How Not to Mislead with Your Data-Driven Story

is in all places. There are numerous books, articles, tutorials, and movies, a few of which I’ve written or created.

In my expertise, most of those assets are likely to current information storytelling in an overwhelmingly constructive gentle. However recently, one concern has been on my thoughts:

What if our tales, as an alternative of clarifying, mislead?

The picture above reveals one of many house buildings in my neighborhood. Now, check out the photograph on the left and picture one of many residences within the white constructing is up on the market. You’re contemplating shopping for it. You’d possible deal with the fast environment, particularly as offered within the vendor’s images. Discover something uncommon? In all probability not, not less than not immediately.

Ought to the fast setting be a dealbreaker? For my part, not essentially. It’s not probably the most picturesque or charming spot—only a typical block in a median neighborhood in Warsaw. Or is it?

Let’s take a brief stroll round to the again of the constructing. And… shock: there’s a public bathroom proper there. Nonetheless be ok with the situation? Perhaps sure, perhaps no. One factor is obvious: you’ll wish to know {that a} public bathroom sits slightly below your future balcony.

Moreover, the house is situated within the decrease a part of the constructing, whereas the remainder of the towers rise above it. That is one other issue which may be vital. Each these “points” for positive will be introduced up in worth negotiations.

This easy instance illustrates how simply tales (on this case, utilizing images) will be misinterpreted. From one angle, the whole lot appears effective, even inviting. Take a number of steps to the suitable, and… whoops.

The identical scenario can occur in our “skilled” lives. What if audiences, satisfied they’re making knowledgeable, data-backed selections, are being subtly steered within the incorrect path—not by false information, however by the way in which it’s offered?

This put up builds on an article I wrote in 2024 about deceptive visualizations [1]. Right here, I wish to take a bit broader perspective, exploring how the construction and movement of a narrative itself can unintentionally (or intentionally) lead folks to incorrect conclusions, and the way we will keep away from that.

Information storytelling is subjective

We frequently wish to consider that “information speaks for itself.” However in actuality, it hardly ever does. Each chart, dashboard, or headline constructed round a dataset is formed by human decisions:

• what to incorporate,
• what to go away out,
• tips on how to body the message?

This highlights a core problem of data-driven storytelling: it’s inherently subjective. That subjectivity comes from the discretion now we have in proving the purpose we wish to make:

• selecting which information to current,
• deciding on applicable evaluation approach,
• deciding on arguments to stress,
• and even what to to make use of.

Subjectivity additionally lies in interpretation — each ours and our viewers’s — and of their willingness to behave on the knowledge. This opens the door to biases. If we’re not cautious, we will simply cross the road from subjectivity into unethical storytelling.

This text examines the hidden biases embedded in information storytelling and the way we will transition from manipulation to significant insights.

We want tales

Subjective or not, we’d like tales. Tales are important to us as a result of they assist make sense of the world. They carry our values, protect our historical past, and spark our creativeness. By way of tales, we join with others, be taught from previous experiences, and discover what it means to be human. Irrespective of your nationality, tradition, or faith, now we have all heard numerous tales which have formed us. Instructed us by our grandparents, dad and mom, academics, mates, and colleagues at work. Tales evoke emotion, encourage motion, and form our identification, each individually and collectively. In each tradition and throughout all ages, storytelling has been a robust technique of understanding life, sharing information, and constructing group.

However whereas tales can enlighten, they’ll additionally mislead. A compelling narrative has the ability to form notion, even when it distorts details or oversimplifies complicated points. Tales usually depend on emotion, selective element, and a transparent message, which may make them persuasive, but in addition dangerously reductive. When used carelessly or manipulatively, storytelling can reinforce biases, obscure fact, or drive selections based mostly extra on feeling than cause.

Within the subsequent a part of this text, I’ll discover the potential issues with tales — particularly in data-driven contexts — and the way their energy can unintentionally (or deliberately) misguide our understanding.

Narrative biases in data-driven storytelling

Bias 1. Information is way, far-off from interpretation

Right here’s an instance of a visible from a report titled “Kentucky Juvenile Justice Reform Evaluation: Assessing the Effects of SB 200 on Youth Dispositional Outcomes and Racial and Ethnic Disparities.”

The graph reveals that younger offenders in Kentucky are much less more likely to reoffend if, after their first offense, they’re routed by way of a diversion program. This program connects them with group assist, similar to social employees and therapists, to deal with deeper life challenges. That’s a robust narrative with real-world implications: it helps decreasing our reliance on an costly prison justice system, justifies elevated funding for non-profits, and factors towards significant methods to enhance lives.

However right here’s the issue: except you have already got sturdy information literacy and topic information, these conclusions are usually not instantly apparent from the graph. Whereas the report does make this level, it doesn’t accomplish that till practically 20 pages later. It is a traditional instance of how the construction of educational reporting can mute the story’s influence. It outcomes from the truth that information is offered visually in a single part and interpreted textually in several (and typically distant) sections of the doc.

Bias 2. The Story of the Lacking Map: Choice Bias

Selecting which information factors (cherries 😊) to incorporate (and which to disregard) is without doubt one of the strongest — and sometimes most neglected — acts of bias. And maybe no trade illustrated this higher than Large Tobacco.

The now-famous abstract of their authorized technique says all of it:

Sure, smoking causes lung most cancers, however not in individuals who sue us.

That quote completely captures the tone of tobacco litigation within the late twentieth century, the place firms confronted a wave of lawsuits from prospects affected by illnesses linked to smoking. Regardless of overwhelming medical and scientific consensus, tobacco corporations routinely deflected duty utilizing a sequence of arguments that, whereas typically legally strategic, have been scientifically absurd.

Listed here are 4 of probably the most egregious cherry-picking ways they utilized in court docket, based mostly on this text [2].

Cherry-pick tactic 1: use “exception fallacy” tactic in authorized or rhetorical contexts.

Sure, smoking causes most cancers — however not this one.

• The plaintiff had a uncommon type of most cancers, like bronchioloalveolar carcinoma (BAC) or mucoepidermoid carcinoma, which they claimed weren’t conclusively linked to smoking.
• In a single case, they argued the most cancers was from the thymus, not the lungs, regardless of overwhelming medical proof.

Cherry-pick tactic 2: Spotlight obscure exceptions or uncommon most cancers varieties to problem common epidemiological proof.

It wasn’t our model.

• “Positive, tobacco might have prompted the illness — however not our cigarettes.”
• In Ierardi v. Lorillard, the corporate argued that the plaintiff’s publicity to asbestos-laced cigarette filters (Micronite) occurred outdoors the slim 4-year window after they have been used, though 585 million packs have been bought throughout that point.

Cherry-pick tactic 3: Deal with model or product variation as a technique to shift blame.

In a number of instances, similar to Ierardi v. Lorillard and Lacy v. Lorillard, the protection admitted that cigarettes could cause most cancers however argued that the plaintiff:

• Didn’t use their model on the time of publicity,
• Or didn’t use the precise model of the product that was most harmful (e.g., Kent cigarettes with the asbestos-containing Micronite filter),
• Or didn’t use the precise model of the product that was most harmful (e.g., Kent cigarettes with the asbestos-containing Micronite filter),
• window years in the past, making it unlikely the plaintiff was uncovered.

This tactic shifts the narrative from

Our product prompted hurt.

to

Perhaps smoking prompted hurt—however not ours.

Cherry-pick tactic 4: Emphasize each different potential threat issue — no matter plausibility — to deflect from tobacco’s position.

There have been different threat components.

• In lots of lawsuits, firms pointed to various causes of sickness: asbestos, diesel fumes, alcohol, genetics, weight-reduction plan, weight problems, and even spicy meals.
• In Allgood v. RJ Reynolds, the protection blamed the plaintiff’s situation partly on his fondness for “Tex-Mex meals.”

Cherry-picking isn’t all the time apparent. It could conceal in authorized defenses, advertising copy, dashboards, and even tutorial stories. However when solely the info that serves the story will get advised, it stops being perception and begins turning into manipulation.

Bias 3: The Mirror within the Forest: How the Identical Information Tells Totally different Tales

How we phrase outcomes can skew interpretation. Ought to we are saying “Unemployment drops to 4.9%” or “Hundreds of thousands nonetheless jobless regardless of beneficial properties”? Each will be correct. The distinction lies in emotional framing.

In essence, framing is a strategic storytelling approach that may considerably influence how a narrative is obtained, understood, and remembered. By understanding the ability of framing, storytellers can craft narratives that resonate deeply with their viewers and obtain their desired objectives. I current some examples in Desk 1.

Bias 4: “The Dragon of Design: How Magnificence Beguiles the Reality”

Visuals simplify information, however they’ll additionally manipulate notion. In my older article [1], I listed 14 misleading visualization ways. Here’s a abstract of them.

1. Utilizing the incorrect chart kind: Selecting charts that confuse somewhat than make clear — like 3D pie charts or inappropriate comparisons — makes it tougher to see the story the info tells.
2. Including distracting components: Stuffing visuals with logos, decorations, darkish gridlines, or litter hides the vital insights behind noise and visible overload.
3. Overusing colours: Utilizing too many colours can distract from the main focus. With no clear coloration hierarchy, nothing stands out, and the viewer is overwhelmed.
4. Random information ordering: Scrambling classes or time sequence information obscures patterns and prevents clear comparisons.
5. Manipulating axis scales: Truncating the y-axis exaggerates variations. Extending it minimizes significant variation. Each distort notion.
6. Creating pattern illusions: Utilizing inconsistent time frames, selective information factors, or poorly spaced axes to make non-trends look vital.
7. Cherry-picking information: Solely exhibiting the elements of the info that assist your level, ignoring the complete story or contradicting proof.
8. Omitting visible cues: Eradicating labels, legends, gridlines, or axis scales to make information arduous to interpret, or arduous to problem.
9. Overloading charts: Packing an excessive amount of information into one chart will be distracting and complicated, particularly when vital information is buried in visible chaos.
10. Displaying solely cumulative values: Utilizing cumulative plots to indicate clean progress whereas hiding volatility or declines in particular person durations.
11. Utilizing 3D results: 3D charts skew notion and make comparisons tougher, usually resulting in deceptive details about measurement or proportion.
12. Making use of gradients and shading: Fancy textures or gradients shift focus and add visible weight to areas that may not deserve it.
13. Deceptive or obscure titles: A impartial or technical title can downplay the urgency of findings. A dramatic one can exaggerate a minor change.
14. Utilizing junk charts: Visually overdesigned, complicated, or overly creative charts which might be arduous to interpret and simple to misinterpret.

Bias 5: “The Story-Spinning Machine: However Who Holds the Thread?”

Fashionable instruments like Energy BI Copilot or Tableau Pulse are more and more producing summaries and “insights” in your behalf. To not point out crafting summaries, narratives, or complete shows ready by LLMs like ChatGPT or Gemini.

However right here’s the catch:
These instruments are skilled on patterns, not ethics.

AI can’t inform when it’s making a deceptive story. In case your immediate or dataset is biased, the output will possible be biased as effectively, and at a machine scale.

This raises a vital query: Are we utilizing AI to democratize perception, or to mass-produce narrative spin?

A latest BBC investigation discovered that main AI chatbots often distort or misrepresent present occasions, even when utilizing BBC articles as their supply. Over half of the examined responses contained vital points, together with outdated details, fabricated or altered quotes, and confusion between opinion and reporting. Examples ranged from incorrectly stating that Rishi Sunak was nonetheless the UK prime minister to omitting key authorized context in high-profile prison instances. BBC executives warned that these inaccuracies threaten public belief in information and urged AI firms to collaborate with publishers to enhance transparency and accountability.[3]

Feeling overwhelmed? You’ve solely seen the start. Information storytelling can fall prey to quite a few cognitive biases, every subtly distorting the narrative.

Take affirmation bias, the place the storyteller highlights solely information that helps their assumptions—proclaiming, “Our marketing campaign was a hit!”—whereas ignoring contradictory proof. Then there’s end result bias, which credit success to sound technique: “We launched the product and it thrived, so our strategy was excellent,”—even when luck performed a significant position.

Survivorship bias focuses solely on the winners—startups that scaled or campaigns that went viral—whereas neglecting the various that failed utilizing the identical strategies. Narrative bias oversimplifies complexity, shaping messy realities into tidy conclusions, similar to “Vaping is all the time safer,” with out enough context.

Anchoring bias causes folks to fixate on the primary quantity offered—like a 20% forecast—distorting how subsequent info is interpreted. Omission bias arises when vital information is not noted, for example, solely highlighting top-performing areas whereas ignoring underperforming ones.

Projection bias assumes that others interpret information the identical method the analyst does: “This dashboard speaks for itself,”—but it might not, particularly for stakeholders unfamiliar with the context. Scale bias misleads with disproportionate framing—“A 300% improve!” sounds spectacular till you be taught it went from only one to a few customers.

Lastly, causality bias attracts unfounded conclusions from correlations: “Customers stayed longer after we added popups—they need to love them!”—with out testing whether or not popups have been the precise trigger.

How you can “Unbias” Information Storytelling

Each information story is a alternative. In a world the place consideration spans are quick and AI writes sooner than people, these decisions are extra highly effective — and harmful — than ever.

As information scientists, analysts, and storytellers, we should strategy narrative decisions with the identical stage of rigor and thoughtfulness that we apply to statistical fashions. Crafting a narrative from information isn’t just about readability or engagement—it’s about duty. Each alternative we make in framing, emphasis, and interpretation shapes how others understand the reality. And on the finish of the day, probably the most harmful tales are usually not the false ones—they’re those that really feel like details.

On this a part of the article, I’ll share a number of sensible methods that can assist you strengthen your information storytelling. These concepts will deal with tips on how to be each compelling and credible—tips on how to craft narratives that have interaction your viewers with out oversimplifying or deceptive them. As a result of when achieved effectively, information storytelling doesn’t simply talk perception—it builds belief.

Technique 1: The Clever Wizard’s Rule: Ask, Don’t Enchant

On the planet of information and evaluation, probably the most insightful storytellers don’t announce their conclusions with dramatic aptitude—they lead with considerate questions. As an alternative of presenting daring declarations, they invite reflection by asking, “What do you see?” This strategy encourages others to find insights on their very own, fostering understanding somewhat than passive acceptance.

Think about a graph exhibiting a decline in check scores. A surface-level interpretation may instantly declare, “Our colleges are failing,” sparking concern or blame. However a extra cautious, analytical response could be, “What components might clarify this transformation? Might it’s a brand new testing format, modifications in pupil demographics, or one thing else?” Equally, when gross sales rise following the launch of a brand new function, it’s tempting to attribute the rise solely to the function. But a extra rigorous strategy would ask, “What different variables modified throughout this era?”

By main with questions, we create house for interpretation, dialogue, and deeper considering. This methodology guards towards false certainty and encourages a extra collaborative, considerate exploration of information. A powerful narrative ought to information the viewers, somewhat than forcing them towards a predetermined conclusion.

Technique 2: The Mirror of Many Truths: Provide Counter-Narratives

Good information storytelling doesn’t cease at a single interpretation. Advanced datasets usually enable for a number of legitimate views, and it’s the storyteller’s duty to acknowledge them. Presenting a counter-narrative—“right here’s one other method to take a look at this”—invitations vital considering and builds credibility.

For instance, a chart might present that coronary heart illness charges are declining general. That looks as if a hit. However a more in-depth look might reveal that the advance is concentrated in higher-income areas, whereas charges in rural or underserved communities stay excessive. Presenting each views—progress and disparity—offers a extra complete and trustworthy image of the difficulty.

By providing counter-narratives, we guard towards oversimplification and assist our viewers perceive the nuance behind the numbers.

Technique 3: The Curse of Crooked Charts: Keep away from Misleading Visuals

Visuals are highly effective, however that energy have to be used responsibly. Deceptive charts can distort notion by way of delicate tips, similar to truncated axes that exaggerate variations, unlabeled items that obscure the dimensions, or ornamental litter that distracts from the message. To keep away from these pitfalls, all the time clearly label axes, begin scales from zero when applicable, and select chart varieties that finest match the info, not simply their aesthetic attraction. Deception doesn’t all the time come from malice—typically it’s simply careless design. However both method, it erodes belief. A clear, trustworthy visible is way extra persuasive than a flashy one which hides the small print.

Take, for instance, the 2 charts proven in Picture 7. The one on the left is cluttered and arduous to interpret. Its title is obscure, the extreme use of coloration is distracting, and pointless components—like heavy borders, gridlines, and shading—solely add to the confusion. There aren’t any visible cues to information the viewer, leaving the viewers to guess what the creator is attempting to say.

In distinction, the chart on the suitable is way more practical. It strips away the noise, utilizing simply three colours: gray for context, blue to spotlight key info, and a clear white background. Most significantly, the title conveys the primary message, permitting the viewers to know the purpose at a look.

Technique 4: Converse Truthfully of Shadows: The Knowledge of Embracing Uncertainty

Uncertainty is an inherent a part of working with information, and acknowledging it doesn’t weaken your story—it strengthens your credibility. Transparency round uncertainty is a trademark of accountable information communication. Once you talk components like confidence intervals, margins of error, or the assumptions behind a mannequin, you’re not simply being technically correct—you’re demonstrating honesty and humility. It reveals that you just respect your viewers’s skill to interact with complexity, somewhat than oversimplifying to take care of a clear narrative.

Uncertainty can come up from numerous sources, together with restricted pattern sizes, noisy or incomplete information, altering circumstances, or the assumptions inherent in predictive fashions. As an alternative of ignoring or smoothing over these limitations, good storytellers deliver them to the forefront—visually and verbally. Doing so encourages vital considering and opens the door for dialogue. It additionally protects your work from misinterpretation, misuse, or overconfidence in outcomes. Briefly, by being open about what the info can’t inform us, we give extra weight to what it could. Under, I current a number of examples of how you can embrace info on uncertainty in your information story.

1. Replace on confidence intervals
   As an alternative of: “Income will develop by 15% subsequent quarter.”
   Use: “We mission a 15% development, with a 95% confidence interval of 12%–18%.”
2. Depart a margin of error.
   As an alternative of: “Buyer satisfaction is at 82%.”
   Use: “Buyer satisfaction is 82%, ±3% margin of error.”
3. Lacking information indicators
   Use visible cues, similar to pale bars, dashed traces, or shaded areas, on charts to point gaps.
   Add footnotes: “Information for Q2 is incomplete resulting from reporting delays.”
4. Mannequin assumptions
   Instance: “This forecast assumes no vital change in consumer conduct or market circumstances.”
5. A number of situations
   Current best-case, worst-case, and most-likely situations to replicate a variety of potential outcomes.
6. Probabilistic language
   As an alternative of: “It will occur.”
   Use: “There’s a 70% likelihood this end result happens beneath present circumstances.”
7. Information high quality notes
   Spotlight points like small pattern sizes or self-reported information:
   “Outcomes are based mostly on a survey of 100 respondents and should not replicate the broader inhabitants.”
8. Error bars on charts
   Visually present uncertainty by together with error bars or shaded confidence bands in graphs.
9. Transparency in limitations
   Instance: “This evaluation doesn’t account for seasonal variation or exterior financial components.”
10. Qualitative clarification
    Use captions or callouts in shows or dashboards:
    “Information developments are indicative, however additional validation is required.”

You may surprise, “However gained’t highlighting these uncertainties weaken my story or make me appear not sure of the outcomes?” Quite the opposite, acknowledging uncertainty doesn’t sign a insecurity; it reveals depth, professionalism, and integrity. It conveys to your viewers that you just perceive the complexity of the info and are usually not attempting to oversell a simplistic conclusion. Sharing what you do know, alongside what you don’t, creates a extra balanced and credible narrative. Persons are way more more likely to belief your insights after they see that you just’re being trustworthy concerning the limitations. It’s not about dampening your story—it’s about grounding it in actuality.

Technique 5: Reveal the Roots of the Story: Let Reality Journey with Its Sources

Each story wants roots, and on the earth of information storytelling, these roots are your sources. A gorgeous chart or placing quantity means little in case your viewers can’t see the place it got here from. Was it a randomized survey? Administrative information? Social media scraping? Identical to a traveler trusts a information who is aware of the trail, readers usually tend to belief your insights after they can hint them again to their origins. Transparency about information sources, assortment strategies, assumptions, and even limitations isn’t an indication of weak spot—it’s a mark of integrity. After we reveal the roots of the story, we give our story depth, credibility, and resilience. Knowledgeable selections can solely develop in well-tended soil.

Closing remarks

Information-driven storytelling is each an artwork and a duty. It provides us the ability to make info significant—but in addition the ability to mislead, even unintentionally. On this article, we’ve explored a forest of biases, design traps, and narrative temptations that may subtly form notion and warp the reality. Whether or not you’re an information scientist, communicator, or decision-maker, your tales carry weight—not only for what they present, however for a way they’re advised.

So allow us to inform tales that illuminate, not obscure. Allow us to lead with questions, not conclusions. Allow us to reveal uncertainty, not conceal behind false readability. And above all, allow us to anchor our insights in clear sources and humble interpretation. The aim isn’t perfection—it’s integrity. As a result of in a world stuffed with noise and narrative spin, probably the most highly effective story you may inform is one which’s each clear and trustworthy.

Ultimately, storytelling isn’t about controlling the message—it’s about incomes belief. And belief, as soon as misplaced, isn’t simply gained again. So select your tales fastidiously. Form them with care. And keep in mind: the reality might not all the time be flashy, nevertheless it all the time finds its technique to the sunshine.

And yet one more factor: in case you’ve ever noticed (or unintentionally created) a biased information story, share your expertise within the feedback. The extra we floor these narratives, the higher all of us get at telling information truths, not simply information tales.

References

[1] How not to Cheat with Data Visualizations, Michal Szudejko, In the direction of Information Science

[2] Tobacco manufacturers’ defence against plaintiffs’ claims of cancer causation: throwing mud at the wall and hoping some of it will stick, A number of Authors, Nationwide Library of Medication

[3] AI chatbots distort and mislead when asked about current affairs, BBC finds, Matthew Weaver

Disclaimer

This put up was initially written utilizing Microsoft Phrase, and the spelling and grammar have been checked with Grammarly. I reviewed and adjusted any modifications to make sure that my meant message was precisely mirrored. All different makes use of of AI (for example picture and pattern information era) have been disclosed straight within the textual content.