Вы находитесь на странице: 1из 6

Cole Lechleiter Joseph Lavicka Jason Rozet Professor Pfaff, Business Statistics Are Residential Laundry Facilities at Ithaca

College Adequate?

I. Introduction
a. The study determines the adequacy of the laundry facilities on Ithaca College campus. In order to establish a more thorough conclusion, we have broken up the criteria into two sub-questions: are there enough dryers compared to washers, and how do the performance of different groups of laundry rooms (clusters) compare to the whole? b. It is apparent that there are more dryers in laundry rooms by design, but is that enough to compensate for the longer drying cycle time? Are the usage proportions of washers and dryers equal? The question of clusters tries to examine whether any specific laundry facilities are better or worse compared to the remainder of the campus. Is the ability to access machines equal across the whole campus? c. The first population consists of all the washers in IC laundry room, while the second is comprised of all the dryers on campus. We have defined the parameters for each population as the proportion of available machine. d. This study articulates an accurate representation of ICs laundry facilities. Our primary motivation is to establish reasoning for administrators to consider our results when prioritizing and planning upgrades to the laundry rooms, to better improve their performances.

II. Data Collection Methods


a. For each of three data collectors: i. Sampling period is 7 days (Wednesday Tuesday) ii. Sampling times are 30-minute periods from 10AM to 12AM iii. 6 sampling times per day randomly selected b. One sample reading consists of one Laundry View snapshot (contains occupancy data for all 28 laundry rooms) c. Sample size was 106 snapshots containing data for both washers and dryers for 28 rooms

III.

Results

Raw data was standardized by converting number of available machines in each laundry room to a proportion of available machines (For example, an availability of 1 washer in any Terrace represents the same proportion as an availability of 14 washers in the Circles, so each is expressed as 0.50 instead of absolute machine counts). After standardization, these proportions were grouped for to perform statistical tests by arithmetic averaging, so T-tests (parameter, )were performed instead of Proportion Z tests (parameter, p). Dryers versus Washers

Descriptive Statistics
Washers Mean Interval high Interval low Standard Error Confidence Level(95.0%) Standard Deviation Sample Variance Min Q1 Med Q3 Max Total Count Mode Number 0.714688163 0.721699839 0.707676487 0.007011676 0.013748248 0.381734077 0.145720905 0 0.5 1 1 1 2964 1 1647 Dryers Mean Interval high Interval low Standard Error Confidence Level(95.0%) Standard Deviation Sample Variance Min Q1 Med Q3 Max Total Count Mode Number 0.693016944 0.699164581 0.686869306 0.006147637 0.01205407 0.33475 0.112057562 0 0.36786 0.75 1 1 2965 1 1273

Table 1: Descriptive Statistics for Washer and Dryer Availability. At 95% confidence, there is no overlap for these two proportion means, and this difference is deemed significant.

Figure 1: Percentage Frequency of Available Dryers. Independent axis represents different availability proportions from 0 to 1.

Figure 2: Percentage Frequency of available washers. Independent axis represents different availability proportions from 0 to 1.

Washers vs Dryers (Two-sample T)


Variable Washers Dryers Calculation (Washers) - (Dryers) 95% CI for difference: T-Test of difference = 0 T-Value = P-Value = DF = N 2964 2965 Min 0.00339 2.32 0.02 5827 Estimate 0.02167 Mean 0.715 0.693 SE Mean 0.007 0.0061 Max 0.03995 StDev 0.382 0.335

Table 2:Two Sample T Test & Confidence Interval Comparing Mean availability of washers and dryers. H0: There is no difference between the mean availability ( = 0); Ha: there is a difference

Clusters versus Whole


Cluster Quads Terraces EmGar Towers Circles

Washer Availability by Cluster


(Quads) - (Campus) < 0 > 0 > 0 or < 0 > 0 or < 0 > 0 P value 0 0 0.111 0.685 0 Conclusion Underperforming overperforming no conclusion no conclusion overperforming

Dryer Availability by Cluster


P value Conclusion 0 underperforming 0 overperforming 0.553 no conclusion 0.15 no conclusion 0 overperforming

Cluster (Cluster) - (Campus) Quads < 0 Terraces > 0 EmGar > 0 or < 0 Towers > 0 or < 0 Circles > 0

Table 3. Summary data of T-Interval results for Washer and Dryer Availabilities by cluster.

IV.

Discussion

Dryers versus Washers Key parameters: washer proportion, dryer proportion Confidence interval (C=0.95) for = 12 H0: There is no difference between 1 and 2 ( = 0) Ha: There is a difference between 1 and 2 ( > 0 or < 0) i. ii. The observed difference () between the availability of washers versus dryers was 0.02167. The 95% confidence interval ranged from 0.00339 to 0.03995. Because the entire interval is above zero, we can conclude that there is a significant difference between mean availability of washers and dryers, and the mean availability of washers is greater. iii. P-value for the test statistic was 0.02. Since this figure is below the level of significance (=0.05) we reject H0. Clusters versus Whole Key parameters: cluster proportion, campus proportion Confidence interval (C=0.95) for =12 H0: There is no difference between 1 and 2 ( = 0) Ha: There is a difference between 1 and 2 ( > 0 or < 0)

i.

We considered the mean proportion of available machines for each cluster and compared its value to the proportion mean for the remainder of campus. The test was run once for washers and again for dryers.

ii. iii. iv.

We created a 95% confidence interval of for each cluster. For the clusters that had mean differences either entirely above zero or entirely below zero ( > 0 or < 0), we were able to make a confident conclusion about its performance. A mean difference interval below zero ( < 0) provided us with a conclusion of underperforming, while a mean difference interval above zero ( < 0) concludes the cluster was overperforming compared to the rest of campus.

Concerns of data integrity i. Some randomly-generated times were skipped due to scheduling convenience ii. Lost one-third of Wednesdays data iii. Potential confounding variables may influence the differences between clusters. Is the difference in laundry usage in the Circles and Gardens simply related to apartment-style living?

V.
a.

Conclusion
Our study has shown us that there is little concern over the inadequacy of the laundry facilities at IC. Even with the means of the washer and dryer availability having a difference of about 0.02, we cannot safely conclude this difference is significant to claim the facilities inadequate. It does show the dryers, as a whole, being less available, however not at a large enough difference to show concern.

b.

The clusters have revealed what specific facilities are inadequate compared to the rest of the campus. We can safely conclude that the Quads laundry rooms are underperforming, while the Terraces and the Circles are over-performing. No conclusion could be reached with the Towers and EmGar because the confidence interval lay around zero, having a potential negative or positive outcome.

c.

Afterperforming this study, some additional research could be met to expand upon these ideas. With a follow up study we could reach a conclusion of which days would be better to do laundry (weekend versus weekday). Further studies could include data collecting that is longer than a week in order to reach more concrete conclusions.

d.

With this data and our conclusion we can help to make recommendations to students as to what laundry facilities are better equipped for their needs. It seems like the Terraces and Circles are the best laundry rooms available on campus. We can also provide this as evidence to administrators of possible improvements needed in their facilities, like the underperforming rooms in the Quads.

Вам также может понравиться