I experienced this firsthand as an idealistic 24-year-old graduate student. Realizing that I needed to develop close personal ties with the policy makers I eventually wanted to help, I began moonlighting as a journalist while in school. During a trip, when I moderated a conference panel with a congressman about how technology can improve the democratic process, I received an email from a well-intentioned professor scolding me for taking time away from research. "You really need to focus on your academic work and stop wasting time on these sorts of excursions," he wrote.
As a result, even the potentially useful research gets overlooked by policymakers who have little contact with experts from the discipline. During a dinner my university threw for a distinguished Harvard professor who also served as a United Nations consultant, I asked our guest if she ever witnessed any actual impact of political-science research. She literally laughed out loud, and regaled the now-perturbed table of academics about her experience with Afghan President Hamid Karzai, who had apparently ignored all the academic experts during his country's transition to democracy and, instead, decided on the structure of government in a tent with his peers.
Part of the problem is that the culture of political science exalts researchers who discover timeless, fundamental precepts of good governance and not those who tackle timely issues, such as healthcare or education. As a result, it's exceedingly difficult to research that is both statistically rigorous and has practical relevance.
For instance, one of the most influential pieces of practical research in the last half-century found that building strong communities, where neighbors have frequent informal interactions with one another, dramatically boosts social welfare, as measured by the health, happiness, and governmental effectiveness of a city. The study took Harvard's Robert Putnam 25 years, which was necessary to observe how natural variation in communities all around the entire country of Italy affected the lives of its 50-plus million citizens.
Quite reasonably, most researchers are squeamish about dedicating half their career to one study and simply avoid the challenge of relevant research. Instead, they turn to clever mathematical models that check their research box but do little to aid real-world public policy. Appropriately enough, Putnam himself has been an outspoken critic of his own discipline and once told me that he begs his colleagues to do research that passes the "mother-in-law test": You ought to be able to explain to your mother-in-law why you're working on it.
There are political scientists who do worthwhile research in less than 25 years, but most don't work in academia. The World Bank is helping to develop online tools for direct democracy, which allows third-world citizens to make decisions on their own local budgets and redistribute public services more fairly. Free from the compulsion to publish in peer-reviewed journals, non-academic political scientists are tasked with solving very specific problems.