Methods for Increasing Caregiver Training: Evidence from a Randomized Controlled Trial

How to cite

Cronin, Christopher; Lieber, Ethan. 2021. "Methods for Increasing Caregiver Training: Evidence from a Randomized Controlled Trial." AEA RCT Registry. AEARCTR-0007156. https://doi.org/10.1257/rct.7156-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Christopher Cronin* , Ethan Lieber

* Primary investigator

Registry ID
AEARCTR-0007156
DOI
https://doi.org/10.1257/rct.7156-1.0
First registered
Last updated
Status
In development
Trial dates
2020-10-08 – 2021-10-02
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/7156

Abstract

This randomized control trial will determine whether financial incentives can be used to encourage caregivers to participate in online training programs. We will partner with Helping Restore Ability (HRA), a non-profit that supports thousands of caregivers in Texas, as they roll out a new online training program to improve in-home care quality. To encourage training take-up, treated individuals will receive emails offering a financial incentive for completing an online training program. Members of the control group will receive a standard email with no incentive for training completion. This study will determine the impact of this financial incentive on participation and performance in the training program.

Experimental design

This study will operate as a randomized-controlled trial. Approximately 750 caregivers will be randomly assigned to one of two groups: a control group, which will receive a standard email invitation to complete the training and a monetary compensation group, which will receive $50 for completing the training.

We also want to determine how many caretakers actually read the email we are sending. To determine this, we will include a line at the bottom of some emails that offers the caretaker a $10 gift card if they simply reply to the email. This line will be included in 150 randomly selected individuals in each treatment arm (i.e., 300 individuals total).

Design details

Randomization method
In this study, we use a rerandomization method in which all participants are randomized to treatment assignments 100 times. Of these 100 randomized datasets, we select the randomization assignment dataset that yields the minimum of the maximum t-statistic over select variables to ensure balance across treatment arms. These variables are gender, race, tenure with client, age, relation with client, tenure with HRA, employee relationship with HRA (Personal Attendant Services [PAS] or Consumer Directed Services [CDS]), an indicator variable for whether caregiver was on HRA payroll in September 2020, an indicator variable for whether caregiver is paid through Amerigroup insurance, indicator variables for whether (and how) the caregiver completed a pre-analysis survey, and an indicator for whether the caretaker used a training token (i.e., logged into the training platform) during a prior experiment offering training opportunities. This approach is described in detail by Banerjee et al. (2020). Each of the variables used in the rerandomization process will be controlled for during analysis, as suggested in Bruhn and McKenzie (2009).
Randomization unit
Randomization will occur at the individual-level, using rerandomization method.
Clusters
747 caretakers
Observations
747 caretakers
Arms
224 control, without $10 reply incentive 150 control, with $10 reply incentive 223 treatment, without $10 reply incentive 150 treatment, with $10 reply incentive
Primary outcomes
The primary outcomes of this experiment are: intervention take-up, training program performance, and training program engagement.
Secondary outcomes
Did caretaker "reply" to the email.

Cited by

No citation data available for this registration yet.

Keywords