|
nif:isString
|
-
A survey research firm initially contacted individuals who were representative of San Francisco peninsula residents with respect to sex, income, education, ethnicity, and occupation. Because a different aim of the study focused on aging, subjects were evenly sampled across the life span and screened for dementia (i.e., with Mini Mental scores >26). Seventy-five healthy adults (age range = 20–85) participated (see summary statistics of individual difference variables in Table 1). Written informed consent was obtained from all subjects, and the study was approved by the Institutional Review Board of the Stanford University School of Medicine. An additional seven subjects (not included in the 75 subjects listed above) initially participated, but did not report full socioeconomic, risk preference, and financial data and were excluded from all analyses. Subjects received fixed payment of $20 per hour, as well as cash equivalent to their total earnings in the task. Subjects were also informed that they could lose money on the task, and that any losses they accrued would be deducted from their total earnings. Subjects completed the self-report measures before completing the learning task. To validate self-report measures of assets and debt, credit reports were obtained for approximately half of the sample who agreed to provide their report and for whom credit reports could be obtained after the experiment (n = 37).
Table data removed from full text. Table identifier and caption: 10.1371/journal.pone.0024390.t001 Summary of individual difference variables. The working memory score indexes the number of items that subjects can hold in memory, the cognitive flexibility score represents the additional time required to connect alternating numbers and letters versus sequential numbers, and the numeracy score represents the number of correct answers out of 11 total items. Risk aversion and loss aversion are indices between 0 and 12 that represent the switching point in lottery questions involving choices between sure outcomes and gambles (see Supplementary Methods). Income, debt, and assets are based on ordered categories (e.g., an income rating of 6 corresponds to an average household income of $60,000–$79,000 and a rating of 7 corresponds to $80,000–$99,000; a debt rating of 7 corresponds to $20,000–$39,999 and a rating of 8 corresponds to $40,000–$59,999; and an assets rating of 13 corresponds to $200,000–$499,999).
Monetary Incentive Learning (MIL) Task: Behavioral measures of gain learning and loss learning were elicited with a probabilistic learning task designed to explicitly separate gain and loss conditions (Figure 1). The MIL task was adapted from conventional reinforcement learning tasks [10], [21], [22]. Subjects saw and chose between one of three pairs of fractal cues (gain acquisition, loss avoidance, or neutral) in each run of 12 trials per condition, for a total of 36 trials. After choosing one of the cues from a pair, subjects saw the outcome associated with their choice. On average, one of the cues yielded a better outcome, while the other yielded a worse outcome. In gain cue pairs, the better cue had a higher probability of returning gains (66% +$1.00 and 33% +$0.00) than the worse cue (33% +1.00 and 66% +$0.00); while in loss cue pairs, the better cue had a higher probability of returning nonlosses (66% –$0.00 and 33% –$1.00) than the worse cue (33% –$0.00 and 66% –$1.00). In neutral cue pairs, choice of either cue had no impact on outcomes (100% $0.00). Thus, the only difference between the gain and loss learning conditions involved the valence of the information presented (i.e., gain versus loss). Since the probabilistic learning component was identical across task conditions, differential performance could be attributed to affective gain versus loss framing of different conditions.
Figure data removed from full text. Figure identifier and caption: 10.1371/journal.pone.0024390.g001 Trial structure for Monetary Incentive Learning task gain (top) and loss (bottom) conditions. Within each cue pair, cues appeared randomly and with equal frequency on the left or right side of the screen. The computer randomly assigned each cue to either the better or worse outcome distribution at the beginning of each run in a counterbalanced fashion. Different cue pairs were used for practice and experimental sessions in order to minimize memory-related interference. Subjects were explicitly informed about cue probabilities before the practice session and told to try to maximize their earnings throughout the experiment. Subjects received cash for their performance after the experimental sessions, but not the practice sessions. Measures of gain learning and loss learning performance were assessed by calculating the percentage of choices that matched the “correct” cue (i.e., or had the higher probability of an advantageous outcome) in each condition (see Supplementary Figure S1. Subjects' percentage “correct” choices in the gain and loss conditions (excluding the first trial) were used as primary predictors of life financial outcomes. Based on information that each subject received during the task (i.e., observed outcomes), a measure of “optimal” choice can be computed as the fraction of trials where a subject made the correct ex-ante Bayesian choice (excluding trials in which either option had an equal chance of being optimal, such as the first trial; see Supplementary Methods S1). This “optimal” choice metric was computed for each individual in each condition and used to validate the simpler “correct” choice measure. Supporting the validity of the simpler “correct” choice measures, gain optimal choices (but not loss optimal choices) were associated with gain correct choices, while loss optimal choices (but not gain optimal choices) were associated with loss correct choices (see Supplementary Table S1). Supporting the distinctness of gain and loss learning, gain and loss correct choices were not significantly correlated within subject (r = 0.09, n.s. ). Assets and debt were assessed via self-report in all subjects (n = 75) and validated with credit report information on a subset of subjects (n = 37). Assets were assessed with the question: “What are your approximate current assets? (i.e., portion of home owned, bank accounts, investments, belongings)” using a 16-category ordinal response scale ranging from <+$500.00 in the lowest category to >+$1,500,000.00 in the highest. Debt was assessed with the question: “What are your approximate current debts? (i.e., outstanding home loans, outstanding car loans, outstanding student loans, credit card debt, medical debt)” using a 16 category ordinal response scale ranging from <$500.00 in the lowest category to >$1,500,000.00 in the highest. From credit reports of the subsample, we extracted the overall credit score (also known as the FICO score), which is a proprietary index commonly used in the United States to determine creditworthiness [23]. As expected, the debt-to-asset ratio derived from self reported assets and debt was significantly associated with the credit score (r = −0.48, p<.01) such that subjects with lower debt-to-asset ratios had higher credit scores. We also specifically computed the available credit amount (i.e., the sum of the credit limits of all open accounts) and the percent of credit used (i.e., the sum of credit used divided by the sum of the credit limits of all open accounts) from the information contained in the credit reports. These measures were used to distinguish and validate self-reported assets and debt. Supporting the validity of the self-reported measures of assets and debt, assets were associated with available credit, whereas debt was associated with percent of credit used (see Supplementary Table S2).
Selected neuropsychological tests were administered to assess potential cognitive confounds. The WAIS-III Digit Span Test assessed working memory capacity by requiring subjects to repeat numerical strings forward and backwards. Working memory capacity is highly correlated with and often used to index general intelligence [24]. The Trail Making Test (TMT) assessed cognitive flexibility by requiring subjects initially to sequentially connect circled numbers, and then to connect a series of alternating numbers and letters [25]. Finally, a numeracy inventory (11 items) assessed quantitative skills with basic number problems [26]. Socioeconomic variables including age (years), education (8 ordinal category scale), sex (M/F), ethnicity (open-ended), and income (a 16 level ordinal scale with the same categories used for assets and debt) were also assessed via self-report.
Two sets of questions (12 items each) assessed risk aversion and loss aversion by soliciting subjects' preferences between probabilistic or “risky” gambles and certain or “safe” amounts of money. For both risk aversion and loss aversion measures, a number was assigned (i.e., an integer lower or equal to 12, representing one of the items in descending order) which corresponded to the item on which each subject switched from preferring the safe to preferring the risky option (see Supplementary Methods S2). Neither risk aversion nor loss aversion measures correlated significantly with gain learning or loss learning measures.
Analyses included multiple regression models constructed to test predicted relationships between learning variables and life financial outcomes. Reduced regressions first tested the association between learning variables (i.e., the average of gain and loss % correct choices, gain % correct choices, loss % correct choices) and life financial outcomes (i.e., debt to asset ratio, assets, debt). Full regressions then verified the robustness of these same relationships after controlling for potential socioeconomic (i.e., income, age, education, sex, ethnicity), cognitive (working memory, cognitive flexibility, numeracy), and risk preference (i.e., risk aversion, loss aversion) confounds.
|
![[This page as RDF]](http://data.gesis.org/somesci/static/rdf-icon.gif){kind=icon}