pbcar / pthelper_archive Goto Github PK

Currently, the bounce criterion calculates the number of items in which price Pn+1 > price Pn. It is this number that is then divided by the total number of items in the purchase task, minus the first item.

However, the Stein et al. criterion states that the subsequent prices should be compared to consumption at the first price (intensity), with only consumption 25% higher than consumption at first price being marked as a 'jump' in consumption.

This means that in some instances, the current bounce criterion was more restrictive than the criterion put forth by Stein et al., as it flagged small increases relative to the price immediately prior as a "jump" in consumption. This also means that individuals with small consumption at intensity who subsequently increase their consumption at higher price points, may be included whereas in the current method, they would not be. This can be altered by changing the formula by Stein et al., to flag any values higher than consumption at first price as a "jump".

A request has been made to display the consumption and expenditure of individuals.

A request has been made to summarize the consumption and expenditure plots by grouping variable.

The current pt_qc() function allows all participants to pass except those whose consumption at intensity (first price) is equal to consumption at the last price. This is equivalent to setting delta Q to a value that only removes those with a delta Q of 0. However, users of the package may want to set their own delta Q value, or including using the value suggested by Stein et al. which is 0.025.

Currently, the package relies on the {beezdemand} package to calculate reversals. The next update will include our own in-house code for identifying and removing reversals from purchase task data.

A request has been made to allow users to choose whether the y-axis is log10 transformed, or is left in its original units.