Our Experimental Study
Spacecraft-generated mechanical noise is a major source of disturbance to astronauts trying to sleep on the ISS. Due to a lack of information online about the specific qualitative and quantitative effects of such noise on sleep quality and energy levels, we at SleeperBuild decided to conduct our own primary experimental study with a group of randomly selected volunteers.*
*All participation in the study was fully consensual*
Phase 1: Data collection
In order to simulate ISS-generated mechanical noise, each participant listened to Space station ambience audio on volume setting 3 during the process of falling asleep, for 2 consecutive days. On each morning after (within 10 minutes of waking up), participants would complete a multiple-choice morning survey to assess the quality of their sleep. At bedtime on the same day, they would complete a multiple-choice night-time survey to assess their fatigue levels during the day.
Phase 2: Statistical analysis
To convert qualitative into quantitative data, numerical values were assigned to each answer option for each multiple-choice question in both surveys. A number from 1 to 5 was appropriately assigned to each option, with 1 corresponding to worse quality sleep, 5 corresponding to better quality sleep, and 3 corresponding to a neutral/unaffected sleep (i.e. a control status). (Non-integer values between 1 and 5 were also assigned if necessary) The mean score for each participant for both surveys was calculated using x bar = Σx / n, where Σx is the sum of their scores for that survey, and n is the number of questions.
The population of sleep quality index scores are generally non-normally distributed, and therefore, it would be inappropriate for us to use a parametric test (e.g. student’s t-test) to analyse our results. Instead, we utilised the non-parametric Wilcoxon single-sample signed-rank test, which is suitable as sleep quality index scores are frequently symmetrically distributed. A one-tailed test at 5% significance was used to test the hypotheses:
Null hypothesis H0 : m = 3 (i.e. sleep-quality/energy-levels are unaffected by space station ambience)
Alternate hypothesis H1: m < 3 (i.e. sleep-quality/energy-levels are negatively affected by space station ambience)
Where m represents the population median of mean survey scores under Space Station ambience.
​
Phase 3: Evaluation
Conclusion:
The test statistic from our morning survey was T = 68, whilst for our night-time survey it was T = 70.
The critical value for a one-tailed test at 5% significance with sample size n = 22, is Tc = 75.
Because both 68 < 75 and 70 < 75, then there is sufficient evidence to reject the null hypothesis for both sureys, and there is sufficient evidence to suggest that the mechanical noise of space station ambience does significantly hinder both sleep quality and energy levels throughout the day, to a significant extent.
​
This supported the validation of our innovation's purpose: by proving the significance of mechanical noise disruptions, SleeperBuild recognises the importance of eliminating it through our noise-cancellation element of the SleeperBuild Headset (in partnership with Sony and Bose).
