“NEW PSYCHOMETRIC TOOL “IMPLICITY” HAS IMPLICITY FAILED TO PREDICT DISCRIMINATORY BEHAVIOUR”
BY DELROY CONSTANTINE-SIMMS
IMPLICITY claims to be the first commercially available implicit attitude psychology metric which predicts the risk of discriminatory behaviour against particular social groups. Unlike other I.A.T metrics designed to assess bias, IMPLICITY is suppose to produce non-judgemental feedback reports which 'contextualises' prejudice and guides test-takers through a series of exercises which help them understand their biases and learn to manage implicit attitudes that could manifest them self into prejudicial behaviour.
The concept of implicit attitudes, will be clarified and critiqued before reviewing the metric tool IMPLICITY, which attempts to measure implicit attitudes, while explaining that IMPLICITY, published by Hogrefe, is based on a similar tool called IMPLICT, which was first developed in 1998 by Anthony Greenwald of the University of Washington and Mahzarin Banaji of Harvard.
Implicit Attitudes
Greenwald, A. G., & Banaji, M. R. (1995) defined implicit attitudes as “introspectively unidentified (or inaccurately identified) traces of past experience that mediate favourable or unfavourable feeling, thought, or action toward a social object.” The definition is very compelling, like Ranganath, K and Nosek, B (unpublished) I have concerns, hence the following questions. What does this definition actually mean? The last part of the definition “favourable or unfavourable feeling… toward a social object” links the definition to attitudes – associations between evaluations and concepts. Greenwald, A. G., & Banaji, M. R. (1995) argue that the phrase, “introspectively unidentified” means that implicit attitudes exist outside of conscious awareness.
We cannot just search our minds for them, and in trying to find them, they may be “inaccurately identified.” .Greenwald, A. G., & Banaji, M. R. (1995) also claim that people are in possession of at least two types of attitudes - conscious, explicit attitudes that are experienced as our feelings, and implicit attitudes that are not part of our conscious experience. According to Ranganath, K and Nosek, B (unpublished), this implies that implicit attitudes could be quite different from explicit attitudes. ”Traces of past experience,” in the above definition, that refer to the presumed origins of implicit attitudes.
Implicit attitudes are thought to reflect an accumulation of life experience. For example, a person might regularly be exposed to negative ideas about black people and crime. Consciously, this person might disagree with the negative ideas and maintain a positive explicit attitude toward the Black people and crime. Implicitly, however, this negative information may be stored as associations between black people and violence. As is evident in this example, implicit attitudes are no more ‘real’ than explicit attitudes. Explicit attitudes reflect conscious values, beliefs, and desired responses. Implicit attitudes reflect experience – whether the person agrees with it or not. Both types of attitudes can be important in shaping thought, judgment, or action.
Consequences of Implicit Attitudes
While there is ample research activity that seeks to identify when implicit and explicit attitudes predict behaviour, the existing evidence suggests that explicit attitudes tend to predict deliberate behaviours that are fairly easy to control. For example, one’s explicit attitude toward choosing fruit might predict whether that person chooses an apple or orange when given as much time to make a choice among a bowl of fruit. Implicit attitudes, on the other hand, tend to predict behaviours that are more spontaneous thus difficult to control. So, implicit attitudes might predict the fruit choice when a person is in a hurry and just grabs the first fruit that seems appealing.
Relationship between Implicit and Explicit Attitudes
There is ongoing research investigating when implicit and explicit attitudes will be related or unrelated, and why, the most extensive area of study is area is self-presentation - whether people are motivated to adjust their explicit responses because they are unwanted or they are unwilling to make them public. For example, it is generally not socially acceptable to express negative attitudes about African Americans, or people with disabilities, etc.
So, if people feel negatively about these groups, they may resist reporting those feelings explicitly. However, implicit responses are relatively uncontrollable, so people may express negativity toward some groups implicitly even when they are trying to avoid it, in extreme cases fake responses. Another predictor of consistency between implicit and explicit attitudes is attitude strength.
Domains such as race and sexuality may be considered more important, or other variables that people have thought about a lot, tend to show more consistency between implicit and explicit responses than those that are unimportant or rarely considered. What appears to be missing is the influence of external variables such as socialisation, stereotypes, education, economic climate and the influence media influence in relation to the manner in which domains are perceived implicitly or explicitly.
Unanswered Research Questions
Ranganath, K and Nosek, B.,(unpublished) have suggested that the concept of implicit attitude should be extensively investigated before test developers use the I.A.T. concept to develop diagnostic tools that test publishers claim can be used to screen out people that have a propensity to act in prejudicial manner towards certain social groups. If clients users or researchers intend to use IMPLICTY, they need to have the following questions answered.
1.To what extent are implicit attitude measures assessing the concept “implicit attitude” as it was defined? For example, is it unconscious, or more like a measure of ‘gut feelings’?
2.How are implicit attitudes formed?
3.How stable are implicit attitudes?
4.Can implicit attitudes be changed?
The IMPLICITY Tool and Validity Concerns
The data supplied by Hogrefe the publishers of IMPLICITY, there are eight versions of IMPLICITY which cover the following social categories: Age, Disability, Gender, Sexual Orientation, Ethnic origin, However, the sample group data has not adequately broken down the make up of the sample group used in this research, in terms of sexuality, disability, gender, race or religion. The researchers have ran two pilots studies of IMPLICITY, however, there is no clear information available that describes the diversity profile of the respondents that were initially used in the two UK population samples (N=318 and N=747)
Hogrefe the publishers of IMPLICITY describe the relationship between IMPLICITY as an 84 item self-report questionnaire comprising 28 overtly prejudiced beliefs and behaviours, 28 statements from the modern racism research (where underlying prejudices are expressed in a way that avoids sanction) and 28 distracter items. The research claims that eight factors emerged from the self-report questionnaire data. Factor 1 was ‘Overtly prejudiced beliefs and behaviour’.
The multiple regression statistical tool was used to discover which of a number of algorithms underlying IMPLICITY scores were most predictive by Factor 1. Bi-variate mean correlation between ‘Implicitly’ raw scores and Factor 1 raw scores were calculated to be .45 across two UK population samples (N=318 and N=747). The scores on IMPLICITY are said to be independent of general speed (across tasks) or error proneness. The study also suggests that I.Q. has no impact, but the results do identify some age effects. Older test-takers (>55) showed more prejudicial attitudes against ethnic groups than younger test-takers (<25). A small sample of people with dyslexia (N=34) tended to produce lower scores and benefited less from practice.
There are other validity issues of concern, which calls into question the validity of IMPLICITY as a viable tool for predicting discriminatory behaviour. From my understanding of psychometrics, criterion validity is suppose to be a measure of how well one variable or set of variables predicts an outcome based on information from other variables, and how it will be achieved. This means that a set of agreed measures from a personality test related to a behavioural criterion that could serve as a means of maximising concrete or criterion validity. A typical way to achieve this is to identify the extent to which a score on a personality test can predict future performance or behaviour. Another method involves correlating test scores with another established test similar to IMPLICITY that also measures the same attitude or personality characteristics.
Given that IMPLICITY is the first tool of its kind, the developers of IMPLICITY have not provided the relevant evidence which could clarify if IMPLICITY had been correlated with a similar tool that also measures the respondent’s conscious or subconscious ability to be prejudice or discriminate. This has not been the case, this has been done, if it has I stand corrected, but as far as I am aware there is no such tool available in the market place that can predict discriminatory behaviour, most tools just measure attitudes toward certain variables of equality, none of them predict behaviour.
It could be argued that the test publishers correlated the results of the two pilot studies, and hence formed his conclusions on the findings of that data set. Sadly, Comparing IMPLICITY scores with IMPLICTY scores proves nothing. The results can only be considered valid the researcher had compared IMPLICITY with a similar tool that had been properly validated, since IMPLICTY, is the only such tool on the market, it’s difficult to see how the researchers can make such statistical correlations..
While statistical correlations are useful, the research methodology, by way of two pilot studies makes little sense unless the six strands of equality and other factors such as profession, qualifications, income etc are used as part of independent variables that determines the dependant variable, in this instance the ability to predict discriminatory behaviour. The researchers can always cite the original tool IMPLICT, the American version of IMPLICITY, which the tool IMPLICIT is a derivative, and claim that his newly developed tool IMPLICITY has a high level of predictive validity hence the ability to predict prejudicial behaviour. But in reality, IMPLICITY has a very low criterion / concrete validity.
Predicting Validity With IMPLICITY
How can IMPLICITY have any predictive validity without taking without taking into account the importance of extraneous variables such as time, environment, geography, culture, language, fear, organizational culture, stereotypes and so many other variables? The developers and publishers may have convinced themselves that IMPLICITY as an I.A.T. tool can effectively measure and predict demonstrable and related prejudicial preferences and traits that are considered concrete criteria in the "real" world.
While I understand and applaud the researchers behind IMPLICITY, I am yet to be convinced that IMPLICITY is the right tool for predicting prejudicial behaviour. I am convinced that the developers in their rush to claim the title of pioneering and the first I.A.T. to predict prejudicial behaviour, have compromised certain standards and principles, especially in the methodology and the lack of consultation with other occupational psychologist that could have added value in terms of their particular diversity interest.
I am also of the view that Hogrefe, the publishers of IMPLICITY have not lived up to their impeccable standards of test scrutiny by virtue of endorsing a IMPLICITY which appears to lack sufficient concrete validity in order to capitalize on the notion that they are able to make money out of diversity and equality issues by licensing and marketing IMPLICITY as tool that can predict discriminatory attitudes, sentiments and of course future behaviour. Hence an excellent diagnostic tool for use in workplace:
1.How can IMPLICITY confidently predict future acts of prejudice, when the variables that are supposed to represent intermediate and ultimate criteria are virtually impossible to measure, given that there is no other tool on the market?
2.Furthermore, how can IMPLICITY be a successful tool in highlighting, unconscious prejudice awareness let alone behavioural intentions?
If there is a high correlation between discriminatory attitudes scores extracted from respondents that have taken the IMPLICITY test and prejudiced behaviour, there should be a high criterion-related validity between the intermediate variables, that increase levels of awareness of the respondents attitudes towards members of disadvantaged groups, with the ultimate variable, success in the ability of respondents to acknowledge and accept their subconscious awareness of their own discriminatory attitudes, and thus enabling them to adapt or change their attitudes in a manner that allows them to embrace people, irrespective of age, gender, sexuality, religion, disability.
Diversity and Equality Contradictions and Concerns
Essentially, the scores on the IMPLICITY tool are supposed to predict the respondent’s success in terms of respondent’s ability not to discriminate. While IMPLICITY may attempt to measure a number of things including unconscious prejudice, I fail to see how IMPLICITY can succeed in weeding out any form of prejudice, let alone racists, without the tried and tested method, called diversity and equality training, not the legislative type, but the type of training that deals with issues that many are only too willing to admit in private, but never declare in public.
While the intention of IMPLICITY is admirable and supportive arguments compelling, as a means of raising individual awareness of discriminatory behaviour. However, the developers need to scrutinize their own personal and organizational attitudes before embarking on the process of developing tools that are attempting to predict prejudicial behaviour. Hence the following questions
1.It is understood that Nigerian student Abiodun Okunrinboye, an MSc programme student at London Metropolitan University, who worked with Dr. Pete Jones achieved a prize for ‘Best Poster Presentation’ at the Postgraduate Occupational Psychology (POP) Conference January 2009, using IMPLICITY. His hypothesis was that traditional questionnaires conducted on potential employees with the aim to ascertain an individual’s attitudes and beliefs towards other ethnic groups, often lead to responses being ‘faked.’ Did the researchers behind assume that a study conducted by a Nigerian student, using the IMPLICITY would give this I.A.T. tool more credibility?
2.If the team behind IMPLICITY are so interested in developing tools that identify levels of subconscious or unconscious attitudes, can the lead researcher consciously, explain why his company he represents consists of four white men, and not one of them is a representative from the six strands of equality legislation in relation to race, gender, disability, age, religion or female? If they do represent a few of these variables, I stand corrected; one variable such as homosexuality is highly misguided for a variety of reasons.
3.If this tool is about detecting unconscious prejudice, then why didn’t the company developing IMPLICITY consider employing the services of UK based BME chartered occupational psychologists or people from groups of people that are traditionally disadvantaged to participate in the research and development of this tool?
Given that the test developers are touting IMPLICITY across numerous British organizations for endorsement, surely members of this organization could have at least dealt with their own unconscious discriminatory practices by way of taking their own test and thus identifying the reality of their own discriminatory practices, in terms of how they chose their development team, associates as well as the sample group in their pilot studies, or other studies for that matter by attending diversity and equality courses that cover the six equality strands such as race, gender, sexuality, disability, religion, before they embark on the development of any tool or study that intends to form the precursor to any anti-discrimination intervention tool.
The Perception of IMPLICITY Items
IMPLICITY is a version of the original American tool called IMPLICIT, is useful for detecting unconscious attitudes and studying cognitive processes. Irrespective of who developed the tool first, it is misleading for I.A.T. researchers involved in the development and research of tools such as IMPLICT and IMPLICITY to give individuals ratings like “slight,” “moderate” or “strong” — and advice on dealing with their bias — when there isn’t even that much consistency in the same person’s scores if the test is taken again, if evidence does exist, it’s not very evident.
If bias does exist after taking this tool, one can decrease racial bias scores on the I.A.T. by simply exposing white people to pictures of Black people in Church or at a Barbecue. Better still one can use images of well respected Black people, such as Obama, Mandela, Mohammed Ali, or any of the 20 most inspiring people often featured in Black publications, such as the Voice, Ebony and Essence.
In contrast, there can be images of Lady Diana in Africa, or Bill Clinton in Harlem, or pictures of white people marching in an anti-racist campaign, that could encourage Black people to have more positive views of white people. In reality, Black people tend to have a much diverse view of white people by virtue of the historical relations between Black and white people as consequence of slavery, colonialism, including historical revisionism reinforced through the media and the Eurocentric educational system.
While the latter views have a part to play in the construction of prejudicial attitudes, neither of these I.A.T tools (IMPLICT or IMPLICITY) proves that either group is racist or has other prejudices, let alone acts on these prejudices, consciously or unconsciously. Just because someone has expressed an opinion, or has been judged by this tool to have a preference in relation to age, race, gender, sexuality, disability or religion does not necessarily mean they will act in a manner that is considered preferential or prejudicial.
Issues and Concerns Expressed by developers of IMPLICITY
Dr. Pete Jones admits that “Prejudiced attitudes are incredibly difficult to measure, as in today’s society admitting to racism, sexism or ageism has severe consequences.”Prejudiced attitudes are incredibly difficult to measure, as in today’s society admitting to racism, sexism or ageism has severe consequences.” Dr. Pete Jones continues to make questionable assertions which suggest that speed and response to images and word are indicative of certain attitudes: “The more they struggled or the longer they took to assign these words indicated inner prejudice” This assertion is problematic, respondents may struggle because of a disability such as poor sight, dexterity, or possibly dyslexia.
Dr. Peter Jones has made several generalizations about sexuality: “Without detracting from the seriousness of the prejudice that still exists against people because of their ethnic origin, the results of our study suggest that being gay or lesbian could be ‘the new black’ when it comes to being a victim of prejudice.” Dr. Pete Jones and others that who share such questionable views that gay is the new black appear to forget that one can hide sexuality, but it’s not so easy to hide certain types of disability, gender, and of course skin colour, Michael Jackson tried to hide his ethnicity, but like some many like him, they failed miserably.
The tone of this commentary, reminds me of a radio Deejay that has just announced the new number one record, but in this instance, DJ Dr Pete Jones appears to be enthused by the notion that race has finally been knocked of the discrimination charts, in place of homophobia, a more deserving form of discrimination than racism. His tone sends the wrong message, in that his equating homophobia with racism, by claiming that “being gay is the new black” is offensive, misguided and dangerous.
Interestingly, Rev. Irene Munroe, of Harvard University, reinforces this point as follows:
“If you are African American and gay, and fighting alongside your white LGBTQ brothers and sisters for queer civil rights, the notion that “Gay is the new black” is not only absurdly arrogant, it is also dangerously divisive. In a presumably “post-racial” era with the country’s first African American president-elect, it’s easy for some to assume that race doesn’t matter but in reality race does matter”
I am not suggesting that Dr. Pete Jones or his colleagues, are racist or prejudice or anything like that, but I do have concerns about their levels of understanding of diversity issues. In my opinion, financial gain is the driving force behind the development of this tool, but in order to gain academic credibility, Dr. Pete Jones has gone through the usual paper presenting process, by presenting a paper at the BPS conference in Blackpool, Jan 2009. If I had attended this conference, I would have asked Dr. Pete Jones the following questions:
1.Has this I.A.T., Tool, IMPLICITY been reviewed by the British Psychological Society or peer reviewed in other psychology publications or journals? If so, please list them.
2.How does Dr Pete Jones justify the responses of sixty respondents to justify a generalization which suggest that black is the new gay?
3.In terms of the finished product Dr. Pete Jones has ran two pilot studies with an increased sample size. Given that the Hogrefe are asking individual to endorse this product from a wide variety of socio-political backgrounds, why hasn’t the publishers of IMPLICTY made this diversity profile of the sample group/respondents available on their website?
4.Given that this tool is about identifying unconscious prejudicial traits, where is the evidence to suggest that Black and Asian organisations and other relevant organizations were contacted or involved in the development of this tool?
5.It’s common practice for test publishers to allow students to use their tools for academic research, in return for the data. My question is this, did the researcher behind IMPLICITY consult any BME chartered occupational psychologists, or occupational psychologists representing other aspects of diversity based on the six strands of equality legislation, in the development of IMPLICITY?
6.Who will be allowed to use the IAT/IMPLICITY tool and what level of training will they be given, most importantly, will there be any academic or professional restrictions regarding who can use IMPLICITY ?
My real issue with this tool, is that I strongly fear that BME’s and other disadvantaged groups not familiar with the language of Diversity and Equality training, will assume that I.A.T. tools like “IMPLICITY” are specifically meant to weed out white racists, are likely to find themselves excluded from employment, especially if the findings of their reports highlight a preference that may be incompatible with the culture of the organization, including the value systems of individuals involved in the recruitment and selection process.
Interpretation of Preferences and Attitudes Towards Specific Groups
The licensing of IMPLICITY by any test publisher should be done with extreme caution, especially when the findings appear to suggest that white people still have a preference for their own group, could be considered racist, when this is not necessarily the case. The researchers behind IMPLICIT have not provided any sound evidence that IMPLICITY has identified preferences that are truly indicative of discriminatory tendencies or practices on the part of respondents, which begs the following question:.
1.Does a Black respondent’s preference for Black people make them prejudice towards white people or white people more racist towards Black people because of their preferences for their own respective ethnic groups?
2.Does a Black respondent’s preference for heterosexuality make them prejudice towards gay and lesbians or white people that have a preference for gays and lesbians, less likely to prefer people because they do not agree with homosexuality?
3.Where is the evidence extracted from either IMPLICIT and IMPLICITY which identifies a correlation between unconscious attitudes and action, without taking into consideration the work place, irrespective of organizational culture and of course legal constraints?
I am yet to be convinced, that IMPLICITY can do what it states in relation to identifying prejudicial attitudes, that will are indicative of future behaviour. Hence my assertion that the “IMPLICTY” tool is not acceptable or an appropriate tool for weeding out racists, sexists, ageists etc in the UK work place. .
Conclusion
While IMPLICITY may be considered innovative, pioneering to those unfamiliar with psychometric tools, and how they function, it’s easy to be convinced that such tools are fantastic and fit for purpose. But in reality, the developers behind IMPLICITY must address the issues that have been highlighted in this review. Those unfamiliar with the language of psychometric testing can easily bamboozled by science, hence endorse the product!
My advice to individual, organizations asked to endorse this product is simple, make sure you are qualified to comment, do some background research, and avoid becoming the mouth piece of a test publisher, that has little interest or understanding of diversity issues beyond the bottom line, which is making money! Hence my final point “Buyer Beware”
Delroy Constantine-Simms
Chartered Occupational Psychologist and Psychometrican
www.passingpsychometrictests.blogspot.com
920 E 214st , New York, New York, 10469 USA
DTHINKDOCTOR@AOL.COM
UK Cell 07533 100492
FURTHER READING
Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition: Attitudes, self-esteem, and stereotypes. Psychological Review, 102(1), 4-27.
http://news.bbc.co.uk/1/hi/uk_politics/7836766.stm
http://news.bbc.co.uk/1/hi/uk_politics/7836766.stm
http://ukgaynews.org.uk/Archive/08/Dec/1604.htm
http://www.authorsden.com/visit/viewArticle.asp?idE428
http://www.bps.org.uk/media-centre/press-releases/releases$/division-of-occupational-psychology/prejudice-study-finds-gay-is-the-new-black.cfm
http://www.gaylesbiantimes.com/?id=13595
http://www.londonmet.ac.uk/depts/dops/psychworks/the-application-of-implicit-attitude-theory.cfm
http://www.pamil-visions.net/gay-blac/
http://www.pinke.biz/news/547/British-People-Found-to-Hate-Gays-More-than-People-From-Other-Races/
http://www.telegraph.co.uk/scienceandtechnology/science/sciencenews/4246874/Britons-more-homophobic-than-racist.html
http://www.telegraph.co.uk/scienceandtechnology/science/sciencenews/4246874/Britons-more-homophobic-than-racist.html
http://www.yorkshireeveningpost.co.uk/news/BNP-wins-Yorkshire-Euro-seat.5343062.jp
https://implicit.harvard.edu/implicit.
Wilson, T. D., Lindsey, S., & Schooler, T. Y. (2000). A model of dual attitudes. Psychological Review, 107(1), 101-126.
-
