The following is taken from a lecture given by Luuk Sengers. He discusses the techniques he uses for gathering and managing research material, and looks at how to build a good hypothesis.
The first technique is not a computer tool, but a tool for the mind. The computer lacks imagination and life experience, the two things needed to build a good hypothesis. Consider your investigation a project. A project you can manage by setting targets and ensuring that the targets are met.
Here’s an example. A study by the Danish Technical University in Copenhagen revealed that the plastic pipes used for the transport of drinking water contained potentially hazardous chemicals that seep into the water. People with plastic water pipes in their homes could be drinking, cooking with, and showering in, contaminated water.
Kaare Gotfredsen, Marleen Teugels and I met at the Global Investigative Journalism Conference (in Copenhagen) and decided to look into the plastic pipes story together. Kaare had already written about it in the Danish media and now we wanted to have a closer look.
Working to a hypothesis proved to be a fantastic tool. It was provable. And it kept our eyes on the ball for nearly a year. Once we’d agreed on the hypothesis we found it easier to do the research together.
the scene of the crime
How did we go about it? First, lets look at how other professionals who investigate – police detectives and scientists – approach their work.
The first thing that the police will do at a crime scene is to take in as much detail as possible. They gather all the evidence they can as quickly as possible. Then they rush back to their offices and start playing the whodunit game. They bring their previous experience to help read the evidence and start building hypotheses. They ask the question: “what could have happened?” They try to draw a picture of the events that led up to the crime.
Their next step is to interview people. And by having another look at the objects and evidence found at the scene. They check every new piece of information against their hypothesis. If the information doesn’t fit the hypothesis, the hypothesis is changed – it is constantly being challenged.
not just reporting the news
Since we cannot use a computer to do this work for us, you need to put your mind into a certain state.
First, you need to use your imagination and your life experience to draw a mental picture of the situation you want to investigate.
Normally journalists react to situations. They report what they see or hear or read. A lot of reporting is done as a follow-up of yesterday’s news.
The investigative reporter has a harder task – trying to reveal something not yet in the news – but instead to make the news. So he has to make a leap into the near future – an uncertain future. How can he possibly report facts that are not in the open yet? How do you search if you do not know what to search for?
The answer is to try to picture the future story and work from that.
Second, you need to be very precise; to examine carefully the words you use. If you use, for example, the word ‘house’ in your hypothesis, ask “is it a villa, or a penthouse?” “Is it a crofter’s house or an apartment?”
And third, you need to be objective. You need to leave out words that express feelings or opinions. If we need information, we need to tread carefully and present ourselves as unbiased.
Sure now you think: I can do that! And I’m confident that you can.
But picture yourself, for a second, at your workplace. Is there a corner where you can sit down quietly and think your hypothesis through for, say, an hour without being disturbed? Will your boss allow you to stay at home one morning to contemplate your plan of attack? Or let me ask you: what is your first impulse when you start researching for a story? Is it not calling people, maybe contacts you have spoken to before, to start asking questions? I am of course not saying that you should not talk to people. What I’m saying is that there are advantages to asking yourself questions before you ask them of anyone else.
Take time to consider your plan of attack and you will save a lot of time later in the investigation. You should ask yourself a set of very important questions about your investigation that one day will be asked anyway; by your editor, before your story is printed or aired or by your readers or viewers afterwards.
If you recognise that that moment of truth is part and parcel of our job, then why not prepare for it? Why not ask the fearsome questions at a stage when you can still change your direction – at the very beginning of your investigation?
look for the conflict
The kind of hypothesis I have in mind comes down to one very short and simple formula.
What is journalism mostly about? What lies at the very core of most of our stories? Conflict.
We reveal unexpected trends, unsafe situations, pollution, fraud, inefficiency, abuse of power, lies, discrimination, unfairness etc. We reveal situations that people don’t approve of because they do not comply with democracy or human rights or taste.
One of my teachers taught me that there are two types of investigative stories:
• rules which are inadequate
• rules which are being neglected
In both cases there is conflict; a breach of confidence, an attack on the status quo, a shock to our beliefs.
So the first step towards a hypothesis is to find a suitable conflict.
don’t sell a topic, sell a story
Don’t tell me the topic you have in mind, but explain to me what is wrong with that topic. Where is the conflict?
Compare theory with practice; plans, rules and procedures with reality, and you will see that the conflicts appears immediately:
• compare costs with revenues
• compare today with yesterday
• compare one place with the other
• compare private needs with public interest
The formula for the hypothesis is simply ‘A does B’.
The A in this formula relates to a person or object. Someone or something that is capable of setting things in motion. The B refers to an action.
It is that simple. In fact, what makes building hypotheses so tiresome is the difficulty of writing simple sentences. To quote one of my idols, the American investigative journalist Eric Schlosser (author of ‘Fast Food Nation‘): “Writing something that is simple and true is the hardest thing.”
Take this example:
Lisa kicks around furiously. (This is a conflict for sure.)
The simplest sentences have only a subject and a verb.
But that is only part of the reason why we choose such a simple form for our hypothesis-formula. The other reason is provability.
What we need, as investigative reporters, is stuff we can prove. Take a closer look at our formula:
The person (Lisa) and the action (kicking) both leave marks. The person has a (often written) history. And the action also leaves a trace (possibly a broken object, marks in the sand, bruises on someone’s body, or a memory in the head of a witness).
Let’s stick to the simple A does B formula, but expand it a little bit.
Let’s introduce an object: “Lisa kicks around furiously with a ball.”
And add a scene: “Lisa kicks around furiously with a ball in the playground.”
And then a victim: “Lisa kicks around furiously with a ball in the playground. As a result Harry gets hurt in his stomach.”
And then a motive: “Lisa kicks around furiously with a ball in the playground. As a result Harry gets hurt in his stomach. She did so because her classmates ignored her.”
Every new item is provable, because it leaves traces: the ball, the playground, the ignoring by the classmates, the wounds on Harry.
So we managed to formulate an event as plain and provable facts.
And spot the similarity with police work: The police also look for plain evidence. They talk to witnesses, study objects and the crime scene and look into the lives of the victim and the (possible) offender. All add up to a reconstruction of the events.
Now you can see why emotions, feelings and verdicts don’t work in hypotheses: they are not provable.
If I stated: “Lisa misbehaved herself” I would not only be non-specific but also add my own subjective opinion.
A similar problem occurs when we use words that state a relative amount, like: too big or not enough.
As in: “Lisa’s mum is not a good mother, as she is away from home too much.”
Who says so? How much is too much? In these cases it is better to compare the data or situations that you’ve found with other data or situations and leave the verdict to the reader/viewer.
For example: “Lisa’s mum worked longer hours than her colleagues.”
This is provable – and the verdict is more subtle, maybe Lisa’s mum has to work so hard, because she has a useless husband that hangs around in the pub all day.
a good hypothesis
Now let’s go back to our drinking water story. The first hypothesis that springs to mind is that plastic pipes are dangerous.
But is this a good hypothesis?
No, it is not. It is not specific enough. A student of mine reacted: “A plastic water pipe is dangerous, because if you hit someone on the head with it, it hurts.”
We tried this: “plastic water pipes leak dangerous chemicals into drinking water.”
But: ‘dangerous’ for whom?
“Plastic water pipes that leak chemicals into drinking water cause health problems in humans.”
This is much better. It shows us what and where to investigate.
But now we also discover two important aspects of a hypothesis:
It puts a fence around your investigation which is exactly what we need. In our later research we discovered that creatures that live in or on water, particularly birds, are also dramatically affected by the chemicals (drinking water goes through the sewer and ends up in open waters, like rivers and lakes). But we decided to leave that aspect out of the main story, to avoid confusion.
Although you mark out your field of investigation you will find that it’s not always easy to make your hypothesis as concrete as you would wish, at least not at the beginning of your investigation, because you lack enough information in this phase to add the necessary details to your hypothesis.
For example: What type of plastic pipes?
And: What kind of health problems?
Just work with the material you’ve got and adapt your hypothesis as you find new details.
“Water pipes of polyethylene leak hormone disrupting chemicals into the drinking water, which causes sex changes in human foetuses and hormonal cancers in adults.”
This was the outcome of six-months of research. The connection between the chemicals and hormone disruption in humans was especially hard to establish.
Let me take you one step further in the art of hypothesis building:
When your hypothesis grows because you add objects, location, victims and motives, you may feel the need to divide it into smaller steps.
This is what we did to our water pipes hypothesis:
The pipe manufacturers put the chemicals in their products: they choose the dangerous chemicals above safer alternatives because they are cheaper.
The polyethylene water pipes leak chemicals into the drinking water.
The chemicals get into the drinking water undetected: the government uses a detection technique that cannot spot the dangerous chemicals.
The chemicals disrupt the human hormone system. Some people with a disrupted hormone system become infertile and/or develop cancer.
What we did was break the hypothesis into small, chronological steps.
We asked ourselves: What happened first and what after that, and then what?
At every step the hypothesis is formulated as: A does B
Since they are part of our research hypothesis, they need to be provable, hence the introduction, in every sentence, of a subject and a verb.
Look at the third partial hypothesis:
The government uses a detection technique that cannot spot the dangerous chemicals.
The formulation looks a bit long-winded. Why not just say:
The government does not detect the dangerous chemicals?
The answer is almost childish but important: because you cannot point to something that is not happening.
In our effort to be as concrete as possible, we have to find a formula that shows proof. In this case, the detecting techniques are the proof. They are obviously not up to the task and we can show that.
Try not to formulate a hypothesis in a negative or denying way. Look for another formula that says the same thing, but in a more provable way.
Look at the last partial hypothesis:
“The chemicals disrupt the human hormone system. Some people with a disrupted hormone system become infertile and/or develop cancer.”
It states that ‘some people, as a result, become infertile and/or develop cancer’. This is another well thought through statement. Scientists would likely say (and indeed they did in interviews with us) that people with a disrupted hormone system have an increased chance of becoming infertile and/or developing cancer.
The problem is that we all have chances and in the end we’ll all die of something. News reports like this ‘drinking water from your tap increases the chances of becoming ill’ cause readers to shrug. ‘Life itself is deadly’, is an often heard reaction to these panic reports. You need to make the danger as real to the reader as possible. Therefore we kept the word ‘chance’ out of our stories. Instead we state that ‘some people’ who drink tap water from polyethylene pipes will develop cancer and/or will never have babies. We pick a virtual group of people who for certain will be affected, and not by chance. And the reader thinks: ‘Maybe I am in this unlucky group.’
Another tip concerning hypotheses is to start your investigation with the partial hypothesis that is easiest to prove.
“The polyethylene water pipes leak the chemicals into the drinking water.”
If this partial hypothesis proves to be wrong, you haven’t lost a lot of time by following the wrong hypothesis.
The partial hypothesis shows you the way to potential sources, when you ask yourself: ‘What do I need to prove them?’
And if you can draw up a list of potential sources, you are probably also able to make a good estimate of the time it will take you to get to these sources, allowing you to create a timetable and to plan – something to focus on and something to convince your editor with.
telling the story
Finally, although the hypothesis works great as a planning devise, the ultimate goal is a good story. Let’s go back to the list of partial hypotheses.
“The pipe manufacturers put the chemicals in their products: they choose the dangerous chemicals above safer alternatives because they are cheaper.”
“The polyethylene water pipes leak the chemicals into the drinking water.”
“The chemicals get into the drinking water undetected: the government uses a detection technique that can not spot the dangerous chemicals.”
“The chemicals disrupt the human hormone system. Some people with a disrupted hormone system become infertile and/or develop cancer.”
There is a potential strong story structure in this list. It is chronologically ordered, but we want it ordered like a story. Than we come up with something along these lines:
1 – Polyethylene water pipes leak chemicals into the drinking water = the shocking news.
2 – The chemicals disrupt the human hormone system. Some humans with a disrupted hormone system become infertile and/or develop cancer = the result of the problem.
3 – The pipe manufacturers choose the dangerous chemicals above safer alternatives because they are cheaper = one of the causes of the problem.
4 – The government uses a detection technique that can not spot the dangerous chemicals = the other cause of the problem.
Don’t you agree that this is a great way to tell the story?