Skip to contents

Planning with the point estimate of \(\rho_{Yf}\) from a small pilot gives an underpowered study about half of the time, because the estimate is as likely to be too high as too low. This function returns the one-sided Fisher \(z\) lower bound $$\tanh\{\operatorname{atanh}(|\hat\rho|) - z_{c}/\sqrt{m - 3}\},$$ truncated at 0 and carrying the sign of \(\hat\rho\), where \(m\) is the number of labeled pairs in the pilot and \(c\) is the confidence level. Square it to get a lower bound for \(R^2\).

Usage

rho_lower_bound(rho_hat, pilot_n, confidence = 0.9)

Arguments

rho_hat

Pilot estimate of the correlation between the outcome and the predictions.

pilot_n

Number of labeled pairs in the pilot (more than 3).

confidence

One-sided confidence level (default 0.9).

Value

The lower bound, with the sign of rho_hat.

Details

The power functions power_ppi_mean(), power_ppi_paired(), and power_ppi_ttest() apply this bound when a pilot size is given.

Examples

rho_lower_bound(0.7, pilot_n = 50)
#> [1] 0.5917579
rho_lower_bound(0.7, pilot_n = 50)^2   # lower bound for R^2
#> [1] 0.3501775