Prepping Data for Z Curve (Best Approach?)

[peclayson]

Hi @FBartos,

I have extracted test test statistics, degrees of freedom, and test values in preparation for a zcurve analysis. I'm trying to understand what the "best" approach is to prepare the statistics for the z curve. The data have been formatted to be put into zcurve_data.

The function provides the precise p values. Then I can use zcurve(p = df_zcurve_data$precise$p) to run a z-curve analysis. Does this make sense?

I couldn't find an example of using zcurve_data and then the approach for passing this to zcurve, so I wanted to double check.

Thanks for your help! Peter

[FBartos]

Hi Peter,

it really depends on exactly what information you have at hand. If you can compute p-values for each data point directly, then there is nothing stopping you from using zcurve(p = p). If you have some censored p-values, then using zcurve(data = df_zcurve_data) might be advantageous (extracting the precise p-values from the zcurve_data object directly does not make that much sense to me though---maybe there is something that I'm missing?) .

Cheers, Frantisek

[peclayson]

I'm sure you're not missing anything. I likely just confused both of us.

I have 100+ observations of test statistics and test values (no p values though), formatted as follows: c("F(2,38)=0.789", "F(2,46)=7.55", "F(1,61)=8.33", "F(1.83,49.4)=3.32", "F(1,52)=4.83", "chi(1)=28.72") etc.

Since I do not have z scores or p values, I am using the zcurve_data function to calculate the p values and then plug those into zcurve for analysis.

zcurve_prepped <- zcurve_data(zcrv_stats, stat_precise = 1)

fit <- zcurve(data = zcurve_prepped)

That is the intended application data = zcurve_prepped, correct? Not zcurve(p = zcurve_prepped$precise$p)?

I apologize. I think I confused myself thinking I need to operate directly on the p values, rather than using data = zcurve_prepped.

Peter

[FBartos]

Hi Peter,

No worries; this is indeed the function's intended use!

I will close it as solved. Please, reopen it if you have any further questions. Cheers, Frantisek