Vous êtes sur la page 1sur 3

Helicopter Flights Experiment

Sahil Chinoy, Tim Hu, Jessica Zhou, Sean Solis April 7, 2013

Part I

Experimental Design
We plan to conduct 10 trials each at four dierent drop heights  8 ft, 6 ft, 4 ft, and 2 ft  for a total of 40 dierent runs. Although we will conduct the drops in sequential order by height, we'll randomize the helicopter we use for each trial using a random number generator. One person will drop the helicopters consistently and accurately  time will be measured by two stopwatches and if they disagree, the trial will be conducted again. Possible sources of variation include human error (stopwatch reaction time), variability in the helicopter performance, and wind or air abnormalities. To reduce these, we'll randomize the selection of helicopters while making them as similar as possible and also shield the drop area from air disturbances.

Part II

Data Collection
1 8 ft 6 ft 4 ft 2 ft 2.2 1.5 1.0 0.5 2 2.2 1.5 0.8 0.8 3 2.3 1.7 1.1 0.5 4 2.2 1.4 1.1 0.5 5 2.4 1.4 1.2 0.5 6 2.1 1.5 1.1 0.5 7 2.1 1.5 1.2 0.6 8 2.2 1.6 1.2 0.6 9 2.3 1.5 1.2 0.5 10 2.0 1.5 1.0 0.5

Part III

Descriptive Statistics

A linear model apppears to be appropriate for this data, as it seems to t the trend in data well  there are no obvious outliers.

The residual plot conrms that the linear model is appropriate, as the scatter above and below the line is relatively random. The slope of the line, 0.268, means that for every foot increase in the drop height, the time of descent increases by, on average, 0.268 seconds. The y-intercept of -0.0016 means that if the helicopter was dropped from a height of 0 feet, it would take -0.0016 seconds to descend  this is clearly due to experimental error, as the y-intercept should be 0.

Part IV

Inferential Statistics
1. Histograms:

Mean: 2.2; Standard Deviation: 0.110

Mean: 1.51; Standard Deviation: 0.0831

Mean: 1.09; Standard Deviation: 0.122

Mean: 0.55, Standard Deviation: 0.0922 The assumptions of normality and equal variance appear to be valid  although the sample sizes for each drop height are relatively low, the histograms do not show strong skew and the standard deviations across the distributions are roughly equal. Other assumptions: the scatterplot shows that the relationship is roughly linear, each trial and height is independent of the others, and random samples and assignment were used. 2.

= a = 0.0016 = b = 0.268 = SEslope = 0.00868 s = 0.129

3.

H0 : = 0; HA : > 0; df = 38 b = 0.268

There is no linear relationship between drop height and time of descent. There is a positive linear relationship between drop height and time of descent.

Signicance level:

= 0.05

SEslope = 0.00868 t=
0.2680 0.00868

= 30.744

P val = P (t > 30.744) = 9.373 1029


4. Signicance level:

= 0.05, df = 38 t = 2.024

b = 0.267 SEslope = 0.00868 0.268 2.024 0.00868 = (0.249, 0.285)


5. Signicance test: Since

P val = 9.373 1029

is less than

= 0.05,

we have sucient evidence to reject the null

hypothesis and conclude that there is a positive linear relationship between drop height and time of descent. We can say with 95% condence that as drop height increases, so does the time of descent. Condence interval: We can say with 95% condence that each foot increase in drop height is associated with between a 0.249 and 0.285 increase in time of descent.

Vous aimerez peut-être aussi