Have AI evaluate your working hypothesis with this prompt (fill in your hypothesis on the fourth line):

# GOAL:

Help me formulate a newsworthy and investigable journalistic hypothesis.

# INPUT:

My working hypothesis is: [HYPOTHESIS]

# INSTRUCTIONS:

Evaluate my working hypothesis according to the principles under the “Method” heading.

Pay specific attention to the concreteness, scope, and investigability of the words used.

Use reliable, up-to-date internet sources to add relevant factual context. Always make an explicit distinction between:

- **Knowledge:** verifiable and sourceable facts.
- **Working assumption:** plausible, but not yet proven assumptions that need to be tested.

Do not secretly improve or rewrite my wording. Provide suggestions and briefly explain why they are more concrete, neutral, better defined, or easier to investigate.

# OUTPUT:

Present the analysis in this table:

| Element / word from my hypothesis | Strength | Weakness or investigative risk | Suggestion | Why the suggestion is more investigable |
| --- | --- | --- | --- | --- |
| [word or phrase] | [what is already usable] | [what is vague, normative, too broad, unproven, or hard to test] | [direction for sharpening, without implicitly rewriting my hypothesis] | [which documents, data, comparisons, or witnesses this enables] |

Evaluate in any case:

- the **actors**: who exactly is doing something?
- the **action**: what is happening concretely and demonstrably?
- the **effect or benefit**: for whom, how, and in what measurable way?
- the **location and time period**: what delimitation is needed?
- vague or judgmental words, such as "large", "many", "benefit/profit", "guilty", "fail", "insufficient", "unjust", or "corrupt".

## Knowledge and assumptions

- **Knowledge:** factual context that can be supported by reliable sources, with citations.
- **Assumptions:** assumptions that may be correct, but for which evidence is still required.

## Proposed hypothesis

Formulate one hypothesis in one or two sentences. It must contain a well-defined actor, a concrete demonstrable action, a relevant location and time period, and where possible a measurable comparison. Avoid judgments and unproven conclusions.

## Why this is investigable

Conclude with a maximum of three bullets.

**Do not show a timeline, a full research plan, or a fully developed story.**

# METHOD:

## How to create a hypothesis?
The first, absolutely crucial step at the beginning of an investigation is to express your hunch in a hypothesis.
If you have a hunch and want to investigate whether it is correct, try to formulate that hunch as concretely as possible. What do you think is going on? What do you hope to find? What do you want to expose? By starting your search with an answer to these questions, you make things easier for yourself and increase your chances of success.
The answer is a hypothesis. According to the dictionary, a hypothesis is a ‘supposition or proposed explanation made on the basis of limited evidence as a starting point for further investigation’ (Oxford Dictionary).
A hypothesis is therefore not a wild idea; it is a plausible explanation for a real phenomenon. You formulate your hunch in a powerful sentence and then ask yourself: what do I need to prove this? Which sources and which facts? By projecting the story, as it were, the path toward it becomes visible.

## FORMULATE A HYPOTHESIS
A hypothesis is your hope or expectation expressed in a single sentence. By formulating this sentence as concretely as possible on paper or on your screen, the sources and questions with which you can start the investigation reveal themselves naturally. For a workable hypothesis you need: a subject, a verb, and possibly nouns, all of which refer to things you can see, verify, or prove with sources, such as people, objects, and actions. The first task is to formulate such a concrete, investigable hypothesis about your investigation. You can do this in the following six steps:

### 1. Brainstorm about a main question. 
Every investigation begins with a question. First, make a list of questions (plural!) that come to mind in connection with the topic. In chapter 3 it was explained what you can do to arrive at original research questions (angles). Then choose one question from that list that you consider the most important: the main question. For example, the most important for your readers: what affects their interests the most? Or the most urgent in connection with current events: which question must be answered with priority? In any case, use the words ‘who’ and ‘what’ or ‘how’ in your main question. For example: ‘Who are the biggest methane emitters in the Netherlands and what are they doing to reduce their emissions?’

### 2. Answer this main question yourself. 
What do you think the answer is? What do you believe is probably going on? What do you hope to discover? Or what do others believe is happening? (You can also ‘borrow’ a hypothesis from someone else.) Your answer is a hypothesis.

### 3. Who are you going to investigate? 
The most effective way to make a hypothesis concrete is to begin the sentence with a person or organization. Do not ask yourself ‘what’ the problem is, but ‘who’.
This is not an obvious question; our initial reflex is often to explain what we want to investigate. ‘Who are you going to investigate’ lets you look at things in a different, more direct way. Journalistic stories, after all, are about people. Not about abstract things like problems, systems, or policies. It is human decisions that cause things to happen the way they do.
However, the question ‘Who is responsible?’ often cannot be answered unambiguously. Usually there are multiple parties that each carry a share of the responsibility. Whom should you choose? Pick the one who has a direct influence on the outcome, who is in a position to directly change the situation, who holds real power, who makes the key decisions.
Be as precise as you can at this early stage: do you only suspect which organization it is, or can you already picture which department within that organization is involved, or what the official title of the responsible person is? Or do you even suspect which individual it is, what their name is? For simplicity, we call the responsible party in the hypothesis ‘the perpetrator’.

### 4. What does the perpetrator concretely do? 
Describe the hypothesis as a concrete action: someone does something that causes a visible or noticeable change in the environment. A thought or feeling is not a concrete action because it has no effect on the environment. Why is it so important that something physically changes? Because you want to be able to demonstrate the hypothesis later. And you can only show something if you can see it yourself. So always ask yourself: what can I see of this?

Take the sentences below: which are concrete and which are not?

- She feels very unsafe
- She no longer dares to go outside
- She is miserable
- She has her groceries delivered to her home every day
- She had five locks installed on her door
- She receives intensive mental health care

Only the last three statements are visible in the person's surroundings and therefore concrete and demonstrable.
A human being can alter their environment through physical actions, language and communication, the use of tools or technology, the application of substances and materials, social behavior toward others, and interventions in nature; all these ways lead to visible or noticeable changes in the direct or wider world.
Think of things people can do with their bodies (pushing, pulling, building, making, demolishing, hitting…), with their language (giving instructions, protesting, writing), with tools (instruments, machines, software…), with substances and materials (cooking, shaping, moving, leaving behind…), with other people (collaborating, helping, harming, entertaining…), and with nature (sowing, planting, cutting down, damaging, protecting…)
The best scenario is when you can see the change yourself. But that is often not possible. In that case, you can describe what others see (provided their observation is reliable, of course). Furthermore, by ‘seeing’ we also mean ‘reading’, because (reliable) documents also count as concrete proof.

## STEP OVER YOUR BIASES

A hypothesis is therefore an expression of your expectations. These expectations are fed by your experiences, but also by your prejudices! Especially when you discuss your hypothesis with others, you may become aware of your own biases and eliminate them in time.
Scientists distinguish many types of cognitive biases (for an extensive list see: en.wikipedia.org/wiki/List_of_cognitive_biases). The four most important for us are:

- Confirmation bias: you ignore relevant and accurate information because it does not align with what you already think or know. You are not interested in counter-evidence. You can prevent this by consciously searching for other facts and perspectives, and by deliberately trying to disprove your hypothesis.
- Status quo bias: you do not want to hear about things that need to change. Instead, you justify the problems and act cynical about the chances of improvement. You can prevent this by asking yourself: what will the situation look like in six months if we do nothing?
- Overconfidence bias: you rely too much on your own judgment because you have been right before. You therefore take unnecessary risks and reckless decisions. Ask yourself whether the situation is truly the same as last time. And say out loud to yourself: ‘I can always make a mistake.’
- Just world bias: you assume the world is fundamentally just, that everyone gets what they deserve, and that an individual's fate is a consequence of their own behavior or character.

Think carefully about whether you might suffer from one of these biases in relation to your topic. Has this bias made you blind to other possibilities? Is the opposite of what you suspect perhaps also possible and plausible? How does your ‘sentiment’ relate to the facts you already know?

## AVOID THE PITFALLS

Check your hypothesis for the following pitfalls:

- Do not investigate things that do not exist / are absent. This is tempting because we often investigate issues where someone falls short. For example: ‘The government is not doing enough to accommodate refugees.’ But how do you prove that? Therefore, avoid words like ‘not’, ‘no’, or ‘never’. Instead, investigate what is happening (‘the government responds dismissively to complaints from refugees’), what happens elsewhere (‘the German government invests more in refugees than the Dutch government’), or what actors admit to doing (‘the minister confirms that there is an unofficial halt on new refugees’). Also try to avoid words like ‘much’, ‘too much’, ‘little’, and ‘too little’. Replace them with comparisons — ‘more than’, ‘less than’ — so they become measurable.
- Do not say what should happen. This is tempting because we often investigate matters that we believe need to change. For example: ‘The municipality must ban highly polluting cars from the city.’ But whether they will actually do so is something no one knows yet. Therefore, avoid words like ‘must’, ‘should’, and ‘will’. Instead, investigate what has happened so far (What indications are there that the municipality wants to tighten policy?) or what is happening elsewhere (Which municipalities are already banning dirty cars from the city?).
- Do not pass judgment. The temptation to let your personal opinion resonate in your hypothesis is strong, but how do you prove a judgment? ‘Top bankers are crooks.’ You may agree or disagree, but proving it is difficult. When is someone a ‘crook’? Instead, show what top bankers do that makes you personally feel they are crooks. In a concrete hypothesis, it is not about what someone *is*, but about what someone *does*.
- Assume the worst. Suppress the urge to tone down your hypothesis in advance. Leave that to your sources. Assume the worst-case scenario. Not because you are a sensationalist, but because experience has shown that sources find it easier to soften a claim than to harden it. Who is going to volunteer to tell a journalist that things are actually far worse? That is why the hypothesis about the drinking water pipes (see earlier in this chapter) states ‘kill people’ and not ‘make people sick’.
- Set not only an upper limit, but also a lower limit. What is the minimum you want to try to prove? It is good to reflect on this. In other words: what is your minimum hypothesis? When will you stop spending more time and money on this investigation? In the water pipes investigation, we set as a minimum hypothesis that people should at least feel unwell for a while from the phenols in the drinking water. We imagined that in that case we would have no chance of landing on the front page of the newspaper, but we could still write a decent piece for the science section. Such a lower limit also sends a reassuring signal to your editor or commissioner that you will not waste time on a dead-end project.

## KEEP AN OPEN MIND

Some journalists hesitate to work with hypotheses because they fear it compromises their objectivity. Hopefully, however, we have shown that a hypothesis is precisely an excellent tool for uncovering objective truth, provided you handle it wisely. A hypothesis is neither opinion nor speculation.
The greatest danger of working with hypotheses is tunnel vision. Anyone who desperately wants to prove something may be tempted to sweep inconvenient facts under the rug. Scientists have been caught doing this, as have criminal investigators (leading to innocent people spending years in prison).
Another danger is that you become so fixated on that one story that you miss signals of something much larger going on. In other words: you are chasing a rare butterfly with your net, and as a result, you fail to see the last Bengal tiger lurking in the tall grass nearby.
These two things help prevent tunnel vision, in addition to an ethical working method:

- Do not try to prove your hypothesis. Test it instead. Testing implies a more neutral attitude: it does not matter what you encounter. For many hypotheses, that is truly the case: whether the statement is true or not, in either case you have a story. The claim comes from somewhere; apparently there are people who believe that is how things are. So if you can demonstrate that reality is different from what is assumed, you also have news.
- Elicit criticism. Tell people you trust what you are working on. Let them respond to your proposition. Do so with an open mind, because you can only win: you will find clues for a good story, or for an even better story, or for a story no one — not even yourself — expected. Or you find out in time that you shouldn't waste time on the chosen hypothesis.
- Actively look for an alternative explanation. What else could be going on? Is that alternative explanation not much more plausible? Often, the simplest explanation is the most likely. It doesn't hurt to formulate a second, alternative hypothesis and keep it in reserve as a Plan B.

## EXAMPLES

Here are some examples of hypotheses that led to published investigations.

### Danger from the tap
The investigation into plastic water pipes, which opened this chapter, was conducted by Danish, Belgian, and Dutch journalists from the investigative collective Investigative Reporters Network Europe (IReNE). They met in Copenhagen in 2005 with a plan to further investigate the findings of chemical scientists. The first thing they did was formulate a hypothesis:

> Pipe manufacturers produce their goods using deadly chemicals that leak into drinking water.

The scientific paper that prompted the investigation into the toxic pipes already revealed key details. It stated that the leakage only occurred in polyethylene pipes, and not other plastics, and that the main group of chemical substances they had discovered were phenols. The hypothesis could therefore be sharpened to:

> Pipe manufacturers produce polyethylene pipes containing deadly phenols that leak into drinking water.

Subsequently, the journalists — Marleen Teugels and Chris Vermeire in Belgium, Kaare Gotfredsen in Denmark, and Luuk Sengers in the Netherlands — needed three months (full-time) to gather all the puzzle pieces and discover that their hypothesis… was spot on! In 2005, the investigation resulted in a series of stories across several countries, including ‘Danger from the tap’ by Luuk Sengers in *Algemeen Dagblad* and ‘Tap water can damage health’ in *Knack*. It was one of the first cross-border investigative stories in Europe, and it was picked up by Members of the European Parliament who used it in debates on new European legislation for hazardous chemicals.

### Murdered babies
An editor at *The American* heard someone say at a party: ‘Doctors kill prematurely born babies to prevent them from growing up severely disabled.’ The source was themselves a parent of a prematurely born, disabled child. The editor passed the sensational idea to Mark Lee Hunter, who worked for him as a freelancer. Mark immediately had reservations: how do you kill babies in a hyper-controlled environment like a hospital without anyone noticing? Nevertheless, he began with the hypothesis:

> Doctors kill prematurely born babies.

The first element of this hypothesis that he examined more closely — because it was easy to verify — was the number of extremely premature infants alive in the US. That number, it turned out, was not decreasing, but rising! Exactly the reverse of his hypothesis appeared to be true.
Further reporting revealed: between 1982 and 1984, the conservative Reagan administration had pushed a law through Congress that stripped away a traditional family right: the right to decide whether a severely disabled newborn should be allowed to die in peace. But subsequently, the government washed its hands of the children: in a 1996 welfare reform, all federal assistance to survivors was eliminated. In other words, the US government had created a population of a quarter of a million disabled children and then turned its back on them.
Although the initial hypothesis turned out not to be true, it put Mark on the track of another major story. A story that won national awards in the US, though unfortunately did not lead to policy reform. However, the article remained a comfort and a reference point for parents confronted with the results of this unprecedented national experiment.
‘The Miracle and the Shame: How an Unknown Law Made a New World of Pain’ can be read at: markleehunter.free.fr/babydoe.html.

### Blind girls
A student at a Parisian university arrives in Mark Lee Hunter's class on a Tuesday afternoon with a bandage over one eye. That afternoon, the students have an assignment to brainstorm investigative story ideas. The student with the bandage doesn't need to think long: she wants to know how many young women her age have gone blind after wearing cosmetic contact lenses.
She tells the class that over the weekend she almost lost her eyesight. On Saturday she had gone out and, as part of her makeup, like so many of her friends, changed her eye color with contact lenses. After partying, she fell asleep exhausted without removing the lenses. On Sunday morning she woke up with sharp pain in one eye. It took her all day Sunday in Paris to find a doctor. By the end of the day, by now in a panic, she was finally examined by an ophthalmologist, who told her: ‘If you had arrived here just a bit later, I might not have been able to save your eye!’
The student's initial hypothesis was:

> Dozens of girls per year in France go blind after wearing cosmetic lenses.

This was quickly verifiable: the national ophthalmological association informed her that in France the previous year, seventy girls had suffered partial or total blindness as a result of wearing such contact lenses. She then shifted her investigative focus to the root causes:

> Lens manufacturers choose not to warn about the risk of blindness in their package inserts and advertisements.

### Polluting livestock barns
In 2017, Greenpeace Europe did something it had never done before: it called in the help of journalists. With a reform of the European Common Agricultural Policy on the horizon, the environmental organization wanted to know how much European agricultural subsidies contribute to rural pollution. Farmers receive funding from the EU based on the acreage of their land; how they use that land is hardly scrutinized.
At that time, we were already in contact with colleagues across Europe regarding agricultural subsidies (the renowned Dataharvest conferences (dataharvest.eu) in Mechelen grew out of farm subsidy hackathons organized since 2009) and with Greenpeace, where we provided training. That is how the connection was made. Mark Lee Hunter brought together seven colleagues in as many member states to conduct data analyses and interviews. Greenpeace funded the effort, provided expertise and sources, but did not interfere with the editorial content of our stories. The hypothesis we started with was:

> Livestock farmers in Europe use European agricultural subsidies to build larger industrial barns that pollute air and water through manure.

That assumption later proved accurate and was substantiated with data. The collaboration between Greenpeace and the journalists proved fruitful for both parties, not least because the advocacy group adhered to the same ethical, legal, and professional standards for the investigation as we did.