The Use of Implicit Association Testing Applied to the Relationships of Individuals Who Are Visually Impaired

By Stacy M. Kelly and Cody M. Laplante

Stacy M. Kelly, Ed.D., TVI, COMS, CATIS, is anassociate professor in the Visual Disabilities Program at Northern Illinois University. She can be contacted by email at

Cody M. Laplante, B.S.Ed., is a Teacher of Students with Visual Impairments in the General Academic Program at the Maryland School for the Blind. He can be contacted by email at


This study used the implicit association test (IAT) to measure attitudes of undergraduate and graduate-level university students toward people with blindness or visual impairments in regard to romantic relationships. The IAT is a computer-generated psychometric assessment that measures unconscious biases. The overall results showed that this sample of sighted participants demonstrated a moderate automatic preference for people with sight over people with blindness in romantic relationships. Implications for families and practitioners are discussed along with future research that builds upon this initial experiment by applying the IAT methodology to more complex topics in the field of blindness and visual impairment.


Blind, implicit association test, social skills, visual impairment


When discussing the disability-specific skills of people with blindness or visual impairments, specific attention must be placed on certain aspects of development. The specific area of social interaction skills is of interest because without explicit instruction, many students with visual impairments may not recognize particular social cues. Thus, they may have decreased success in social interactions, such as being accepted by a peer group and forming close friendships as well as romantic relationships (Pinquart & Pfeiffer, 2013). Unfortunately, people with visual impairments often remain socially isolated (Sacks & Wolffe, 2006). For example, Kelly and Smith (2008) found that a nationally representative sample of adolescents and transition-aged youth with visual impairments received telephone calls from friends significantly less often than those with other disabilities, such as learning disabilities, speech impairments, and emotional disturbance.

To close this gap, researchers often suggest that youth with visual impairments should receive higher quality and more frequent instruction in the areas of social skills, emotional awareness, and dating in particular (Kapperman & Kelly, 2013; Kelly & Kapperman, 2012; Kelly, Wild, Ryan, & Blackburn, 2015; Wild, Kelly, Blackburn, & Ryan, 2014). However, little research investigates the attitudes of the general public towards people with visual impairments and how they may also affect that later involvement in romantic relationships.

One example of the limited research in this area was implemented by Rowland and Bell (2012) when these researchers conducted a study of attitudes toward visual impairments among undergraduate and graduate-level university students. The study used an assessment to measure attitudes toward people with visual impairment that included both Likert scale and true/false items. There was nothing computer-generated involved in this methodology and all data were self-reported. The study found that the more exposure the participants had to visual impairment, the more likely it was that the participants had a positive attitude about blindness and visual impairment. Also, this study found that people who are visually impaired have a more positive attitude about visual impairment than sighted people (Rowland & Bell, 2012).

The present study used assessment methods different from Rowland and Bell (2012) to also analyze attitudes among a similar population of participants. Specifically, the present study investigated the unconscious attitudes of undergraduate and graduate-level university students towards people with visual impairments in regard to romantic relationships using an implicit association test (IAT). Implicit association testing is an online computer-generated psychological test originally developed to test unconscious biases regarding minority groups (Greenwald, McGhee, & Schwartz, 1998).

The IAT was first developed in 1998 by researchers at the University of Washington to measure biases towards race and gender unaffected by what the participants think they “should” believe (Banaji & Greenwald, 2013; Greenwald et al., 1998). The major benefit in using the IAT is that this testing method is specifically designed to show unconscious attitudes towards specific groups of people without the social pressures and judgment usually associated with such issues. This is accomplished through the requirement that participants must respond to stimuli as quickly as possible. Banaji and Greenwald (2013) found that the accuracy of the test increases as response time decreases. In a sense, the quicker the participant responds, the more accurately unconscious biases are measured. If the participant has time to reason through the responses, they are more likely to succumb to social pressure (Banaji & Greenwald, 2013; Greenwald et al., 1998).           


Thus, we selected the longstanding and renowned IAT methodology for this study and comprehensive IAT psychometrics software to administer it. Specifically, Inquisit (2016) was the specialized IAT software used to implement the computerized test involved in this study. We used Inquisit Version 5.0. Ethical clearance for this study was obtained from the Northern Illinois University Institutional Review Board (IRB) and there were two researchers who were both IRB approved to be involved in data collection for this study.

The Systematic Process of the IAT

The IAT is a computerized psychometrics test using words and pictures to determine implicit biases. In the first section of the test, participants were asked to sort pictures into two groups using the left and right arrow keys on the keyboard. In the case of this study those were, (1) pictures of people who were blind, and (2) pictures of people who were sighted. The pictures used were combinations of clip art depicting stick figures either with or without white canes, and pictures of real people with and without white canes. If the responses provided by the participants were incorrect, the participant would be asked to try again. If the participant generally responded with too much hesitation (i.e., based on the automatically calculated length of the time interval for the response), the section would be repeated.

After sorting the pictures, participants were asked to sort words into two groups (1) words synonymous with attractive, and (2) those synonymous with unattractive. The words used that were synonymous with attractiveness were beautiful, cute, good looking, partner, hot, sexy, lovely, and wonderful; and those synonymous with unattractiveness were ugly, plain, nasty, alone, unappealing, unpleasant, tragic, and awful (Banaji & Greenwald, 2013). The purpose of the first two sections was to allow the participants to become accustomed to the testing processes by first introducing scenarios that were less complex (Greenwald et al., 1998). After sorting each factor separately, participants entered a section of the test in which both factors were tested together. For example, in the left group were pictures of people who are blind and words synonymous with attractiveness, while in the right group were pictures of sighted people andwords associated with unattractiveness. After a number of trials, the groups would switch. For example, in the left group were pictures of sighted people andwords synonymous with attractive, while in the right group were pictures of blind people andwords synonymous with unattractive.

The IAT Software Measurement and Coding System

Participants’ biases were measured based on the speed and accuracy of responses in each section of the test (Greenwald et al., 1998). For instance, if a participant responded faster and more accurately in the section of the test in which blind people and attractiveness were paired than the section of the test in which sighted people and attractiveness were paired, then that participant showed a higher automatic association with people who are blind and attractiveness than with sighted people and attractiveness. The output scores range from -1.5 (which shows a strong automatic preference for people who are blind) to 1.5 (which shows a strong automatic preference for people who are sighted). A score of 0 shows no automatic preference for either group. Scores closer to zero indicate automatic preferences that are less extreme. Scores farther from zero indicate automatic preferences that are more extreme.

Administration of the IAT

As was the case in the Rowland and Bell (2012) study, all participants were recruited by making announcements in their undergraduate and graduate-level university classes. All of the participants were volunteers who reported to a computer lab on the campus of Northern Illinois University to complete this online test. Results were collected devoid of identifying information by giving each participant a unique participant identification number. Participants whose major was specific to visual impairments (e.g. pursuing certification as a teacher for the visually impaired, orientation and mobility specialist, or vision rehabilitation therapist) were given an identification number between 600 and 699. Participants whose majors were disability-related (e.g., special education, occupational therapy, rehabilitation therapy, physical therapy) were given identification numbers between 700 and 799, and participants whose area of study did not relate to disability (i.e., unrelated) were given identification numbers between 800 and 899. This allowed for differentiation of results based on background information, knowledge, and interest. All data were sorted, coded, and analyzed using descriptive statistics features available within SPSS Inc. Software Version 24 (2016). The data were imported directly into SPSS using a .csv file format.

Demographic Questionnaire

A short five question demographic questionnaire was completed by each participant prior to starting the IAT process. The demographic questionnaire contained items pertaining to degree program, area of study, age, and gender. There were no demographic questionnaire items that were forced response. All demographic survey items were optional. Results of the demographic questionnaire were collected devoid of any identifying information as well. The same participant identification number assigned for the IAT testing purposes was used in the completion of the demographic questionnaire.


Participant Demographics

In all, data for 129 participants were collected. All of those who participated in this study were sighted students at Northern Illinois University. Eighty-six percent (86%) of participants were undergraduate students and fourteen percent (14%) were graduate students. Participants ranged from 18 to 40 years of age. The vast majority of participants were between the ages of 18 and 24 (approx. 81%), while twelve percent (12%) of participants were between the ages of 25 and 30, and seven percent (7%) were between the ages of 31 and 40. The majority of participants were female (76%) as opposed to male (24%). Of the 129 participants who participated in this study there were 18 participants with a visual disabilities-specific major, 35 participants with a disability-related major, and 83 participants with an unrelated major.

Descriptive Analysis

The median of each category was calculated, as well as the percentage of both negative and positive scores. Negative scores showed an automatic preference for people with blindness in romantic relationships and a positive score indicated an automatic preference for sighted people in romantic relationships.

The median of all scores generated by the IAT software was calculated at 0.55 which indicates a moderate automatic preference for people with sight over people with blindness in romantic relationships. In all, approximately 92% of participants received scores above zero indicating at least a slight preference for people with sight over people with blindness in this context. Only about seven percent (7%) of participants received scores below zero. This means that only seven percent (7%) of participants tested had an automatic preference for people with blindness over people with sight.

In all differentiated categories, the standard deviations were similar: unrelated major = 0.45, disability-related major = 0.42, and specific major in visual impairments = 0.33. These results show scores that were similarly distributed. In addition, the medians of each category were similar: unrelated major = 0.51, disability-related major = 0.51, and specific major in visual impairments = 0.58. Thus, background knowledge or interest in the participant did not seem to affect results generated by this IAT software experiment. In addition, between 90% and 100% of participants in each group received scores that indicated an automatic preference for people with sight as romantic partners.


Although numerous sighted participants were recruited (n = 129), the sample was a convenient sample of university students available to voluntarily participate on the campus of Northern Illinois University. The results of such convenience samples cannot be generalized to the broader population beyond the scope of participants included in this study. Future studies should include even larger samples with more balanced ratios of male participants compared to female participants. Also, this study enabled exploration of the presence of stigmas or biases toward people with visual impairments within this context of romantic relationships. It must be noted that this was a very specific context within which to test for bias and there are many additional areas that should also be explored. Also, there is debate among psychologists about the IAT needing more psychometric development (Azar, 2008). However, the same critics also explain that further development is not necessary. Instead, feedback should be provided by experts to the general public about the meaning of specific scores generated by the IAT (Azar, 2008). Thus, this study addressed this known limitation of the IAT instrument by providing explanation of the IAT coding system and resulting IAT scores reported in this article.

Discussion and Implications for Practitioners and Families

The issue in question is the gap that exists between people with blindness or visual impairments and people with sight in regard to romantic relationships. The use of an implicit association test experimental procedure enabled exploration of the presence of possible stigmas or biases within this context of romantic relationships.

The results of the present study suggest that at least part of the problem may be caused by the ignorance of the public. The importance of awareness of this sort of ignorance on behalf of practitioners and families cannot be underestimated. Gordon, Tschopp, and Feldman (2004) note a consistency of negative attitudes among the general population in regards to people with disabilities. The possibility exists that people with sight may not see blind or visually impaired people as suitable mates either socially or romantically (Bolt, 2005). This possibility is supported by popular fictional literature in which characters with visual impairments are commonly seen as asexual or sexually deviant, neither of which are seen as suitable characteristics for a potential mate (Bolt, 2005). These sorts of unrealistic and incorrect portrayals must be explicitly addressed. Kef and Bos (2006) acknowledge the possibility that a delay in romantic behavior by people with visual impairments may be affected by a disinterested general public. For the first time, this hypothesis was tested by the participants in this study using a well-established psychological experiment.

Future Research

This study outlined the IAT, how it is administered, and how it may be applied to the field of blindness and visual impairment. Since this is a new application of the IAT, it was introduced on a smaller scale and in the context of a simple study. The IAT can be applied to address more complex topics in the field.

For example, further investigation is warranted with a more diverse sample of participants in other areas of social skills and within the context of each of the disability-specific skill areas, such as expectations of sighted people toward the careers, recreation and leisure pursuits, and independent living skills of people with visual impairments (Hatlen, 1996). Taking this a step further, the implicit association test procedure has the possibility of serving as the ideal testing mechanism from which to learn a great deal about the general public’s attitude toward all aspects of life experienced by people with visual impairments. The ability to implement any IAT experiment entirely online further underscores the opportunity for a wide range of large-scale studies that can build upon this initial experiment.


The funding for this study was provided by the Undergraduate Special Opportunities in Artistry and Research (USOAR) award at Northern Illinois University.


Azar, B. (2008). IAT: Fad or fabulous? Monitor on Psychology, 39(7), 44. Retrieved from

Banaji, M. R., & Greenwald, A. G. (2013). Blindspot: Hidden biases of good people. New York, New York: Delacorte Press.

Bolt, D. (2005). Castrating depictions of visual impairment: The literary backdrop to eugenics. Journal of Visual Impairment & Blindness, 99(3), 141-150.

Gordon, P. A., Tschopp, M. K., & Feldman, D. (2004). Addressing issues of sexuality with adolescents with disabilities. Child and Adolescent Social Work Journal, 21(5), 513-527. doi:

Greenwald, A. G., McGhee, D. E., & Schwartz, J. K. L. (1998). Measuring individual differences in implicit cognition: The implicit association test. Journal of Personality and Social Psychology, 74(6), 1464-1480. doi:

Hatlen, P. (1996). The core curriculum for blind and visually impaired students, including those with additional disabilities. RE:view, 28(1), 25-32. Retrieved from ERIC database. (EJ532379)

Inquisit (Version 5.0) [Computer Software]. (2016). Retrieved from

Kapperman, G., & Kelly, S. M. (2013). Sex education instruction for students who are visually impaired: Recommendations to guide practitioners. Journal of Visual Impairment & Blindness, 107(3), 226-230.  

Kef, S., & Bos, H. (2006). Is love blind? Sexual behavior and psychological adjustments of adolescents with blindness. Sexuality and Disability, 24(2), 89-100. doi:

Kelly, S. M., & Kapperman, G. (2012). Sexual activity of young adults who are visually impaired and the need for effective sex education. Journal of Visual Impairment & Blindness, 106(9), 519-526.

Kelly, S. M., & Smith, T. J. (2008). The digital social interaction of students with visual impairments: Findings from two national surveys. Journal of Visual Impairment & Blindness, 102(9), 528-539.

Kelly, S. M., Wild, T. A., Ryan, C. L., & Blackburn, M. V. (2015). Looking back at the service delivery models of sex education in the United States: Adults with visual impairments report on their experiences. British Journal of Visual Impairment, 33(2), 138-145. doi:   

Pinquart, M., & Pfeiffer, J. P. (2013). Does visual impairment lead to lower or higher levels of success in solving developmental tasks? — A longitudinal study. Journal of Developmental and Physical Disabilities, 25(6), 579-595. doi:

Rowland, M. P. & Bell, E. C. (2012). Measuring the attitudes of sighted college students toward blindness. Journal of Blindness Innovation and Research, 2(2). Retrieved from doi: 

Sacks, S. Z., & Wolffe, K. E. (Eds.). (2006). Teaching social skills to students with visual impairments: From theory to practice. New York, New York: AFB Press.

SPSS Inc. (2016). PASW Statistics for Windows, Version 24.0. Chicago, Illinois: SPSS Inc.

Wild, T. A., Kelly, S. M., Blackburn, M. V., & Ryan, C. L. (2014). Adults with visual impairments report on their sex education experiences. Journal of Blindness Innovation and Research, 4(2). Retrieved from doi:

The Journal of Blindness Innovation and Research is copyright (c) 2018 to the National Federation of the Blind.