Category Archives: Statistics and Methods

Science: not a religion, just a way of truth-seeking

This is good. However, smarter science deniers will say that you can do all of these things and still, knowingly or unknowingly, be creating biased/untrue findings; they say all of science is just a huge data cherry picking contest<–which is true, to a large degree! –> All the more reason to be thorough, reproduce, and #openscience, #data, #code, etc. Don’t walk away from or doubt #science, just do it better yourself, dig deeper, and we should train better and more student, citizen and professional scientists…. this is a no-brainer in my opinion.


Public Health and Environmental Advocacy Groups: YOU NEED CITATIONS!

Citation Needed!This post is for everyone, but is aimed at public health and environmental advocacy organizations. First, I want to say that what you do is crucial to the success of individuals, communities and societies as a whole. I commend your hard work (most of it).

However, you often shoot yourself in the foot by failing to provide adequate citations and references in websites, flyers, radio and TV ads, PSAs, etc. Citations are often missing entirely, are incomplete, or only cite one of your own documents which is also missing the reference. This makes it impossible to find the article and will result in emails like this:

“Hello, I would like to know the facts behind your webpage regarding air pollution monitoring. Who is doing the monitoring? How regularly? What are the results and what standard is used? Does your group provide factual information or just distorted hype? Please have the fortitude to respond to simple factual inquiries. I have inquired multiple times without response.”

Painful, but he has a point. If you cannot provide the reference for a scientific study, you should not be using a specific number. For example, You can’t say that a certain law saved 100,000 lives unless there was some study or formulation done that came to that conclusion. Do not make claims that go beyond the scope of the research. Similarly, saying that a certain law or program would prevent “many” deaths, is a weak argument and is unlikely to garner support from funding agencies or politicians.

Failing to provide references and citations for actual scientific research that supports your arguments completely undermines an individual’s or organization’s ability to advocate for change.

The knee-jerk response to my point is often “Who cares, you can make statistics say anything you want.” That is true, to some extent, but that does not mean you should stop citing or conducting research. Providing a citation and reference enables someone from the outside your organization, outside the research, to examine the methods used to gather the data, look at the results, and see whether or not those methods and the claims that follow are appropriate.

This may remind you of high school science classes when you had to use at least 2 citations for a 3 page paper. However, the entire health and environmental advocacy field has shifting towards this standard, which it should! The way governments and organizations spend their money is being scrutinized at a higher level. That means that advocacy organizations need to step up their game in order to maintain relevance, earn respect from the public and public officials, as well as provide the most good for the most people using the least amount of money. Being more steadfast about basing advocacy work/products/literature on scientific research and using citations will not only improve your work, but will also preclude naysayers from claiming that environmental groups are radical zealots making unfounded claims. Like it or not, advocacy, especially in today’s world, requires some scientific prowess and familiarity with research methodologies.

If you, or your organization is uncomfortable with all of this, please, don’t avoid that discomfort. Instead embrace it as a new way for your organization to achieve it’s mission. Depending on your staff’s education, it may only mean that they need to be reminded to provide citations whenever possible. Or, it could mean that your organization needs to look at hiring individuals with a different skill set.

I know, it may be painful. Science and research are scary things to some people. Yet, they are some of the most powerful tools we have as health and environmental advocates. Good research is good advocacy. Good advocacy is based on good research.

As a side note, unfortunately, many scientific articles are not available for free. However, many are! You just need to look. Open-access journals are out there. For example: Environmental Health. Also, many government websites provide their articles for free. Most other articles are available through college and university library systems. Institutions of higher learning spend millions of dollars annually to give their faculty and students access to the largest and latest research databases. HINT: Maintain staff or interns that are currently enrolled.
If you have no idea what I am talking about, here is an example:

In your text:
“Research has shown that air pollution levels are higher in metropolitan areas (Dockery et al., 1993).”

At the bottom of your document/website
Krewski, D., Burnett, R., Jerrett, M., Pope, C.A., Rainham, D., Calle, E., Thurston, G., Thun, M. (2005) Mortality and long-term exposure to ambient air pollution: ongoing analyses based on the American Cancer Society cohort. Journal of Toxicology and Environmental Health, 68, 13-14.

For most purposes, don’t worry about the different citation styles, or having everything formatted perfectly. That would be great, but you must provide at least enough information for someone else who wants to find the article.
A couple recommended links:

Google Scholar

Structural Equation Modeling (Temple University Dept. of Public Health)

Structural Equation Modeling
Fall 2011 – Dr. Adam Davey
Temple University – Department of Public Health

Selected quotes and my random thoughts as I take the SEM journey.

Here’s the text:

“Unlike multiple regression, though, SEM allows the analysis of curvilinear or interactive relations of both observed and latent variables.”
Whoa! Bring it.

“SEM readily allows the representation and estimation of possible
spurious associations because of common causes.”

On missing data (which is crucial to a proper SEM analysis)
“Also note that there is no magical statistical “fix” that
will remedy systematic data loss. About the best that can be done is to
attempt to understand the nature of the underlying data loss mechanism
and take preventive measures.”