top of page

Doing Research When You Don't Know What You're Doing

Some research projects are a pretty straightforward journey. You have a clear question, you know how to start, and you can look for relevant patterns in your data right away.

Often, none of these apply.

Someone comes over, dumps a dataset on you, and says ‘find me the key themes in that’ before disappearing in a cloud of ambiguity. I’d argue this is actually a good way of doing research.[1] But it can be intimidating.

I’m going to explore the process of doing that kind of research project, considering both emotional and practical elements. I’ll do so through an extended analogy. Imagine you have just won an all-expenses-paid holiday to a foreign country. You’re keen to get as much out of it as possible. [2]

But one issue: you have never heard of the country, and neither have any of your friends. For the purposes of this exercise, let’s also imagine you like to have a good solid itinerary sorted out well before the holiday. (If you happen to know my dad, imagine you’re him).

So, how are you going to plan this holiday?

(Throughout the below I’ll use italics to break out of the analogy and discuss research projects more directly).

Stage 1: Scanning, and Panicking

You start by feeling excited and intrigued, but that’s quickly replaced by panic. It’s a whole country, full of all sorts of places and things. There’s maps and travel guides and TripAdviser, but they’re just overwhelming you with information.

But you have to start somewhere, so you scan the map. You see that the towns and cities are all concentrated on the east coast, lots of green space in the west, and some mountains up t’north. You also see where the airports and train lines are. But that’s about it.

That doesn’t seem to help the panic very much. Look at all those cities! Can you visit them all? Is it feasible to do both city and country? Are the cities even worth visiting? Is the green space full of beautiful rolling forests, or simply acres of sheep? [3]

Ultimately the central question – what are you going to do? – is still completely unanswered.

At this stage you have data, but no useful direction or detail. Everything you can see seems obvious and uninteresting – a bit like if you were reading the Bible and found lots of references to God, or noticed there are lots of pictures on Instagram.

You may feel a frustrating overwhelm, a directionless confusion. But it’s a necessary stage to get through whenever you meet a new load of data, and don’t (yet) have specific questions to guide you.

Stage 2: Collecting, and Feeling Overwhelmed

At this stage you do have some vaguely useful information, even if it doesn’t feel like it. The name of the capital, for example, or the places which would be easy/hard to move between. That helps with more Googling – you can now specifically search for information about the capital city, or that big forest out west.

After doing this for a while, you’ll probably be experiencing a slightly different form of panic. Before, you had just a vague, formless mass of information. Now you’re collecting more detailed pockets of information, but it isn’t clear how they fit together.

For example, you’ve discovered that the country has a form of cooking called ‘Yum Cuisine’. You’ve also learned that the cities are very different to the countryside. Both possibly useful, but does it mean you won’t find good Yum Cuisine in the cities?

So now you may feel like you are actually learning something – but it isn’t clear how you’re going to link up all the learning in a way that actually helps your decisions.

At this point if your boss came over and asked for some themes, you could list off some random things you’ve noticed. But this is the point at which vague questions like ‘look for key themes in this data’ start to get annoying – anything could be a ‘key theme’, and the boss (presumably) wants something a bit more insightful than a list of things you’ve spotted.

At this stage in my own PhD, I think I jotted down around thirty different possible themes. Some were sort-of linked (for instance ‘how people use evidence’ and ‘what websites do people link to?’) but others were really quite separate. It was a mess.

Stage 3: Selecting, and Calming

It’s decision time. You’ve only got a few days in the country, and you can’t do everything. You have to choose what to do and not do, and try to ignore the rising worry that you’re going down a wrong path.

The decisions ultimately have to come from what kind of holiday you want. You’re a real foodie, so you decide you want this holiday to be a pursuit of the best Yum Cuisine – trying it in both the countries and the cities. It would have been nice to visit the mountains, but it’s just not feasible.

This is where stages 1 and 2 start to pay off. You could have just taken a package holiday, and passed on all that decisionmaking. But that may not have satisfied your foodie needs. Or you could have jumped straight to Googling the best restaurants and booked into all of them. But without broader knowledge of the country, that could have resulted in you spending the whole holiday sitting on long train journeys between really boring cities. Or finding out that the country doesn’t even have good food.

You’ll still be doing research at this stage, googling the best restaurants and the ways Yum Cuisine varies between regions. But it’s a lot more directed, while also informed by your broader knowledge of the country. It may even be fun.

Lots of textbooks about research talk about themes ‘emerging’ as you look at data. I think that can be quite unhelpful – at some point you need to actively make decisions. Otherwise you just float around in information overload forever.

While your holiday decisions came from the kind of holiday you want, your research decisions at this stage should be guided by what you are trying to offer for an audience. If it’s an academic piece, you should probably focus on the routes which fill gaps in the existing literature. If it’s for a non-specialist audience, go for themes which can be interesting without needing lots of prior knowledge. And so on.

All this will help you construct some specific research questions. You could have come up with such questions right at the start of the project. [4] But after stages 1 and 2, you’re doing it with a better understanding of all the possible and sensible routes open to you.


You’ve landed in the country, the weather’s great, and you’ve got your plan of action. You know roughly what each day will look like, and you feel like you’re definitely making something out of your trip.

But even a foodie like you only spends a few hours a day eating, so you’ve got the rest of the time to fill. Thanks to your preliminary research, you know some good things to try. But you’ll also stumble on unexpected curiosities as you wander round. You may even develop an entire new theme to the holiday, and spend as much time in museums as you do restaurants.

The point here: even though you created a structured plan, that doesn’t completely restrict you. Indeed, having that pre-planned path from restaurant to restaurant can help you find unexpected things. At the very least it ensures that, even if you do go end up wandering aimlessly for an afternoon, you have something definite to return to.

You may even discover that you don’t like Yum Cuisine at all, and decide to re-think your plans entirely. In which case, it’s lucky you did that earlier research and know the good alternatives.

This stage of research is when you’ve set specific question(s) and are now applying them to data. You may gather new data based on your questions. Or you may just re-read the stuff you looked at in stages 1 and 2 – now you know what you’re actually looking for, you’ll be able to spot details and patterns which you didn’t earlier.

At this point the analogy with the holiday breaks down a bit. In particular, changing direction during research is much more tempting, and much riskier. I said about the holiday that “You may even discover that you really don’t like Yum Cuisine at all, and decide to re-think your plans entirely”. This is likely to happen during research, and should be treated with caution.

I have spoken to many PhD students who had a strong feeling, a couple of years into their research, that the whole thing was going in the wrong direction. They (we) usually found the original plan worked best – but only after wasting time doing stage 1 and 2 all over again. If you get this feeling, it’s definitely worth talking to supervisors or other senior researchers before doing anything drastic.

Though less risky, a researcher should also be careful of going on too many mini-detours. In a holiday, mini-detours to museums are good. In a research project, it’s likely to lead to a confusing final product for a reader. A good data set always offers potentially interesting detours. But often it’s best to just make a note and move on, saving them for a future project.

Conclusion: Try Something A Bit Different This Year

This post was about doing a research project where you’ve got information to look at, but no clear guidance on what exactly to do with it. You’re unlikely to emerge from such a research project with a tan (often, in fact, the reverse). But, somewhat like the end of a holiday, you should:

  • Be able to look at that mass of information you encountered at the start of stage 1 without feeling a sense of confusion and overwhelm. Instead, you can pick out patterns and details.

  • If someone asks for a brief summary, you can provide an overview of one key theme – such as ‘they have something called Yum Cuisine which was really good’, or ‘online science enthusiast groups like talking about things not people’ [5]. Perhaps dusted with one or two examples, for colour.

  • If they ask for more details, you have a slideshow at the ready.

  • You may even want to go back and explore some more.


[1] If a research project has a very clear question and direction from the start there’s less chance of finding something really surprising. It can mean the research reflects the interests of the researcher much more than whatever is actually interesting in the data. More worryingly, it can sometimes reflect a subconscious belief that you can probably guess what’s interesting anyway – a mindset that’s important to catch and criticise. Sometimes it’s a necessity, particularly if you have very short deadlines or a very specific task to accomplish. But I wouldn’t say that a more specific brief leads to a better research.

[2] It’s a little far-fetched, I’ll admit. But I did once have an unexpected holiday when I flew into Slovenia for a conference that was a) in Austria and b) started a day later than I’d thought. That was, incidentally, one of the best holidays I’ve had.

[3] Acres of sheep can, I’ll agree, contribute to a good holiday – hence the existence of Google Sheepview – but I’m assuming you want more than that. Though, y’know, each to their own.

[4] See footnote 1 for fuller argument.

[5] That sounds like some really great research doesn’t it. Wonder what awesome person did that research. Blog post coming soon, or click here if you fancy reading the whole PhD thesis.

29 views0 comments

Recent Posts

See All


bottom of page