Monday, July 27, 2020

GPT-3 is like looking into the future... What a time to be living!

Here are two poems by Wallace Stephens, except one is written by AI. Guess which one was written by Stephens?

1
Barque of phosphor 
on the palmy beach,

Move outward into heaven
Into the alabasters,
And night blues

Foam and cloud are one
Sultry moon monsters
Are dissolving.

Fill your black hull 
With bright moonlight.

There will never be an end
To this droning of the surf.

2
I must have shadows on the way
If I am to walk I must have
Each step taken slowly but alone
To have it ready made

And I must think in lines of grey
To have dim thoughts to be my guide
Must look on blue and green
And never let my eye forget
That color is my friend
And purple Must surround me too

The yellow of the sun is no more
Intrusive than the bluish snow
That falls on all of us. I must know
Grey thoughts and blue thoughts walk with me
If I am to go away at all.

AI continues to do things we never thought possible. GPT-3 does what your brain is doing now or more accurately what your brain is doing when you write. It is producing the next word in the sentence. Except it’s not. It is using brute force and a huge amount of data from the internet, to process what your brain does with relative ease. Nevertheless what GPT-3 does is astounding. What GPT-50 will do may be really astounding.

GPT-2 was impressive, a Natural Language Processing model that took a massive amount of text from the internet, at least the quality text, and it performed well, really well. So well, they didn’t release the software, for fear of its power. Open AI kept on going and created GTP-3, which is 117 times bigger that GTP-2. Bigger appears to be better. How much better it can become no one really knows.

Benchmarking GPT-3, on writing articles, humans can distinguish ‘human from AI’ created content accurately 52% of the time. In other words only as good as guessing, in other words, indistinguishable. Whether it’s a poem by Wallace Stephens or an essay, this is a stunning, if not frightening, achievement. Basically, AI is producing texts that are often hard to distinguish from human texts. Let that sink in, as the consequences of that 'competence' are massive.

The researchers claim that GPT-3 has learnt how to learn and that that the bigger the model, the better its adaptive abilities. This hints at the possibility that we are moving slowly towards AGI (Artificial General Intelligence) here. The trick is to focus on what they call the 'few shot' problem. In standard machine learning, you give it loads of data. However, for many real-world problems, the data is small, only a few examples. An example of a few shot learning problem, its digital ID. You give it only a few photos or image of your face, and it recognises your face to give you access to your phone. The larger models, like GPT-3, make better use of fewer examples in the data. What’s more, when it gets things wrong, it tends to get them wrong in plausible ways. Now that is interesting. We may not only get right answers but common misconceptions, which are useful in teaching and learning. 

In learning it can create questions and provide answers to questions, so that quick assessments are possible - lots of them. There's an application of GPT-3 that allows you to speak to people from the past - have a conversation with Marx, Freud... whoever. Its ability to create structured text means that future iterations will write good answers and essays. This could blow the whole 'essay' model in Higher Education.

It is easy to anthropomorphise GPT-3 but the deeper you delve and extrapolate, the more scary it becomes. It feels like a child, a smart, fast learning, precocious child, feeling its way forward in the world. It's not, of course, as a child never gets access to this much language when learning to speak. It may even work without the absence of semantic power. Such power, creeping towards AGI, from just predicting the next word (or token) in a sentence. As the efficacy of the model is still trending upwards, it may still have a long way to go, a very long way. It would seem that ‘competence without comprehension’ is with us in a big way. We could even just stumble upon AGI if the internal ‘learning how to learn’ feature, that seems to be emerging, gets exponential. What a time to be living.

Sunday, July 19, 2020

Diversity training, the damning evidence that $8billion a year is wasted

Harvard Professor of public policy Iris Bohnet says "I did not find a single study that found that diversity training in fact leads to more diversity”. This is not surprising, as decades of research have shown that diversity training is literally a waste of time and money. It is really hard to change minds and attitudes and harder still to get. People to take positive or avoid negative actions to see these things through. So no matter that $8 billion is spent a year (2016) and that figure increased dramatically in the following years when the unconscious bias movement was sprung upon us. The Black Lives Matter movement has raised awareness of the problem further, on a global scale but we would be ill-advised to see ‘diversity training’ as the solution to such a serious problem.

History of diversity training

Diversity training tends to come in waves, often a reaction to some event in the real world. As we will see, first it was the assassination of Martin Luther King in 1968, then in 2018 it was an incident in Starbucks when two black men waiting on a business meeting found that the police had been called on them, now it was the George Floyd incident that exploded into the Black Lives Matter movement. In every case, and one hopes that this will not happen after the BLM protests, rather crude training was designed, developed and delivered.
The most visible figure in the history of diversity training for decades was Jane Elliot. As an unknown schoolteacher in Iowa, she delivered her first lesson on race relations to her class in 1968, the day after the assassination of Martin Luther King. So angry was she at what she witnessed that she designed and delivered her own, home-grown lesson on race relations to the children in her class. I suppose it is what we’d now call a learning experience.
She split her all-white class into blue and brown-eyed groups, then chose the brown-eyed group, stating that they were superior giving them lots of privileges such as sitting at the front of the class, more food at lunch, linger breaks and not allowing the groups to use the same fountain. The blue-eyed group was even given blue collars by the brown-eyed group. The brown-eyed kids were told to play only with other in their group, justified by the explanation that melatonin in brown-eyed people made them more intelligent. This was followed up with chastisement of certain blue-eyed students. She claimed that the born-eyed kids scores rose, while those in the blue-eyed group fell. After a day, the process was reversed. 
Even in her own school, many felt uncomfortable with Elliot’s methods Elliot left teaching, and a school and town that was divided about her experiment and fame, to become a sort of celebrity diversity trainer. She became a regular guest on TV programmes such as The Tonight Show Starring Johnny Carson and The Oprah Winfrey Show, as well as the subject of many films and documentaries. She left teaching to become  a spokesperson and trainer in schools, Universities, as well as corporates and large public organisations. It was widely practiced for decades in education and in corporate and public organisations.
It doesn’t take long to see the ethical problems with this simulation or experiment. Some claim that it was, and still is, wrong to experiment with people in this way, especially children. Others see the flaws in such an accusatory method, which may be counterproductive, if bad ideas and behaviours are reinforced. Even more worrying is the lack of evidence that it worked. Subsequent research shows a mixture of moderate changes in attitudes, but there is no long-term evaluation of any impact on organisations, only self-reported feelings about how people felt about the course, what we would call Kirkpatrick Level 1 evaluation, which says little about whether anything changed or had any impact. 
In one study by Tracie L. Stewart, Do the “Eyes” Have It? A Program Evaluation of Jane Elliott's “Blue‐Eyes/Brown‐Eyes” Diversity Training Exercise, there was evidence that the participants were made more aware of their attitudes and behaviours but, as the authors were at pains to point out, whether this has any positive or negative impact was not established. On the downside, researchers also report ethical concerns, along with stress, anger, even resentment towards other groups on the basis of feeling falsely accused. 
The BBC researcher Munira Mirza, the daughter of Pakistani immigrants, investigated diversity training for the BBC and uncovered some destructive training, including the Jane Elliot’s ‘blue eyes/brown eyes’ classroom courses. She is the Director of the Number 10 Policy Unit and in her book The Politics of Culture: The Case for Universalism (2012), warnedof the dangers of using the wrong approaches, such as training, to diversity. After the Black Lives Matter protests in June 2020, she was asked to establish a Commission on racial inequalities for the UK Government.
These issues that dogged the ‘blue eyes/brown eyes’ approach continue to plague diversity training to this day. There is a common assumption that training is he solution to the problem, when the evidence suggest that it is not. These problems were also surfaced in Diversity Inc.: The Failed Promise of a Billion-Dollar Business, where Pamela Newkirk, an award-winning, black journalist and a Professor of journalism at New York University, took a long, hard look at diversity in corporates, academia and Hollywood. Her focus was on African American/Blacks, Hispanics/Latinx and Asian Americans, and her findings were shocking. Despite all the talk and gargantuan sums spent over many years on diversity training, consultants and management jobs, nothing much had changed. 
The problem is clear. From 1985 to 2016 the proportion of black men in management at all US companies with one hundred employees or more barely budged, from 3 percent to 3.2 percent. Despite massive sums of money having been spent, little progress seems to have been made. Google, after spending hundreds of millions on diversity found that their 2% black employee figure remained rock solid. It is not all bad news. After a large class action lawsuit, Coca Cola made real progress and the Rooney Rule, named after Pittsburgh Steelers chairman Dan Rooney, in the National Football Association, where least one minority candidate must be considered for every coaching job, has had positive results. Even there people have found loopholes and ways to circumvent the rule.
Her conclusion is that tick-box, training initiatives are not the solution. They fall short because the process of delivering diversity is a process not an event. Training events tend, in fact, to be counterproductive, as they give the illusion of action, but rarely result in actual impact and change. 

Why diversity training does not work

One of the problems of diversity training is its initial framing. There is a tendency to see groups in a homogeneous fashion, ignoring diversity within that group, not seeing them as individuals. Peter Wood's The Invention of a Concept, Rumy Hasabn's Multiculturalism and Yehudi O Webster's Against the Multicultural Agenda are worth considering as they make these and other arguments against the underlying framework behind diversity training.
We have to be careful that the rhetoric that surrounds diversity in itself may censor debate, a diversity of views being the victim. As employers and employees we should no expect to accept that we are in a state of original sin regarding racism and. diversity and must go through some sort of confession process, facing up to our deficits, through ‘diversity’ courses, to absolve our sin. What we need is research and evaluation.
The evidence shows that it is often an end-in-itself, rather than a means to an end. The vast amount of time and money spent on diversity training, when evaluated, is found wanting, mostly ineffective, sometimes counter-productive. With evidence from large-scale studies, as well as many other focused pieces of research, you would have thought that the message would have had impact. The truth is that few on either the supply or demand side, know the research and whether what they practice works or not. Diversity training is commonly an article of faith rather than a researched and reasoned response.
Companies worldwide spend many hundreds of millions of dollars each year on diversity training. The tragic truth is that much of this is wasted. Groupthink seems to be at the heart of the matter. Groupthink within HR and Learning & Development, also among compliance training companies, who simply do what they do without supporting evidence and tout unproven ‘courses’. This is all part of a culture that finds it easier to just run ‘courses’ rather than tackle real social, organizational and business problems.
Goran Adamson’s The Trojan Horse is a leftist critique of multiculturism and diversity training and initiatives. It is a searing account of the failure of the diversity driven agenda as currently implemented. His detailed examination of diversity in Sweden he calls it out as wrong-headed, counterproductive and conservative. It makes one think deeply about the subject, especially the ‘diversity’ industry, touting ‘diversity courses’. Several dimensions of the diversity agenda are identified as wanting, even dangerous. 

Research evidence – no effect

The major problem is that most diversity training is not evaluated or languishes in the Kirkpatrick Level 1 land of ‘happy sheets’. When research is conducted, it is damning. Major studies from Dobbin, Kalev and Kochan show that diversity training does not increase productivity. Most do not know if it actually works, as evaluations are as rare as unicorns. Thomas Kochan, Professor of management at MIT’s Sloan School of Management’s five year study had previously come to the same conclusions, "The diversity industry is built on sand," he concluded. "The business case rhetoric for diversity is simply naive and overdone. There are no strong positive or negative effects of gender or racial diversity on business performance." 
Harvard’s Frank Dobbin conducted the first major, systematic study of diversity programmes across 708 private sector companies, using employment data and surveys on employment practices. His research concluded that, “Practices that target managerial bias through…diversity training, show virtually no effect.” The research is a very thorough piece of work, and well worth reading, which is why it was completely ignored. 
So check out Alexandra Kalev’s study from the University of Arizona on 31 years of data from 830 companies – how’s that for a Level 4 evaluative study. Her study found, after the delivery of diversity training, a 7.5% DROP in women managers, 10% DROP in black women managers and a 12% DROP in black men in senior management positions. There were similar DROPS among Latinos and Asians. Kochan found that none of the companies he contacted for his study had carried out any systematic evaluation of diversity training. Evidence around productivity is mostly anecdotal and repeated as a mantra by interested parties. The strength of this study comes from the quantity and integrity of the data. It relies on compulsory federal EEOC (Equal Employment Opportunity Commission) filings on the number of women and people of colour in management, along with details of diversity training programmes.
The bottom line, as explained by Dobbins and Kalev, is that “Statistical analyses of time-series data on the effects of corporate diversity measures reveal several patterns. Initiatives designed to quash managerial bias, through diversity training, diversity performance evaluations, and bureaucratic rules, have been broadly ineffective”.

Backlash effects

At a more universal level, some argue that training people to believe that racism is in the eye of the beholder. In other words, if you think some act is racist, then it automatically is. This can result is an accusatory grievance culture.
Dobbins stated that “Research to date suggests that… training often generates a backlash.” Many other studies show that diversity training has activated, rather than reduced diversity; Kidder et al 2004, Rynes and Rosen 1995, Naff and Kellough 2003. Louise Pendry of Exeter University also claims that there is no evaluative evidence showing that these programmes work. Even worse, many may do more harm than good. Tracie Stewart, a professor at Georgia University, has identified "backlash" or "victim blame", after some courses, where the learners harbour resentment against other minority groups for the way they are made to feel. Rather than bringing people together, it may be reinforcing differences.

Legal drivers

The bottom line is that the vast majority of diversity courses should be dropped, especially if it is simply driven by HRs perception of avoiding prosecution. Diversity courses are “more symbolic than substantive" says University of California LAW Professor, Lauren Edelman, She independently reviewed Kalev's study and concluded that the problem was training in "response to the general legal environment and the fact that organizations copy one another." The problem centres around courses run in response to legislative and external pressures. Kalev found that, "Most employers….force their managers and workers to go through training, and this is the least effective option in terms of increasing diversity. . . . Forcing people to go through training creates a backlash against diversity." Organisations must be careful not to compound the problem, as two wrongs may not put things right. Racists and sexists in the workplace are unlikely to be rooted out by punishing many because of the actions of a few.

An unregulated mess

A fundamental problem is that training in this area is not regulated, has no real guidance on best practice, no real quality control and is open to anyone who builds and sells a course. This ‘training’ can be anything, including; pure social activism, he imaginary delving into one’s unconscious, divisive and arguably unethical role-playing and attempts at behavioral change. It can also be tick-box training, designed merely to claim that one is doing something. It is mostly a mixture of the dubious, ineffective and destructive. The idea that one can solve such problems with short training courses seems in itself flawed. Deeply ingrained processes, beliefs and behaviours are unlikely to be much affected by a few hours in a classroom course. To conclude, the world is littered with courses on diversity, racism and sexism, it is NOT littered with evidence that it works. 

Drop 'unconscious bias' training

Racism and sexism are serious problems but not all training efforts are serious solutions. The latest solution is training courses that purport to tackle ‘unconscious bias'. Note that I'm not attacking training on conscious racism and sexism, only the idea that training should focus on the unconscious. Starbucks led the charge, largely as a PR campaign to protect their share price,  but it is everywhere. There is something rather creepy about HRs move on the unconscious. Since when did it become acceptable to see an employees ‘unconscious’ as an addressable area for ‘retraining’? It’s Orwellian.
First, the unconscious is wrong target. Apart from the dedicated racist, few will admit to being racist in surveys. Many may hold light or even strong views on race without admitting it to anyone, certainly not researchers, who would almost certainly be seen as judgemental. This has led the HR and training world to turn and target the unconscious, a big mistake. Explicit, conscious racism and sexism, may actually be the true focus for training, not the diversion of ‘unconscious bias’, all on the basis of seriously flawed psychometric tests. 
In truth, they are not measuring unconscious bias at all. The Banaji and Greenwald IAT (Implicit Association Test), created in 1994, is one of a number that are being foisted upon millions of employees. Just because people select words from pairs does not mean that this taps into their unconscious. This paper sends several cannonballs over the bow of the supposed ship sailing into the uncharted sea of the unconscious. Just because someone can’t explicitly explain something does not mean that it has its origins in the unconscious. There are plenty of alternative explanations with more plausible causality. You may simply be registering familiarity (not bias) in matching words with images. Alternatively you may be using conscious but instantaneous recognition, not the unconscious, to links the words and images. As Tony Greenwald, one of the creators of IAT said, "I see most implicit bias training as window dressing... After 10 years working on this stuff and nobody reporting data, I think the logical conclusions is that if it was working, we would have heard about it."
NR and training simply use the wrong language. The mutual exclusivity of conscious and unconscious bias is far from proven and psychologists are wary of even using the word ‘unconscious’. One can add the prefix ‘un’ to the word ‘conscious’, and assume this is something clear, the ‘unconscious’, a place where hidden biases are stored in little Pandora’s boxes. But the ‘unconscious’ is problematic in psychology. What is the difference between a memory and an unconscious event? If you read the literature in this field you will find the word ‘unconscious’ is largely absent. Psychologists tend to use the terms ‘implicit’ and ‘explicit’, which cuts loose from the terminology of psychotherapy, to bring in a wider range of phenomena. Psychologists really are wary of this binary opposition between unconscious and conscious - but not HR. Of course, selling a course called ‘Implicit beliefs’ may not bring in the expected sales.
Going back to the tests for ‘unconscious bias’, reliability matters in tests. You don’t want a test that get very different results on same person when they retake the test. Guess what? The IAT test is unreliable, so it should NOT be used as a test, as there is not enough evidence that it predicts your behaviour. To be precise, the desired retest reliability should be above 0.7. It is, in fact, 0.44 for racism and 0.5 for IAT tests overall. Even if we assume the unconscious has some status, the causality of beliefs and behaviour can still be studied, that’s why this is bad news - four separate meta-analyses show weak predictive behaviour from such tests. This is a real problem, as even if one encounters the unconscious bias, as it has almost no causal effect, all that work is largely pointless.
Even the people who work in this area warn against the inference that reducing unconscious bias reduces racist or sexist behaviour. In fact, a meta-study in 2017, that looked at 494 previous studies, showed no evidence for the reduction of unconscious bias having an effect on biased behaviour. Let’s be clear, if true, then what is claimed by those who sell this training and much of the training, that they change behaviour, is quite simply, untrue. 
Training in ‘diversity’ and ‘unconscious bias’ fits nicely with the zeitgeist. At best, it is a costly waste; at worst, falsely accusatory and counterproductive. If the identification of unconscious bias is a waste of time, as is training around that concept, that still leaves us with conscious bias.

A way forward

All is not lost. Rather than unconscious bias one needs to focus on conscious racism and bias, not psychobabble. Who better to turn to than the world’s acknowledged expert in ‘bias’, Daniel Kahneman, who won the Nobel Prize for his work in the field. His book Thinking Fast and Slow is essential reading if you are interested in how bias works in the mind. Note that if you’re interested in less academic book that explains it in a more readable form The Undoing Project by Michael Lewis, is excellent. Coming back to Kahneman, in the last two pages of the book he addresses the issue of combatting bias and starts by saying that… “System 1 is not readily educable”. So don’t look to changing System 1, and thinking that you can eliminate unconscious bias, where the supposed ‘unconscious bias is said to exist. His recommendation is… “The way to block errors in System 1 is simple in principle: recognise the signs that you are in a cognitive minefield, slow down, and ask for reinforcement  from System 2.” This is good advice, so how do we do this? Kahneman suggests that organisations use process and “orderly procedures”, such as “useful checklists… reference forecasts… premortems”. I agree. Much is to be gained through organisational checks and balances, not falsely accusatory training based on unreliable, supposedly diagnostic tools.
One of the problems, that Dobbin, Kalev and others found, was the focus on ‘sensitivity training’ where people are often forced to focus on interpersonal conflict. These were the training courses that produced a backlash, as they were intrinsically accusatory. Blaming and shaming doesn’t work and training should not be presented as ‘re-education’. 
One bright spot was the finding that some diversity initiatives, namely those that were voluntary and aligned with business goals, were successful. The problem with voluntary initiatives is that they attract the already converted, as the miscreants don’t volunteer. 
We know that diversity can be a good thing, although this is a more complex issue than many imagine. The good news is that a solid, strategic effort can solve this problem but it needs a strategic approach with planning, action and perseverance. This is similar to Professor Frank Dobbin’s study at Harvard, who showed, in his massive study that ‘training’ was not the answer, and that other management interventions were much better, such as process. Around recruitment, promotion and mentoring.

So what works?

The evidence suggests that enormous sums are spent. On wasted training, money and time that would be better spent on more strategic initiatives that look at processes, rather than training events. The trick is to drop diversity courses and look at direct actions. 

Data-driven approach

Harvard Professor of Public Policy Iris Bohnet’s recommendation is that we “use data on what works to inform our decision". This is correct, data that leads to decisions and actions is the likeliest solution to these problems. We now have the mindset and tools to use ‘data’ in all aspects of organisational development from recruitment to promotion to the very top in ways that unblock barriers.

Diversity manager

A Diversity manager will be able to identify a strategy and implement subtler approaches, as well as having the authority to effect change. Choose a diversity manager who does not have a focus on training and who has the skills to implement and use a data-driven approach to the management, promotion and evidence-base on diversity. It will give initiatives power and focus.

Diversity programmes

Strategic programmes, with measurable outcomes, driven by data, that focus on management interventions and strategic progress, work better than crude, compulsory courses. It is important that everything cascades out of a strategic intent, stated by those at the top. It is only if he strategy has a powerful group behind it that the change management will happen.

Recruit openly to boards

Recruitment through personal contacts and recommendations tends to produce homogeneous boards. Open recruitment through advertisements, especially if online, will get to a wider target audience. Again, anonymous selection processes will work. Even headhunters will have a limited range of candidates, often selected with bias by the headhunters themselves. If you do use headhunters or recruitment agencies check their own diversity mix.

Anonymised recruitment

This starts with recruitment. In general, anonymise recruitment as much as possible. Where one can anonymise the process, eliminate off-putting  language, names, gender, age and photographs. Blind evaluation and more attention to actual tasks, competences and performance, as opposed to those that are assumed, is necessary for fair recruitment and subsequent development of employees. 

Drop graduate recruitment

Except in cases where it is necessary, drop the ‘graduate’ requirement. Many large organisations have dropped their general graduate recruitment requirement, such as Google and most major consultancies. Demand that apprenticeship schemes be adopted and that competences rather than paper qualifications are assessed. The trick is to have a more open door of opportunity, not a closed door of pre-determined resume-tick outcomes.

Appraisal process

Actions can be taken throughout the development process to make sure that factors that may induce bias are eliminated as far as possible. Annual appraisals, in some organisations, have given way to more continuous forms of management feedback. This allows for more careful and continuous attention to be given to the issue of diversity.

Process and  procedures

Organisations invariably have processes and procedures. These are marbled like fat into the meat of the organisation. Attention to these minor processes and procedures will produce results as that is where most of the problems lie. There may be areas such as internal communications, office environments, décor and social events, where some reflection on diversity would work.

Mentoring

Performance appraisals show poor results, with managers showing bias in outcomes. These need to be redesigned or abandoned, as some have doe, to be replaced with continuous feedback. Whereas mentoring seems to have a positive, beneficial effect on actual diversity outcomes.

Do nots…

Do not rely on grievance procedures and processes to effect change. Do not run one-off diversity courses. Do not run ‘unconscious bias’ courses. Do not treat diversity as an event, treat it as a process, an on-going process that need vigilance and constant effort.

Conclusion

To genuinely build a constructively fair and meritocratic organisation that values everyone, regardless of race, gender and socio-economic background, most diversity strategies have to change. It needs a subtler, less compulsory and more managerial, strategic approach that eschews glib, compulsory courses and grievance procedures, in favour of more voluntary and actionable approaches, along with positive interventions. 

Bibliography

Bloom G. Stephen, 2005. Lesson of a Lifetime Smithsonian Magazine 
Stewart, T.L., 2003. Do the “Eyes” Have It? A Program Evaluation of Jane Elliott's “BlueEyes/BrownEyes” Diversity Training Exercise 1. Journal of Applied Social Psychology33(9), pp.1898-1921.
Mirza, M., 2011. The politics of culture: The case for universalism. Springer.
Newkirk, P., 2019. Diversity, Inc: The Failed Promise of a Billion-dollar Business. Bold Type Books.
Wood, P. and Wood, P., 2003. Diversity: The invention of a concept. San Francisco, CA: Encounter Books.
Hasan, R., 2010. Multiculturalism: Some inconvenient truths. Politico's Publishing.
Webster, Y.O., 1997. Against the multicultural agenda: A critical thinking alternative. Greenwood Publishing Group.
Adamson, G., 2015. The Trojan Horse: A leftist critique of multiculturalism in the West. Arx Förlag Malmö.
Dobbin, F. and Kalev, A., 2013. The origins and effects of corporate diversity programs.
Dobbin, F. and Kalev, A., 2016. DIVERSITY why diversity programs fail and what works better. Harvard Business Review94(7-8), pp.52-60.
Kalev A, Dobbin F, Kelly E. Best practices or best guesses? Assessing the efficacy of corporate affirmative action and diversity policies. American sociological review. 2006 Aug;71(4):589-617.
Kochan, T., Bezrukova, K., Ely, R., Jackson, S., Joshi, A., Jehn, K., Leonard, J., Levine, D. and Thomas, D., 2003. The effects of diversity on business performance: Report of the diversity research network. Human Resource Management: Published in Cooperation with the School of Business Administration, The University of Michigan and in alliance with the Society of Human Resources Management42(1), pp.3-21.
Kidder, D.L., 2004. Backlash toward diversity initiatives: Examining the impact of diversity program justification, personal and group outcomes. International Journal of Conflict Management15(1).
Rynes, S. and Rosen, B., 1995. A field survey of factors affecting the adoption and perceived success of diversity training. Personnel Psychology48(2), pp.247-270.
Naff, K.C. and Kellough, J.E., 2003. Ensuring employment equity: Are federal diversity programs making a difference?. International Journal of Public Administration26(12), pp.1307-1336.
Pendry, L.F., Driscoll, D.M. and Field, S.C., 2007. Diversity training: Putting theory into practice. Journal of Occupational and Organizational Psychology80(1), pp.27-50.
Stewart, T.L. et al., 2012. White privilege awareness and efficacy to reduce racial inequality improve White Americans’ attitudes toward African Americans. Journal of Social Issues68(1), pp.11-27.
Gawronski, B., Morrison, M., Phills, C.E. and Galdi, S., 2017. Temporal stability of implicit and explicit measures: A longitudinal analysis. Personality and Social Psychology Bulletin43(3), pp.300-312.
Oswald, F.L., Mitchell, G., Blanton, H., Jaccard, J. and Tetlock, P.E., 2013. Predicting ethnic and racial discrimination: a meta-analysis of IAT criterion studies. Journal of personality and social psychology105(2), p.171.
Forscher, P.S., Lai, C.K., Axt, J.R., Ebersole, C.R., Herman, M., Devine, P.G. and Nosek, B.A., 2019. A meta-analysis of procedures to change implicit measures. Journal of Personality and Social Psychology117(3), p.522.
-->
Kahneman, D., 2011. Thinking, fast and slow. Macmillan.