Jan 10, 2020
29
14
Status
Pre-Veterinary
In a research-reading heavy course, our professor taught us to read the figures first. I tried that on a paper I'm trying to read in a new field (not for the class) and it wasn't working. The figures are almost all really long, dense data tables, some with p-values, and some without. I tried reading the background and introduction first, then taking another look at the tables. That helped somewhat, but the tables without p-values just look like number-soup to me. Any advice or tips? Any mistakes I might be making? I was taught (by both aforementioned professor and my PI, so I'm assuming it's normative) not to read the author's own conclusions too much (lumped in with the analysis section of the manuscript in this article's case) before looking at the data and coming to my own conclusions.
 

LetItSnow

Skipping the light fandango
7+ Year Member
Jan 13, 2011
19,263
18,492
Plymouth, MN, USA
animaltracks.wordpress.com
Status
Veterinarian
Perhaps try thinking about what question the paper is asking, and then ask yourself what information (data) they'd need to make conclusions, and then go look for that data. So do it the other way around from what you are doing - instead of looking at the data and trying to make sense of what it is and why it might be there, get some idea of what data is important to the hypothesis, and then go look to find it.

Just a thought. I don't know. It's an interesting question.

One of the real bummers of modern medicine is that we are taught to be evidence-based ... but ... I don't know about other practicing clinicians, but I sure don't have time to stay super up on literature. Even in my own area of practice I can barely keep up on new developments, much less check out all the relevant literature.
 
  • Like
Reactions: hachamor_persists

that redhead

10+ Year Member
Feb 26, 2010
11,140
10,836
One of the real bummers of modern medicine is that we are taught to be evidence-based ... but ... I don't know about other practicing clinicians, but I sure don't have time to stay super up on literature. Even in my own area of practice I can barely keep up on new developments, much less check out all the relevant literature.
I mean I stay up to date through conferences and reading a ton of VIN threads and the like but honestly, it’s hard to forge new paths based on data that came from a study with a sample size of like, five, yanno?
 
  • Like
Reactions: hachamor_persists
About the Ads

Musicandhorses

2+ Year Member
Mar 28, 2017
29
59
Status
Veterinarian
Often times heavy data tables will star or bold or italicize the important info like certain p values they want to point out to you. Are there any starred or bolded?
 
  • Like
Reactions: hachamor_persists

Musicandhorses

2+ Year Member
Mar 28, 2017
29
59
Status
Veterinarian
Adding to previous post...if there’s nothing emphasized on the tables I would then read the abstract. It should very briefly summarize what they found. Then look at the figures and the methods to see if their data backs that up.
 
  • Like
Reactions: hachamor_persists

LetItSnow

Skipping the light fandango
7+ Year Member
Jan 13, 2011
19,263
18,492
Plymouth, MN, USA
animaltracks.wordpress.com
Status
Veterinarian
I mean I stay up to date through conferences and reading a ton of VIN threads and the like but honestly, it’s hard to forge new paths based on data that came from a study with a sample size of like, five, yanno?
Fact.
 
  • Like
Reactions: hachamor_persists
Jan 10, 2020
29
14
Status
Pre-Veterinary
:(I've taken several more parses through on the paper. I understand what the authors think the data says, and I sort of understand why... Just really grateful that I don't have to present it; I don't think I could explain it without loads more work, nor do I feel like I understand all the logic. It's a multifactoral paper with several cohorts and the data is still dizzying (I'd love to see those p-values because a lot more than 20 tests were run with a 95% threshold...). If I really had to, I could maybe spend more time looking it over, but at this point, I'd probably read the surrounding literature (it was done to validate some studies performed in other models, and better explore its own model). If I get a chance, I'm going to at least skim some other papers in the field to see if they're all this hard to read. But for now, vet school essays and job applications are taking priority.

Often times heavy data tables will star or bold or italicize the important info like certain p values they want to point out to you. Are there any starred or bolded?
Nope :( Just lots of footnotes. Which I could barely follow. They're looking at several factors, and which measures are statistically significant with each factor. I understand that they're doing this because even in a laboratory model, organisms are complicated. I'm glad they considered the complexity and planned lots of analysis into the study to reduce animal use, but it doesn't make it easy to parse though...

Adding to previous post...if there’s nothing emphasized on the tables I would then read the abstract. It should very briefly summarize what they found. Then look at the figures and the methods to see if their data backs that up.
I did read the abstract (just not taking it at face value with the idea that I should be doing my own critical thinking). However, there was a lot going on in the paper and the abstract (& intro) only did so much in terms of giving me enough direction and insight to prep me to parse through the data and figure out the analysis goals. Maybe if I was more used to papers in the field then conventions would have done the rest?

Perhaps try thinking about what question the paper is asking, and then ask yourself what information (data) they'd need to make conclusions, and then go look for that data. So do it the other way around from what you are doing - instead of looking at the data and trying to make sense of what it is and why it might be there, get some idea of what data is important to the hypothesis, and then go look to find it.
I ended up doing something similar: taking a look at what they said, and then going back to the tables to confirm. It looks in-line with their p-value footnotes (unsurprising, but always good to check), and the experimental design is solid (standard for the lab, field, etc.), there was just A LOT going on.

If I have time to read more research while I'm taking time to work before vet school, maybe I'll make a research paper organizer to help with complicated papers like these? If I'm going to bother, I'd want to make one that will help me with a variety of "hard" papers and even just reading some more of those would be a big project. Either way, I'll keep y'all's considerations in mind, and if I make something useful I'll put it back up on this thread.
But yeah, I can see how clinicians don't have the time to read the research. Especially more basic science papers like this one. Maybe I'm weird, but I find drug trial studies relaxing to read :rofl: I don't want to put the cart before the horse, but I really hope I can (long-term) get a clinical job where I can incorporate helping with and reading research into my work.
 
About the Ads