Increasing saving behavior through age-progressed renderings of the future self (Hershfield et al. 2011)
Hershfield, H., Goldstein, D., Sharpe, W., Fox, J., Yeykelis, L., Carstensen, L., & Bailenson, J. (2011). Increasing saving behavior through age-progressed renderings of the future self. Journal of Marketing Research, 48(SPL), S23–S37.
By clicking this link, you are leaving CLEAR and are subject to the privacy and security policies of the owners/sponsors of the external site. CLEAR does not control or guarantee the accuracy, relevance, timeliness, or completeness of information contained in a linked site. We also do not endorse the organizations or individuals maintaining sites that we link to, any views they express, or any products/services they offer.
When last checked, this publication was not available for free online. You may be able to locate it by conducting a Google search for the citation.
- The study’s objective was to determine whether contact with an age-progressed visual representation of the self led participants to allocate more money to savings for retirement. The authors presented four random assignment studies relying on varying degrees of contact with these age-progressed visual representations in an attempt to pinpoint their effect on retirement savings behavior and the mechanism through which such representations alter savings attitudes.
- In the four experiments discussed here, the authors collected data directly from participants by administering post-intervention surveys or collecting hypothetical asset allocation information.
- The study found that participants who saw age-progressed pictures of themselves chose to allocate a significantly higher proportion of hypothetical income to retirement compared to a control group; the participants in this study were drawn from a group of adults ages 18 to 35. Results were similar in three other experiments using undergraduate students.
- The quality of the causal evidence presented in this study is high. This means we are confident the differences in outcomes observed between treated and control groups resulted from the interventions tested, and not other factors.
The authors conducted four related experiments examining the effect of interactions with age-progressed virtual representations of the self on participants’ hypothetical savings and spending behavior. In each study, participants were randomly assigned to one of two research groups: participants who were assigned to the treatment group interacted with an age-progressed digital version of themselves, and participants assigned to the control group interacted with a digital version of their current appearance.
In the first experiment, participants in the treatment group conversed with other avatars through virtual versions of themselves transformed to appear around age 70. Participants in the control group performed the same tasks with non-aged virtual versions of themselves. Immediately after performing tasks in the virtual reality environment, respondents were asked to allocate $1,000 among four options: buying a nice gift for someone special, planning a fun occasion, saving for retirement, and saving in a checking account.
In the second experiment, treatment and control participants interacted with digital avatars as in the first experiment. Rather than immediately performing an asset allocation task, the participants completed three short surveys after a short time delay. The first required them to choose between an immediate payout and a short-term payout, defining short-term as occurring in 10 to 75 days. The second required participants to choose between immediate and long-term payouts, where a long-term payout would occur 35 to 45 years from survey administration. The final task was a retirement spending questionnaire asking participants to choose between various combinations of monthly budgets during their working and retired lives.
In the third experiment, researchers photographed participants making happy, sad, or neutral expressions, and created a continuum of 11 facial expressions based on these images. Three to four weeks after submitting photographs, participants completed an online task in which they allocated money to current consumption and consumption in retirement. Participants in the treatment group saw pictures of their face, age-progressed to appear about age 65, making happier faces as the allocation of salary to retirement increased and sadder faces as the allocation to retirement decreased. Participants in the control group observed the same phenomena based on their current (non-aged) appearance, but these faces were happier with less allocated to retirement.
In the fourth experiment, participants completed a similar task but only saw a single picture of their face, with a neutral expression. The treatment group saw an aged picture and the control group saw the current picture.
- In the first study, undergraduate students in the intervention group allocated more funds to the retirement account. However, this result was not statistically significant at standard levels using a two-sided test.
- In the second study, students in the intervention group exhibited a greater tendency to delay consumption or receipt of funds when measures were tested jointly, although the differences were not individually statistically significant at standard levels using a two-sided test.
- In the third study, undergraduate students in the intervention group exhibited a tendency to allocate a significantly higher percentage of pay to retirement.
- In the fourth study (the only experiment conducted on a working population), participants in the intervention group allocated a significantly higher percentage of pay to retirement than the control group, 6.17 percent vs. 4.41 percent.
Considerations for Interpreting the Findings
The first three experiments used data from undergraduate students, with each containing 21 to 50 participants. The fourth used data from 40 adults across the United States who were recruited to participate in the study through Amazon Mechanical Turk. Because none of the experiments’ participants were currently in the workforce full time, their actions may not reflect the actions of employed adults (the population of interest for this topic area). In addition, the results in this study should be interpreted carefully, as the outcomes analyzed relate to hypothetical, and not actual, savings behavior.
Causal Evidence Rating
The quality of causal evidence provided in this study is high because all four experiments were well executed. This means we can be confident the estimated effects are attributable to the interventions, and not to other factors.