Editor’s Note: Samuel Wilder King II is a candidate for OHA trustee at-large.
The Office of Hawaiian Affairs has spent millions of dollars developing voter rolls in an attempt to push federal recognition, but very little thought seems to have gone into collecting data on the economic condition of native Hawaiians. OHA appears to be ignoring the margin of error in its datasets.
On OHA’s Homeownership Indicator Sheet 2016, OHA interprets the movements of the lines on a graph, re-created below, directly, as if they can be taken literally: “Although there was a large decrease in 2014 to 52.9 percent, the increased change to 57.7 percent in 2015 puts OHA on track,” they claim.
OHA obtains its data from the American Community Survey, “an ongoing statistical survey by the U.S. Census Bureau sent to approximately 295,000 addresses monthly (or 3.5 million per year),” according to OHA. What OHA does not explain, however, is that this data includes fewer than 3,000 Native Hawaiians per year. That results in a margin of error so large it makes it impossible to measure whether OHA is achieving its goal.
This is disheartening to say the least. As explained in the biography for CEO Kamana’opono Crabbe, who was the research director from 2009 to 2012 (when he was made CEO), the data is “gathered” to help OHA “make sound decisions that allow it to engage policymakers in its efforts.”
Margin Of Error
Our campaign team recreated the above table following the citations provided by OHA, but we added the margin of error to the chart. The margins of error are provided by the original data source, right alongside the numbers OHA cites.
As you can see, once the margin of error is added, it is clear there is no evidence of any real change between 2013 and 2015, up or down. All the changes are well within the margin of error:
What this shows is that the policy goal promulgated by OHA is too small to ever be detected with the data they are using. They propose to raise homeownership from 56.62 percent to 58 percent within 10 years, from 2008 to 2018. As you can see from the graph above, homeownership in 2008 may have already been 58 percent, or even higher. The data is not good enough to tell us one way or another; the samples are too small.
Extending the graph to include 2016 — which is now available from American FactFinder and on OHA’s website — shows why OHA’s approach is deeply problematic. The 2016 estimates are sharply lower than those for 2015. Not only did homeownership probably decline in 2016, as we can see from the graph, the 2015 figure is likely an overestimate.
It is important to note that re-surveying the original population in the same year could give an answer anywhere inside the margin of error. The evident increase in 2015 is quite possibly just noise due to sampling error.
Moreover, OHA continues to cite the evident “success” from 2015 in the most-recent version of its indicator sheet, despite the fact that measured homeownership dropped sharply in 2016.
Are Native Hawaiians on track to reach the 2018 goal? Undetermined. Of the 72,940 Native Hawaiian housing units, 50.3 percent were owner-occupied. The percent of Native Hawaiian owner-occupied units decreased 7.4 percentage points, while the number of Native Hawaiian housing units increased 5.8 percent from 2015 to 2016. In 2015, the Native Hawaiian owner-occupied percent of 57.7 percent indicated that Native Hawaiians were on track to reach the 2018 goal.
In other words, OHA is using inappropriate data to track progress toward its economic goals and misinterpreting that data. One would think that the CEO, as the former head of the research division, would have remedied that by now. The Trustees overseeing him should have been asking these questions already.
The same issues seem to permeate OHA’s data.
OHA knows who its beneficiaries are and can organize surveys of Native Hawaiians.
OHA’s “Median Family Income Indicator Sheet 2016” results come from the same source as the homeownership results (the one-year ACS samples) and are missing the margins of error from the ACS. The peaks and valleys in that graph are completely implausible — they imply massive progress on the intractable issue of poverty in a single year.
The sheet claims that Native Hawaiians experienced a $15,000 increase in family income in a single year. The changes in the statewide median income shown in that graph are much more believable — they indicate slow progress over time, not wild jumps in the time series.
The solution to all this is 1) for OHA to at least show the margin of error on its charts and 2) for OHA to start collecting its own data on the indicators it claims to be tracking, rather than trying to source data from national survey samples that are not designed to study this population.
OHA knows who its beneficiaries are and can organize surveys of Native Hawaiians. Doing so would be an invaluable resource for its own planning and also for a range of other organizations dedicated to improving the welfare of Native Hawaiians.
The problem is that OHA’s trustees and the CEO are too busy fighting with each other and arguing over sovereignty and federal recognition instead of focusing on the day-to-day issues facing Native Hawaiians. Instead of surveying Native Hawaiians to investigate economic conditions, OHA has only spent money trying to create voter rolls. This lack of quality data has been an issue for years. OHA’s leadership has failed to notice in fantastic form.
I have said OHA should develop an overwhelming obsession with programs that will improve the daily lives of Native Hawaiians. That starts with the people at the top of OHA actually paying enough attention to know that their data is inadequate.
Thoughts on this or any other story? Write a Letter to the Editor. Send to news@civilbeat.org and put Letter in the subject line. 200 words max. You need to use your name and city and include a contact phone for verification purposes. And you can still comment on stories on our Facebook page.
The post Campaign Corner: OHA Uses Misleading Data About Hawaiians appeared first on Honolulu Civil Beat.