Log in
Register|
graveendranath ☆ 2009-03-18 06:18 (6298 d 22:19 ago) Posting: # 3376 Views: 6,624 |
|
|
Dear sir, For ANOVA calculation some articles showing 10% level of significane for testing sequence effect, 5% for Period and Treatment effects and some articles showing 5% for all sequence, period and treatment effects. I have confusion which one is correct? Please suggest me the correct Level of Significance for Sequence. |
|
sinhmar ☆ India, 2009-03-21 12:31 (6295 d 16:07 ago) @ graveendranath Posting: # 3388 Views: 5,516 |
|
|
❝ For ANOVA calculation some articles showing 10% level of significane for ❝ testing sequence effect, 5% for Period and Treatment effects and some ❝ articles showing 5% for all sequence, period and treatment effects. ❝ I have confusion which one is correct? Please suggest me the correct Level ❝ of Significance for Sequence. Dear graveendranath, Both are correct it's on our wish, how much stringent we want to be. Generally, in BE studies as per regulatory requirement, to test the carryover effect in 2 way crossover design, the sequence effect is tested at the 10% level of significance using the subjects nested within sequence mean square as the error term. And all other main effects(treatment, period etc.) are tested at 5% level of significance against the residual error(MSE) from ANOVA model as error term. Sinhmar |
|
KR ★ India, 2009-03-28 06:46 (6288 d 21:52 ago) @ sinhmar Posting: # 3407 Views: 5,476 |
|
|
sinhmar ☆ India, 2009-03-30 13:53 (6286 d 15:45 ago) @ KR Posting: # 3414 Views: 5,399 |
|
|
Dear KR ❝ If we use different significance level for difference effects then how can we do this using PROC MIXED or PROC GLM using sas? Using PROC MIXED or PROC GLM in SAS we can calculate F value and P-value for different effects and we need level of significance to find tabulated F value which we can find from F-distribution tables. To check that effect is significant or not, we see calculated F-value against tabulated F- value (at level of significance= 0.05 or 0.10 or...) or if we have SAS output we simply see level of significance against p-value of the effect. Regards, Sinhmar |
Ing. Helmut Schütz