By Lauren Cohen / Posted on October 31st, 2014
The Green House Project has partnered with the Robert Wood Johnson Foundation’s THRIVE (The Research Initiative Valuing Eldercare) collaborative to learn more about the Green House model as well as other models of care. Supported by the Robert Wood Johnson Foundation, the THRIVE team is conducting a series of interrelated research projects that together will comprise the largest research effort undertaken to date in Green House homes. Each quarter, a member of the THRIVE team will contribute a blog post to the Green House Project website.
Data collection for the THRIVE projects is now complete, and the research team is analyzing the results. The THRIVE team will share research findings in upcoming articles in a special issue of the journal Health Services Research, and through conference and webinar presentations and blog posts. In 2014, conference presentations will include those at annual meetings of LeadingAge (October), and the Gerontological Society of America, and the Green House (both in November). This blog post is part of our series devoted to explaining research terms so that non-researchers can better understand these articles, presentations, and posts. This post focuses on quantitative research – research based in numbers – and explains the important topic of “significance.”
Quantitative research findings are often discussed in terms of their statistical significance. What does it mean to say a finding is significant?
Let’s consider an example. A researcher thinks that there may be more female than male elders living in Green House homes. This hunch is called a hypothesis. The researcher visits all the Green House homes in the state, tallies the numbers of females (85) and males (15) and performs a statistical test to compare males and females. The statistical test will result in a p-value (probability value) expressing whether the difference is large enough to indicate that it isn’t just by chance.
To better understand what it means to have a “large enough” difference, think of it this way: if the number of females was 52, and the number of males was 48, the difference between these numbers is pretty small, and it’s not likely statistically significant. The question is, is the difference between 85 and 15 large enough to suggest that there are statistically more females than males living in Green House homes? A difference of 85 to 15 is probably large enough to not be by chance (i.e., it is statistically significant), whereas a difference of 52 to 48 is so small that it quite likely occurred by chance.
It’s also important to realize that findings that are statistically significant may not be clinically significant. Clinical significance means that the information is important for clinical care. In terms of care, does it matter that there are more females than males residing in Green House homes? It does matter, for example, if women tend to be more depressed than men, or to have more family members. However, if there are no clinical implications related to the difference, than they are statistically, but not clinically, significant.
The bottom line is that it’s important to carefully consider the meaning of all findings, and use your knowledge and judgment to interpret when differences matter and when they don’t.
Stay tuned for the next THRIVE blog post. In the meantime, if you have questions about this post, or suggestions for future ones, please let us know.
Questions about THRIVE can be directed to Lauren Cohen (email@example.com or 919-843-8874).