The man who changed a fortune cookie and started a revolution - Notes on attitude - Why Aren't We Saving the Planet?: A Psychologist's Perspective - Geoffrey Beattie

Why Aren't We Saving the Planet?: A Psychologist's Perspective - Geoffrey Beattie (2010)

Part I. Notes on attitude

Chapter 4. The man who changed a fortune cookie and started a revolution

If attitudes were measured easily and uniformly for the first sixty years of their existence, in the 1990s there was something of a revolution in the measurement of attitudes and indeed in our whole approach to the topic. What drove this revolution was what psychologists like to call the ‘attitude- behaviour problem’ - the fact that attitudes and behaviour often simply fail to match up, and that attitudes measured in this way often do not predict behaviour in quite the way that many had hoped.

In an article in 1990 Anthony Greenwald attacked the then current theoretical understanding of attitudes and developed an argument that we should reconceptualise what we mean by an attitude in the light of new research in cognitive rather than social psychological research (the principal domain for this type of work). He cited the conclusion of Myers (1987), who had come to the view that the models of the attitude-behaviour relationship only really worked by ‘limiting the scope of the attitude concept’. Thus:

Our attitudes predict our actions (1) if other influences are minimized, (2) if the attitude is specific to the action, and (3) if, as we act, we are conscious of our attitudes, either because something reminds us of them or because we acquired them in a manner that makes them strong. When these conditions are not met, our attitudes seem disconnected from our actions. (Myers 1987:45)

Greenwald, clearly not a man to mince his words, wrote: ‘Myers’ conclusion is decidedly embarrassing as a summary of the predictive power of social psychology’s major theoretical construct’ (1990:256). What he did in the remainder of this short paper was to review new research in cognitive psychology on unconscious cognitive processes to provide a new theoretical basis to the work on attitudes. The fully formed Gordon Allport, the psychologist who had rejected Freud’s attempt to bring the role of the unconscious even into his brief meeting with his younger self, would have turned in his grave. One consequence of Greenwald’s devastating review was to ‘call into question the appropriateness of the presently most favoured techniques of attitude measurement’ (1990:256). The rest, as they say, is history.

So who was this Anthony Greenwald? Greenwald was an American professor who had obtained his PhD in Social Psychology from Harvard in 1963, interestingly under the direct tutelage of Gordon Allport. As something of a rising star, Greenwald became known for his ability to take old concepts and theories and breathe new life into them through his singleminded determination (according to his colleagues). Mahzarin Banaji, who has worked with him for many years, illustrated Greenwald’s eye for improvement with a telling anecdote during her congratulatory speech for Greenwald when he was honoured with the ‘Recipient of the Distinguished Scientist Award, 2006’. Banaji recalled, ‘I was there when you opened that famous fortune cookie message. It read innocuously, “There is nothing which cannot be improved.” Not good enough for you!’ Greenwald carefully reached for his pen, and drew a line through the word ‘which’, replacing it with the word ‘that’, so the fortune cookie now read, ‘There is nothing that cannot be improved.’ It takes a certain chutzpah to correct a fortune cookie in front of your colleagues, especially to correct the grammar of a fortune cookie in public. Greenwald was to demonstrate this chutzpah repeatedly in the development of this new field, a field that is so closely linked to this day with the name of this one eminent researcher.

Greenwald started with Myers’ conclusion. Myers’ paper was in many people’s eyes a devastating summary of the attitude-behaviour relationship, but one bit of the conclusion was wrong according to Greenwald, and that was his third conclusion. Greenwald claimed that some of the most reliable and robust findings of attitudes predicting behaviour are exactly in those domains where the actor is not attentionally focusing on the attitude. One example of this, used by Greenwald, is the halo effect. The halo effect is the tendency to make new positive (or negative) judgements about a person when a positive (or negative) attitude towards that person already exists. In a famous study Landy and Sigall (1974) found that male participants judged the quality of a poor essay more favourably when the female author was attractive than when she was unattractive. This seemed to occur without any attentional focus on the underlying attitude. The fact that a photograph of the author was in the folder with the essay was designed to be almost coincidental.

Greenwald did not attempt to excuse himself here. He used an example from his own life as an academic to illustrate that implicit attitudinal forces are in constant operation in our everyday lives:

As a manuscript reviewer, I often cannot help noticing an initial warm, positive reaction when I review a manuscript that cites my work favourably (or maybe just cites it at all), and sometimes I notice the opposite - a colder reaction when some of my work that might have been cited is not mentioned. I know that these reactions interfere with the way my work as reviewer should be done, but it is difficult to avoid these reactions - and it is difficult not to do the review by searching for virtues that will justify the initial warm reaction, or for flaws that will justify the initial cold reaction. (Greenwald 1990:257)

This was the phenomenon that he was trying to pin down and understand - this initial reaction full of emotional overtones well under the radar of consciousness but ultimately controlling our behaviour. Nobody, but nobody, escaped his gaze. With Eric Schuh and Katharine Engnell he analysed the citation patterns for authors whose names were selected from the 1987 Social Sciences Citation Index on the basis that they could be classified unambiguously as Jewish or Anglo-Saxon in origin. They found that Jewish-named authors cited 6% more authors with Jewish names than did Anglo-Saxon authors, and conversely Anglo-Saxon authors cited 7% more authors with Anglo-Saxon names than did Jewish authors (Greenwald, Schuh and Engell 1990) His conclusion was that ‘Social scientists (who are widely regarded as being relatively free of prejudice) might display ethnic prejudices’ (Greenwald 1990:258). It seemed that even his friends and colleagues were not safe from his emerging theoretical views. This was putting colleagues on the spot just as pointedly as did Freud exactly seventy years earlier.

Greenwald saw an emerging pattern in all of this.

The subject is in a situation that requires a response to some object; attitude towards an attribute of the object influences the response, and it does so without the subject’s being aware that an attitude is being activated. These situations amount to indirect memory tasks for which the response has an evaluative component. (1990:259)

Greenwald’s goal was to elucidate the processes underlying implicit cognition and to demonstrate why they were critical to a reformulation of how we thought about attitudes and how we should measure them.

For Greenwald and Banaji in 1995, ‘The signature of implicit cognition is that traces of past experience affect some performance, even though the influential earlier experience is not remembered in the usual sense - that is, it is unavailable to self-report or introspection’ (1995:4-5). This paper provides a simple example of implicit cognition in operation. Consider an experiment where the participants have to generate a complete word in response to an incomplete letter string (a word stem or a word fragment). The words that the participants generate here are more likely to be words that they have been ‘casually’ exposed to earlier in the experiment than words that they have not been exposed to. Greenwald and Banaji write that ‘This effect of prior exposure occurs despite subjects’ poor ability to recall or recognize words from the earlier list’ (1995:5). In other words, even though the participants do not recognise certain words as being on a list that they have seen previously, these words have been ‘primed’ in their memory and come out more readily than words that have not been primed.

Similar to this is the experimental work he had already conducted on what he called ‘detectionless processing’, one of the main areas of implicit cognition, in which stimuli of which people have little conscious awareness can be demonstrated to have an impact on behaviour. In detectionless processing, he demonstrated that words can be processed in terms of meaning, and therefore have an impact on our subsequent thoughts, even in situations where the words are presented below the threshold of conscious perception. What the participants in his study had to do was to decide whether each of a series of target words meant something good (e.g. ‘fame’, ‘comedy’, ‘rescue’) or bad (‘stress’, ‘detest’, ‘malaria’). Half a second before each of these target words was presented, a priming word was presented briefly to the non-dominant eye. The priming words themselves were either good (‘happy’, ‘joy’, ‘peace’, ‘love’, ‘excellent’, ‘pleasant’) or bad (‘evil’, ‘grief’, ‘sad’, ‘gloom’, ‘ugly’, ‘horrid’). The priming word was followed a matter of milliseconds later by a pattern mask (a complex visual stimulus) that interferes with the perception of the word that precedes it such that participants were unable to report what position the priming words were in on a computer screen (i.e. whether they were to the left or right of a fixation point). So even though the participants seemed to have little conscious information about the priming words and did not know even where these words had been presented, these ‘invisible’ words affected processing of the subsequent target words, such that target words that were preceded by a congruent prime (for example, a positive target word preceded by a positive prime, or a negative target word preceded by a negative prime) were identified significantly more quickly than target words that were preceded by a incongruent prime (see Greenwald, Klinger and Liu 1989). In other words ‘invisible’ words (with extremely brief presentations followed by a pattern mask) can influence the workings of the human mind. Suddenly, the unconscious was back in vogue.

The article that Greenwald wrote with Mahzarin Banaji in 1995 is in many senses of the word a classic paper. It has been cited by social scientists over 1200 times. The thesis of the paper articulated Greenwald’s view of the unconscious in determining our behaviour. It successfully links the growing body of research on implicit cognition with the research on attitudes and behaviour. It argued that ‘Recent work has established that attitudes are activated outside of conscious attention, by showing both that activation occurs more rapidly than can be mediated by conscious activity … and that activation is initiated by (subliminal) stimuli, the presence of which is unreportable’ (1995:5).

Greenwald was keen to demonstrate the role of implicit processes in a series of domains. He employed the ‘false-fame procedure’ in an attempt to find experimental evidence for implicit stereotyping, including sex-role stereotyping which associates gender with achievement. The method is simple yet produces highly significant results. It was based on a procedure used by Kelley and his colleagues (Jacoby, Kelley, Brown and Jasechko 1989) to uncover implicit memory. In the Kelley studies, participants would read a list of both famous and non-famous names on day 1. The next day the same participants were presented with previously seen non-famous names from the first list or new non-famous names that they had not seen before mixed in with previously seen and new famous names. The question that they were asked was ‘Is this person famous?’ The researchers hypothesised that although the non-famous names should fade from memory over the twenty-four-hour period (from day 1 to day 2), the fact that the participants have seen some of the non-famous names before on the first list might lead them to the conclusion that some of these names were actually famous. And that is exactly what they found: there was a higher false-alarm rate for the previously seen non-famous names than for the new ones. Some non-famous names had quite literally ‘become famous overnight’ because of the extra exposure and the resulting familiarity. This was a demonstration of a significant unconscious influence on memory.

What Greenwald and Banaji did next was to introduce gender into this simple paradigm. They asked what happens if you use a set of male and female names here and consider gender as a principal variable. They found that the false-alarm rate for (previously seen) non-famous names was significantly greater for male than for female names. In other words, the participants were more likely to assume that the male names were famous (because of the familiarity due to repeated exposure) when they were not actually famous, compared with female names. So the authors concluded that this was clear evidence for implicit gender stereotyping in everyday life where maleness is associated with achievement.

In their summing up in this article, they concluded that:

considerable evidence now supports the view that social behaviour often operates in an implicit or unconscious fashion. The identifying feature of implicit cognition is that past experience influences judgment in a fashion not introspectively known by the actor. The present conclusion - that attitudes, self-esteem, and stereotypes have important implicit modes of operation - extends both the construct validity and predictive usefulness of these major theoretical constructs of social psychology. Methodologically, this review calls for increased use of indirect measures - which are imperative in studies of social cognition. (1995:4)

In other words, this review article was arguing for a major reconceptualisation of how we should view both attitudes and social behaviour more generally. And, of course, if we accept the argument about the role of implicit or unconscious factors in controlling both our attitudes and our behaviour then we must accept that we will need a new methodology for studying some of these processes: a new methodology that does not involve self-report and is sufficiently sensitive to detect some of the very rapid processes involved. The development of the new method was crucial here because Greenwald recognised that some of what he was saying had been said before.

Implicit social cognition overlaps with several concepts that were significant in works of previous generations of psychologists. Psychoanalytic theory’s concept of cathexis contained some of the sense of implicit attitude, and its concept of ego defense similarly captured at least part of the present notion of implicit self-esteem … At a time when the influence of psychoanalytic theory in academic psychology was declining, its conceptions of unconscious phenomena that related to implicit social cognition were being imported into behaviour theory (Dollard and Miller 1950; Doob 1947; Osgood 1957). The New Look in Perception of the 1950s focused on several phenomena that are interpretable as implicit social cognition. The developing Cognitive approach to these phenomena can be seen in Bruner’s (1957) introduction to the concept of perceptual readiness. (Greenwald and Banaji 1995:20)

But the punch line came next. Here after all was the man with all the chutzpah to correct the grammar of a fortune cookie in public.

Importantly, the psychoanalytic, behaviourist, and cognitive treatments just mentioned all lacked an essential ingredient, that is, they lacked reliable laboratory models of their focal phenomena that could support efficient testing and development of theory. The missing ingredient is now available. (Greenwald and Banaji 1995:20)