The disposition effect in a scopic regime: Data from a laboratory experiment☆

This article presents a new dataset on the disposition effect in a scopic regime, collected in a laboratory experiment reported in “Framing and the disposition effect in a scopic regime” [6]. 81 subjects were recruited and asked to participate in an incentivized stock trading game insprired by Weber and Camerer [7], which was computerized, programmed using oTree [1]. Subjects were able to monitor the trading performance of randomly selected peers in comparison to their own performance. Two sets of rankings were employed to display the comparisons. The data allow to test theories on the impact of social comparisons on investors’ decision making.


Value of the data
• The data can be used to test theories on the impact of social comparisons on decision making or to set a benchmark for formulating a new theory of decision making in a scopic regime. • The data can be used to test existing disposition effect theories.
• This data can be combined with other experimental data on the disposition effect to find undiscovered findings.

Data description
The data is stored in one file, containing the entire raw data collected during the experiment. The data contain observations from an incentivized experiment, conducted in June 2019 at the BaER-Lab at Paderborn University. 81 subjects were recruited through ORSEE [4] . The experiment was programmed using oTree [1] . The subjects were incentivized as we paid cash proportional to the profit earned in the experiment.
Participants completed the multi-period stock trading task inspired by Weber and Camerer [7] . The data contain every decision of investors across periods. Participants were able to monitor the performance of two randomly selected peers during the experiment and compare their peers' performance to their own. Two different types of ranking were employed to frame the comparison with peers. The data contain detailed information on the randomly composed peer groups and the type of ranking per treatment condition. The dataset also includes information on the participants' risk preferences [2,5] and loss aversion [3] as well as standard demographics such as age and gender.
The data can be found online at https://data.mendeley.com/datasets/jfg8s32xdm . Definitions of all variables in the dataset are given in Table 1 .

Experimental design, materials and methods
A total of six sessions of the experiment were conducted. Each session was conducted in two parts. Part I included the trading task, part II included the additional collection of participants' demographics, risk preferences, and loss aversion.
In part I, a total of 14 periods of the trading task based on Weber and Camerer [7] were run. Subjects were endowed with fictional wealth of 10,0 0 0 Talers. With their endowment, subjects were able to trade six different stocks with uncertain return characteristics over 14 pe-  [5] player.holt_laury"j" Participant's decision on "j"th lottery (j = 1 to 10) Risk preference task [2] player.lottery Participant's selected lottery Loss preference task [3] player.coin_"j" Participant's decision on "j"th coinflip (j = 1 to 10) Initial stock holdings "i" player.quantity_pre_"i" Quantity of stock "i" at the beginning of period; i = A,B,C,D,E,F Stock sales "i" player.quantity_sell_"i" Quantity of sold stocks "i" Stock purchases "i" player.quantity_buy_"i" Quantity of bought stocks "i" Final stock holdings "i" player.quantity_"i" Quantity of stock "i" at the end of period Price "i" player.price_"i" Price of stock "i" Portfolio value player.result_final Participants' portfolio value Guess task player.total_count_guess Sum of correct stock type guesses riods. Stock prices followed distinct random processes and were not influenced by investment decisions of participants. The six stocks represented different stock types with different chances to increase their prices. The stock with the best expected performance had a 65% chance of a positive price change; the second best stock had a 55% chance of a positive price change; two neutral stocks had even probability of a positive or negative price change; the fifth stock had a 55% chance of a decrease in stock price; the stock with the worst expected performance had a 65% chance of a decrease in stock price. Participants were informed about the existence of different stocks and their characteristics, but did not get any information which stock possesses which characteristics. After the direction of stock price changes had been determined, as second stochastic process determined the magnitude of the price change. Both stochastic processes were independent. Price increases were measured in terms of absolute values and could be either 1, 3, or 5 Talers. Participants were be informed about the two-stage pricing process before the experiment. Participants received information on the stock prices of four periods (-3, -2, -1, 0) before period one started. To evaluate whether subjects had a good understanding of the stock types, they had to guess the stock types after periods seven and fourteen. After rounds five, ten and fourteen, participants were shown rankings with their own performance and the performance of two randomly selected peers. All three rankings were shown with respect to the same peers. Importantly, rankings did not have any implications for participants' payoffs and this fact was pointed out to participants. In treatment condition spt1 , participants were ranked based on their percentage of profitable trades. The fraction of profitable trades denoted the percentage of trades that were closed at a price, which was higher than the entry price. In addition, the rankings information displayed the raw performance (current portfolio value) of participants. In treatment condition spt2 , participants were ranked based on their raw performance (current portfolio value). No information about the percentage of profitable trades were given in this treatment condition.
In part II of the experiment, information on participants' gender, age, university major, stock market experience, and major were collected. Additionally, participants had to complete the risk elicitation tasks of Eckel and Grossman [2] and Holt and Laury [5] as well as the loss aversion elicitation task of Gächter et al. [3] . The risk elicitation tasks and the loss aversion elicitation task were incentivized.