This blog originally appeared on June 6, 2014. It was re-posted on August 25, 2021.

Bias. It seems such a simple, innocuous word at first; yet, it is one that history educators at many levels have come to loathe and despise. We grimace when we see it in essays, and grumble when students use it in class. Yet, most of us would agree that understanding ‘bias’ is an important part of historical education. So, how did four little characters come to cause so much pain and consternation? The word itself is not to blame, of course. It a long-standing member of the English language and should be afforded the same respect and deference that we grant ‘perspective’, or ‘opinion’, or any other word depicting the great inner workings of the human mind. No, the reason the word has become so reviled in history circles is because of its misuse.

It is not entirely clear from where this misuse stems. Many university lecturers, accosted with page after page of uncritical dismissals of sources under the grand heading of ‘bias’, blame the secondary education system. Those teaching Highers, A-Levels and AP History blame those earlier on, those in middle schools or preparing students for Standard and GSCE-level exams. Perhaps the blame lies in primary school. Perhaps nursery. Perhaps the womb.

But, I hear you ask, how is the term, indeed, the concept of bias misused? The first and most easily rectified misuse is the confusion of the noun with the adjective. Far too many of my students have consistently, and often stubbornly, referred to a person or a source as ‘bias’, despite years of angry annotations on returned work. In this case, the correct form of that word is ‘biased’, or having characteristics that suggest a bias in perception. Bias is a noun, biased is a (predicative) adjective. I am biased. I have a bias. Simple.

Although this grammatical error is easy to mock (and has recently become the bane of UK Twitter grammarians), it is not the core issue. Instead, it is the seemingly obstinate misunderstanding of the word’s meaning that renders student essays so painful. Saying, for example, that ‘Thomas Jefferson was biased’ is akin to saying ‘Thomas Jefferson breathed air’. It is true, but it is not particularly useful to state it in your essay on the Virginia Statute for Religious Freedom. All creatures possess biases. Indeed, I would argue that all things are biased, especially clothing, but that is an argument for another day. Instead, this post will focus very narrowly on biases within human perception and reasoning.

When I began this line of enquiry, I thought that bias was not a loose or fuzzy term, that it has a very specific epistemological meaning. That it, surely, did not mean any of the following:

  1. Lying
  2. Having an opinion
  3. Arguing a point
  4. A non-specific synonym of racist, sexist or classist

These four definitions, the most commonly used by my students, were absurdly vague. Indeed, taken to their natural extreme (often by students themselves), the use of the term bias was simply a way of discounting all of what has been previously written by the essayist. For example

Thomas Jefferson believed that different races could never successfully integrate into a single society, but, he owned slaves, so he was biased.

On a basic, gut-reaction level, this sentence makes perfect sense. But let us examine the second clause more closely. Rephrased, all it argues is that because he owned slaves, he was racist, and because he was racist, we can ignore him. That, dear reader, is not critical analysis. It is not even analysis. It is uncritical dismissal, and it is the very shoddiest of historiographical argument. This is not something we should allow our students to get away with — and thankfully, most teachers do attempt to correct this obvious deviation from reason.

But we are constantly undermined by some unknown, sinister force. No matter how often we discuss bias with our students, at university and secondary school levels, the term keeps popping up in this much maligned form. Who is the culprit? We are.

Every year, across the world, students are coming into university with the very best intentions. Most genuinely want to do well at their degree. They and their parents do their best to prepare for the rigours of university history by buying the entire first-semester reading list alongside at least one ‘student guide’ to history or history skills. There are dozens of these thin paperback volumes sitting on Amazon’s virtual shelves, from the entire spectrum of academic publishers. They are crafted by historians from across the world, peer-reviewed and given a lecturer’s stamp-of-approval through university book store stockings and course reading lists. These are the books we encourage, even demand, our students purchase and live by. But have you ever really sat down and read them? Really?

In my quest to understand what has gone wrong in our view of the word bias, I looked through a large stack of these slender guidebooks, hoping for a succinct, effective explanation of source bias to which I could point my students. I was shocked and indeed horrified by what I found. Many made no use of the word ‘bias’ at all. Sections on how to analyse primary or secondary material are brief, worryingly so, as most of the pages were spent discussing margin sizes and note taking. Mary Abbot’s essay on ‘sources’ in History Skills (Routledge, pp. 23-26), for example, briefly explains that ‘A good historian is a sceptic who takes nothing at face value, who is ever open to new ideas and who is prepared to modify her views in light of new evidence or persuasive argument’ before offering up a variant of the 5 Ws.

Jeremy Black and Donald Macraild, in their Palgrave Guide to Studying History are more thorough, offering 5 pages on bias (p. 12-15, 40-1), in which they explain to readers various possible ideologies espoused by historians, paying particular note of feminist and women’s history. This section offers some heady abstractions about historical conceptions of time, and some lambasting of ‘the first draft of history’ by newspaper journalists (by Orwell, not the authors of the guidebook). It provides interesting food for thought, and could open up fruitful discussions as a required reading, but a useful guide to the practical implications of bias in student reading and writing it is not.

The late Richard Marius, who provided my own Short Guide to Writing about History (Longman) as an undergraduate, offers a lengthy account of ‘Questioning your Sources’ (p. 32-45) with special attention given to statistical inquiry. It offers a detailed look at each of the 5 Ws, in turn, and how they overlap and corroborate each other in helping students understand their materials. It also explores the fallacies that often strike students in determining causation and correlation. Overall, it is a very good look at how to examine material, but bias itself is only implicitly touched upon when he concludes that ‘Skepticism is one of the historian’s finest qualities. Historians don’t trust their sources, and they don’t trust their own first impressions. They question everything.’

Looking at these guides, all of which have slightly different audiences and aims, it would be hard to outright disagree with the material they put forth. I use the 5Ws (with my good friend H) every day and I am a happy skeptic, but this is simply not enough to initiate history novices into the field. It leads to being critical (if not petulant) for the sake of not looking naive, and rather mind-numbing lists of the answers to those 5 Ws.

I turn now to the two guidebooks that discussed bias at greatest length. One made me furious, the other relieved.

The former is Jules R. Benjamin’s 2010 Student Guide to History. He introduced his section on bias in the follow manner:

One of your most important concerns when reading a primary source is determining its reliability. Primary sources can be fraudulent, inaccurate or biased [emphasis his]. Eyewitness accounts may have been deliberately distorted in order to avert blame or to bestow praise on a particular individual or group. Without intending to misinform, even on-the-scene judgements can be incorrect. Sometimes, the closer you are to an event, the more emotionally involved you are, and this involvement distorts your understanding of it. We can all recall events in which we completely misunderstood the feelings, actions, and words of another person. Historians have to weight evidence carefully to see whether those who participated in an even understood it well enough to describe it accurately. (p. 31-32)

This is a perfectly sensible way to ease students into the idea. However, once the his specific advice begins to flow, I begin to take exception. When discussing bias within secondary sources, he explains that

Any bias [emphasis his] on the part of an author or work that you are considering will tell you a lot about a work’s seriousness. Bias — an author’s perspective on a topic — can be modest or very strong. Strong bias is indicated by an unwillingness to consider or acknowledge other interpretations and by the use of harsh language to characterize authors with different perspectives. A book with a strong bias is unlikely to be a useful source, unless your research requires an examination of biased works, as would a paper on the antisemitism expressed in nineteenth-century French accounts of the Dreyfus trial. (p.110)

In all honesty, I was taken aback by the absoluteness of these statements (though, surprisingly, some of my colleagues were not). Bias as a determinant of seriousness (by which I believe he meant our ability to take it seriously, not its own intentions of serious argument)? All sources contain biases, and these can themselves be incredibly useful when teasing out information about the period, not just when discussing the bias itself. Arbitrarily discounting it as ‘unlikely to be a useful source’ based on ‘strong bias’ seems ludicrous and a very bad (even lazy) precedent to set. The second point, defining bias as ‘an author’s perspective on a topic’ is equally absurd (for reasons I will discuss at more length below), yet he continues this train of thought with

Modest bias, however, is a characteristic of all scholars. Every author has a perspective on his or her subject and, like you, is making an argument in support of a thesis. A history of World War II written by an English author is likely to have a different viewpoint from one written by a German author. Many historical developments and their interpretations are topics of profound controversy, and it is almost impossible for a historian to investigate of these controversial areas without being affected by his or her own biases (p. 110).

Here bias is established as being the same as ‘argument’, or even ‘focus’, as well as perspective. Although I would concede that bias does affect argument, focus and perspective, it is not the same thing (nor are these three things the same as each other) and confusing these terms is doing much harm. Such a loose definition of such a seemingly ubiquitous word leads us down the path of Godwin’s Law, and towards a society (in which I sadly fear we already live) that you can discount another’s argument by crying ‘bias!’ and stomping away.

It was here, dear reader, that I was going to enlighten you all with my elevated and flawless definition of bias and how it should be taught to students. But as I began to write, I remembered I had one more student guide to examine, one I had mistakenly left at work and, in fairness, should consult before posting. It reaffirmed my faith that I was not mad, that the true nature of bias was recognised not only by other historians, but by historians writing student history guides. The man I refer to is I. W. Mabbett, and the book is Writing History Essays (Palgrave). Unlike Black and Macraild’s more general history guide, Mabbett focuses specifically on writing essays, and therefore has more time to devote to bias itself, which he allots an impressive 11 pages (p. 44-55), though his index mistakenly suggests a mere 2. Here he discusses problems of establishing authority, distinguishing between fact and opinion, the difference between prejudice and bias, and how to recognise and manage (rather than discount) authorial interest, tendentious language and that oft-dreaded concept, balance. I will not quote extensively from Mabbett, as to do so would be a disservice to the hard work he has done in creating an engaging the helpful flow to this section, but I would highlight this, his definition of bias.

A bias is a built-in tendency to lean to one side, a preference that inclines one to favour one side in an argument. Again, what matters is whether this inclination prevents us from being influenced by evidence to the contrary. This amounts to much the same thing as being unwilling to take account of anything that conflicts with one’s prejudice. It is important to avoid confusing prejudice or bias with the mere possession of an opinion. We all have opinions; what matters is the extent to which we are ready to let our opinions be changed by examination of the evidence. (p. 49)

This is followed by a lengthy set of examples in which he explores how language implies (sometimes unsubtly) what the author’s opinion is, and then explains when this is prejudice, when it is bias, and when it an argument based on previously stated evidence and in which its one-sidedness is justified. I must applaud Mabbett on his work here, as I think that it comes nearest to my own attempts to explain bias to students and how best to manage and use it, rather than wield it as a mighty stick of disqualification.

Yet, as admirable as I find his analysis, my own views of bias stem not from history but cognitive science. Here, bias has a very specific meaning, with a host of sub-definitions for the various aspects and manifestations of it, and it is this specificity that has made the concept far more useful to me than the woollier idea of ‘perspective’. In brief, it is the systematic nature of cognitive biases that makes them an effective analytical tool. A bias is not simply a pre-conceived answer to a particular question, but a systematic mistake, an identifiable pattern, in logic or decision making across a length of time or sample group. Given the same criteria and the same context, the person will systematically make the same mistake, or come to the same incorrect conclusion. Why is this so helpful? The reason is twofold.

First, systematic mistakes can be tracked. A lie, an outright deception, is rarely systematic. People slip up, forget, or have different narratives in different contexts. Moreover, different source types, such as oral histories, photographs, letters, newspaper reports and so on, hide or highlight biases to different extents. If a person has made an error, in fact or logic, this conception of bias encourages me to look at a range of different materials, at different times and in different forms, across a long period or geographic base. I can then determine if something is appearing again and again. If it is systematic, knowing this can help me understand and analyse sources in which the bias is less visible, even invisible without outside knowledge. If it does not appear consistently, even within a very short frame of time or space, then I must consider if the mistake is a lie, and find evidence to that effect, or if the person has obtained misinformation in good faith. Both these answers lead to very interesting conclusions about the source and the individual.

Second, exploring cognitive biases prevents generalisation. Wider, all encompassing terms such as ‘racist’ are very common in everyday speech, but fail to account for cognitive dissonance or different responses to race by the same person, other than to revert to the idea that they are lying. The clichéd argument that person could have a genuine friendship with someone of a different race, but unabashedly condemn that race for economic distress (often unemployment), is difficult to reconcile under the wider term ‘racist’. With cognitive bias, however, it can be explained. A systematic mistake, such as Ingroup Bias, may affect a person differently in two different situations, leading to seemingly conflicting results. In one case, the outgroup is another race, who are seen to be ‘stealing jobs’ or ‘lowering wages’ or engaging in other behaviours that will ultimately destroy that person’s individual chance of employment success. In the other case, the outgroup of that same person may be individuals from another class or region, or even a different football team, and a person of a different race, who happens to work, play or live alongside that person, is now in the ingroup and the assessment of their behaviour is completely altered. The bias is technically the same, and all most humans to some degree, but has different outcomes in different situations. By framing bias in this way, I am not only prevented from discounting a source out-of-hand, but I am encouraged to look more deeply at the internal logic and external context of the source in a critical and ultimately useful way.

Of course, these brief examples are themselves very simplistic views of cognitive bias, but the core idea is also very simple. A bias is a consistent tendency (to use Mabbet’s language) or systematic mistake in logical deduction. It is not merely a reference to a person’s race, gender or educational background, though these things can affect which biases they hold. It is something measurable, and something that should encourage further, deeper and more careful research, rather than discount sources and arguments as irrelevant.

But, then again, I may be bias.