"How can the design of the Gacha system affect player payment?"
About
This is a solo project researching the impact of Gacha system design on player in-game payment. The game was designed on StudyCrafter.
Hypotheses
3 hypotheses were posted:
Hypothesis 1 (H1): Players would pay more if the Gacha item in the system has a higher value.
Hypothesis 2 (H2): Players would pay more if the Gacha system has a probability announcement.
Hypothesis 3 (H3): Players would pay more if the Gacha system assures players could get an item after several times of Gacha (a guarantee mechanic).
Hypothesis 1 (H1): Players would pay more if the Gacha item in the system has a higher value.
Hypothesis 2 (H2): Players would pay more if the Gacha system has a probability announcement.
Hypothesis 3 (H3): Players would pay more if the Gacha system assures players could get an item after several times of Gacha (a guarantee mechanic).
Methods
5 different Gacha systems were designed (including 1 control group and 4 experiment groups) on StudyCrafter to verify the hypotheses. 30 players played the game. The money they spent was logged. Paired T-test and Turkey HSD were used to analyze the money players spent in the game.
Results
The distributions of spent money in 5 different Gacha systems were summarized as CG (M = 330.67, SD = 279.92), EG1H1 (M = 111.67, SD = 259.54), EG2H1 (M = 782.67, SD = 268.87), EGH2 (M = 388.33, SD = 460.97), and EGH3 (M = 696.33, SD = 227.32).
Tukey HSD
A Tukey HSD test was performed to compare the effect of the Gacha item value on money spent by players (H1). The test revealed that there was a statistically significant difference in money spent by players between the 3 groups. H1 was accepted.
Paired T-Test
The test revealed that there was no statistically significant effect of the absence of the probability notations on money spent by players at the p < .05 level (T(30) = -0.70, p = 0.49). H2 was rejected.
However, the money spent in the Gacha system with a guarantee mechanic compared to the money spent in the control group that did not have the guarantee mechanic demonstrated a significant difference by another paired T-test (T(30) = -7.04, p = 4.8e-8). H3 was accepted.