In: Math
12.26 Does haptic feedback improve performance?
A group of technology students is interested in whether haptic feedback (forces and vibrations applied through a joystick) is helpful in navigating a simulated game environment they created. To investigate this, they randomly assign 20 students to each of three joystick controller types and record the time (seconds) it takes to complete a navigation mission. The joystick types are (1) a standard video game joystick, (2) a game joystick with force feedback, and (3) a game joystick with vibration feedback. Using the table below of group means and standard deviations, construct an ANOVA table. Based on the F statistic and degrees of freedom, compute the P-value. What do you conclude?
Joystick |
Group mean |
Group sd. |
n |
1 |
279 |
78 |
20 |
2 |
245 |
68 |
20 |
3 |
258 |
80 |
20 |
Since the significance level is not mentioned, we use the default which is 0.05.
The Hypothesis:
H0: There is no difference in the mean times of the three types of joysticks used.
Ha: There is a difference in mean times between at least 2 types of joysticks used.
The Calculations for ANOVA are below:
The overall mean = [(20 * 279) + (20 * 245) + (20 * 258)] / 60 = 1564060 = 260.67
SS treatment = SUM n* ( - overall mean)2 = 20 * (279 - 260.67)2 + 20 * (245 - 260.67)2 + 20 * (258 - 260.67)2 = 11773.33
df1 = k - 1 = 3 - 1 = 2
MSTR = SS treatment/df1 = 117733.33 / 2 = 5886.67
SSerror = SUM [(n - 1) * Variance] = 19 * 782 + 19 * 682 + 19 * 802 = 325052
df2 = N - k = 60 - 3 = 57
Therefore MS error = SSerror/df2 = 325052/57 = 5702.67
F = MSTR/MSE = 5886.67/5702.67 = 1.03
The ANOVA table is as below.
Source | SS | DF | Mean Square | F | Fcv | p |
Between | 11773.33 | 2 | 5886.67 | 1.03 | 3.16 | 0.3628 |
Within/Error | 325052.00 | 57 | 5702.67 | |||
Total | 336825.33 | 59 |
The p value is calculated for F = 1.03, df1 = 2, df2 = 57 and p value is = 0.3628
The Decision Rule: If p value is < Alpha , then reject H0.
The Decision: Since p value is > Alpha , we fail to reject H0.
The Conclusion. Fail to reject the null hypothesis. There is insufficient evidence at Alpha = 0.05 to conclude that there is a difference in the mean times between at least 2 of the joysticks used.