Introduction: Fishing for Cures?
A new study finds that eating fatty fish may decrease your likelihood of getting Rheumatoid Arthritis! Are researchers just fishing for cures? Lets take a look....
http://thechart.blogs.cnn.com/2013/08/12/eating-fatty-fish-may-help-prevent-rheumatoid-arthritis/
Step 1: Step 1: the Basics
As we have learned, there are many facts, figures and studies that have been done and misreported to the readers in order to prove their hypothesis or a point. To start with analyzing these types of articles we have to understand the following:
1. What are we looking at?
2. Where is the data coming from?
3. How was the research done?
4. Does the data make sense?
Step 2: Step 2: What Are We Looking At?
This step is pretty simple, just read the article! We are pretty intelligent students here at Chapman University, so use that $50k education to examine the text.
What do we have?
1. The article tells us what its all about pretty plainly. It's goal is to convince us that eating fatty fish 2 times a week will lower womens' chances of getting Rheumatoid Arthritis.
2. What RA? Again, the text does a good job explaining what we are looking at. RA is an accidental autoimmune attack to joints, which destroys bone and cartilage.
OK, so now what?
Step 3: Step 3: Where Is the Data Coming From?
1. WHO?? Annals of Rheumatic Diseases
- BOOM - a little bit of research shows the intentions and credentials of this journal.
- "Annals of the Rheumatic Diseases (ARD) is an international peer review journal committed to promoting the highest standards of scientific exchange and education."
- Owned out of Europe - Explains why the study was done in Sweden!
2. Although we can get the actual study, lets stick with what they report in CNN's article for now.
Step 4: Step 4: How Was the Research Done?
This is where we can get our hands dirty. Don't worry, we are ready for this!
Lets break it down:
Lets start with Variables. What can we assume from the information:
1. IV = Eating fatty fish twice a week. Yes/No (Nominal)
2. DV = Disease at certain age? Yes/No (Nominal)
What can we assume from here? This looks to be a Chi Square test from the surface. Since we are only looking at what we are given, we can only assume how the research was done.
Lets stick with our gut on this. What about Chi Squares do we know?
1. Both variables have to be nominal/ordinal
2. Equal distribution in all squares
3. Random Sampling
4. Independence
So let take what we know and compare it to what we are given.
Step 5: Step 5: Does the Data Make Sense?
So, we established variables, assumptions of Chi-Square analysis and we have limited data.
Lets break it down:
1. Both variables have to be nominal/ordinal
2. If it were not the study would be flawed from the start!
2. If its not it makes it quite complicated to find where the change is actually coming from.
2. Food questionnaires in the 1980s and then in the 1990s.
Does this work for our random sampling & Variance?
2. However, we are only talking about Swedish women. Their country has different eating habits than others around the world, which means there could be difference circumstances causing the change.
3. For equal variance, we have to assume that there are equal fish and non fish eaters. In addition we have to assume that the non-fish eaters never eat fish, otherwise its flawed.
- We established earlier that their culture could be part of the reason they get RA less, not necessarily the eating of fatty fish. So we may have an issue here.
Looks like we might have some assumptions broken!
Lets see how we can address some of these on the next step.
Step 6: Step 6: Comm. 300 Strikes Back!
Being the studious students that we are, how would we do it?
1. Sampling
Because this is a broken assumption, we should fix it! Here I would suggest using a more diverse population from multiple countries. This would lower our chances of error when sampling.
2. Controlling the Study With this type of test we have to look at longevity, such as these researchers have done, however because of the sampling assumptions I would like to see a stronger control group and a stronger test group.For this study the researchers heavily rely on the women's memory that spans 10 years. How accurately can you remember when and how much you ate fish? So in this case it would be important to find a test control group that does not eat fish and then establish a group that does eat fish, and heres the key, tracks their habits over the years.
Having stronger controls over the study could result in stronger outcomes!
So we have fixed (2) of the issues. What about the other?
1. Independence - we have slightly addressed this issue to some degree by adjusting the sampling, but are we sure that the actual intake of fish is causing the deviation in RA symptoms? Medically this seems almost impossible to do since they are assuming Omega-3 fatty acids cause the effect and RA technically speaking is an accident from your immune system.
To address this area we could do an additional study trying to find the correlation between Omega-3 and RA disease.
Step 7: Step 7: Conclusion
Because of the assumptions being broken and the fact that we are not really provided with any actual hard data to show that the results are significant or even related, I say NO.
We have looked through enough examples now to know that a conclusion that says,
"Women who consistently ate at least one serving of fatty fish each week for a period of about 10 years developed rheumatoid arthritis at half the rate of women who ate little or no fish. The researchers say the results held up even after taking lifestyle factors into account such as how often people smoked or drank alcohol."
probably needs more information to argue the point.
Lets sum this up:
1. Too many assumptions were broken to be convinced that this study has merit.
2. Changes such as,
- Correlation study on Omega 3 and RA disease to analyze whether these have anything in common.
