Four key challenges of a great research project

Your research project is about to begin. You’ve checked out a bunch of resources and listened to a bunch of advice. You know all the tricks and techniques for knowledge gathering necessary to make this project great. What you might not have thought about are the challenges you’re likely to face.

From unquestioned assumptions to imperfect data, research projects can throw a variety of challenges up in front of an intrepid researcher. Knowing which ones you’re likely to encounter and how to deal with them can transform the outcome of your project.

Identifying your assumptions

Even the most impartial researcher will have certain assumptions about how they think their project might work out. For instance, you might think that gathering data from demographic a will probably reveal trends x and y. Letting assumptions like these go unexamined is dangerous, as you may, consciously or unconsciously, arrange your study to increase the likelihood of arriving at those conclusions.

Your research project should begin by making a logbook of your assumptions. Make a list of hypotheses and note the assumptions in them. Include everything from established theories in your subject area to methods of data collection you presume will be most effective for you. If you’re embarking on an academic research project, you should also take careful note of assumptions you find in journal articles during a literature review. Assumptions left unquestioned by prior researchers can be opportunities for further research for you. And if you don’t believe us, believe Einstein.

If your research project is for a business exercise, kick things off by getting your stakeholders together to address everyone’s assumptions head-on. Take note of everyone’s assumptions about your research topic and where they come from — instinct, prior data, previous research, or elsewhere. Existing knowledge (or, as it’s sometimes known, conventional wisdom) is a great place for assumptions to hide.

Your aim should be to test these assumptions throughout your project. For a business-oriented research project, you can test assumptions about user needs or behaviors against focus groups. For academic research projects, you can test them against a figure of trust, like an editor, a project supervisor, or a mentor.

Finding your data

From scientific research to UX research, quantitative data is key to any modern field of study. But finding good quality data can be tricky. Some very worthwhile areas of interest are obscure and difficult to access, either incidentally ( like a poorly documented online subculture) or intentionally (such as anything to do with an individual's medical history). Even in more accessible subject areas, you might find that data is available but poorly organized or not easily available in readable formats like CSVs.

Your search for data begins by establishing the personas or communities you’re interested in. Finding your data can be tough, but it’s impossible if you don’t really know who you’re looking to gather information from. For example, targeting a very broad audience like “millennials” is a surefire way to incur avoidable problems finding good data. But targeting something more specific audience like “millennials earning more than $50,000 per year with a history of using product x” will make data gathering easier.

Familiarize yourself with basic web scraping techniques before you kick off the data collection phase of your project. This will make it easier to collect information from sites that display readily available data but that don’t offer easily downloadable CSVs.

If your project requires you to talk with hard-to-reach communities, begin building a network in that community before your project begins. Consult existing connections you might have. Approach community leaders on social media. And be aware that ethical safeguards must be put in place if your research project concerns vulnerable populations.

Working with bad data

An absence of data is a challenge, and bad data is a completely different challenge. If you’re using a methodology based on secondary data analysis, you might find the data available to you is not of sufficient quality for you to use in its current state.

The first step to overcoming this challenge involves understanding what the problem is with the data.

• Does the data have poor metadata or logging conventions, making it hard to locate or parse?
• Is the data incomplete, with multiple entries missing?
• Is the data gathered from a sample size that’s too small?
• Is the data gathered without sufficient ethical safeguards in place?

From there, you must determine if the bad data is salvageable. Data gathered unethically is an automatic no-go, as your project could have legal or moral ramifications. Data that’re incomplete or drawn from an unrepresentative sample size can be salvaged if its metadata demonstrates how the data is unrepresentative or incomplete. For example, if the data is biased by age and the metadata indicates what age it’s biased toward, it may be possible to build upon that data with an additional targeted study.

Datasets that have poor metadata can also be salvaged if the metadata is not so bad as to make the overall set insensible. It’s also salvageable if the set is not so large that manually amending the metadata is impossible.

If you determine the data you’re working with can’t be saved, scrape new data from available open databases. Otherwise, you may have to switch to a primary data collection methodology. It might cost more and require a longer time frame, but it will save your project.

Securing stakeholder buy-in

Whether you’re embarking on an academic-minded- or business-driven research project, managing stakeholder expectations is a common challenge. Stakeholders can often be a tremendous asset to a project’s final outcome, but you can just as often find them resistant to your research direction, focus, or even the project itself.

Business-oriented research projects might encounter resistance from stakeholders who lack belief in R&D or team members who don’t want to invest time and input. Academic research programs can also find themselves falling afoul of the sensibilities of their supervisor or grant funder if the project is interdisciplinary.

The best research projects often come from bold ideas and the desire to try new things. Learn to sell your project in your stakeholder’s language. For example, an academic should situate their proposal in the context of a well-executed literature review that includes content that advisors or grant providers know and credit. They can also draw stakeholder attention to the similarities between their research proposal and previous projects in their field of study that successfully delivered value.

For business research projects, a project leader faced with skeptical stakeholders should demonstrate the cost of going ahead with a project without the right research data. They should also provide examples of companies whose bold approach to R&D paid off big time, though it can be just as effective to cite companies who totally missed the boat on big innovations. If you’re encountering particular opposition from a particular team in your organization, explain to them how success in this research project will make their lives easier.

Trying to do it all alone

Whether the posited outcome of your study is a groundbreaking research paper or a splendid new product, having a proactive approach to potential challenges makes for a more fruitful research process. No more getting blindsided at the 11th hour by the realization you’ve been operating under a false assumption all this time! No more bad data!

And whatever the nature of your research project, you need the right tools to help secure the basics — like finding and saving useful sources so they remain accessible to you. And we’ve got just the tool to make keeping track of the important stuff a whole lot easier.