Member Info for jackdaw99

It's just stats but they are objective and help make an informed decision. If you do the calculations yourself you will get exactly the same result, so you can hardly accuse the statistical procedures of any bias one way or the other. The objective trend is VERY clear - strongly upwards - the only subjective element is whether you believe the trend will continue. I don't see any reason why not whereas you think it won't. Fair enough. The answer lies in the biology of this theraputic and, for that, we need more data.

It's the best we can do with the available data. The real problem is the lack of information from 6-9 months (and the "condensed" way they have pooled the data until that point) , so we have to wait. The curve may well flatten, with a law of diminishing returns, but there is no indication of that in the 0-6 months data, so while it's not impossible I would be surprised if that were to turn out to be the case. What I am confident about is that no statistician would EVER base a prediction on a single data point so this 9 month datum is meaningless for me at the moment.

IF the regression predicts reliable from the first 6 months of data, the value at 9 months would not be your +12 as per your table, but would be +37 (ie about 8 higher than the 28 at 6 months in your table).

No problem. Thanks for the mature reply!

Your analysis is misleading. First, the datum at 9 months, as has been pointed out already, is a single point, not a mean. For that reason I would exclude it as unreliable. Using the rest of the data (while conceding we cannot do the ideal, repeated-measures analysis of variance on the data as presented) it's possible do to a rough-and-ready regression and correlation analysis of mean increase versus time since treatment (ie the rest of the data in your table) ...... which gives an indication that the relationship between time and improvement is highly significant with 92% of the variance explained and that each additional month since treatment adds about 3 more letters of improvement. If you add the 9-month observation into the mix, obviously the strength of the effect declines but, as I said before, it's not really possible to weight the analysis properly for sample size given the information we have. My judgement is that the 9 months point, given the strong relationship shown by the rest of the data (linear and not showing any sign of peaking, by the way), is an outlier and should be included only when other 9-month points are available. In summary, the data so far gives me confidence this works and I would expect to see a mean value for the 9-month datas, once more samples are in, of about 37 letters. Let's hope so.

sheltie, happens all the time. Isn't it just down to that strange quirk (the logic of which I don't understand) that the Autos determine the percentage displayed? In this case the only Auto trade of the day.

Thanks PR. Very useful summary.

I think he's angry at himself but wants to blame someone else?

Yep, buys about to overtake sells soon.

His mind.

... trades today.

Good finish and NAS up 7% ATM.

https://www.theguardian.com/society/2019/sep/11/bacteria-developing-new-ways-resist-antibiotics-doctors-warn

Consider your duty done. Now get lost.

Scott, Ever seen that Alan Partridge programme where he is confronted by a slightly unbalanced (male) fan, pictures of Alan plastered all over his bedroom walls, and has to make a panicked retreat? To quote another comedian, "That's you, that is."

Well yes, his first post today provided us with the reason! I think he needs more training....

No internet in Devon?

Non-sequitur, but anyway. You're right, which is why answers to your question are flooding in.

Yes, lots of people know. They're just not saying. Because people are notoriously reticent on bulletin boards. Obviously.

Well, those looking for 4 million as convincing volume are nearly there.