Tips

[gdcumming]

0.1.9

Line 2: precision for planning: "Find the smallest N that will give a 95% CI with MoE no more than chosen target MoE, either on average, or with 99% assurance"

Tab 1: "Two Independent Groups Design, each group of size N"

Tab 2: "Paired Design", with group of size N"

1 Target MoE of 95% CI: "Use large slider below figure to choose a value for target MoE" [target MoE value]: "Value of target MoE, marked by position of cursor" ...set by slider below figure: "Use large slider below figure to choose a value for target MoE"

2 N curves: "Curves show the relation between N and target MoE" (o) N curve for target MoE, on average: "Black curve shows how N varies with target MoE on average" (o) N curve for target MoE, with 99% assurance: "Red curve shows how N varies with target MoE with 99% assurance (grey curve: on average)"

3 Display values: "Turn on or off the display of N values at each point on displayed curve(s)" Display all N values [..]: "Turn on or off the display of N values at each point on displayed curve(s)"

4 Truncate display at target MoE value of: "Use slider to choose where to truncate the left end of curve(s)" [slider]: "Select a value within 0.05 and 0.30" [value]: "Left-truncate curve(s) at this value"

[gfmoore]

I think we need something for correlation in paired?

[gdcumming]

0.1.12

Sorry for the rho lacuna!

The long statement "Find the smallest..." I'd intended as a tip on 'precision for planning', but I prefer it the way you have done it! stet

Tweaks please:

Ind Groups tab:

Panel 2: All three tips, italic N

Trunc slider: "... between 0.05 and 0.30" is better than "within". My error.

Paired Design tab: Remove double quote. My error, sorry.

2 Population correlation: "Use slider to set population correlation ρ" Correlation ρ: "Use slider to set population correlation ρ" Slider: "Set ρ between .00 and .99"

Panels 3, 4, 5: Same as for Panels 2, 3, 4 of Ind Groups tab.

[gfmoore]

Done.

[gdcumming]

0.1.17 Thanks, closing...