public institutions. In view of this, their perception would help to understand whether they would include individuals with ID in national development. Thus, the findings of this study could reflect what patterns in the society with respect to how the society perceive individuals with ID.

We formulated four study hypotheses: first that the CLAS-ID will yield a suitable structure for its use in sub-Saharan Africa; second, that there will be no difference between students on the basis of their country of origin; third, that female students will hold more positive attitudes toward individuals with IDs than their male counterparts; and fourth, that having a relative or contact with an ID will be positively associated with attitudes toward persons with IDs. To confirm or disprove our hypotheses, this cross-national study was guided by the following research questions: 1) What is the underlying factor structure of the CLAS-ID in sub-Saharan African countries such as Cameroon and Ghana? 2) What is the association between students’ profile and attitudes toward individuals with IDs in Cameroon and Ghana? 3) What are the predictors of attitudes toward individuals with IDs across Cameroon and Ghana? 4) Does country (as a variable) moderate the relationship between students’ profile and attitudes toward persons with IDs?

Study participants

The participants of this study were university students pursuing various undergraduate programs in both Cameroon and Ghana. In the view of the authors, this study is the first of its kind to be conducted in sub-Saharan Africa. As such, we recruited students exclusively from public universities where the influence of the state can impact course developments and school policies. Therefore, one public university from each country was selected based on convenience and ease of access of the research team. While the university in Cameroon is mainly social science oriented, the university in Ghana is a medical school for training health professionals. The recruitment process was guided by the following inclusion criteria: The participant a) had to be enrolled in the institution selected for the study; b) had to be at any of the various stages of undergraduate study; c) had a fair knowledge or understanding of IDs; and c) had to be at least 18 years and able to read and sign the consent form.

Printed questionnaires were distributed to the students in both countries. Out of 950 questionnaires distributed, 770 were received by the research team, representing a response rate of 81%. However, 29 questionnaires were excluded because they were partially incomplete or did not indicate the country of data collection. Consequently, 741 questionnaires were retained for this study (see Table 3 for details), of which 413 were recruited from Cameroon and 328 from Ghana. Fifty-one percent of the participants were female compared to 49% male; 71% were between the age of 18 and 21 years compared to 29% who were at least 22 years. While 86% indicated that they had no relative with IDs, 14% indicated that they did.

Instruments

A two-part questionnaire was used for the data collection. The first part of the instrument collected information about the participants’ demographic characteristics, such as country, gender, age, year of study, whether they had a relative with IDs, frequency of contact with people with IDs, and religion. The decision regarding which demographic variables to include in this study was based on a review of the literature (e.g., [18,34,35]).

The second component of the questionnaire was the revised CLAS-ID [63]. The scale comprised 17 items anchored on a five-point scale, with responses ranging from 1 (strongly disagree) to 5 (strongly agree). The scale consisted of four main factors: empowerment, exclusion, sheltering, and similarity. The scale was given to two academics each in Cameroon and Ghana to verify whether the content was appropriate for data collection in the two countries. All the academics responded that the content was appropriate for data collection however, some iterations were suggested which were incorporated in the final draft used for data collection. Some of the items on the scale included the following: “A person would be foolish to marry a person with an intellectual disability,” “People with intellectual disabilities should live in special facilities because of the dangers of life in the community,” and “People with intellectual disabilities can be productive members of society.”

The scale yielded an appropriate reliability score, which was computed using Cronbach’s alpha (CLAS-ID = .80). Mean scores—the sum scores divided by the number of items—were reported in this study to facilitate understanding of the results. Thus, a mean score of at least 4 was interpreted as a positive attitude toward persons with IDs.

Procedure

The study received institutional approvals from the University of Health and Allied Health, Ghana, and the Institute of Governance, Humanities and Social Sciences of the Pan African University, University of Yaounde II, Cameroon, before the data collection. After the approval, formal letters were sent to the various departments in the schools requesting permission to collect data from the students. While author three collected data in Cameroon, author five collected data in Ghana. Before the lecture, the authors informed the class about the study and waited until the end of the classes to distribute the questionnaires to students who met the inclusion criteria and consented to take part in the study. Before distributing the questionnaire, the study focus, objectives, and the inclusion criteria were explained to the students. They returned the questionnaire one week later when the authors attended the same classes to pick up the questionnaires directly from them.

With respect to the Cameroonian data, the questionnaires were both in English and French since both languages are the medium of instruction and officially recognized languages. Each document given to the students had both English and French versions and the students had an option to either complete the English or the French version. The questionnaire administered in Ghana was in English. The students spent about 20 to 25 minutes to complete each questionnaire. The data were collected between March 2018 and November 2019, and the participants were informed that their participation was voluntary and that there was no financial reward or incentive for completing the questionnaire. They were also told that their identity, department, and class would not be included in the reporting of the study. The questionnaires were distributed to those who met the inclusion criteria and signed the written consent form.

Data analysis

The questionnaires were first entered into Microsoft Excel for cleaning before being transferred to SPSS, version 26, for analysis. The instrument was then transferred to AMOS, version 27, for computation of the factor analysis. In the first stage, the authors assessed the reliability of the data before proceeding with the data analysis. Using boxplots and histograms, the authors concluded that the data were normally distributed based on a visualization of the outputs.

The authors then proceeded to answer the research questions. To answer research question 1, factor analyses were computed to understand the structural validity of the CLAS-ID. In order to ascertain the validity of the instrument, we followed the format of Su et al. [36] using both confirmatory factor analysis (CFA) and principal component analysis (PCA). According to Worthington and Whittaker [64], the use of CFA and PCA is appropriate to get an accurate understanding of the validity of an adapted scale in a new context. CFA, which was performed using AMOS (version 27), was computed to assess the unidimensionality and validity of the latent constructs of the CLAS-ID. According to Schumacker and Lomax [65], in CFA, the researcher specifies the number of factors and estimates the factor loadings and model parameters before testing the model. For PCA, SPSS was used for the computation. The assumption of inter-reliability correlations was assessed to ensure the suitability of the data for PCA. We then computed mean scores for both the overall attitude scale and sub-scales.

To answer research question 2, independent samples t-tests and analysis of variances (ANOVA) were computed for the demographic variables, with two and at least three levels, respectively. The t-tests and ANOVAs were computed to understand the association between the demographic variables and attitudes toward persons with IDs. Here, we used Levene’s test to ensure that the assumption of homogeneity of variance was not violated. For the t-tests, in the event of violation, the results of the equal variances assumed, which were supposed to be reported, were not considered. However, the equal variances not assumed were reported. For the ANOVAs, in the event of violation, the results of the Welch statistics were reported. To determine the weight of both the t-tests and ANOVAs, we computed partial eta-squared. The effect sizes were interpreted as follows: small = .01 to .05; moderate = .06 to .9; and large = at least 1 [66].

To answer research question 3, a linear regression was computed to ascertain the predictors of attitudes. Before computing the linear regression, we used Pearson’s moment co-efficient correlations to understand the associations between the sub-scales. Although our intention was not to determine causative effect, we sought to understand the strength of the associations. The correlations were interpreted as follows: small (r = .10 to .29); medium (r = .30 to .49); and large (r = .50 to 1.0) [66]. Afterwards, we proceeded to compute the predictors of attitudes using linear regressions. An initial analysis showed that there was no violation of the following assumptions: assumptions of normality, linearity, multicollinearity, and homoscedasticity.

To answer research question 4, a two-way factor analysis was used to determine the moderation effect of country on attitude toward persons with IDs. We checked the output for Levene’s test to ensure that the homogeneity of the variances was not violated. Again, we assessed the effect size using partial eta-squared.

Structural validity of CLAS-ID

First, CFA was computed for 17 observed variables used to assess attitudes toward persons with IDs in Cameron and Ghana. As shown in Fig 1 and Table 1, the results point to four latent factors in reflecting the associated observed variables. These factors were exclusion, similarity, empowerment, and sheltering.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. CFA of CLAS-ID scale.

https://doi.org/10.1371/journal.pone.0257482.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Standardised regression weights of CLSD scale.

https://doi.org/10.1371/journal.pone.0257482.t001

The standardized regression weights presented in Table 1 ranged from -.821 to .773, suggesting that some of them were far below the acceptable threshold of 0.5 (see [67]). In addition, the model produced a squared multiple correction ranging from 0.000 to 0.675. As reported by Awang [67], in CFA, all items should have a squared multiple correlation of 0.4 or higher. This suggests that some of the items in the present model were below standard.

Several goodness-of-fit indicators can be used to test the adequacy of a CFA model (e.g., [62,67–69]). For example, Awang [67] reported that goodness of fit is evaluated by employing the chi-square (χ2) (P > .05), the comparative fit index (CFI > .90), the Tucker–Lewis Index (TLI > .90), and the root mean-square error of approximation (RMSEA < .08).

The chi-square value was significant (χ² (113) = 837.36, p = .001). The other model fit indices were as follows: CFI = 0.80; TLI = .70; and RMSEA = .080. These results suggest that the model did not fit the data acceptably in the population since all the indicators fell outside the commonly agreed threshold.

Second, the 17-item CLAS-ID was subjected to PCA. The assumptions of inter-reliability correlations were assessed to ensure its suitability for PCA. The output inspection showed that most had a correlation co-efficient of .30. Also, the Kaiser-Meyer-Oklin value was .77, suggesting its appropriateness. Further, the Bartlett test of sphericity was statistically significant at .001.

The PCA showed that four components had eigenvalues exceeding 1. An inspection of the scree plot showed four clear breaks. Since at least three items loaded on each factor, we decided to retain the four factors of the CLAS-1D. The four-component factors explained a total of 54% of the variance, with factors I, II, III, and IV contributing 21%, 16%, 10%, and 7%, respectively. The communality values for all the items were above .03, which means that all the items loaded properly on the factors.

To further support our analysis, an oblimin rotation was conducted to compare or support our interpretation. The rotation supported the four above-identified factors. For the oblimin rotation, the Kaiser-Meyer-Oklin was .77, and Bartlett’s test of sphericity was significant at .001. The four factors loaded strongly, with all exceeding the .3 threshold. The communality values also exceeded .3 for all the items.

We determined the names of the factors based on their comparability with the original scale (See Table 2). Thus, Factor I was called exclusion (7 items = 1, 2, 7, 8, 9, 10, 12), Factor II similarity (3 items = 13, 14, 15), Factor III empowerment (3 items = 3, 4, 5), and Factor IV sheltering (4 items = 6, 11, 16, 17).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Summary of factor loading of CLAS-ID.

https://doi.org/10.1371/journal.pone.0257482.t002

Student profile and attitudes

Independent samples t-tests were conducted to compare the association between the demographic profile of students and their attitudes toward persons with IDs (see Table 3). On overall attitudes, significant differences were found between attitudes and two demographics variables, country and age. However, students in Cameroon were more receptive of persons with IDs than their counterparts in Ghana, t (739) = 3.02, p = .002, with a small effect size (partial eta-squared = .01). With respect to age, younger students in both countries reported more positive attitudes toward persons with IDs than older students, t (739) = 3.00, p = .003, with a very small effect size (partial eta-squared = .01).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Association between demographics and attitudes.

https://doi.org/10.1371/journal.pone.0257482.t003

Other interesting differences were identified in the sub-scales. With respect to gender and having a relative with an ID, there were differences between the participants in terms of exclusion and similarity. Regarding exclusion, male students had more positive attitudes than female students. However, the female students were more positive than their male counterparts in terms of similarity.

With respect to exclusion, the participants who indicated that they had a relative with an ID were more positive than those who indicated otherwise. However, regarding similarity, those who said that they did not have a relative with an ID were more positive than those who indicated otherwise. Further, one-way between-groups ANOVAs were conducted to explore the association between three levels of demographics and attitudes toward persons with IDs (see Table 3). On overall attitudes, there were significant differences between the participants only in terms of religion, F (2, 738) = 7.09, p = .001. However, the effect size was small (partial eta-squared = .02). A post-hoc comparison using the Tukey HSD test indicated that students who were African traditionalists were more positive than those who were Christians and Muslims.

Predictors of attitudes

The associations between the sub-scales were computed using Pearson’s correlation co-efficient. While there was a correlation between exclusion and similarity (r = -232, p = .001) and exclusion and sheltering (r = .19, p = .001), there was no correlation between exclusion and empowerment (r = .04, p = .35). Also, there was a positive but small correlation between similarity and empowerment (r = .25, p = .001) and similarity and sheltering (r = .23, p = .001). In addition, while there was no correlation between empowerment and sheltering (r = .03, p = .42), there was a very small correlation between sheltering and exclusion (r = .19, p = .001).

The predictors of attitudes toward individuals with IDs were assessed using linear regression (see Table 4). The demographic variables contributed only two percent to the variance in attitudes, F (7, 739) = 2.25, p = .03. In terms of the demographics, only gender (beta = -.25, p = .04) and age (beta = -2.26, p = .02) contributed significantly to the variance in attitudes.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Summary of demographics regressed on attitude.

https://doi.org/10.1371/journal.pone.0257482.t004

Effects of country and demographics on attitudes

A two-way between-groups ANOVA was computed to assess the impact of country and other demographics on attitudes, as measured by the CLAS-ID (see Table 5). Country interacted only with religion. The interaction effect between country and religion was statistically significant, F (2, 735) = 4.64, p = .01, with a very small effect size (partial eta-squared = .004). Overall, the African traditionalists (M = 3.32; SD = .58) were more positive than those in other religions (Christians, M = 3.26, SD = .57; Muslims, M = 3.02, SD = .60). However, a post-hoc comparison using Tukey HSD showed a significant difference between Muslims and the other two religions, which did not differ from each other. In both countries, the African traditionalists appeared slightly more receptive toward individuals with IDs than those in the other religions.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Summary of two-way analysis of variance.

https://doi.org/10.1371/journal.pone.0257482.t005

Study limitations

A number of study limitations could affect the interpretation of the findings reported here. First, the data comprised self-evaluations from participants who responded to a standard international tool used for measuring communal attitudes toward individuals with IDs. It was beyond the scope of this study to ascertain or observe the actual reaction of the participants in their encounters with persons with IDs. However, adequate information was provided to the participants before they took part in this study, and we are confident that they provided appropriate responses to the items on the scale. Second, the CFA did not confirm the structural validity of the CLAS-ID in this study. Although it supported the four-factor structure, the measurement indices did not reach the required thresholds. Previous studies have encountered similar situations where they used only PCA to validate the factor structure of the CLAS-ID [29]. Nonetheless, our study confirmed the four-factor structure of the CLAS-ID, and we recommend its use as a valid tool to measure communal attitudes toward individuals with IDs. Third, one institution was selected in each country. Therefore, the results may not be representative of the views of all university students in both countries. With the internationalization of education and diversity among students, we contend that the profile of the students may not be different from what may obtain in other institutions. Indeed, this study has filled a major gap in the literature through the use and validation of the CLAS-ID in the sub-Saharan African context (Cameroon and Ghana).

Declarations

The authors wish to thank all the students who took part in this study.