In the "Poker Experiment," 144 participants played a simplified poker-like game in pairs. The games rules are as follows: first each player places an initial "ante" bet, next each player is dealt two cards, the players bet who has the highest card and are given the opportunity to fold. The players compete for a common pot and their implicit goal is to maximize their personal reward. Players play two rounds of the game, a practice and one which determines their compensation. In all cases, players make use of a web-based interface to play the game.

The 72 pairs of participants were divided into two groups of 36 depending upon the motivator to which they were assigned. These groups were evenly split between Charity Gains and Charity Loses motivators. In the Charity Gains motivator, a charity of the participants choice received a reward equal to the amount the participant won in the poker game. In contrast the Charity Loses motivator gave a reward equal to the amount the participant lost in the poker game. There were also 3 conditions: no sensors, a game involving sensors on one player, used to present information related to emotion to the other player, and a game in which one card was shown by one player to the other. The latter two conditions were deliberately designed to be unfair to one player.

As with the Quiz and Interview Experiments, in the Sensors condition one subject was wired using 3-lead electrocardiogram sensors and the HandWave skin conductance sensor. Additionally, mouse pressure and coordinates were collected. A face-tracking camera was also used to collect video of the participant's facial expressions. Participants with sensors in this condition also encountered an extra screen explained how the sensors would be attached. These subjects were paired with a second interviewee subject who received information from these sensors using the MixedEmotions system described in the apparatus section.

The No Sensors and Visible Card conditions performed the same task and experienced the same questionnaire instruments as the Sensors condition. The difference was the absence of sensors and information related to attaching sensors.

In all conditions, participants were told: "You are about to play a poker-like game. You will begin with a practice round. Following this you will play a real round in which your compensation will be determined. The player who has the highest card wins (i.e. Ace beats King, Jack beats 10). If two players have the same high card, then the highest suit wins (Clubs beats Spades beats Hearts beats Diamonds). The winner gets the pot plus any of their own remaining chips which were not put into play. The loser gets only their own remaining chips which were not put into play. During the game players may: raise bets (counter an opponent's bet by betting more), call a bet (answer an opponent's bet by putting in an equal amount, ending the game), or fold (forfeit the game, keeping whatever chips were not put into play). Your goal is to win as much money as possible. When the game starts, each player automatically places $1.00 into a common pot. In addition, you have $5 to bet. You may bet as much or as little as you like, depending on the hand you are dealt."

Taking a pessimistic view, we will consistently call the player in the room with the sensors the "disadvantaged" player, even though there is no obvious disadvantage in the No Sensors condition when the player is merely accompanied by the sensing apparatus and not wearing it or communicating information through it. More disadvantage is hypothesized when that player is in the Sensor condition (communicating affective information to their opponent) and when that player is in the Card Visible condition, where the opponent can see one of their cards.

Disadvantaged participants in the Charity Gains motivator were additionally told that the charity of their choice "will be given funds that match your chips in the case that you are the winner." Disadvantaged participants in the Charity Loses motivator, on the other hand, were told that the charity of their choice "will be given funds that match your remaining chips in the case that you lose" and (as a consequence) nothing if they won. These motivators were identical in conditions both with and without sensors.

Disadvantaged players were placed in the office containing the sensing apparatus. The sensors were attached to participants only in the Sensors condition. Their opponent was seated in a similar office, but without any sensing apparatus. By asking the experimental subjects for their email address I was able to consult the random assignments of the experiment scheduling system, which had determined which subject should be seated in which office. At this point I would launch web browsers to view the experimental instructions.

Figure 4.1. Initial page of the Poker Experiment

The first web page encountered by subjects in the poker experiment was used to verify their demographic information. Disadvantaged subjects were also asked to chose a "favorite charity you'd like to see a donation made toward." Subjects in the sensor condition were also asked to choose the "hand with which you use the mouse." After the consent form was signed and the identity of the participant verified, subjects were instructed to move on to the next page.

Figure 4.2. Sensor preparation information (disadvantage)

In the case that participants were assigned to a sensor condition, they were shown a page that informed them that sensors would be used. In the case that subjects were assigned to the No Sensors condition, the system skipped forward to a page showing the instructions for their task.

Figure 4.3. Your task (disadvantage, Charity Loses motivator)

After viewing the sensor preparation information, a page displaying the instructions for the disadvantaged player was displayed. This page explained the game rules to participants and also the compensation they would receive (which was dependent upon their motivator).

Figure 4.4. Your task (advantage, same across two motivators)

Individuals paired with participants at a disadvantage also received similar instructions on their task. The instructions differed by not including any information about a charity. This page also explained the game rules and compensation.

Figure 4.5. Your task - sensors (advantage)

In the case that the participant at an advantage was paired with a disadvantaged participant in the sensor condition, the participant with the advantage received additional information about how to interpret the data coming from MixedEmotions. This took the form of instructions comparing a "calm" state with a "stressed" state.

The goal of these instructions was to help the advantaged participant spot disadvantaged participants who might appear to be stressed. For instance, a participant who is bluffing about their cards might change facial expressions, exhibit irregular heart rate intervals, and an increase in skin conductance. In practice, many of these signals are noisy and even throughly trained polygraph tests may have difficulty perfectly interpreting such data.

Please note that there was an error in these instructions informing interviewers that candidates with "irregular heart rate" may be more stressed. Quite the contrary, normal resting individuals often show irregular heart rates. This should be corrected in future uses of this protocol.

The experimenter would then verbally verify that both participants had understood the instructions. The experimenter would also explain to both participants in their respective office how to contact the experimenter once the experiment was completed. At this point the experimenter would leave the participants to play the practice and real poker-like games.

Figure 4.6. Your task - sensors (advantage)

At this point, one of the two participants was randomly chosen as the dealer for both the practice and actual games. As per the rules of the game, each player places $1 onto the table to form a pot to induce betting and bluffing. As with poker, this player would give out cards and then bet after the other player had bet. However, since this is an online game the action of giving out cards is automated, so the player first sees a status message saying "Waiting for your opponent..." and then after their opponent has read the instructions and clicked next page "Waiting for opponent's bet..." The dealer would bet following the other player's initial bet.

Figure 4.7. Dealer's initial screen

The player who is not chosen as the deal, on the other hand sees a screen showing their cards and the actions they can perform initially: "Place Bet" along with a drop down containing the amount of their bet and "Fold." Additionally, the current state of the pot "on table," "your chips," as well as any cards shown as part of a disadvantage associated with the visible card condition are displayed.

The poker screens are laid out as follows. In large text at the top is a title describing the action associated with the screen (i.e. "waiting for opponent's bet"). Below this is a section titled "on table" which shows the dollar value of the money at stake (which can increase following betting). Following this is a section entitled "your hand" which shows the two cards dealt to the participant. Below this is another section entitled "your opponent's hand" which shows the backs of the two cards dealt to the participant's opponent. In the case that the opponent is in the disadvantaged visible card condition, the participant would see the face of one of their opponent's cards. Below this is the "your opponent sees" section in which the backs of two cards are visible. In the case that the participant is assigned to a disadvantaged visible card condition then one of the two cards in their hand is shown here.

Figure 4.8. Non-dealer's initial screen

Suppose the non-dealer bets $1, in this case the dealer will then see a screen stating this and allowing them to either "Raise Bet," "Call Bet ($1)," or "Fold." In the mean time, their opponent would see a screen similar to the dealer's initial screen.

Figure 4.9. Your bet screen

If the dealer were to call the $1 bet, they would see a screen revealing their opponent's cards and stating the outcome of the game. Additionally the screen informs the participant that the practice round is over and they will now play the actual round. Their opponent would see a similar screen relating complementary information (viz. Bet Called You Lose, in this case).

Figure 4.10. Bet called screen

At this point both players would see a screen which re-iterated the rules of the game and would inform them that the "actual game" was commencing. This screen would also remind the players of the compensation scheme at work in the game. In the case that a participant was in a disadvantaged condition, they were reminded either that the charity of their choice "will be given funds that match your remaining chips in the case that you lose" in the Charity Loses motivator or "will be given funds that match your chips in the case that you are the winner" in the Charity Gains motivator.

In the sensor condition, the opponent in the non-sensor room (advantage) saw the MixedEmotions window on the computer, displayed side by side with the screens just described. In the visible card condition, the opponent in the room without sensors saw the first of the opponents cards face up on their screen.

Figure 4.11. Shuffling deck screen

The player selected as the dealer in the practice game remained the dealer in the actual game. So they would again see a webpage similar to "Dealer's Initial Screen" above, except with new cards.

Figure 4.12. Dealer's initial screen for actual game

The dealer's opponent would again see a screen prompting them to bet. Since this is a new game, the pot on the table as well as the amount of chips each player has is reset. Namely, there is $2 on the table as well as $5 in chips.

Figure 4.13. Bet screen for dealer's opponent

Suppose that in this situation, the dealer's opponent chooses to fold. The game would then display a fold screen with information about how many chips the player won.

Figure 4.14. Fold screen for dealer's opponent

At this point both players completed the same questionnaire form as the participants in the Quiz and Interview Experiments (figure 3.17). As with the Quiz and Interview Experiments, the survey questions on this page were shown as eight-point Likert-scale questions with a "No Opinion" option. Each of the questions opposed a category related to an ethically positive term like "respectful" or "moral" against ethically negative terms like "invasive" or "immoral."

Following this was a brief page asking participants what "were the intentions with which you played the game." Participants were asked to choose between "Bad" and "Good." This question served as a manipulation check to see if the compensation structure induced intentions we had hoped. Note that the question is on a 7-point Likert scale without the no-opinion option. In future versions of the experiment to be consistent with the style of questionnaire used in other portions of the experiment, an 8-point Likert scale with no opinion ought to be used.

Figure 4.15. Follow up question

Finally, participants encountered a page thanking them for participating. The page also prompted participants to contact the experimenter using an instant messaging client so that the experiment could be completed. Informal discussions about the purpose of the experiment were then conducted by myself.

Figure 4.16. Final page