# Sedentary behavior & reduced medial temporal lobe thickness

1. Apr 15, 2018

### Buzz Bloom

My wife keeps telling me that I need to avoid sitting at the computer so much. She sent me the following link.
I confess that I don't find the data presented in the article very convincing, and I would appreciate comments regarding the statistical analysis from knowledgeable participants.

As I see it, the data consists of two scatter diagrams (Figs. 1 and 2) and two data summaries (Tables 1 and 2). Table 1 describes the characteristics of the population, and Table 2 (shown below) gives the statistical results. As I look at the scatter diagrams, the steepness of the least mean square fit seems to me to be strongly influenced by a few data points I would call outliers. The data in Table 2 gives values for β and p-value. I do not understand what β represents, but the range of 95% CI seems seems odd that it is so much larger than the value of β.

2. Apr 15, 2018

### Ygggdrasil

3. Apr 15, 2018

### Buzz Bloom

Hi Ygggdrasil:

You are probably right about that, but I would like to understand, if I can, whether the data in the Table 2 reliably implies that there is good support for the premise that avoiding excessive sitting improves memory. In particular:
(1) Do you know what β represents?
(2) For example looking at the first row, do the numbers {-0.02, (-0.04,-0.002)} mean that:
(a) β=-0.02 is the mean or median of a probability distribution of possible values,
(b) the probability that the true value of β ε {-0.04,-0.02} is approximately 0.475, and
(c) the probability that the true value of β ε {-0.02,-0.002} is approximately 0.475?​
(3) Is the p-value of 0.03 a reliable predictor of the premise, or it it only an indication (weak?) that the data is not consistence with the null hypothesis? Am I correct that this support against the null hypothesis implies only that the two correlated variables were not likely to have been generated by random numbers from a single distribution? Or is there some stronger implication?

I would appreciate any comments you would care to make.

Regards,
Buzz