Attitudes towards individuals with disabilities as measured by the Implicit Association Test: A literature review

Research in Developmental Disabilities 35 (2014) 294–321

Contents lists available at ScienceDirect

Research in Developmental Disabilities

Review article

Attitudes towards individuals with disabilities as measured by the Implicit Association Test: A literature review Michelle Clare Wilson *, Katrina Scior Department of Clinical Educational and Health Psychology, University College London, Gower Street, London WC1E 7HB, UK

A R T I C L E I N F O

A B S T R A C T

Article history: Received 24 June 2013 Received in revised form 5 November 2013 Accepted 6 November 2013 Available online 4 December 2013

Research investigating attitudes towards individuals with disabilities has largely focused on self-reported explicit attitudes. Given that factors such as social desirability may influence explicit attitudes, researchers have developed tools which instead assess less consciously controllable implicit attitudes. Considering research on implicit attitudes thus seems pertinent. A review of studies measuring implicit attitudes towards individuals with physical disabilities (visual, motor or hearing) or intellectual disabilities via the Implicit Association Test (IAT; Greenwald, McGhee, & Schwartz, 1998) was carried out. Systematic searches of PsycINFO, CINAHL, EMBASE, ERIC, MEDLINE, PUBMED, Scopus and Web of Science databases identified relevant articles published between January 2000 and September 2012. Seventeen articles (reporting on 18 studies that employed the IAT) were identified. These investigated implicit attitudes towards individuals with; physical disabilities (N = 13), intellectual disabilities (N = 3), both physical and intellectual disabilities (N = 1), and ‘unspecified disabilities’ (N = 1). Across all studies, moderate to strong negative implicit attitudes were found and there was little to no association between explicit and implicit attitudes. Individuals’ beliefs about the controllability of their future, sensitivity to the concept of disease, and contact with individuals with disabilities appear to be associated with implicit attitudes. A consistent pattern of moderate to strong negative implicit attitudes towards individuals with disabilities was evident. These studies provide a starting point, but methodological issues related to sampling and the employed IATs limit the generalizability of these results. Further research investigating implicit attitudes towards specific disability types, with a wider subject pool are necessary as well as further investigation of factors that contribute to these attitudes. ß 2013 Elsevier Ltd. All rights reserved.

Keywords: Implicit attitudes Physical disabilities Intellectual disabilities Disability Implicit Association Test Review

Contents 1.

2.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Attitudes towards individuals with disabilities 1.1. Implicit attitudes . . . . . . . . . . . . . . . . . . . . . . . . 1.2. Aims and objectives . . . . . . . . . . . . . . . . . . . . . . 1.3. Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Search strategy . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Inclusion and exclusion criteria . . . . . . . . . . . . . 2.2.

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

* Corresponding author. Tel.: +44 20 7679 1897; fax: +44 20 7916 1989. E-mail addresses: [email protected] (M.C. Wilson), [email protected] (K. Scior). 0891-4222/$ – see front matter ß 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ridd.2013.11.003

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

295 295 295 296 296 296 296

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

3.

4.

5.

2.3. Definition of disability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Measures of implicit attitudes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. Results of search strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Overview of studies included in the review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Study characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Study location, design and sampling strategies employed . . . . . . . . . . . . . . . . . . . 3.2.1. Sample characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2. IAT characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.3. Quality rating of the studies included in the review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Implicit attitudes towards disabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Implicit attitudes towards individuals with physical disabilities . . . . . . . . . . . . . . 3.4.1. Implicit attitudes towards individuals with ID . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.2. Implicit attitudes towards individuals with ID and physical disabilities . . . . . . . . 3.4.3. Implicit attitudes towards ‘disabilities in general’ . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.4. Relationship between explicit and implicit attitudes towards individuals with disabilities. 3.5. Limitations of the studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Limitations of the review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

295

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

296 297 297 297 297 297 297 312 313 314 314 315 316 316 316 316 318 318 319 319 319

1. Introduction The recently published World Report on Disability (World Health Organisation and World Bank, 2011) suggests there are approximately one billion people across the globe with some form of disability, representing around 15% of the world’s population. This is a sizeable increase from the last World Health Organisation estimate of around 10% in the 1970s (World Health Organisation and World Bank, 2011). Over the last few decades there has been considerable work towards achieving equality for individuals with disabilities, e.g., inclusion within local communities, education and equal opportunities within work contexts. Many national and international strategies and pieces of legislation have these principles at their core (e.g., Americans with Disability Act, 1990; Disability Discrimination Act of Australia, 1992; Disability Discrimination Act of the UK, 1995, 2005; United Nations Convention on the Rights of Persons with Disabilities, 2008). There is evidence to suggest that to some extent these principles have been implemented within communities, education settings and workplaces (e.g., Favazza, Phillipsen, & Kumar, 2000; Riches & Green, 2003; Schwartz & Armony-Sivan, 2001). One factor which remains a barrier to these goals being fully achieved concerns negative attitudes held towards individuals with disabilities (e.g., Department of Work and Pensions, 2002; National Coordinating Agency for Population and Development, 2008; National Disability Authority, 2011; Office of Disability Issues, 2011). 1.1. Attitudes towards individuals with disabilities Within the last 30 years, there has been increasing interest in researching attitudes towards individuals with disabilities, factors that influence such attitudes, and potential approaches to changing negative attitudes (e.g., Brillhart, Jay, & Wyers, 1990; Daruwalla & Darcy, 2005; Hunt & Hunt, 2004; Paris, 1993; Strohmer, Grand, & Purcell, 1984). Researchers have largely focused on measuring explicit attitudes, i.e., attitudes which are consciously accessible and controllable (Prestwich, Kenworthy, Wilson, & Kwan-tat, 2008). These are often measured via self-report questionnaires. Public awareness of the need for equal opportunities and rights for individuals with disabilities has been heightened in campaigns by disability rights groups. Given this greater awareness one may question how accurate self-reported attitudes are and whether other factors may be influencing the attitudes reported in such studies. Antonak and Livneh (2000) suggest that measuring explicit attitudes towards individuals with disabilities poses several risks to validity. One such risk is respondent reactivity, where respondents realise their attitudes are being measured and attempt to alter their responses. This can result from a number of factors, but perhaps one of the most important to consider here is the effect of social desirability, i.e., when an individual is motivated to endorse responses which they believe to be the most socially appropriate (Antonak & Livneh, 2000). 1.2. Implicit attitudes In view of the risks of response biases, such as socially desirable responding, researchers have developed a variety of measures for assessing implicit attitudes, i.e., attitudes which are automatically activated and occur without effort or intention (Prestwich et al., 2008). One such measure is the Implicit Association Test (IAT; Greenwald et al., 1998). The IAT is a computer based task that measures the relative strength of association between pairs of concepts and works by using four different groups of images/words presented on the screen (Lane, Banaji, Nosek, & Greenwald, 2007).

296

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

Participants have to categorise the presented image/word using one of two keys on the keyboard. Two of these groups are target concepts (e.g., white and Asian) and the remaining two, a pair of attribute concepts (e.g., pleasant and unpleasant). Initially one group from the target concepts and one group from the attribute concepts are paired to one response key and the others to the second response key. The pairings are then swapped over. Responses are facilitated (and hence faster and more accurate) when closely associated categories share the same response key compared to when they do not (Lane et al., 2007). The difference in speed of response to the different pairings indicates the strength of the association between the two different pairings. Therefore if the first set produces a faster response than the second, this is taken to indicate a respondent’s relative implicit preference for the first set of pairings. Studies suggest that the IAT is a valid and reliable measure of implicit attitudes (e.g., Cunningham, Preacher, & Banaji, 2001; Egloff & Schmukle, 2002; Greenwald & Nosek, 2001). In addition research has repeatedly shown that there is often little to no correlation between explicit attitude measures and the IAT, and that the IAT is fairly robust against the effects of faking (e.g., Banse, Seise, & Zerbes, 2001; Steffens, 2004). When one considers all these issues together, it would seem pertinent that available data on implicit attitudes towards individuals with disabilities be considered as separate from that of research solely examining explicit attitudes. This body of evidence may potentially render a more accurate insight into attitudes towards people with disabilities. 1.3. Aims and objectives The aim of this review was to summarise the findings of studies that have investigated implicit attitudes towards individuals with disabilities. The review: 1. Examined (where possible) the relationship between implicit and explicit attitudes towards individuals with disabilities. 2. Examined any potential differences in the strength and extent of implicit attitudes towards individuals with different types of disabilities. 2. Method 2.1. Search strategy Studies published between January 2000 and September 2012 were identified by electronic searches of the following databases: PsycINFO, CINAHL, EMBASE, ERIC, MEDLINE, PUBMED, Scopus and Web of Science. Search terms used were: ‘Attitudes’, ‘Implicit’, ‘Disabilit*’, ‘Indirect’, ‘Attitudes towards’. These were combined in various ways using the Boolean terms ‘OR’ and ‘AND’. Of the 540 articles produced from the initial search, the titles and abstracts were individually examined. Articles that were duplicated or not relevant were deleted (N = 502). Full text copies of the remaining 38 articles were obtained and thoroughly reviewed against the inclusion and exclusion criteria. In addition to this the reference lists of the articles of interest were reviewed to identify any additional articles for inclusion (N = 44). See Table 1 for detailed numbers of articles identified through each database. 2.2. Inclusion and exclusion criteria For an article to be included it had to have been published in English between January 2000 and September 2012, with the primary focus being the investigation of implicit attitudes towards individuals with disabilities. Both published journal articles and unpublished dissertations were considered for inclusion. 2.3. Definition of disability The International Classification of Functioning, Disability and Heath (World Health Organisation, 2001) define a disability as the interaction between an individual’s health condition (e.g., paralysis, depression, deafness) and pertinent Table 1 Number of papers retrieved from each database search. Database

Number of papers

PsychInfo CINAHL EMBASE ERIC MEDLINE Pubmed Scopus Web of science

105 37 16 38 43 55 151 95

Total

540

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

297

environmental and personal factors (e.g., negative attitudes and self-esteem). This classification does not distinguish between specific types or causes of disability such as ‘mental’ and ‘physical’ health conditions and thus includes a very wide range of disabilities. The inclusion of all these disabilities would be beyond the scope of this article. Therefore for this review, the following disability categories were included: intellectual disabilities (ID) and physical disabilities (including visual, hearing and motor). This decision was informed by the two very broad categories of disability to which attitudes have been fairly extensively studied. 2.4. Measures of implicit attitudes There are several different ways of measuring implicit attitudes including evaluative priming tasks (e.g., Hornstra, Denessen, Bakker, van de Bergh, & Voeten, 2010), error choice methods (e.g., Clarke & Crewe, 2000) and prototypes (e.g., McCaughey & Strohmer, 2005). As discussed by Lane and colleagues (2007) however, several studies have identified weak correlational relationships between different measures of implicit attitudes (e.g., Bosson, Swann, & Pennebaker, 2000; Fazio & Olson, 2003). They suggest that this poor relationship may partly relate to the often low reliability of at least some of the measures used, resulting in underestimation of the relationship between them. In addition, they suggest that the weak relationship may in fact be related to the potential heterogeneity of cognitive processes tapped by different implicit attitude measures (Lane et al., 2007). Given the potential difference in measurement by the various implicit attitude tools available, it was decided to include only studies that employed the IAT (Greenwald et al., 1998). This decision was reached because the IAT is a well-researched, valid and reliable measure of implicit attitudes that has frequently been used to investigate implicit attitudes towards various social groups (e.g., Cunningham et al., 2001; Egloff & Schmukle, 2002; Greenwald & Nosek, 2001). By only including studies that used the IAT, it was hoped that this would allow for more comparable evaluations of the results obtained by each study as they would all have been derived from the same measurement tool. Interested readers may wish to consult the above references for examples of studies that have employed other methods for the evaluation of implicit attitudes. 3. Results of search strategy An overview of the search strategy results as well as explanations for exclusion of articles at each stage is provided in Fig. 1. 3.1. Overview of studies included in the review In total 17 articles were included. These articles reported the results of 20 studies, 18 of which used the IAT. All articles reported at least one study that employed an IAT. See Table 2 for an overview of all articles included. 3.2. Study characteristics The following section considers various study characteristics and examines methodological factors pertinent to the use of the IAT across all studies included. 3.2.1. Study location, design and sampling strategies employed The majority of studies were carried out in the USA (N = 12), two each in China, Italy and the UK, and one each in Germany and France. The majority of studies employed a cross-sectional quantitative design (N = 16), three an experimental betweensubjects design, and one a qualitative design (focus group). Of the studies employing experimental designs, one manipulated type of information presented prior to the presentation of the IAT; one manipulated type of IAT presented; and one used a quasi-experimental repeated measures design with two groups. All 20 studies included in the review sampled volunteers. One study (Proctor, 2012) initially attempted to obtain a random sample of all potential participants in the population of interest (child protection workers), however due to very low initial response rates (less than 8%) the remainder were recruited via convenience sampling. 3.2.2. Sample characteristics Sample sizes ranged from 30 to 245 participants (M = 100, SD = 67); the total number of participants across all studies being 2009. Over half of these participants were taken from undergraduate and postgraduate student samples (N = 1059). The remaining participants were health care profession students (N = 271), social care workers (N = 249), health care educators (N = 143), members of the general public (N = 107; of whom 51 were children), individuals who worked with (or were in close relation with) individuals with disabilities (N = 85), parents of individuals with disabilities (N = 38), ‘specialised’ (N = 19) and ‘non-specialised’ teachers (N = 26), and parents of typically functioning children (N = 13). Nineteen of the 20 studies sampled participants aged 18 years or older, and one study participants aged between five and six years old. Although participants were drawn from various different populations, some participant characteristics may have introduced bias in the results reported. The first potential source of bias lies in the fact that over half of participants were undergraduate and postgraduate students. Previous explicit attitude research has suggested individuals with higher

298

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

540 papers generated from the electronic database search

372 duplicate references removed leaving 168 papers

Titles and abstracts for all remaining 168 papers reviewed

130 irrelevant papers removed leaving 38 papers for detailed review against inclusion and exclusion criteria

From these 38 papers, 44 additional papers identified from screening of reference lists

65 references removed: 6 not available in English, 2 book chapters, 1 conference abstract only, 2 dissertations whose results were published in 2 papers to be included in the review, 9 did not use the IAT to measure implicit attitudes, 3 evaluated attitudes towards a disability category not included in this review, 38 did not measure implicit attitudes and for 3 primary focus not measuring implicit attitudes towards disability

Total number of papers included in the review: 17 Fig. 1. Flow diagram of search strategy procedure.

educational attainments tend to have more positive attitudes towards individuals with disabilities than those with lower educational attainments (e.g., Office of Disability Issues, 2011; Yazbeck, McVilly, & Parmenter, 2004). Of the studies that included students (N = 14), 13 sampled students only and hence were not able to compare results between individuals from different populations. The remaining study compared the results relating to students with those for caregivers and nonstudents (Enea-Drapeau, Carlier, & Huguet, 2012). Their results identified a statistically significant difference between caregiver attitudes and those of students and non-students, with caregivers holding more positive implicit attitudes than students and non-students. Unfortunately it was not reported whether the difference between students and non-students was statistically significant. A gender bias was also noted as the vast majority of participants were female (N = 1209), with only 422 male participants across all studies.1 Previous research into explicit attitudes has suggested that women often report more positive attitudes towards individuals with disabilities than men (e.g., Antonak & Livneh, 1988; Hergenrather & Rhodes, 2007; Office of Disability Issues, 2011). Given these findings, consideration of possible gender differences in implicit attitudes seems pertinent. Of the studies that included information on the gender of participants (N = 17), seven examined for gender differences in implicit attitudes. Of these, six found no significant differences. The one remaining study did identify a difference, in that women held more positive implicit attitudes than men (Chen, Ma, & Zhang, 2011). Therefore, contrary to explicit attitude research, most studies that considered the effect of participant gender found no difference in implicit attitudes towards individuals with disabilities between male and female participants. The lack of examination of gender differences in the majority of studies is problematic though. It provides the potential for an unidentified bias in reported attitudes given the much higher proportion of female participants across all studies. It is also worth noting that the ratio of females to males was not equal across studies, ranging from approximately 1:1 to 9:1. This brings into question the comparisons made, which in some cases were between two very different sized groups.

1 These figures do not tally with the total number of participants across all studies included as not all studies provided information on participant characteristics.

Table 2 Overview of studies included in the review. Study

Location and sample characteristics

Disability investigated

Aaberg (2012)

USA

Physical disabilities (visual and motor)

Disability related IAT categories and items used

Implicit

Explicit

Categories

Items

Disability IAT (DA-IAT; Greenwald et al., 1998)

None

‘Disabled’, ‘Able bodied’, ‘Good’, ‘Bad’

- ‘Disabled’: 3 symbols related to physical disabilities

Nurse educators from BSc Nursing degree courses

- ‘Able bodied: 4 symbols representative of ‘able bodied’

N = 132. No further sample characteristic data available

- ‘Bad’: 6 words (rotten, angry, terrible, bomb, nasty, hate) - ‘Good: 6 words (joy, love, pleasure, peace, wonderful, excellent)

USA

Physician assistant students

Physical disabilities (visual and motor)

Disability IAT (DA-IAT; Greenwald et al., 1998)

None

‘Disabled’, ‘Able bodied’, ‘Good’, ‘Bad’

- ‘Disabled’: 3 symbols related to physical disabilities

- ‘Able bodied’: 4 symbols representative of ‘able bodied’

Method

Main findings including (where available) D-scores on the IATa

Cross sectional design. Participants contacted via email and completed the IAT online on a computer

- Mean IAT scores (D = 0.76) indicated a strong implicit preference for able bodied individuals (i.e., negative implicit attitudes towards individuals with disabilities) - Mean IAT score for this sample was considerably greater than that of the average DA-IAT score reported on the Project Implicit website (N > 38,000; D = 0.45)

Cross sectional design. Completed IAT on computer 2 weeks before class, submitted results to lecturer. Did the PBL lesson, discussed impact of bias on clinical decisions, and usefulness of IAT in learning. Follow-up IAT at 3 months

- At Time 1 the majority of participants had strong (47%) or moderate (33%) preferences for abled bodied individual (suggesting moderate to strong negative implicit attitudes towards physical disabilities)

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

Archambault et al. (2008)

Implicit and explicit attitude measures used

- At Time 2, there were still large numbers of participants with strong (33%) or moderate (31%) negative implicit attitudes 299

300

Table 2 (Continued ) Study

Location and sample characteristics

Disability investigated

Implicit and explicit attitude measures used

Disability related IAT categories and items used

Implicit

Categories

Explicit

Method

Main findings including (where available) D-scores on the IATa

Items - ‘Bad’: 6 words (rotten, angry, terrible, bomb, nasty, hate)

39 female, 10 male; mean age 26.8years

- ‘Good’: 6 words (joy, love, pleasure, peace, wonderful, excellent)

- Qualitative results suggest that using the IAT helped participants consider the impact of implicit bias on clinical decisions

Time 2: N = 27 Chen et al. (2011)

China

Physical disabilities (visual and motor)

A disability IAT developed by the researcher

AIDS Attitude measure (Neumann, Hulsenbeck, & Seibt, 2004) with wording adapted for this study

‘Disabled’, ‘Nondisabled’, ‘Pleasant’, ‘Unpleasant’

- ‘Disabled’: 4 symbols related to physical disabilities

Undergraduate students

- ‘Non-disabled’: 4 symbols representative of ‘able bodied’

N = 56; 39 female, 17 male; 23 rural and 33 urban residents; mean age 20 years

- ‘Pleasant’: 6 words (honour, lucky, freedom, happy, jolly, love)

Cross sectional design. Students volunteered and individually completed first the IAT, then explicit measures on a computer

-Individuals from rural settings and men held more negative implicit attitudes than individuals from urban settings and women respectively - Mean explicit attitudes were positive towards individuals with disabilities - No correlation between implicit and explicit attitude scores

- ‘Unpleasant’: 6 words (evil, sad, poverty, disaster, dirty, worthless)

Enea-Drapeau et al. (2012)

France

ID (Trisomy 21 (T21) aka Down syndrome)

2 T21 IATs developed by the researcher: ‘Strong typical T21’ and ‘Weak typical T21’

Developed for this study: Rating how appropriate certain attribute words were for each picture used in the IATs (6 point Likert scale: strongly agree to strongly disagree)

‘Trisomy’, ‘Normal’, ‘Positive’, ‘Negative’

- ‘Trisomy’: photographs of the faces of children (7– 14 years) with T21. ‘Strong typical T21’ IAT used 6 faces strongly typical of T21. ‘Weak typical T21’ IAT used 6 faces weakly typical of T21; 3 female and 3 male faces for both

- IAT results suggested strong negative implicit attitudes towards disabilities (D = 0.94)

Cross sectional design. Each participant completed the IAT then the explicit attitude task. IATs were counterbalanced between participants

- For all three groups of participants IAT scores suggested negative implicit attitudes towards T21, both for facially weakly typical and strongly typical T21

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

Time 1: N = 48 of possible 49;

- ‘Normal’: photographs of the faces of typically developing children (7– 14 years), 3 female and 3 male faces

N = 165; mean age for university students = 20.6 years, general population = 38.3 years professional caregivers = 39.1 years. Male/female numbers not given

- ‘Positive’: 6 positive trait words (humane, affectionate, educable, endearing, sociable, intelligent)

- Caregivers had significantly less negative implicit attitudes than students and nonstudents - For caregivers, the longer the experience as a caregiver (in years), the less negative the implicit attitudes held

- When rating photos, all participants had a preference for positive traits over negative (positive explicit attitudes to T21), but across all participants preferences for positive traits were always slightly less for T21 pictures. - Both students and non-students made fewer positive explicit evaluations of strongly typical T21 than weakly typical T21 faces (slightly less positive explicit attitudes)

- ‘Negative’: 6 negative trait words (mean, stupid, ugly, annoying, aggressive, deformed)

Federici and Meloni (2008)

Italy

Parents of students with/without disabilities, teachers, special needs educators, socialhealth workers

Not specified, simply stated as ‘Disabilities’

Experiment 2 only:

Experiment 2 only:

All 3 IATs: ‘Disabled’, ‘Normal’

- ‘Disabled’: images recognisable as ‘disabled’ (no examples given)

Cross sectional design

3 disability IATs developed by the researchers (Good/Bad, Healthy/Ill, Outcast/ Integrated)

3 semantic differential tasks developed by the researchers

Good/Bad IAT: ‘Good’, ‘Bad’

- ‘Normal’: images recognisable as ‘normal’ (no example given)

Experiment 1: 17 focus groups carried out with volunteer participants.

- For all participants there were negative implicit attitudes for disability across all 3 IATs, the strongest being for the Bad/ Good IAT -Small correlation between the ‘outcast/ integrated’ sematic differential and the ‘outcast/integrated’ and ‘healthy/ill’ IATs

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

Under- and postgraduate students, general population, professional caregivers of individuals with ID

301

302

Table 2 (Continued ) Study

Disability investigated

Implicit and explicit attitude measures used

Disability related IAT categories and items used

Implicit

Main findings including (where available) D-scores on the IATa

Categories

Items

Experiment 1: N = 90; 89 female, 21 male; mean age 47 years.

Healthy/Ill IAT: ‘Healthy’, ‘Ill’

- For all the remaining categories (‘Good’, ‘Bad’, ‘Healthy’, ‘Ill’, ‘Outcast’ and ‘Integrated’) the researchers did not report the specific items used

Experiment 2: Volunteer participants completed the 3 computer based IATs first and then the 3 semantic differential tasks

-Relatives of individuals with disabilities (N = 13), were more likely to associate disabilities with disease than nonrelatives (i.e., they had a stronger preference for ‘Normal’ and ‘Healthy’ in the healthy/ill IAT than individuals without a relative with a disability)

Experiment 2: N = 36; 31 female, 5 male, mean age 41 years

Outcast/Integrated IAT: ‘Outcast’, ‘Integrated’ - ‘Disabled’: photographs of individuals with ID

Cross sectional design. Participants completed all measures on a computer, firstly a selfactivation task, followed by the IAT, and finally the explicit attitude questionnaire

- IAT results were indicative of negative implicit attitudes towards individuals with ID (D = 0.62).

Germany

Undergraduate students doing special education major courses N = 47; 36 female, 11 male; mean age 24 years

ID

ID IAT developed by the researcher

Explicit

Method

The Multidimensional Attitudes towards Disabled Persons Scale (MAS; Findler, Vilchinsky, & Werner, 2007)

‘Disabled’, ‘Nondisabled’, ‘Pleasant’, ‘Unpleasant’

- ‘Non-disabled’: photographs of individuals without ID - ‘Pleasant’: 8 words (good, luck, pleasure, pretty, happiness, admirable, nice, love)

- ‘Unpleasant’: 8 words (tragic, disgust, harm, pain, horrible, horror, mean, evil)

- Neither quantity nor quality of contact predicted implicit attitudes - Explicit attitudes (affective, cognitive and behavioural components) were predicted by contact quantity - Implicit attitudes were not significantly correlated with explicit attitudes

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

Hein et al. (2011)

Location and sample characteristics

Ma et al. (2012)

China

Physical disabilities (not specified)

Disability IAT developed by the researchers

None that measured attitudes towards individuals with disabilities

‘Disabled’, ‘Nondisabled, ‘Positive’, ‘Negative’

University students

Physical disabilities (not specified)

2 IATs developed by the researchers: DiseaseDisability IAT (D-D IAT) and Disability IAT

None that measured attitudes towards individuals with disabilities

D-D IAT: ‘Disabled’, ‘Health’, ‘Able bodied’, ‘Disease’

- ‘Disease’ and ‘Health’: words associated with these categories (words not specified in paper).

- ‘Disabled’: photographs of well-known individuals with physical disabilities

- IAT results suggested negative implicit attitudes towards individuals with disabilities

- No gender differences in implicit attitudes - There was a significant negative correlation between the fate control subscale of the SAS and implicit attitudes – individuals with greater belief in fate control had less negative implicit attitudes towards people with physical disabilities Experimental between-subjects design. Participants randomly allocated to read one of two ‘news articles’: one heightened fear of disease and the other was neutral. Then completed the IATs and several questionnaire measures

- European participants: those more sensitive to situations where germs are transmitted showed more negative implicit attitudes towards physical disabilities than those who were less sensitive (stronger associations between disability with disease or unpleasantness) - East Asian participants: those with greater disgust reactions towards physical disabilities had more negative implicit attitudes towards physical disabilities.

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

N = 101; 71 female, 30 male; mean age 21 years; 57 from urban areas, 43 from rural.

USA

Cross sectional design. Participants volunteered and completed measures on the computer. No information given on order of presentation of the IAT and the Social Axioms Survey (SAS; Leung et al., 2002)

- ‘Nondisabled’: symbols representative of able bodied individuals - ‘Positive’ and ‘Negative’: the researchers used ‘positive’ and ‘negative’ words but no example items were provided

Undergraduate students

Park et al. (2003)

- ‘Disabled’: symbols representative of disabilities

303

304

Table 2 (Continued ) Study

Location and sample characteristics

Disability investigated

Disability related IAT categories and items used

Implicit

Categories

Items

Disability IAT: ‘Disabled’, ‘Able bodied’, ‘Pleasant’, ‘Unpleasant’

- ‘Able bodied’: photographs of wellknown individuals with no physical disabilities - ‘Unpleasant’ and ‘Pleasant’: unpleasant and pleasant words (words not specified in paper)

S-IAT: ‘Good parenting’, ‘Bad parenting’

- ‘Good parenting’: 5 words (nurturing, aware, responsible, encouraging, available)

Child protection service workers

A-IAT: ‘Good’, ‘Bad’

- ‘Bad parenting’: 5 words (abandoning, inattentive, inconsistent, unsafe, ignoring)

N = 230; 197 female, 33 male; mean age 43.7 years in the ID condition and 41.7 in the non-ID condition

Both had: ‘People with ID’, ‘People without ID’

- ‘Good’: 5 words (marvellous, superb, pleasure, beautiful, joyful)

Explicit

N = 74; East Asian heritage (N = 45) and European heritage (N = 29)

Proctor (2012)

USA

ID

Stereotype IAT (S-IAT) and Attitude IAT (AIAT) developed by the researcher

No specific explicit attitude measure

- ‘Bad’: 5 words (tragic, horrible, agony, painful, terrible)

Method

Main findings including (where available) D-scores on the IATa

Experimental betweensubjects design. Completed online. Participants randomised to ID or non-ID vignette group. First read neglect vignettes describing parents as either ‘mild mentally retarded’ or ‘intellectually average’; then completed several questionnaire measures (order counterbalanced). After this they completed the two IATs (order counterbalanced)

- No significant differences between the groups on all demographics

- For ID group IAT scores on both IATs not associated with risk, removal or help decisions - Across both ID and non-ID groups, stronger implicit stereotypes were associated with higher risk and removal of child ratings - Ratings of pity were higher in the group reading ID vignettes, with anger and disgust being higher in the group reading Non-ID vignettes

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

Implicit and explicit attitude measures used

- Future risk ratings were higher for ID group than the non-ID group

- ‘People with Intellectual Disabilities’: 5 words (retarded, impaired, deficient, slow, limited) - ‘People without Intellectual Disabilities’: 5 words (average, sharp, intelligent, normal, smart) Pruett and Chan (2006)

USA

Physical disabilities (motor and visual)

Disability IAT (DA-IAT; Greenwald et al., 1998)

‘Disability’, ‘Nondisabled’, ‘Good’, ‘Bad’

- ‘Disability’: 4 symbols associated with disability

Rehabilitation counselling students (under- and postgraduates)

- ‘Nondisabled’: 4 symbols associated with ‘able bodied’

N = 223; 187 female, 36 male; median age 24 years; 173 reported no disability, 50 reported having a disability

- ‘Good’: 4 words (good, terrific, happy, love)

- ‘Bad’: 4 words (poison, evil, bad, vomit)

Cross sectional design. Volunteer participants completed IAT then the explicit attitude measure and various other questionnaires. IAT was then completed a second time. Both IATs completed were paperbased versions of the IAT

- Results were suggestive of negative implicit attitudes towards physical disabilities

- Explicit attitude scores were suggestive of positive attitudes towards physical disabilities - No significant correlation between the IAT scores and ATDP scores - Demographic variables (age, gender, socioeconomic status, ethnicity), contact with individuals with disabilities and rehabilitation education received did not predict IAT scores - Contact with individuals with disabilities accounted for a very small amount of variance in IAT scores

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

Attitudes towards disabled people scale (ATDP-A; Yuker & Block, 1986)

305

306

Table 2 (Continued ) Location and sample characteristics

Disability investigated

Implicit and explicit attitude measures used

Disability related IAT categories and items used

Implicit

Explicit

Categories

Items

Robey et al. (2006)

USA

Physical (motor) and ID (including cerebral palsy)

2 IATs developed by the researchers (infantilisation IAT and Evaluative IAT)

Two tasks developed by the researchers: a semantic differential task and two hypothetical scenarios of interacting with individuals with/ without physical disabilities

Infantilisation IAT: ‘Disabled’, ‘Adult’, ‘Nondisabled’, ‘Child’

- ‘Disabled’: 4 words (disabled, palsy, handicapped, impaired)

Nursing staff, nursing administrators, therapists, therapy aids, therapy administrators, personal care attendants N = 30; 27 female, 3 male; mean age 43 years; average tenure 8 years

- ‘Nondisabled’: 4 words (able-bodied, normal, nondisabled, nonhandicapped)

Evaluative IAT: ‘Disabled’, ‘Good’, ‘Bad’, ‘Nondisabled’

- ‘Child’: 4 words (infant, child, kid, young)

- ‘Adult’: 4 words (adult, mature, grownup, parent)

- ‘Good’: 4 words (happy, joy, love, peace)

- ‘Bad’: 4 words (sad, awful, failure, war)

Method

Main findings including (where available) D-scores on the IATa

Cross sectional design. Volunteer participants completed the IATs on a computer and the explicit measures on paper. The order of administration of the IAT and explicit measures was counterbalanced between participants.

- Evaluative IAT scores suggested negative implicit attitudes towards individuals with disabilities were held (D = 0.57).

- Infantilisation IAT scores suggested participants were more likely to associate disabilities with childlike qualities (D = 0.2) - No significant correlation between the two IATs - Imagined social interactions with individuals with disabilities suggested positive explicit attitudes - For the semantic differential tasks the ‘non-disability’ item was rated more positively than all three ‘disability’ items (explicit attitudes) - Evaluative IAT score were not significantly correlated with explicit attitudes

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

Study

- A small significant correlation was found between Infantilisation IAT scores and one semantic differential on the childlike/adult like continuum. Rojahn et al. (2008)

USA

Physical disabilities (motor and visual)

Disability IAT developed by the researchers

‘Disabled’, ‘Able’, ‘Good’, ‘Bad’

- ‘Good’: words (e.g., joy, love, pleasure)

Undergraduate students

- ‘Bad’: words (e.g., Evil, Angry, Terrible)

N = 41; 33 female, 20 male; mean age 27 years

- ‘Disabled’: symbols associated with disability (e.g., crutches, a seeing dog, a wheelchair)

Cross sectional design. Participants completed the IAT first on a computer and then the explicit tasks (including the reading of 16 vignettes)

- IAT results suggest moderately strong negative implicit attitudes towards physical disability

- No differences in IAT scores between males and females - No difference in IAT scores for participants with/without family members with a physical disability

- ‘Able’: symbols associated with able bodied (e.g., walking pedestrian, a skier) Stone and Wright (2012)

UK

Undergraduate and postgraduate students

Experiment 1: N = 55; 39 female, 16 male; mean age 32 years

Physical disabilities (motor) and facial disfigurement

Experiment 2 only:

Explicit evaluation of photographs used in the IATs on various attribute traits

Disfigurement IAT: ‘Distinctive’, ‘Typical’, ‘Good’, ‘Bad’

- ‘Distinctive’: 4 photographs of individuals with a facial disfigurement

- ‘Typical’: 4 photographs of individuals without a facial disfigurement

Disfigurement IAT and Wheelchair IAT developed by researcher

Wheelchair IAT: ‘Wheelchair’, ‘Nonwheelchair’, ‘Good’, ‘Bad’

- ‘Good’ and ‘Bad’: words (number and specific words for each were not specified)

Cross sectional design. Volunteer participants completed tasks on a computer. Presentation of the Disfigurement IAT and the Wheelchair IAT were counterbalanced across participants

307

-IAT results were suggestive of moderate to strong negative implicit attitudes towards individuals with facial disfigurement and individuals with motor disabilities - Implicit attitudes towards individuals with facial disfigurement were significantly more negative than towards wheelchair users - Explicit attitudes were suggestive of positive attitudes towards wheelchair users in the areas work competence, social potency and warmth

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

None measuring explicit attitudes towards individuals with disabilities

308

Table 2 (Continued ) Study

Location and sample characteristics

Disability investigated

Implicit and explicit attitude measures used

Disability related IAT categories and items used

Implicit

Categories

Explicit

Method

Items

- ‘Non-wheelchair’: 4 photographs of individuals sitting but not in wheelchairs - ‘Good’ and ‘Bad’: as above Thomas et al. (2007)

USA

Undergraduate psychology students

Physical disabilities (motor)

4 Disability IATs developed by the researchers (Paraplegia IAT; Alcoholism IAT; Cancer IAT; Mental Illness IAT)

For Experiment 2 only: Interaction with disabled persons scale (IDP; Gething, 1994)

Only categories for Paraplegia IAT reported: ‘Paralyzed’, ‘Nonparalyzed’, ‘Pleasant’, ‘Unpleasant’

Words were used in each of the categories for each IAT. Researchers did not report how many or which specific words were used for each category

Cross sectional design. Different participants completed Experiments 1 and 2. For both, the IAT was completed on a computer. For Experiment 1 order of administration: Paraplegia, Alcoholism, Cancer, Mental Illness. For Experiment 2 order: IATs (Alcoholism, Cancer, Mental Illness, Paraplegia), IDP then a social desirability scale

- For both experiments, IAT scores suggested negative implicit attitudes towards all disability categories

- In Experiment 1 implicit attitudes were most negative towards individuals with paraplegia, but in Experiment 2 most negative to alcoholism. These differences were linked to order effects of individual IATs

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

- There were small significant correlations between Wheelchair IAT scores and explicit ratings on work competence and social potency for wheelchair users (less negative IAT scores were associated with more positive explicit attitude ratings)

- ‘Wheelchair’: 4 photographs of individuals in a wheelchair

Experiment 2: N = 72; 51 female, 20 male, 1 unspecified; mean age 30 years

Main findings including (where available) D-scores on the IATa

- No significant difference in IAT responses between men and women in both experiments - Significant correlation between total IAT score and the IDP - Significant correlation between social desirable responding and IAT scores

Experiment 1: N = 129; 48 male, 74 female; mean age 20.2 years

Experiment 2: N = 69; 51 female, 18 male;

USA

Physical disabilities (motor)

Child IAT (Baron & Banaji, 2006) with some stimuli items and category labels changed for this study

The Acceptance Scale for Kindergartner (ASK; Favazza & Odom, 1996)

‘Children with disabilities’, ‘Children without disabilities’, ‘Good’, ‘Bad’

‘Children with disabilities’: 4 photographs of children with physical disabilities

Kindergarten and elementary school children

‘Children without disabilities’: 4 photographs of children without physical disabilities

N = 51; 27 in ‘inclusive classrooms’, 24 in ‘noninclusive classrooms’; ages 5–7 years; 29 males, 22 females

‘Good’: 4 words presented by speaker (happy, fun, nice, good)

‘Bad’: 4 words presented by speaker (mean, sad, bad, yucky)

Quasi-Experimental design. The IAT was completed first on a computer, followed by paper and pen completion of the explicit attitude measure. Both measures were presented at both time points

-IAT scores at Time 1 (4–6 months into the school year) and Time 2 (9 months into the school year) indicate negative implicit attitudes towards children with disabilities for both classroom groups - Trend for implicit attitudes to become less negative for inclusive classroom children and more negative for noninclusive classroom children at Time 2 - Explicit attitudes were positive at both time points for both classroom groups - No correlation between IAT scores and explicit attitudes

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

Thurneck (2008)

309

310

Table 2 (Continued ) Study

Location and sample characteristics

Disability investigated

Vaughn et al. (2011)

USA

Physical disabilities (motor)

Disability related IAT categories and items used

Implicit

Explicit

Categories

Items

Multiple disability IAT developed by the researchers

IDP (Gething, 1994)

Only categories for the Paraplegia IAT reported: ‘Person with paraplegia’, ‘Person without paraplegia’, ‘Pleasant’, ‘Unpleasant’

Only stimuli for the Paraplegia IAT reported:

Psychology undergraduates

- ‘Person with paraplegia’: 4 words (immobile, restricted, disabled, challenged)

N = 245; 181 female, 64 male; mean age 20.2 years

- ‘Person without paraplegia’: 4 words (mobile, active, healthy, functional) - ‘Pleasant’: 4 words (vacation, sunrise, diploma, love) - ‘Unpleasant’: 4 words (hatred, devil, bomb, poison)

USA

Physical disabilities (motor)

Disability IAT developed by the researchers

Attitudes towards disabled persons scale Form O (Yuker, 1988; Yuker & Block, 1986)

‘Abled-bodied athletes’, ‘Disabled athletes’, ‘Pleasant’, ‘Unpleasant’

- ‘Disabled athletes’: 10 photographs of athletes with disabilities playing 5 different sports (soccer with crutches, wheelchair tennis/track/basketball, skiing with one pole)

Method

Main findings including (where available) D-scores on the IATa

Cross sectional. Volunteer participants were first briefed on the 4 disability categories, then completed demographics questions, followed by attitude measures (counterbalanced between participants). Then completed the Wonderlic personnel test-revised (WPT-R; EF Wonderlic Associates, 1983) and then the Balanced inventory of desirable responding (Paulhus, 1994)

- Across all 4 IATs, scores suggestive of moderate to strong negative implicit attitudes towards individuals with alcoholism (D = 0.6), cancer (D = 0.68), mental illness (D = 0.65) and paraplegia (D = 0.64)

- This effect was found across all 4 IATs, even though results still suggested impact of order effect of IAT (with IATs scores decreasing with each IAT presented) - There was no significant correlation between the IDP scores and IAT scores - Cognitive ability and social desirability did not predict IAT scores

Cross sectional design. Volunteer participants completed tasks on a computer, with IAT presented first, then explicit attitude measures

- Results from IAT suggest negative implicit attitudes towards athletes with physical disabilities

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

White et al. (2006)

Implicit and explicit attitude measures used

University students

N = 115; 53 male, 62 female; mean age 21.2 years

Scale of Attitudes Towards Disabled Persons Form R (Antonak, 1982, 1985)

- ‘Abled-bodied athletes’: 10 photographs of athletes without disabilities playing 5 different sports (sports as above) - ‘Pleasant’: 10 words (honour, lucky, diamond, loyal, freedom, gift, happy, jolly, rich, love)

- Explicit attitudes as measured by the ATDP suggested positive attitudes (higher than the norms for this measure) - Explicit attitudes as measured by the SADP suggested slightly less positive attitudes than those of normative data but still suggestive of positive explicit attitudes - Men generally had less positive explicit attitudes than women (on the ATDP but not on the SADP) - No correlation between IAT scores and the SADP - Small correlation between ATDP scores and IAT scores

a For most studies, a positive D-score represents a negative implicit attitude towards individuals with disabilities. For some studies however, D-scores were reversed and are hence reported as negative D-scores. Please take note the narrative explanation of the provided D-scores for each study.

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

- ‘Unpleasant’: 10 words (evil, rotten, poverty, disaster, vomit, hatred, sad, greedy, dirty, worthless)

- No differences between IAT scores for male and female participants

311

312

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

3.2.3. IAT characteristics In order to evaluate the quality of the IATs employed in the studies included, various characteristics of the IATs were examined against methodological factors considered important when using an IAT (Lane et al., 2007). This included characteristics of the stimulus items for each category and the category labels; number of items per category; counterbalancing of the congruent and incongruent blocks of the IAT; counterbalancing the presentation of explicit attitude measures and the IAT; and the scoring procedure used to calculate the IAT scores for participants. These factors are considered in turn below. 3.2.3.1. IAT item characteristics and category labels. Nosek, Greenwald, and Banaji (2005) highlight several factors of importance in relation to the items used within any given IAT. They report that stimulus items can be presented as words, sounds or pictures and that using combinations of these modalities is acceptable. The studies included in the review used a variety of stimuli within their IAT tasks, including words, photographs, symbols and audio recordings. These were used in various combinations including using words for all four categories (Proctor, 2012; Robey, Beckley, & Kirschner, 2006; Thomas, Vaughn, & Doyle, 2007; Vaughn, Thomas, & Doyle, 2011); using words for two categories and photographs for the other two categories (Enea-Drapeau et al., 2012; Hein, Grumm, & Fingerle, 2011; Park, Faulkner, & Schaller, 2003; Stone & Wright, 2012; White, Gordon, & Jackson, 2006); using words in two categories and symbols/images in the other two categories (Aaberg, 2012; Archambault, Van Rhee, Marion, & Crandall, 2008; Chen et al., 2011; Ma, Chen, Zhou, & Zhang, 2012; Federici & Meloni, 2008; Pruett & Chan, 2006; Rojahn, Komelasky, & Man, 2008); and using photographs in two categories and auditory presented words in the remaining two categories (Thurneck, 2008). All studies were therefore seen to employ item types that were suitable for use in the IAT. A second factor concerns the need for careful consideration of the categories and stimuli used in the IAT. In order to ensure the IAT measures attitudes towards the attitude object of choice, it is important to ensure categories are clearly defined, and that items used for each are clearly identifiable to their relevant category. As mentioned previously, within the IAT there are generally four categories included, two ‘target categories’ (e.g., ‘disabled’ and ‘able bodied’) and two ‘attribute categories’ (e.g., ‘bad’ and ‘good’). Within each category a researcher would present several items that are representative of it (e.g., ‘good’ words: joy, love, pleasure, wonderful). Researchers can access previously used stimulus items and category labels from a variety of sources including the Project Implicit website (http://www.projectimplicit.net/stimuli.html) where the creators of the IAT (Greenwald et al., 1998) provide lists of items that can be used. These items have been used on numerous occasions and have been found to be representative of the respective categories under which they fall. Alternatively, appropriate pilot studies can be carried out to identify suitable stimulus items. Considering these issues, it is important for studies to either assess the representativeness of items included in each category (e.g., via a pilot study) or use items that have been shown to be appropriately representative from previous research. Of the 18 studies that employed an IAT, two did not report the items they used in the target or attribute categories, nor did they make reference to how the items were selected (e.g., from a pilot or from previous research). For the remaining 16 studies, for the target categories, five used items from previous research studies, seven conducted pilot studies to obtain the most representative items, and four did not detail how the target category items were selected. With regards to the attribute category items, eight studies used items from previous research, four conducted pilot studies to obtain the most representative items and four did not state how attribute category items were selected. It therefore appears that for the majority of studies the stimulus items used in the various IATs were appropriately representative of the category in question. Incomplete reporting on how category items were selected was a draw back for only a limited number of studies. Overall this suggests that for the most part, the studies included considered the importance of the categories and stimulus items that were utilised in their respective IATs. 3.2.3.2. Number of items per category. Of the 18 studies employing IATs, all reported using between four and ten items per category in their respective IATs. As detailed by Nosek and colleagues (2005), results of an IAT are relatively unaffected up to as low as two items per category. Given that all the IATs used at least a minimum of four items per category, it can be assumed that the results reported were unlikely to have been affected by the number of items used in each category. 3.2.3.3. Counterbalancing of congruent and incongruent blocks. As mentioned earlier, in the IAT participants categorise items presented on the screen as quickly as possible using two keys on the keyboard. The IAT consists of seven blocks of trials. In the first block participants categorise the target category items, with one target category paired to one response key and the other to the second response key e.g., ‘able bodied’ items to the left key and ‘disabled’ items to the right key. The second block is similar, except participants now categorise the two attribute categories, e.g., ‘good’ items to the left response key and ‘bad’ items to the right response key. In the third and fourth blocks, these pairings are combined, e.g., ‘good’ and ‘able bodied’ to the left response key and ‘bad’ and ‘disabled’ to the right response key. In block five the position of the target categories switches (e.g., ‘disabled’ moves to the left response key and ‘able bodied’ moves to the right response key). In this block only the target category items are presented such that participants can practice these new positions. In the sixth and seventh blocks, this new positioning of the target categories is paired with the previous attribute category positions, e.g., ‘good’ and ‘disabled’ to the left response key and ‘bad’ and ‘able bodied’ to the right response key. It is the difference in response time between these different pairings that gives an indication of the implicit attitudes of participants. The pairings that occur in blocks three and four and in blocks six and seven are either referred to as ‘congruent’ or ‘incongruent’. Congruent blocks refer

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

313

to a pairing or association that one would expect participants to make implicitly, e.g., ‘disabled’ paired with ‘bad’. Incongruent blocks would be the opposite pairing, e.g., ‘disabled’ paired with ‘good’. Therefore one may present the congruent pairings first (in blocks three and four) and the incongruent pairings second (in blocks six and seven), or the other way around. Lane and colleagues (2007) note that performance on blocks three and four (i.e., the first combined pairing) tends to have an effect on the performance on blocks six and seven (i.e., the second combined pairing). They make several suggestions to combat this effect, including increasing the number of trials in block five and counterbalancing the presentation of the above described congruent and incongruent pairings between participants. Therefore for half of the participants, blocks three and four would be the congruent pairing and blocks six and seven the incongruent pairing; and for the other half of participants, the opposite. This allows researchers to identify any potential order effects and where necessary account for them statistically. Of the 18 studies that employed IATs, eight reported counterbalancing congruent and incongruent pairings, four reported that they did not, and the remaining six studies failed to report whether counterbalancing was carried out. Although about half of the studies counterbalanced congruent and incongruent blocks, the apparent lack of counterbalancing across the remaining studies may have provided the potential for bias in the results reported by those studies where it was not carried out. 3.2.3.4. Counterbalancing presentation of explicit attitude measures and the IAT. The potential impact of the order of presentation of explicit attitude measures and the IAT on participant responses was noted by Lane et al. (2007). Although they reported that several studies have shown little to no effect related to the order of presentation of explicit measures and the IAT, they suggested presentation be counterbalanced between participants to allow for examination of any potential differences in responses (unless there is a good reason to present them in a set order). Of the 18 studies that employed an IAT, 11 also included a measure of explicit attitudes towards individuals with disabilities in the same study. Of these, only three reported counterbalancing the presentation of explicit attitude measures and the IAT. Two studies reported presenting the IAT first to avoid priming participants with the explicit attitude measure. The remaining six studies provided no explanation for the lack of counterbalancing. Although previous research suggests only a small chance of order effects, the relative lack of counterbalancing of the IAT and measures of explicit attitudes or of explanations for lack of counterbalancing is a methodological issue that may have introduced bias into the results of the studies in question. 3.2.3.5. Method used for calculating IAT scores for participants. Since the creation of the IAT, several developments have been made to improve the reliability and validity of this tool and its associated results. A significant development is related to the way the IAT score, which gives an indication of the implicit attitude of the individual participant, is calculated. A more robust scoring algorithm to calculate IAT scores was put forward by Greenwald, Nosek, and Banaji (2003) (the D6 score). This scoring algorithm aims to improve scoring in relation to several factors, including participant previous experience of the IAT and the impact of differences in response speed across the task. It is thus the preferred method of calculating IAT scores in research studies. All the studies in this review were published between 2003 and 2012 and hence (with the potential exclusion of the one paper published in 2003) had access to the improved scoring algorithm. Of the 18 studies that used the IAT, the majority (N = 10) employed the improved scoring algorithm; two based their results on the original (Greenwald et al., 1998) scoring algorithm, one study (Pruett & Chan, 2006) used a paper based IAT and hence employed a scoring algorithm specific for paper based IATs. The remaining five studies did not report which algorithm was used to calculate IAT scores. Although the original algorithm still provides interpretable scores, caution should be taken as it does not take account of factors such as prior IAT experience or differences in response speed across the task. This failure to use the improved scoring algorithm or to state which method was used to calculate scores is a methodological issue for a sizable number of studies included in this review. 3.3. Quality rating of the studies included in the review In order to assess the quality of the articles included in the review, a search for critical appraisal tools and checklists was carried out. Although a wide range of tools were identified, the majority were designed for evaluating articles reporting clinical intervention outcomes (e.g., Downs & Black, 1998) or epidemiological type studies (e.g., STROBE; Vandenbroucke et al., 2007). While these were comprehensive and detailed, and some items were relevant to the current review, given the nature of the studies included (survey based attitude research), these tools were judged to be largely inappropriate. The search did however identify the QualSyst (Kmet, Lee, & Cook, 2004), a tool developed to assess the quality of primary research articles using a variety of designs. Although it is largely relevant to assessing the quality of clinical outcome research, thorough investigation of the scoring system identified the scoring criteria to be inclusive enough to be used to assess the quality of the type of articles included in this review. The QualSyst provides quality criteria for quantitative (14 items) and for qualitative (10 items) research articles. For this review, only the quantitative criteria were required. Each article was scored on the 14 criteria, with scores ranging from 0 to 2 (0 = No; 1 = Partial; 2 = Yes; and N/A = item not relevant to the article being rated). Criteria items include ratings of the study design, sample size and analyses used. See Fig. 2 for full details of the areas assessed by the quantitative items. Inter-rater agreement of between 40% and 100% was reported for the 14 quantitative items by Kmet et al. (2004).

314

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

Items 1, 2, 3 & 4 evaluate the description of the study question / objective, study design, subject selection method and subject characteristics. They also evaluate the appropriateness of the study design and method of subject selection.

Items 5, 6 & 7 evaluate the description and explanation of the random allocation to treatment group and the interventional blinding of investigators or subjects.

Quality Rating Score (0 – 1)

Items 8 and 9 assess measures used and sample size

Items 10, 11, 12, 13 and 14 assess the analysis used, results reported and conclusions drawn from the study.

Fig. 2. Overview of areas of evaluation captured by the QualSyst quantitative studies items.

For each article a quality score was calculated by summing the scores for the article, and dividing this by the possible total score (i.e., 28  (number of ‘‘n/a’’  2)), generating a score ranging from 0 to 1. Although the authors do not provide cut-offs for quality scores, the ratings were used to provide some indication as to the quality of the articles included in this review. The scoring system has been used by other authors (e.g., Gravel, Le´are´, & Graham, 2006) and has been shown to have good inter-rater reliability (Kmet et al., 2004). Table 3 provides details of the scores for each article and overall quality ratings. Quality ratings of articles included ranged from 0.60 to 0.95, suggesting the quality of articles included to be reasonably high. 3.4. Implicit attitudes towards disabilities In the following section, the evidence on implicit attitudes towards individuals with disabilities will be presented and discussed. The evidence is distinguished by the type of disability investigated. In total, 13 studies investigated implicit attitudes towards individuals with physical disabilities; three towards individuals with ID; one towards both individuals with physical disabilities and individuals with ID; and one towards individuals with ‘disabilities’ of an unspecified type. 3.4.1. Implicit attitudes towards individuals with physical disabilities Of the 13 studies that investigated implicit attitudes towards individuals with physical disabilities, five examined implicit attitudes towards visual and motor disabilities, six towards motor disabilities only; and two stated that they investigated implicit attitudes towards physical disabilities but did not specify the type of physical disability. Across all 13 studies, the results of the IATs employed were suggestive of moderate to strong negative implicit attitudes towards individuals with physical disabilities. It is noteworthy that one of these 13 studies employed a paper based IAT (Pruett & Chan, 2006), whereas the rest employed computer based IATs. Some of the studies attempted to identify factors associated with implicit attitudes (N = 9). Three of these examined how much contact participants reported having with individuals with physical disabilities. Allport’s contact hypothesis suggests that increased contact with out-group members can help to improve attitudes towards them (Allport, 1954). Previous investigations into implicit attitudes towards other social groups have suggested that implicit attitudes are to some extent associated with contact with an out-group (e.g., Aberson & Haag, 2007; Lemm, 2006; Prestwich et al., 2008). For the three studies that considered contact, results were mixed, with only one finding contact to account for a significant but small amount of the variance in IAT scores (Pruett & Chan, 2006). The results from Thurneck (2008), although not significant, were suggestive of contact potentially playing a role in implicit attitudes as early as childhood. This study investigated changes in implicit attitudes of children towards individuals with physical disabilities over time. Implicit attitudes of children in

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

315

Table 3 Quality assessment of studies included in the review using the QualSyst criteria for quantitative studies. Study

Aaberg (2012) Archambault et al. (2008) Chen et al. (2011) Enea-Drapeau et al. (2012) Federici and Meloni (2008) Hein et al. (2011) Ma et al. (2012) Park et al. (2003) Proctor (2012) Pruett and Chan (2006) Robey et al. (2006) Rojahn et al. (2008) Stone and Wright (2012) Thomas et al. (2007) Thurneck (2008) Vaughn et al. (2011) White et al. (2006) a b

QualSyst criteria item scores (0, 1, 2, n/a)a 1

2

3

4

5

6

7

8

9

10

11

12

13

14

Quality score (0–1)b

2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

2 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

1 1 1 1 1 1 2 1 2 1 2 2 2 1 2 2 1

1 2 2 2 2 2 2 1 2 2 2 2 2 2 2 2 2

n/a n/a n/a n/a n/a n/a n/a 1 1 n/a n/a n/a 1 n/a n/a n/a n/a

n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a

n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a

2 1 2 2 1 2 1 1 2 2 2 1 2 1 2 1 2

2 1 1 2 1 1 2 2 2 2 1 1 2 2 1 2 2

1 1 2 2 2 2 1 1 2 2 2 2 2 2 2 2 2

2 0 0 2 1 2 0 0 2 2 0 0 0 0 2 2 2

n/a n/a n/a 0 0 n/a 1 1 2 n/a n/a n/a n/a n/a 1 n/a n/a

2 2 2 1 2 2 2 1 2 1 2 2 2 2 2 2 2

2 1 1 2 1 2 2 2 2 2 2 2 2 2 2 2 1

0.85 0.60 0.75 0.81 0.68 0.90 0.77 0.63 0.95 0.90 0.81 0.80 0.86 0.80 0.91 0.95 0.90

2, yes, 1, partial, 0, no, N/A, not applicable. Quality score calculated by summing scores of relevant items and then dividing this number by the total possible (i.e., 28  (number ‘‘n/a’’  2)).

inclusive and non-inclusive classrooms (i.e., classrooms that had at least one child with a disability or those that did not) were compared. Results suggested a trend of implicit attitudes of children in inclusive classrooms becoming more positive over time (i.e., as they had continued contact with individuals with physical disabilities) whereas those in non-inclusive classes became more negative. Other factors suggested to be significant in implicit attitudes towards individuals with physical disabilities include evolutionary developed responses to disease (Park et al., 2003) and greater belief in fate control (Ma et al., 2012). Park et al. (2003) reported that heightened fear of disease was associated with more negative implicit attitudes towards individuals with physical disabilities. They explained this as being an indication that evolutionary disease avoidance systems may be over inclusive and may become activated when confronted by visible physical disabilities. The consequence of this are more prejudiced reactions towards individuals with visible physical disabilities than those without. Ma and colleagues (2012) reported that individuals with stronger fate control beliefs, i.e., those with a greater belief in their ability to influence their predetermined future, had less negative implicit attitudes towards individuals with physical disabilities. They suggested that fear of the uncontrollability of disability may be an explanation for stronger negative implicit attitudes in individuals with a weaker belief in their ability to influence their ‘fate’. Other factors such as cognitive ability and ethnicity were reported to have no association with implicit attitudes (Thomas et al., 2007; Vaughn et al., 2011). Overall, results across the studies included were suggestive of a preference for able bodied individuals over individuals with physical disabilities. Furthermore, factors which may play a significant role in understanding these implicit attitudes include beliefs around the controllability of one’s future, sensitivity to the concept of disease, and contact with individuals with physical disabilities. 3.4.2. Implicit attitudes towards individuals with ID Of the three studies in this category, two investigated implicit attitudes towards individuals with ID, and one specifically towards individuals with Down syndrome. Across all three studies, results of the IATs were suggestive of moderate to strong negative implicit attitudes towards individuals with ID. As with some of the studies that investigated implicit attitudes towards individuals with physical disabilities, findings on the relationship between implicit attitudes towards individuals with ID and contact with them were mixed. One study in this subset found that greater contact with individuals with Down syndrome was associated with less negative implicit attitudes towards this group (Enea-Drapeau et al., 2012). In contrast Hein et al. (2011) found that neither contact quality nor quantity were associated with implicit attitudes towards individuals with ID. What was surprising was that for all three of these studies, the participants were paid caregivers of people with ID and yet their reported implicit attitudes were still generally negative. One may have speculated that individuals who work in close relation to individuals with ID may have either no preference or have somewhat positive implicit attitudes towards individuals with ID. This was clearly not the case given the results reported in these studies. In addition, and of some concern, were the results of Proctor (2012). In this study child protection service workers were presented with vignettes describing neglect of children. Half the participants received vignettes where the parents were described as having ‘mild mental retardation’ and the other half parents ‘of average intelligence’. Participants exposed to neglect vignettes describing parents with mild ID were more likely to predict greater future risk to children from the parents than those exposed to vignettes of parents of average intelligence, even though the vignettes were identical in all other aspects. Although these risk ratings

316

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

were not associated with decisions about removal of the child from the family, they provide some evidence of apparent prejudice against parents with ID among these child protection staff. Overall, these studies suggest that negative implicit attitudes towards individuals with ID are present, even among individuals who work in close relation to them. 3.4.3. Implicit attitudes towards individuals with ID and physical disabilities One study included in the review also reported moderately negative implicit attitudes towards individuals with ID and physical disabilities among formal caregivers of individuals with ID (Robey et al., 2006). Interestingly this study also included an IAT evaluating infantilising implicit attitudes towards individuals with disabilities, i.e., whether participants were more likely to associate child-like rather than adult-like qualities with individuals with disabilities. The results suggested there was a slight preference for associating the category of ‘child-like’ with individuals with disabilities as opposed to the category of ‘adult-like’. These results add to previous research that simply investigates positive/negative implicit attitudes towards individuals with disabilities and highlights another possible type of implicit attitude held towards this group, i.e., that of associating the concept of ‘child-like’ with the concept of disability. 3.4.4. Implicit attitudes towards ‘disabilities in general’ The final study (Federici & Meloni, 2008), examined implicit attitudes towards individuals with ‘disabilities’ of an unspecified type. The results were similar to those described above in that they suggested participants held negative implicit attitudes towards individuals with disabilities. They also identified that individuals who had a family member with a disability were more likely to associate ‘disability’ with the concept of disease than those without. These results add to those of Park et al. (2003) by highlighting the potential association between the concept of disability and notions of disease. It is worth noting though that the sample size in this study was small (N = 36), and hence comparisons between individuals with (N = 13) and without a relative (N = 23) with a disability were being made between two very small groups of uneven sizes. 3.5. Relationship between explicit and implicit attitudes towards individuals with disabilities Of the 18 studies that employed an IAT, 11 also included measures of explicit attitudes. The measures ranged from well-established ones (e.g., The Interaction with Disabled Persons Scale; Gething, 1994), to tasks developed by the researchers. Of these 11 studies, five found no significant correlation between implicit and explicit attitudes. The remaining six found only small significant relationships between implicit and explicit attitudes, the magnitude of which were within the range generally found in other attitude research which employs the IAT (Hofmann, Gawronski, Gschwender, Le, & Schmitt, 2005). 3.6. Limitations of the studies In addition to the above quality ratings, all articles were examined for specific limitations. These are summarised in Table 4 and discussed below. As discussed earlier, there were several methodological factors pertinent to the use of the IAT that were considered for each study included in the review. Overall, studies appeared to make use of appropriate stimulus item types; items that were appropriately representative of the category in question; the appropriate number of stimulus items per category; and the appropriate algorithm for obtaining IAT scores. However, there was a relative lack of counterbalancing of congruent and incongruent blocks in the IATs and presentation of explicit attitude measures and the IAT. This may have introduced bias into the results reported and is a significant limitation of some of the studies included. It is worth noting though that these limitations may just be due to incomplete reporting, as several authors for example simply do not state if they had counterbalanced congruent and incongruent blocks. Reporting of the specifics of the IAT employed was also in some cases incomplete. Some studies failed to provide clear explanations of the stimulus items used in their IATs. Although they all provided information on the category labels, this makes replication difficult, if not impossible, and does not allow for examination of how representative the items used were of each category. Another significant limitation of the studies included in the review is that of sampling. Although several different subject populations were sampled, over half of all participants were university students. The number of individuals from the general public and indeed children was relatively low. In addition, females significantly outnumbered males, with many studies not assessing for gender differences in implicit attitudes. The generalizability of the results is therefore possibly lower than would be hoped when one considers these potential biases. It is also noteworthy that some of the studies had very small sample sizes (N < 40), which again compromises the generalizability of the results. With regards to the disability categories investigated, there was a complete lack of studies that investigated implicit attitudes towards individuals with hearing impairments. In addition, of the studies investigating attitudes towards individuals with physical disabilities, the majority combined visual and motor disabilities within the same IAT task. Although this provides insight into the implicit attitudes held towards individuals with physical disabilities, the lack of specificity of physical disability makes the results somewhat ambiguous. It may have been preferable for studies to

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

317

Table 4 Limitations of studies included in the review. Study

QualSyst Quality Score (0–1)

Study limitations

Aaberg (2012)

0.85

No information on participant characteristics provided Lack of reporting on counterbalancing of congruent and incongruent blocks in the IAT

Archambault et al. (2008)

0.72

Somewhat small sample size (N < 50) Follow-up sample at Time 2 significantly smaller than that at Time 1, making meaningful comparisons difficult Lack of reporting on counterbalancing of congruent and incongruent blocks in the IAT

Chen et al. (2011)

0.75

Sample predominantly female Lack of counterbalancing of explicit and implicit attitude measures with no rationale provided

Enea-Drapeau et al. (2012)

0.90

Study used photographs of children, therefore ambiguous if attitudes assessed specific to children with Trisomy 21 or towards individuals with Trisomy 21 generally No investigation of participants’ knowledge of Trisomy 21, which may have contributed to attitudes Lack of counterbalancing of explicit and implicit attitude measures with no rationale provided Contact not formally measured but effect of contact on implicit attitudes assessed

Federici and Meloni (2008)

0.68

Disability category ambiguous Small sample size (N < 40) Comparisons between participants with/without relatives with disabilities based on very small groups of different sizes Lack of reporting on counterbalancing of congruent and incongruent blocks in the IAT Lack of reporting of stimulus items used Lack of counterbalancing of explicit and implicit attitude measures with no rationale provided

Hein et al. (2011)

0.90

Vignette used to measure explicit attitudes described an individual with Down Syndrome, but unclear if ‘pictures of individual with (ID)’ in the IAT also of individuals with Down Syndrome or not. If not, explicit attitudes may have tapped attitudes towards individuals with Down Syndrome and the IAT attitudes towards individuals with ID in general. Sample size fairly small (N < 50) Lack of reporting on counterbalancing of congruent and incongruent blocks in the IAT Lack of counterbalancing of explicit and implicit attitude measures, but no comment on why

Ma et al. (2012)

0.77

Type of physical disability investigated not reported Lack of reporting on specific items used in the IAT The sample was predominantly female.

Proctor (2012)

0.95

Lack of counterbalancing of congruent and incongruent blocks in the IAT Exact IAT scores not reported

Pruett and Chan (2006)

0.90

Employed a paper based version of the IAT which has been shown to be less reliable than computer based versions University student only sample Lack of counterbalancing of explicit and implicit attitude measures with no rationale provided

Robey et al. (2006)

0.90

Small sample size (N = 30) Sample predominantly female Lack of reporting on counterbalancing of congruent and incongruent blocks in the IAT

Rojahn et al. (2008)

0.80

Small sample size (N < 45) Sample was predominantly women University students only however they report specifically wanting to sample university students Lack of reporting on counterbalancing of congruent and incongruent blocks in the IAT

Stone and Wright (2012)

0.86

University students only sample Failure to report attribute category stimulus items Sample predominantly female

Thomas et al. (2007)

0.80

Due to apparent practice effects with multiple IATs, analysis completed with overall IAT scores. Somewhat problematic as the 4 disability categories are very different rendering overall IAT score questionable. Lack of reporting on counterbalancing of congruent and incongruent blocks in the IAT Lack of counterbalancing of explicit and implicit attitude measures with no rationale provided

Thurneck (2008)

0.91

Two quasi-experimental comparison groups quite small (each group N < 28) Lack of counterbalancing of explicit and implicit measures, however reported this was to avoid explicit measures priming the participants

Vaughn et al. (2011)

0.95

Significant practice effects for multiple presentation of IATs, even though IATs were counterbalanced between participants University student only sample

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

318 Table 4 (Continued ) Study

QualSyst Quality Score (0–1)

White et al. (2006)

0.90

Study limitations Sample predominantly female Lack of counterbalancing of explicit and implicit measures, however reported this was to avoid explicit measures priming the participants Failure to report target category stimulus items

investigate attitudes to different physical disability categories separately. These outcomes may potentially have given more accurate insight into the implicit attitudes held towards individuals with different forms of physical disability. There was also a significant lack of studies that investigated implicit attitudes towards individuals with ID. The few studies in this area, where appropriate, appeared to try and investigate implicit attitudes towards specific ID categories, e.g., by examining only attitudes towards individuals with Down syndrome. In addition, all the IATs employed across the studies included were in line with the original IAT paradigm (Lane et al., 2007), i.e., they included two target categories (e.g., disabled and non-disabled), and two attribute categories (e.g., good and bad). Therefore in all studies attitudes were assessed in relation to a comparison group (e.g., individuals with motor disabilities vs. individuals without motor disabilities). None employed more recent variations of this paradigm whereby implicit attitudes are assessed without using a comparison group (e.g., the Single Category IAT; Karpinski & Steinman, 2006). This method has been suggested to access more clearly the attitudes towards a particular attitude object and is an approach that could have been employed by studies included in this review. This is not necessarily a significant limitation of the studies included but highlights another potential avenue for future research. Taking all these issues into consideration, although the studies included in the review provide some insight into implicit attitudes held towards individuals with physical disabilities and/or ID, the limitations described above decrease the weight that can be placed on these results and thus the conclusions that can be drawn from them. 4. Discussion This review considered the results of 18 studies which employed the IAT to measure implicit attitudes towards individuals with physical disabilities (motor, visual or hearing) and/or ID. Overall, the results suggest that implicit attitudes towards individuals with these disabilities are negative, i.e., there was a moderate to strong preference for individuals without disabilities over those with disabilities. These negative implicit attitudes were evident across a wide range of subject populations and cultural groups (e.g., American, Chinese, and German), over the period of nearly a decade (2003–2012). The majority of studies were carried out in the USA (N = 12), most used a cross-sectional design (N = 16) and all employed convenience sampling methods. A significant proportion of participants were university students and more than half were female. There was a notable lack of evaluation of gender differences in implicit attitudes and with the vast number of studies employing student only samples, comparisons between students and other subject populations were often not possible. Both of these factors limit the generalizability of the results reported. There was a relative lack of studies investigating implicit attitudes among the general public, and only one study investigated implicit attitudes of children towards individuals with disabilities (Thurneck, 2008). Several of the studies had methodological limitations specifically linked to the IATs employed, including the lack of counterbalancing of the congruent and incongruent blocks in the IATs administered and lack of use of the improved scoring algorithm. As noted, these limitations may be due to incomplete reporting rather than genuine weaknesses. Some of the studies attempted to identify factors that were associated with implicit attitudes. The results suggest possible associations with an individual’s beliefs about the controllability of their future, sensitivity to the concept of disease, and contact with individuals with disabilities. Although nearly all the studies aimed to evaluate positive versus negative implicit attitudes towards individuals with disabilities, interesting variations were also explored, such as whether child-like rather than adult-like qualities are attributed to individuals with disabilities (Robey et al., 2006). Overall, notwithstanding methodological limitations, the studies reviewed provide a helpful starting point for research into implicit attitudes towards individuals with disabilities. They add to the existing literature on explicit attitudes towards this population and indicate some marked differences in self-reported attitudes. 4.1. Limitations of the review Although this article has attempted to provide a systematic review of literature assessing implicit attitudes towards individuals with disabilities, potential for bias needs to be acknowledged. Thorough assessment of identified articles against the inclusion/exclusion criteria was carried out by the primary author alone. Although final judgements of whether articles would be included in the review were made in discussion with the second author, this was done after this initial sort and thus must be acknowledged. Similar limitations apply to the quality assessments of included articles. While additional reference list searches of included articles were made, meticulous hand searches of all relevant journals was not feasible. This may have resulted in relevant articles being excluded from this review.

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

319

4.2. Recommendations Research into the attitudes held towards individuals with disabilities is both valuable and important. It gives insight into currently held attitudes, allows for the examination of factors that can contribute to them, and may highlight potential interventions to improve attitudes. This review indicates that, on a whole, implicit attitudes towards individuals with disabilities are negative and as such further research is called for. The significant lack of research into implicit attitudes towards individuals with ID is evident and further studies in this area are needed. Although more research investigating implicit attitudes towards individuals with physical disabilities was available, further research in this area is also required (i.e., studies focusing on specific physical disability types e.g., motor disabilities or visual disabilities only may be particularly useful). In addition, no studies were identified that assessed implicit attitudes towards individuals with hearing impairments. Future research should also investigate the influence of implicit attitudes on behavioural intentions towards individuals with ID or physical disabilities. This might help to highlight potential discriminatory behaviour towards these groups and further expand on literature investigating the link between implicit attitudes and actual behaviour. In addition, given the relative lack of exploration of factors associated with implicit attitudes towards individuals with disabilities, such as contact or emotions, future studies should examine these. It may be helpful for such studies to be longitudinal (e.g., as in Thurneck, 2008), with the aim of gaining an even deeper understanding of the relationship these variables have with implicit attitudes. To enhance validity of the reported results, future studies should carefully consider and implement methodological issues pertinent to the use of the IAT e.g., counterbalancing of congruent and incongruent blocks. This will allow for potential bias in reported scores to be accounted for and considered when interpreting the findings. In addition, thorough reporting on all aspects of the employed IATs should be provided, including the stimulus items. This will allow for clearer interpretation of the reported results and replication of these tasks by other researchers. Finally, further research is needed into implicit attitudes held towards individuals with disabilities among the general public and children. 5. Conclusions This review provided an overview of implicit attitudes held towards individuals with disabilities. It is concerning that the results suggest that relatively strong negative implicit attitudes prevail, and were consistently present over the period of nearly a decade. Research into explicit attitudes towards individuals with disabilities suggests that these have become less negative over time (e.g., Office of Disability Issues, 2011). It would appear however from the results of the studies included in this review that relatively strong negative implicit attitudes remain. Researchers have suggested that an individual’s behaviour is partly predicted by implicit attitudes (e.g., Greenwald, Poehlman, Uhlmann, & Banaji, 2009) and that they have the potential to affect behaviours that are not consciously controllable yet are likely to be important to the everyday experience of people with disabilities, such as eye contact and body language (Dovidio, Kawakami, & Gaertner, 2002). These behaviours may impede communication and trust, factors considered vital in the process of building long-term positive relations (Dovidio et al., 2002). Hence where negative implicit attitudes prevail, people with disabilities are likely to continue to experience social encounters as negative and discriminatory. Therefore factors that influence and explain implicit attitudes should be further investigated such that mechanisms that can help tackle negative implicit attitudes can be better understood. This would also provide important evidence to inform strategies aimed at decreasing discrimination and improving the quality of lives of individuals with disabilities in various contexts. References2 *Aaberg, V. A. (2012). A path to greater inclusivity through understanding implicit attitudes toward disability. Journal of Nursing Education, 51(9), 505–510. Aberson, C. L., & Haag, S. C. (2007). Contact, perspective taking, and anxiety as predictors of stereotype endorsement, explicit attitudes, and implicit attitudes. Group Processes and Intergroup Relations, 10(2), 179–201. Allport, G. (1954). The nature of prejudice. Reading, MA: Addison-Wesley. Americans with Disability Act. (1990). United States of America. Retrieved from: http://www.ada.gov/archive/adastat91.htm#Anchor-33869. Antonak, R. F. (1982). Development and psychometric analysis of the scale of attitudes toward disabled persons. The Journal of Applied Rehabilitation Counseling, 13(2), 22–29. Antonak, R. F. (1985). Construct validation of the scale of attitudes toward disabled persons. Journal of Applied Rehabilitation Counseling, 16(1), 7–10,. 48. Antonak, R. F., & Livneh, H. (1988). The measurement of attitudes toward people with disabilities: Methods, psychometrics, and scales. Springfield, IL: Charles C Thomas. Antonak, R. F., & Livneh, H. (2000). Measurement of attitudes towards persons with disabilities. Disability and Rehabilitation, 22(5), 211–224. *Archambault, M. E., Van Rhee, J. A., Marion, G. S., & Crandall, S. J. (2008). Utilizing implicit association testing to promote awareness of biases regarding age and disability. The Journal of Physician Assistant Education, 19(4), 20–26. Banse, R., Seise, J., & Zerbes, N. (2001). Implicit attitudes towards homosexuality: Reliability, validity and controllability of the IAT. Experimental Psychology, 48(2), 145–160. Baron, A. S., & Banaji, M. R. (2006). The development of implicit attitudes: Evidence of race evaluations from ages 6 and 10 and adulthood. Psychological Science, 17(1), 53–58. Bosson, J. K., Swann, W. B., & Pennebaker, J. W. (2000). Stalking the perfect measure of implicit self-esteem: The blink men and the elephant revisited? Journal of Personality and Social Psychology, 79(4), 631–643.

2

References marked with an asterisk indicate studies included in the systematic review.

320

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

Brillhart, B., Jay, H., & Wyers, W. E. (1990). Attitudes toward people with disabilities. Rehabilitation Nursing, 15(2), 80–85. *Chen, S., Ma, L., & Zhang, J.-X. (2011). Chinese undergraduates’ explicit and implicit attitudes toward persons with disabilities. Rehabilitation Counseling Bulletin, 55(1), 38–45. Clarke, N. E., & Crewe, N. M. (2000). Stakeholder attitudes toward ADA Title 1: Development of an indirect measurement method. Rehabilitation Counseling Bulletin, 43(2), 58–65. Cunningham, W. A., Preacher, K. J., & Banaji, M. R. (2001). Implicit attitude measures: Consistency, stability and convergent validity. Psychological Science, 12(1), 163–170. Daruwalla, P., & Darcy, S. (2005). Personal and societal attitudes to disability. Annals of Tourism Research, 32(3), 549–570. Department of Work and Pensions. (2002). Great Britain. ‘Disabled for life?’ attitudes towards, and experiences of, disability in Britain. Retrieved from: http:// research.dwp.gov.uk/asd/asd5/rrep173.pdf. Disability Discrimination Act of Australia. (1992) Retrieved from: http://www.comlaw.gov.au/Details/C2013C00022. Disability Discrimination Act of Great Britain. (1995, 2005). Retrieved from: http://www.legislation.gov.uk/ukpga/1995/50/contents. Dovidio, J. F., Kawakami, K., & Gaertner, S. L. (2002). Implicit and explicit prejudice and interracial interaction. Journal of Personality and Social Psychology, 82(1), 62–68. Downs, S. H., & Black, N. (1998). The feasibility of creating a checklist for the assessment of the methodological quality both of randomized and non-randomized studies of health care interventions. Journal of Epidemiology and Community Health, 52(6), 377–384. EF Wonderlic Associates. (1983). Wonderlic personnel test manual. Northfield, IL: Author. Egloff, B., & Schmukle, S. C. (2002). Predictive validity of an implicit association test for assessing anxiety. Journal of Personality and Social Psychology, 83(6), 1441– 1455. *Enea-Drapeau, C., Carlier, M., & Huguet, P. (2012). Tracing subtle stereotypes of children with Trisomy 21: from facial-feature-based to implicit stereotyping. PLoS ONE, 7(4), e34369. Favazza, P. C., & Odom, S. L. (1996). Use of the acceptance scale to measure attitudes of kindergarten-age children. Journal of Early Intervention, 20(3), 232–249. Favazza, P. C., Phillipsen, L., & Kumar, P. (2000). Measuring and promoting acceptance of young children with disabilities. Exceptional Children, 66(4), 491–509. Fazio, R. H., & Olson, M. A. (2003). Implicit measures in social cognition research: Their meaning and use. Annual Review of Psychology, 54(1), 297–327. *Federici, S., & Meloni, F. (2008). Making decisions and judgments on disability: The disability representation of parents, teachers, and special needs educators. International Multi-Conference on Society, Cybernetics and Informatics, II, 149–155. Findler, L. N., Vilchinsky, N., & Werner, S. (2007). The multidimensional attitudes scale toward persons with disabilities (MAS): Construction and validation. Rehabilitation Counseling Bulletin, 50(3), 166–176. Gething, L. (1994). The interaction with disabled persons scale. Journal of Social Behavior and Personality, 9(5), 23–42. Gravel, K., Le´are´, F., & Graham, I. D. (2006). Barriers and facilitators to implementing shared decision-making in clinical practice: A systematic review of health professionals’ perceptions. Implementation Science, 1(1), 1–16. Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring individual differences in implicit cognition: The implicit association test. Journal of Personality and Social Psychology, 74(6).. 1646-1480. Greenwald, A. G., & Nosek, B. A. (2001). Health of the Implicit Association Test at age 3. Experimental Psychology, 48(2), 85–93. Greenwald, A. G., Nosek, B. A., & Banaji, M. R. (2003). Understanding and using the implicit association test: I. An improved scoring algorithm. Journal of Personality and Social Psychology, 85(2), 197–216. Greenwald, A. G., Poehlman, T. A., Uhlmann, E. L., & Banaji, M. R. (2009). Understanding and using the Implicit Association Test III: Meta-analysis of predictive validity. Journal of Personality and Social Psychology, 97(1), 17–41. *Hein, S., Grumm, M., & Fingerle, M. (2011). Is contact with people with disabilities a guarantee for positive implicit and explicit attitudes? European Journal of Special Needs Education, 26(4), 509–522. Hergenrather, K., & Rhodes, S. (2007). Exploring undergraduate student attitudes toward persons with disabilities: Application of the disability social relationship scale. Rehabilitation Counseling Bulletin, 50(2), 66–75. Hofmann, W., Gawronski, B., Gschwender, T., Le, H., & Schmitt, M. (2005). A meta-analysis on the correlation between the Implicit Association Test and explicit self-report measures. Personality and Social Psychology Bulletin, 31(10), 1369–1385. Hornstra, L., Denessen, E., Bakker, J., van de Bergh, L., & Voeten, M. (2010). Teacher attitudes toward dyslexia: Effects of teacher expectations and the academic achievement of students with dyslexia. Journal of Learning Disabilities, 43(6), 515–529. Hunt, C. S., & Hunt, B. (2004). Changing attitudes toward people with disabilities: Experimenting with an educational intervention. Journal of Managerial Issues, XVI(2), 266–280. Karpinski, A., & Steinman, R. B. (2006). The single category implicit association test as a measure of implicit social cognition. Journal of Personality and Social Psychology, 91(1), 16–32. Kmet, L. M., Lee, R. C., & Cook, L. S. (2004). Standard quality assessment criteria for evaluating primary research papers from a variety of fields. Edmonton: Alberta Heritage Foundation for Medical Research (AHFMR), HTA Initiative #13. Lane, K. A., Banaji, M. R., Nosek, B. A., & Greenwald, A. G. (2007). Understanding and using the implicit association test: IV. What we know (so far). In B. Wittenbrink & N. S. Schwarz (Eds.), Implicit measures of attitudes: Procedures and controversies (pp. 59–102). New York: Guilford Press. Lemm, K. M. (2006). Positive associations among interpersonal contact, motivation, and implicit and explicit attitudes toward gay men. Journal of Homosexuality, 51(2), 79–99. Leung, K., Bond, M. H., de Carrasquel, S. R., Munoz, C., Hernandez, M., Murakami, F., et al. (2002). Social axioms: The search for universal dimensions of general beliefs about how the world functions. Journal of Cross-Cultural Psychology, 33(3), 286–302. *Ma, L., Chen, S., Zhou, M., & Zhang, J. (2012). Social axioms and implicit attitudes about people with disabilities. Social Behavior and Personality, 40(2), 251–258. McCaughey, T. J., & Strohmer, D. C. (2005). Prototypes as an indirect measure of attitudes toward disability groups. Rehabilitation Counseling Bulletin, 48(2), 89–99. National Coordinating Agency for Population and Development, Kenya. (2008). Kenya national survey for persons with disabilities – Preliminary report. Retrieved from: http://www.afri-can.org/CBR%20Information/KNSPWD%20Prelim%20Report%20-%20Revised.pdf. National Disability Authority. (2011). Northern Ireland. Literature review of international evidence on attitudes to disability. Retrieved from: http://www.nda.ie/ cntmgmtnew.nsf/0/B89C8098F9D7A0C8802573B800430A9B/$File/Literature_Review.pdf. Neumann, R., Hulsenbeck, K., & Seibt, B. (2004). Attitudes towards people with AIDS and avoidance behaviour: Automatic and reflective bases of behaviour. Journal of Experimental Social Psychology, 40(4), 543–550. Nosek, B. A., Greenwald, A. G., & Banaji, M. R. (2005). Understanding and using the Implicit Association Test: II. Method variables and construct validity. Personality and Social Psychology Bulletin, 31(2), 166–180. Office of Disability Issues. (2011). Great Britain. Public perceptions of disabled people. Retrieved from: http://odi.dwp.gov.uk/docs/res/ppdp/ppdp.pdf. Paris, J. M. (1993). Attitudes of medical students and health-care professionals toward people with disabilities. Archives of Physical Medicine and Rehabilitation, 74(8), 818–825. *Park, J. H., Faulkner, J., & Schaller, M. (2003). Evolved disease-avoidance processes and contemporary anti-social behaviour: Prejudicial attitudes and avoidance of people with physical disabilities. Journal of Nonverbal Behavior, 27(2), 65–87. Paulhus, D. L. (1994). Balanced inventory of desirable responding: Reference manual for BIRD, version 6. Vancouver, Canada: University of British Columbia. [Unpublished manuscript]. Prestwich, A., Kenworthy, J., Wilson, M., & Kwan-tat, N. (2008). Differential relations between two types of contact and implicit and explicit racial attitudes. British Journal of Social Psychology, 47(4), 575–588. *Proctor, S. (2012). Implicit bias, attributions, and emotions in decisions about parents with intellectual disabilities by child protection workers. Pennsylvania, United States of America: Pennsylvania State University. [Unpublished doctoral dissertation].

M.C. Wilson, K. Scior / Research in Developmental Disabilities 35 (2014) 294–321

321

*Pruett, S. R., & Chan, F. (2006). The development and psychometric validation of the Disability Attitude Implicit Association Test. Rehabilitation Psychology, 51(3), 202–213. Riches, V. C., & Green, V. A. (2003). Social integration in the workplace for people with disabilities: An Australian perspective. Journal of Vocational Rehabilitation, 19(3), 127–142. *Robey, K. L., Beckley, L., & Kirschner, M. (2006). Implicit infantilizing attitudes about disability. Journal of Developmental and Physical Disabilities, 18(4), 441–453. *Rojahn, J., Komelasky, K. G., & Man, M. (2008). Implicit attitudes and explicit ratings of romantic attraction of college students towards opposite-sex peers with physical disabilities. Journal of Developmental and Physical Disabilities, 20(4), 389–397. Schwartz, C., & Armony-Sivan, R. (2001). Students’ attitudes to the inclusion of people with disabilities in the community. Disability and Society, 16(3), 403–413. Steffens, M. C. (2004). Is the Implicit Association Test immune to faking? Experimental Psychology, 51(3), 165–179. *Stone, A., & Wright, T. (2012). Evaluations of people depicted with facial disfigurement compared to those with mobility impairment. Basic and Applied Psychology, 34(3), 212–225. Strohmer, D. C., Grand, S., & Purcell, Al. M. J. (1984). Attitudes toward persons with a disability: An examination of demographic factors, social context, and specific disability. Rehabilitation Psychology, 29(3), 131–145. *Thomas, A., Vaughn, D., & Doyle, A. (2007). Implementation of a computer based Implicit Association Test as a measure of attitudes toward individuals with disabilities. Journal of Rehabilitation, 73(2), 3–14. *Thurneck, D. A. (2008). The impact of inclusive education on the development of disability attitudes. Virginia, United States of America: James Madison University. [Unpublished Doctoral Thesis]. United Nations. United Nations Convention on the Rights of Persons with Disabilities. (2008). Retrieved from: http://www.un.org/disabilities/documents/ convention/convoptprot-e.pdf. Vandenbroucke, J. P., von Elm, E., Altman, D. G., G;zsche, P. C., Mulrow, C. D., Pocock, S. J., et al. (2007). Strengthening the reporting of observational studies in epidemiology (STROBE): Explanation and elaboration. Epidemiology, 18(6), 805–835. *Vaughn, E. D., Thomas, A., & Doyle, A. L. (2011). The multiple disability Implicit Association Test: Psychometric analysis of a multiple administration IAT measure. Rehabilitation Counseling Bulletin, 54(4), 223–235. *White, M. J., Gordon, P., & Jackson, V. (2006). Implicit and explicit attitudes toward athletes with disabilities. Journal of Rehabilitation, 72(3), 33–40. World Health Organisation. (2001). Geneva. The international classification of functioning, disability and heath. Retrieved from: http://www.disabilitaincifre.it/ documenti/ICF_18.pdf. World Health Organisation & World Bank. (2011). Malta. World report on disability. Retrieved from: http://whqlibdoc.who.int/publications/2011/ 9789240685215_eng.pdf. Yazbeck, M., McVilly, K., & Parmenter, T. R. (2004). Attitudes toward people with intellectual disabilities: An Australian perspective. Journal of Disability Policy Studies, 15(2), 97–111. Yuker, H. E. (Ed.). (1988). Attitudes toward persons with disabilities. New York: Springer. Yuker, H. E., & Block, J. R. (1986). Research with the attitude toward disabled persons scales (ATDP): 1960–1985. Hempstead, NY: Hofstra University.