A government agency was concerned with its rankings on certain metrics in the Federal Employee Viewpoint Survey, a tool that federal agencies use to measure organizational climate. We demonstrated that most high-ranked and low-ranked agencies were in fact quite small. In other words, the rankings failed to take in account one of the most basic ideas in statistical sampling: the standard error of the mean. We helped the agency understand that it needed to understand its results relative to those of agencies of comparable size.
A municipality wished to make a major push to promote owner occupancy. Our policy research demonstrated that there were demographic niches that could be better served. We also identified city policies that were working against its ability to penetrate those markets. But perhaps the most important conclusion we drew was that changing the percentage of owner-occupied housing was probably less important than the need for a guiding vision of what kinds of owner-occupied housing the city should aim to attract. Years before the housing market collapse in 2007, we explicitly warned about the perils associated with promoting owner-occupancy for households that would require sub-prime financing.