Analyzing the data: what is most important for weight loss?
Replies
-
This content has been removed.
-
lemurcat12 wrote: »The problem is that you haven't (and probably can't without tools you don't have easily available) analyzed changes based on fat loss vs. changes based on water fluctuations.
For example, you conclude that binging outweighs actual net calories. But there are some obvious possible explanations: One is that binging is associated with increases in refined carbs or sodium, both of which cause water retention. A second is that it could lead to stomach upset and related issues on average. A third (okay, if you were a woman, so consider this just a hypothetical) could be that binging is more likely to occur at certain times of the month which happen to correspond with water retention or slower weight loss/weight gain.
One thing I noticed in analyzing my own results was that until I was quite close to goal I lost pretty consistently and never experienced a whoosh, but I did almost always have two weeks in the month (immediately after my TOM began, which is the opposite of many) where I lost really well--more than my average--and two other weeks where I lost less well, with the worst loss almost always coming the week before TOM.
Oddly enough, the pattern continued but got less significant as I got closer to goal, and was replaced with some whooshing--2-3 weeks of nothing and then a 1.5-2 lb loss when I was only losing on average about .75 lbs. There was nothing of note about my eating or exercise within the weeks to explain the differences.
Another thing I've noticed is that if a take a "deload" week or simply am super busy and miss workouts and exercise less but also adjust my calories (as when I was eating back exercise) that I tend to lose extra well. When I first add a new exercise or increase the intensity of my exercise routine I tend to loss less. This happens even when the change is more cardio, not simply with weights. (But I certainly would not conclude the exercise did not help me lose or was not important for increasing fitness and motivation, etc.)
I think trying to come up with an explanation for why we lose more in one week than another despite consistent calories and exercise is usually an exercise in futility, although I agree it can be interesting. You have to accept that sometimes it just happens and not assume that you did something right or wrong and that otherwise fluctuations would not happen. They just do.
Thank you. This is the kind of thoughtful and insightful response I was hoping to stimulate. Almost worth reading through all the others to get to this. Thanks for taking the time to write it.
0 -
ILiftHeavyAcrylics wrote: »You may be overthinking this.*
That's what works for some people.
Doesn't work for other people.
It's all good...
The problem is that I don't think the kind of analysis the OP wants is possible. Bodies are complicated and it can be difficult to sort out whether he's looking at differences in water weight, fat loss, or just plain old coincidence. If you recognize things like that binging or increased sugar probably also mean increased carbs in general, which can cause water retention and hide fat loss then it's all good. But OP seems to be suggesting that those things slow or halt fat loss. Binging would, if overall calories were significantly higher, but increased refined sugars most likely would not (unless OP has an undiagnosed medical condition).
It may work for him, and if so, great! I'm a woman who likes to over-complicate things so I understand the tendency. For me personally it eventually just leads to frustration. Since it's nearly impossible to figure in all the variables it's never going to be really reliable. About the best a person can hope for is a list of correlations, some of which may be true cause and effect, and others of which likely aren't.0 -
onelentilatatime wrote: »I just analyzed 6 months of data to understand what leads to the largest weight loss in a given week.
Sounds interesting. I've read many things where weight loss doesn't correlate well with simply calorie deficit.
What type of analysis did you do - some sort of multiple regression ?
0 -
Did you weigh your daily poop, water intake and urine?
Here is an easy experiment.
Every two weeks spend two days eating no carbs. Then eat your normal diet.
You'll come to the conclusion that carbs cause weight gain.
Weight fluctuation =\= weight gain0 -
onelentilatatime wrote: »I just analyzed 6 months of data to understand what leads to the largest weight loss in a given week.
Sounds interesting. I've read many things where weight loss doesn't correlate well with simply calorie deficit.
What type of analysis did you do - some sort of multiple regression ?
Ooh good. Someone who talks my language. Multiple linear regression. I added variables that explained a significant amount of the variance. The final model accounted for 70% of variance with the 4 variables I mentioned, which is pretty good by all accounts.0 -
EvgeniZyntx wrote: »Did you weigh your daily poop, water intake and urine?
Here is an easy experiment.
Every two weeks spend two days eating no carbs. Then eat your normal diet.
You'll come to the conclusion that carbs cause weight gain.
Weight fluctuation =\= weight gain
Interesting. This is the kind of thing we need to do to get past the correlation/causation question. I saw a TV show with twins - one of who cut out sugar and ate lots of fat, the other cut out fat and ate lots of carbs. They both lost weight. They were both miserable.0 -
onelentilatatime wrote: »I just analyzed 6 months of data to understand what leads to the largest weight loss in a given week.
Sounds interesting. I've read many things where weight loss doesn't correlate well with simply calorie deficit.
On a day-to-day basis, that's often the case.
On a long term basis, it's never the case, weight loss always correlates with deficit over the long run.
0 -
Thanks for all your spirited contributions to this thread everyone. In the interests of making the discussion more constructive I thought I would try and be clearer about what I am and am not saying. I am not telling anyone what to do. I am not claiming to have found the laws of weight loss. I am just reporting statistical relationships I have found in my data that warrant explanation. The explanation may or may not be interesting (see below). Also, I am not doing this because I am frustrated with my own weight loss efforts, just curious about what the data can tell us.0
-
For those interested in helping me think more about this I wanted to develop two critiques of my analysis to see if we can do better.
1) Correlation is not causation.
Yes, but it's nearly impossible to identify causal relationships here. We can do experiments by cutting out things from our diet but then that has knock-on effects for our behavior and so doesn't necessarily help isolate important factors. In which case, correlational data may be all we have. I found a correlation between sugar intake and weight loss, even after taking overall calorie intake and exercise into account. This relationship is an empirical fact that needs to be explained (at least if other people find the same thing and the finding becomes a little more robust - which is why I am asking if anyone else has seen/done this kind of analysis). Possibilities are:
a) Sugar calories are worse than other calories (somehow) for weight loss
b) Sugar calories do temporary things to one's weight (e.g. water retention) but make no more difference in the long run
c) There is something else I do on weeks when I eat more sugar that causes me to put on weight
d) anything else?
2) It's so much more complicated than you realize
I'm sure it is. All statistical models are simplifications. That's the disadvantage. The advantage is that they have the potential to illuminate things because they are based on evidence. If you do it right. So while the relationships I looked at are blunt instruments, the findings can still point to interesting insights. I am not sure how to make the analysis reflect the complexity of reality any better except by measuring more stuff. I have data on fat % which I haven't entered yet, so that's a possibility. I could also look at the data over longer time periods (e.g. did you sugar intake 3 weeks ago affect your weight loss this week) but I think it will be hard to find strong relationships.
Very interested in hearing any thoughts about how to use these data better. It may not be possible of course but I've had fun trying so far.0 -
This content has been removed.
-
If the point of looking at the data is just entertainment, then ok. But there is research done on the biological processes involved here. Without contextualizing the data within some of that research in specific ways, it seems like wasted time...for no particularly useful result.
0 -
OP does a study on her/himself without a control group.
No science to see here folks.0 -
onelentilatatime wrote: »Possibilities are:
a) Sugar calories are worse than other calories (somehow) for weight loss
b) Sugar calories do temporary things to one's weight (e.g. water retention) but make no more difference in the long run
c) There is something else I do on weeks when I eat more sugar that causes me to put on weight
d) anything else?
e)Sugar has absolutely no effect and the correlation was random coincidence.
Since your n=1, this has to be considered a very real possibility
0 -
I think the uptick in fiber resulted in this post.0
-
nakedraygun wrote: »OP does a study on her/himself without a control group.
No science to see here folks.
All science starts with an n=1 observation.
0 -
ceoverturf wrote: »onelentilatatime wrote: »Possibilities are:
a) Sugar calories are worse than other calories (somehow) for weight loss
b) Sugar calories do temporary things to one's weight (e.g. water retention) but make no more difference in the long run
c) There is something else I do on weeks when I eat more sugar that causes me to put on weight
d) anything else?
e)Sugar has absolutely no effect and the correlation was random coincidence.
Since your n=1, this has to be considered a very real possibility
Depends on what you mean by "no effect". I have a huge amount of my own data, and there is a very clear correlation between temporary weight gain and uptick in simple carb ingestion. This is under conditions of regular cardio + caloric deficit, so the most likely explanation is temporary replenishment of glycogen stores, with its attendant water retention. This would be covered by the OP's option (b).
Over the long run, deficit prevails, of course - as it has to.
0 -
Irrelevant .nakedraygun wrote: »OP does a study on her/himself without a control group.
No science to see here folks.
All science starts with an n=1 observation.
And if that is the case, science does not begin and end with observations about oneself.
0 -
Irrelevant .nakedraygun wrote: »OP does a study on her/himself without a control group.
No science to see here folks.
All science starts with an n=1 observation.
And even if that is the case, science does not begin and end with observations about oneself.
0 -
nakedraygun wrote: »...science does not begin and end with observations about oneself.
Nobody suggested otherwise.0 -
This content has been removed.
-
nakedraygun wrote: »OP does a study on her/himself without a control group.
No science to see here folks.
Don't forget, a one week study.
Not sure I understand your last comment. In the first post he said he analysed 6 months of data, to try to see if he could correlate to weekly trends.
0 -
Flatlndr, congratulations on the weight loss by the way. Good for you, you could be the next James Bond.
0
This discussion has been closed.
Categories
- All Categories
- 1.4M Health, Wellness and Goals
- 393.9K Introduce Yourself
- 43.9K Getting Started
- 260.3K Health and Weight Loss
- 176K Food and Nutrition
- 47.5K Recipes
- 232.6K Fitness and Exercise
- 430 Sleep, Mindfulness and Overall Wellness
- 6.5K Goal: Maintaining Weight
- 8.6K Goal: Gaining Weight and Body Building
- 153.1K Motivation and Support
- 8.1K Challenges
- 1.3K Debate Club
- 96.4K Chit-Chat
- 2.5K Fun and Games
- 3.9K MyFitnessPal Information
- 15 News and Announcements
- 1.2K Feature Suggestions and Ideas
- 2.7K MyFitnessPal Tech Support Questions