These Data are from an Associated Study, which was approved and implemented as part of Wave 6 and Wave 7 data collection.
Researchers using these data should acknowledge their source in any publications arising from analysis of the data. The citation for the data includes the following information:
“This paper also makes use of risk and time preferences data collected as part of the Future Research Leader project Linking Experimental and Survey Data: Behavioural Experiments in Health and Wellbeing (ES/K001965/1, PI: MM Galizzi), funded by the Economic and Social Research Council”
The study aims to combine survey data from the Innovation Panel with data on risk and time preferences. Risk preference is defined as the attitude for taking a gamble and is operationalized by systematically asking respondents to choose between lotteries yielding different pay-out probabilities and different pay out amounts. Through a series of 18 questions, pay-out amounts and pay-out probabilities were varied thus allowing for risk preference to be quantified.
Time preference is defined as the degree to which time closer to the present is valued more highly than time more distant in the future. It is operationalized by systematically asking respondents to choose between receipt of money closer or further away in time with interest. Through a series of 72 questions, the time differential and interest rate varied thus allowing for time preference to be quantified.
A target sample of around 580 respondents was selected such that only one respondent participated per household. Household within PSUs were randomly selected and then selection of respondents within households was made with a Kish grid of enumerated adults. One-tenth of selected participants were given a payment upon completion of the questions. Among those selected to receive a payment, the amount was based on one of the 91 questions that they answered. If one of the lottery questions was selected, the preferred lottery was played and pay-out was made accordingly. The study was replicated at Wave 7 among the same set of respondents as at Wave 6 but with re-randomization of pay-outs.
At Wave 6, the selection of respondents and items for pay-out are controlled by the variables F_FF_TIMERISKW6, F_FF_TRSEL2, F_FF_TRSEL3, F_FF_TRSEL4, F_FF_TRSEL5, F_FF_TRWINW6, F_FF_TRQW6, and F_FF_TRDW6, on record F_HHSAMP_IP. The variable F_FF_TIMERISKW6 indicates selected households. The variables F_FF_TRSEL2 through F_FF_TRSEL5 were used to create the Kish grid. The variable F_FF_TRWINW6 was randomized within those respondents selected to receive the lottery questions and determine whether they received a pay-out. The variable F_FF_TRQW6 indicates from which question the pay-out was determined. Finally, if a ‘lottery’ question was selected, the variable F_FF_TRDW6 indicates the randomized outcome relevant for determining the lottery pay-out.
Wave 6 substantive variables on record indresp are f_trflag, f_trnotes, f_trpre, f_trprei, f_choice01 through f_choice71, f_riskpre, f_riskpreb, f_trenda, f_choice73 through f_choice91, f_trend, f_selpay, f_dieroll, f_trnowin, f_trwin, f_drawball, f_payout0172, f_payout7390, f_payout91, f_trdie, f_trpayout, f_runtr, f_tmprf3, f_tmprf12, f_tmprf1, f_trriska, f_trhlrisk, f_trflrisk, f_trimpat, f_trimpul, f_trwemwba, f_trsmoker, f_trncigs, f_trevralc, f_tregalco, f_trdklm, f_tr5alcdr, f_trdrnkyr, f_trffdwk, f_trjfd, f_trfrutppd, f_trhlwtr, f_trtrydiet, f_trextype1 through f_trextype97, and f_trpsprt which will appear on record f_indresp_ip.
At Wave 7 the controlling variables are
On record indsamp:
ff_trflag – This is the fed-forward value of the IP6 individual level selection flag calculated in GRIDVARIABLES_IP6. At IP6, households were selected via ff_timeriskw6 but a single individual within each household was selected post enumeration and the value of this individual selection is the variable “TRFLAG”. Since the time individuals are being asked the time-risk preference again in IP7, we are retaining the IP6 individual selection and feeding forward the IP6 value for individuals rather than the household selection mechanism. trflag is Coded 1 = Participant, blank for continuing respondents not selected at IP6, all IP7 new entrants, and the IP7 refreshment sample.
On record HHSAMP:
ff_trwinw7 – Fresh randomization for IP7. Takes the values 1-10 without any value labels, missing for all ff_trflag = blank. Takes values1-10
ff_trqw7 – Fresh randomization for IP7. Takes the values 1-91 without any value labels, missing for all ff_trwinw7 > 1. Takes values 1-91.
ff_trdw7 – Fresh randomization for IP7. Takes the values of 1-10 without any value labels, missing for all ff_trqw7 < 73.
Wave 7 substantive variables on record INDRESP are:
“Time and Risk Preference” module: trpre, choice01-choice91, riskpre, iskpreb, trenda, trend
“CASI Time Preference Control” module: runtr, tmprf3 , tmprf12, tmprf1, trriska, trhlrisk , trflrisk, trimpat, trimpul, trevralc, tregalco, trdklm ,tr5alcdr, trfrutppd, dospertf1, dospertf2, dospertf3, dospertf4, dospertf5, dospertf6, dosperth1, dosperth2, dosperth3, dosperth4, dosperth5, dosperth6
