Heels over head? How order can change a whole story

Journalists invest a lot of time planning the course of an interview, thinking about how to start and end a story, on crafting titles, teasers and final twists or punchlines. Despite this high sensitivity towards the impact of presentation order on their audiences, they may not fully realise how much order of information intake also affects their own journalistic process: from the keywords used for the first online search, to the order in which people are contacted or places are visited, the tone of introducing someone, the arc of an interview, framing controversial topics and putting it all together when writing the story up. Have you ever decided to leave out aspects or entire viewpoints that you came across rather late in your research, because they “didn’t quite fit in anymore”? What angle might your story have had if you had met that person or learned about those statistics earlier in the process? Might the new mall have become a threat to small local businesses rather than a long due investment in infrastructure with job opportunities for residents, or vice versa?

In 1946, Polish-American Gestalt psychologist Solomon Asch ran a now famous series of experiments to study impression formation in a systematic and controlled way (4). In his lab, Asch presented participants with two versions of attributes describing a person and asked them to form an impression. Both versions contained exactly the same attributes, just in reverse order. In one experimental condition, the list went intelligent, industrious, impulsive, critical, stubborn, envious; in the other condition, it said the person was envious, stubborn, critical, impulsive, industrious, intelligent. When later asked for their overall impression, participants in the first group characterised the person much more positively than in the second. This effect became known as the primacy effect in impression formation — the first items had a much stronger impact on the overall picture than the latter ones.

These and similar findings strongly supported Asch’s leading hypothesis that “the whole is more than the sum of its parts”. In line with the general claim of Gestalt psychology, he was convinced that humans actively organise sensory input by connecting it to prior knowledge and drawing conclusions, striving to make sense of our experiences and to achieve a meaningful “gestalt”. With this explicit assumption of dynamic interactive effects between input, context, and mental preconceptions from prior learning, the approach stood for a whole new understanding of the human mind and it is endorsed to date in social cognitive psychology (5). In contrast to strictly behaviourist learning theory, it claims that humans are much more than the result of countless everyday learning experiences; in contrast to psychoanalysis, its hypotheses can be tested by controlled and replicable experiments, delivering solid scientific data.

One could of course argue that the attributes used by Asch weigh differently when integrated into an overall impression, and that the lists thus can hardly be compared, in particular, when it comes to context effects (e.g. the cleverness of a person also described as “warm” gets a different connotation than that of a person also described as “cold”; 4). But even when the information is restricted to one single trait domain, the primacy effect persists: Jones and colleagues (6) let their participants watch a person solve an intelligence test. In all experimental conditions, the person answered 50% of the items correctly, but conditions differed in the distribution of correct and wrong answers over time: In one condition, the person started off with many correct answers and then made more and more mistakes. In another condition, they started off poorly and then got better, and in a third condition, correct and wrong answers were distributed randomly over time. The person was seen as considerably more intelligent if they had had a good start. This was surprising in that it was originally hypothesised that improving performance over time should lead to a more positive judgement than declining performance — but the primacy effect is apparently stronger than any such eventual effect (6). By the way, even before Asch, Lund (7) had shown in 1925 that people are more easily persuaded by arguments when they are presented at the beginning, and this insight has quickly found its way into rhetoric and propaganda.

Applied to everyday life, it is thus literally only fair that the order of appearance at the Eurovision Song Contest and in many jury sports contests or political debates for elections is determined through lottery. Similarly, in negotiations — like on flea markets, in real estate sales or for freelance salaries — the party who gets the chance to suggest a price first usually has a substantial advantage by setting that anchor as a frame of reference (11). In research, primacy effects have been reliably shown in many different studies and for many different contexts and domains, including perceived fairness of policy decisions, processing of political news, assessing the suitability of job applicants, moral behaviour, or sports performance (2, 12). Similarly, there is a preference for political candidates, food and other consumer objects that were presented first in a series (13). Taken together, order is reliably found to have a major impact on impression formation, further information processing and judgement, and primacy effects usually dominate recency effects except when it comes to memory.

Why and how do primacy effects emerge?

There are a number of different explanations for why primacy effects come about. First of all, it is cognitively easy — we quickly grasp the first bits and pieces of information. If in that early stage we can already come up with a reasonably reliable idea, we are unlikely to invest much additional attention or cognitive effort. In line with this, the effect is stronger if people are tired, distracted or not particularly motivated to form an accurate judgement. Also, attention naturally decreases over time, either due to boredom or because we assume that the other followed the norm to put most important aspects first, which, however, only seems to play a minor role, if any (9).

Accounts that focus on motivational processes claim that we basically love to see our ideas about the world confirmed, not only for reasons of cognitive ease, but also in order to maintain or boost our self-esteem (see article #1 of this series). Early information allows us to form an expectancy, which then serves as an anchor against which to compare and integrate whatever information comes in later. Once we have such a preconceived notion, we tend to re-interpret, downgrade, or even ignore information that is not consistent with it. We do not like to admit having been wrong and “kill our darlings” — and preconceptions turn into darlings fast. If necessary, we may also label the source of inconsistent info as not credible or outdated (3, 4, 8), or come up with rather complex explanations about why that “exception” occurred (14). All these processes allow us to maintain our internal picture of the world.

In line with this reasoning, it has been shown that people with a strong need for cognitive closure exhibit stronger primacy effects, as these fulfil their urge for quick, simple and clear answers over ambiguity and confusion (15; see article 8 on self-efficacy and tolerance of ambiguity). With regard to dealing with inconsistency, preconceptions based on first impressions work in similar ways as preconceptions based on other forms of “prior knowledge” such as stereotypes and prejudice about members of particular social groups (16; see article 9 on stereotypes and prejudices).

From a certain angle, one could argue that the way different aspects are presented to media audiences in a feature is rather similar to what experimenters did when they presented their participants with several pieces of information about a politician, an event or a controversial issue. Thus, whatever is chosen as an opener is likely to more strongly affect their impression and judgement of the issue at hand. However, order effects not only work on the audience, but also on the journalists themselves, as we will see in the next section.

Practical impact of order effects in journalistic research and production

Considering the overall picture of findings, it seems fair to assume that the impression journalists form on their topics, sources and protagonists is likewise affected by and will in turn affect:

• the order in which they receive, research and integrate facts and data;
• the order in which they visit relevant places or talk to different stakeholders with different perspectives;
• the interview arc, i.e. the order in which they address different aspects of an issue within an interview; and
• the tone and framing of the story, both language-wise (see article 3 on the power of language) and visually (see article 5 on the power of images), how it is edited and produced in the end.

Order effects in journalistic research: Seek and you shall find

Working on a feature sometimes starts incidentally: You come across an interesting new fact, you are introduced to someone who is affected by an acute problem, you learn about controversial viewpoints on an issue that you never considered before, and sense a story. The fact that this was your initial exposure to the issue is very likely to affect each further step of your research: The keywords you use to explore the topic or dig deeper, and the statistical materials, relevant stakeholders, experts and case stories they yield (e.g. when entering “mall + threat + local business” or “mall + outskirts + ghost town”, rather than “mall + social infrastructure + family spaces” or “mall + economy + employment opportunities”).

As a consequence, you will be more likely to first talk to persons mentioned or quoted in these sources, or to visit “relevant” communities to talk to their inhabitants — “relevant” usually meaning confirmatory for the point to be made. Even when you deliberately search for different views, talk to other experts “from the other side” or go to different places later, the anchor is likely to be set, and the more consistent materials you find, the more will they set the tone of your story.

Of course, order is inevitable: One fact, onsite-trip or viewpoint will always be first, and you need to start your research somewhere. However, being aware of order effects and understanding them better will help you to consider and counterbalance their potential influence on your workflow, which starts with preliminary background reading and selecting potential interviewees and sites to portrait. Everything that works to counter confirmation bias (see article no. 1 of this series) also works here, and additional ideas to prevent bias are to be found in the Tools and Tips section below.

Order effects in designing interview questions

Once you have chosen with whom you will talk, be it in field research or as guests in your show, order concerns continue: Who do you meet (or host) first, what do you mention first when introducing them, again setting anchors and frames for whatever comes up later? And last but not least, how do you arrange your set of questions to create a good interview arc?

Fortunately, survey researchers found themselves facing similar challenges as journalists, and wondered if order arrangement (and wording) of questions could alter response behaviour. In a number of controlled experiments, they revealed that the answers to a question may differ considerably depending on the position of that question within a set, i.e. on the order, even if the wording of the question was perfectly identical (for a good overview of this line of research, see (17) or (18). The effect is not so much determined by the position of a question per se but by the content of previously asked questions.

To illustrate this, let’s take a look at a now classic series studies from the era of the Cold War (18): When first asked in a survey whether they think US-American reporters should be allowed in communist countries, a large majority (70-90%) of the respondents agreed, while only about half as many approved of allowing reporters from communist countries in the US. However, when they had been asked about the respective other party before, the numbers converged, that is, considerably more respondents agreed to — also — let communist journalists report from the US, and considerably less respondents claimed US journalists must be allowed to report from communist countries. It seems that people reassessed their second answer for fairness after having considered “the other side”. As such, this does not come as a surprise, yet agreement rates to these kinds of questions are no longer comparable when embedded within other questions on related issues. In fact, comparing responses to the same question asked in a different survey or interview contexts could, for instance, lead to false conclusions about changes in public opinion (19).

As already mentioned above, primacy effects may in part be driven by a need for consistency of one’s thoughts, attitudes and behaviours. When asked a series of questions in interview or talk show settings, people seek to present coherent (rather than very nuanced and possibly at first glance inconsistent) views on a topic, and will therefore adjust their answers to align with their previous statements. A “yes” to a previous question is likely to lead to agreement to similar follow-up questions as compared to if these later questions were asked alone. People differ in their need for consistency, as well as in their proneness to acquiescence, with some having a greater tendency to agree with whatever is suggested in a question (19).

Digging deeper, survey research on order effects has shown that it also matters how general and more specific questions on the same issue are arranged: For instance, asking people first how happy they are with their life as a whole and then how happy they are with their intimate relationship yields rather strong correlations between both answers. When first asked how happy they are with their relationship, and later how happy they are with life as a whole, the correlation between both answers is considerably lower. The authors of this study concluded that in the latter case, “life as a whole” was reinterpreted as “life apart from marriage”, since this special aspect had already been covered (20).

Taken together, questions obviously can change their meaning and alter the response given — an effect that is very likely to similarly occur in journalistic interviews and talk show formats. Again, order cannot be omitted, but knowledge about such effects can help to craft more balanced and nuanced interview arcs.

Order in storytelling

Storytelling choreography is probably the aspect journalists are already most aware of, in terms of how much order matters. It is taught in journalism classes and on the job training that the opening as well as the closing section of a story must be crafted with great care: Headline, teaser and first paragraph (or first seconds) determine whether the audience stays with you; the closing statement will ideally leave them with an interesting twist and/or a smile that stays in their memory. We have seen that this playbook is well rooted in psychological research. 

Before we get to the Tools and Tips, we would therefore like to draw your attention to potentially less obvious effects, such as how the levels of operation interact and how order effects from prior stages — e.g. research and interviewing — might affect your final storytelling even without you being aware of it. It seems likely that once a primacy effect has emerged on the stage of receiving information, it will procreate in dynamic ways, affecting — as sketched above — not only keywords and other research decisions, such as choice of interview partners, but also the selection and arrangement of interview questions; and that these cumulatively affect, in turn, how you decide to layout and present your story or feature.

 

The relative weight of earlier as compared to later information will affect its perceived relevance, rendering it more likely that it will also get a pole position within the storyline; the overall impression you formed based on those early inputs will have affected the consistency and richness of extra information you researched for, such as statistics, facts or expert voices, rendering it more likely that the theme will make it into the final version and earn a prominent place, such as in a pull quote, a photo, or a graphic, and so on. Thus, a small difference in the beginning (or at any later stage during the process of collecting and selecting materials) can set a switch and subtly or completely change the direction a story is taking.

Once more: Order is inevitable. You obviously also have to present some fact or voice first, some later, some last. However, reflecting on how these decisions might be affected by order effects you were subject to earlier in the process might help to not get drawn into a chain of primacy effects yourself, and also to present your audiences a more balanced basis for them to form impressions and judgements of their own.

Think of one of the stories you are currently working on: What difference might it have made if you learned first what you just learned most recently? What questions would you then have asked, where and how would you have researched for facts and data, whom would you have talked to with priority? Try to reflect how your entire strategy was affected by what you received or whom you were introduced to first. Which parts do you think about cutting out, what conclusions become more prominent once you focus on aspects you learned about later in the process?

During research:

• Apply this idea to places: What difference might it have made if you had visited the place first that you went last? Did your first impression, resulting from the “feel” of a place, change your approach to the places you visited later?

• Apply it to people: What difference might it have made if you had met the person or group first that you met later or last? Did your first impression of that group of stakeholders alter your approach to the people you interviewed later, and if so, how?

When planning and running interviews:

• Consider the order of questions: How do you introduce your topic, what do you imply by the order of questions? Do they cover aspects on the same level of abstraction or are they zooming in or out, inviting for reinterpretation of what they refer to?

• If you host a TV or radio show with more than one guest, whom do you present and address first, what seems a fair order, and how can you counterbalance order effects in your audiences’ impression formation?

In storytelling:

• Try to turn the order of accounts around and try to feel the impact that difference might make for the audience.
• Briefly pitch the two outlines to a co-worker and ask them for their overall impression. Ideally, have several co-workers read both versions in different order (first A, then B, or vice versa), because there can be primacy effects here too.

In editing:

• Challenge each other in your team by turning a story upside down or by approaching a topic you are collaborating on from different directions to yield a more nuanced overall picture.
• Apply it to the newsroom: How do first reports eventually alter the interpretation or impact of later news on a related topic? Does a different arrangement affect the meaning of a notice? This may, for instance, be the case when industry and environmental news are presented in direct vicinity.

In the long-term:

• When you are covering a story that has regular updates, don't forget to consider, from time to time, whether the context you're using from the beginning (based on what you knew then) still makes sense now that you have more information. You might need to reorganise/update the narrative.
• Reflect on when you realised — in your professional or in your personal life — that the first impression you had formed of a person or event was at least in part wrong, unfair, or seriously incomplete. What did you first “try” in order to maintain it, and what did it finally take to disconfirm, overcome and change your initial judgement?

What other ideas do you have to use your knowledge about order effects in your future work in constructive ways?

Margarida Alpuim is a Portuguese psychologist and journalist. She completed her master's in Community Psychology at the University of Miami, where she focused on issues of collective well-being. As a journalist, Margarida wants to explore more constructive ways to tell stories that consider both the audience and media professionals. Margarida currently works from Lisbon on innovative projects that unite psychology and journalism.

Katja Ehrenberg holds a PhD in psychology and is professor at Fresenius University of Applied Sciences in Cologne. For almost 25 years, she has conducted research, taught and published on topics of social, communication, organisational and health psychology. As a freelance systemic consultant, she advises teams and individuals and enjoys applying social science findings to the various challenges of everyday human (work)life.