Question on statistics

Dear all, I am now having some questions on statisics. Anyone of you know how to interpret the Confidence Interval when using paired sample t-test?
Example: if the score of the pre and post test of a treatment was found to have statistically significant difference by using paried sample t-test, say p<0.05 (95%CI, -6.258, -1.235). How to interpret the upper and lower (?) for that case?
1) is there any case that having negative upper and lower limits?
2) what is the meaning of the difference between the upper and lower limits, is it larger represent better, or vice versa?
3) is there any meaning for the actual number of the upper or lower limits, is it smaller (close to zero) represent poor effectiveness of the treatment?
Any good reference to read about that?

That's an excellent book:

http://www.amazon.com/Statistical-Methods-Health-Care-Research/dp/0781748402

May be Andrew is the best person to help. I recall a little bit but can't be sure!!

Eddie 寫到:

Dear all, I am now having some questions on statisics. Anyone of you know how to interpret the Confidence Interval when using paired sample t-test?
Example: if the score of the pre and post test of a treatment was found to have statistically significant difference by using paried sample t-test, say p<0.05 (95%CI, -6.258, -1.235). How to interpret the upper and lower (?) for that case?
1) is there any case that having negative upper and lower limits?
2) what is the meaning of the difference between the upper and lower limits, is it larger represent better, or vice versa?
3) is there any meaning for the actual number of the upper or lower limits, is it smaller (close to zero) represent poor effectiveness of the treatment?
Any good reference to read about that?

無人答就等我亂講吓。 首先知道p<.05即reject null hypothsis囉; 95% CI 即喺Mean +/- 1.96xSE(standard error of mean), +即喺upper limit, 减就喺lower limit, 所以lower limit 可以負數。upper limit 喺負數都有机會掛，如果個mean 喺負數，但我唔清楚，等高人再指點。Upper to lower limit 呢個range代表咗個mean 可能出現喲範圍，應該越細越好??死火，講講吓都有喲亂，都喺睇吓書先 ，不過effectiveness睇個p咪ok囉，至於有幾effective, 就要計吓個effect size 啦。

Colin 寫到:

Eddie 寫到:

Dear all, I am now having some questions on statisics. Anyone of you know how to interpret the Confidence Interval when using paired sample t-test?
Example: if the score of the pre and post test of a treatment was found to have statistically significant difference by using paried sample t-test, say p<0.05 (95%CI, -6.258, -1.235). How to interpret the upper and lower (?) for that case?
1) is there any case that having negative upper and lower limits?
2) what is the meaning of the difference between the upper and lower limits, is it larger represent better, or vice versa?
3) is there any meaning for the actual number of the upper or lower limits, is it smaller (close to zero) represent poor effectiveness of the treatment?
Any good reference to read about that?

無人答就等我亂講吓。 首先知道p<.05即reject null hypothsis囉; 95% CI 即喺Mean +/- 1.96xSE(standard error of mean), +即喺upper limit, 减就喺lower limit, 所以lower limit 可以負數。upper limit 喺負數都有机會掛，如果個mean 喺負數，但我唔清楚，等高人再指點。Upper to lower limit 呢個range代表咗個mean 可能出現喲範圍，應該越細越好??死火，講講吓都有喲亂，都喺睇吓書先 ，不過effectiveness睇個p咪ok囉，至於有幾effective, 就要計吓個effect size 啦。

多謝兩位大師兄, 以你所講, 其實Upper to Lower limits係咪指mean difference可能出現嘅範圍呢(因為我而家compare 緊pre & post 嘅mean)?

i only know that Confidence interval does not include 0 that again confirms the significant results. Peggo Lam only teach this

Adam

Eddie 寫到:

Colin 寫到:

Eddie 寫到:

Dear all, I am now having some questions on statisics. Anyone of you know how to interpret the Confidence Interval when using paired sample t-test?
Example: if the score of the pre and post test of a treatment was found to have statistically significant difference by using paried sample t-test, say p<0.05 (95%CI, -6.258, -1.235). How to interpret the upper and lower (?) for that case?
1) is there any case that having negative upper and lower limits?
2) what is the meaning of the difference between the upper and lower limits, is it larger represent better, or vice versa?
3) is there any meaning for the actual number of the upper or lower limits, is it smaller (close to zero) represent poor effectiveness of the treatment?
Any good reference to read about that?

無人答就等我亂講吓。 首先知道p<.05即reject null hypothsis囉; 95% CI 即喺Mean +/- 1.96xSE(standard error of mean), +即喺upper limit, 减就喺lower limit, 所以lower limit 可以負數。upper limit 喺負數都有机會掛，如果個mean 喺負數，但我唔清楚，等高人再指點。Upper to lower limit 呢個range代表咗個mean 可能出現喲範圍，應該越細越好??死火，講講吓都有喲亂，都喺睇吓書先 ，不過effectiveness睇個p咪ok囉，至於有幾effective, 就要計吓個effect size 啦。

多謝兩位大師兄, 以你所講, 其實Upper to Lower limits係咪指mean difference可能出現嘅範圍呢(因為我而家compare 緊pre & post 嘅mean)?

大セ鬼師兄，"學無前後，達者為先"。我見好多比我junior的OT，都唔知醒過我多多聲，反而好多老鬼及老老鬼(包括我在內) ，成日知少少或唔知， 扮代表 。講翻t test，經我亂嗡，可能使你更confuse, 現作以下補充：

P value and CI 均可用作hypothesis testing，有時p>.05，無significant difference，可以用CI睇吓有無得補救。For CI，可以幫我地知道"Given the observed difference between Tx & control groups, what is the plausible range of difference within which the true difference may lie"。
CI 95% = Mean (diff) +/- CV x SE(diff)
where critical value (CV) = t-value at .05 level with n degree of freedom
當CI 95% includes zero, then the mean difference is not different from zero. In that case the 2 means are not significantly different from each other.(the null hypothesis is: Tx group mean - control group mean = 0) Your cases may have similar intepretation
總結，我都喺知少少，請大家多多指教

Colin 寫到:

Eddie 寫到:

Colin 寫到:

Eddie 寫到:

Dear all, I am now having some questions on statisics. Anyone of you know how to interpret the Confidence Interval when using paired sample t-test?
Example: if the score of the pre and post test of a treatment was found to have statistically significant difference by using paried sample t-test, say p<0.05 (95%CI, -6.258, -1.235). How to interpret the upper and lower (?) for that case?
1) is there any case that having negative upper and lower limits?
2) what is the meaning of the difference between the upper and lower limits, is it larger represent better, or vice versa?
3) is there any meaning for the actual number of the upper or lower limits, is it smaller (close to zero) represent poor effectiveness of the treatment?
Any good reference to read about that?

無人答就等我亂講吓。 首先知道p<.05即reject null hypothsis囉; 95% CI 即喺Mean +/- 1.96xSE(standard error of mean), +即喺upper limit, 减就喺lower limit, 所以lower limit 可以負數。upper limit 喺負數都有机會掛，如果個mean 喺負數，但我唔清楚，等高人再指點。Upper to lower limit 呢個range代表咗個mean 可能出現喲範圍，應該越細越好??死火，講講吓都有喲亂，都喺睇吓書先 ，不過effectiveness睇個p咪ok囉，至於有幾effective, 就要計吓個effect size 啦。

多謝兩位大師兄, 以你所講, 其實Upper to Lower limits係咪指mean difference可能出現嘅範圍呢(因為我而家compare 緊pre & post 嘅mean)?

大セ鬼師兄，"學無前後，達者為先"。我見好多比我junior的OT，都唔知醒過我多多聲，反而好多老鬼及老老鬼(包括我在內) ，成日知少少或唔知， 扮代表 。講翻t test，經我亂嗡，可能使你更confuse, 現作以下補充：

P value and CI 均可用作hypothesis testing，有時p>.05，無significant difference，可以用CI睇吓有無得補救。For CI，可以幫我地知道"Given the observed difference between Tx & control groups, what is the plausible range of difference within which the true difference may lie"。
CI 95% = Mean (diff) +/- CV x SE(diff)
where critical value (CV) = t-value at .05 level with n degree of freedom
當CI 95% includes zero, then the mean difference is not different from zero. In that case the 2 means are not significantly different from each other.(the null hypothesis is: Tx group mean - control group mean = 0) Your cases may have similar intepretation
總結，我都喺知少少，請大家多多指教

Sorry, 你個case 應該有significant different，因為無includes zero

Copied from a web http://www.stats.gla.ac.uk/steps/glossary/confidence_intervals.html

Confidence Interval for the Difference Between Two Means
A confidence interval for the difference between two means specifies a range of values within which the difference between the means of the two populations may lie. These intervals may be calculated by, for example, a producer who wishes to estimate the difference in mean daily output from two machines; a medical researcher who wishes to estimate the difference in mean response by patients who are receiving two different drugs; etc.


The confidence interval for the difference between two means contains all the values of µ1 - µ2 (the difference between the two population means) which would not be rejected in the two-sided hypothesis test of:
H0: µ1 = µ2
against
H1: µ1 not equal to µ2
i.e.
H0: µ1 - µ2 = 0
against
H1: µ1 - µ2 not equal to 0

If the confidence interval includes 0 we can say that there is no significant difference between the means of the two populations, at a given level of confidence.

The width of the confidence interval gives us some idea about how uncertain we are about the difference in the means. A very wide interval may indicate that more data should be collected before anything definite can be said.

We calculate these intervals for different confidence levels, depending on how precise we want to be. We interpret an interval calculated at a 95% level as, we are 95% confident that the interval contains the true difference between the two population means. We could also say that 95% of all confidence intervals formed in this manner (from different samples of the population) will include the true difference.

咁有冇人知道點解include zero就reject唔到null hypothesis 呢

Colin 寫到:

咁有冇人知道點解include zero就reject唔到null hypothesis 呢

因為如果upper limits同lower limits 之間有0的話, 即係兩組mean嘅difference有正有負, 亦即係話有時mean(a)大過mean(b), 有時mean(a)細過mean(b), 咁即係話兩個mean互有高低, 唔可能統一地話(a)好過(b)(or vice versa), ie. null hypothesis cannot be rejected.

Colin 寫到:

咁有冇人知道點解include zero就reject唔到null hypothesis 呢

我答你唔係因為我明白你地討論緊嘅嘢, 只係用邏輯思維答你個問題。

如果H0: µ1 - µ2 = 0 , 咁mean difference includes 0 咪即係話呢條算式係啱, 咁又點會被reject呢?

Hello, 各位 Ching, 新人報到
路過此地，見到好多熟人。

95% CI include 0 is not sign. apply to odd ratio (OR), relative risk (RR)only.

95% CI imply you have 95% of probability to include the mean at this interval range, that so call upper limit and lower limit but actually not.
e.g. if mean is 65 and 95%CI = 65 +/- 1.96SE, so within this interval range, you have 95% of confidence to include this mean 65.
Therefore, in general, no matter they include 0 or not, it need more information to use this value to reject the null. (1.96 is for 95% only, if 99% or 90%, the constant is other number, which is accordign to normal distribution).

For comparing 2 group difference, if 95% CI overlap, the 2 groups most likely are not significant different, but if their 95%CI did not overlap, the two group have chance to be statistically significnat different.

In OR or RR condition, as they are ratio, if 95% CI interval include the OR or RR, but it include 0, it imply the OR or RR MAY BE 0, so should be non significant.

Eddie 寫到:

Dear all, I am now having some questions on statisics. Anyone of you know how to interpret the Confidence Interval when using paired sample t-test?
Example: if the score of the pre and post test of a treatment was found to have statistically significant difference by using paried sample t-test, say p<0.05 (95%CI, -6.258, -1.235). How to interpret the upper and lower (?) for that case?
1) is there any case that having negative upper and lower limits?
2) what is the meaning of the difference between the upper and lower limits, is it larger represent better, or vice versa?
3) is there any meaning for the actual number of the upper or lower limits, is it smaller (close to zero) represent poor effectiveness of the treatment?
Any good reference to read about that?

１）　我想個result即係signitficant la (因為p<0.05){少過5% the result is due to chance}

２）　正如ｋｏｎｇ　所講你過ｍｅａｎ　有９５% of chance fall into the range of -6.258 to -1.235
個ｒａｎｇｅ 應該同你個Ax. tool 有關
　　　好似ｒａｎｇｅ　太大　可能是你ｄ　subject 等分數差得遠／subject 太
　　　少
　　　比如你個ｔｏｏｌ　有　１－１０　分，　９５％CI, 1.4, 9.6) 就好似有
　　　ｄ問題ｌａ


３）　　http://web.uccs.edu/lbecker/SPSS/confintervals.htm