Use first line as column names: yes
Share with everyone: yes
Last edited: Jun 14, 2014
Teamwork survey of registered nurses
For what survey produced it, see http://www.statcrunch.com/5.0/survey.php?surveyid=8178&code=YINVQ
and inputs of all team mates.
Towards the end, some validation was done, deleting data where working hours was less than a work day, or outliers to legally admissible work days. Finally arbitarily long chains which were less likely to be encountered in draws of simulated data (M/F, Degrees etc.. were discarded). A total of 12 observations were thus thrown out.
All Credit goes to Team 3,the Instructor, our unnamed Friends in the Nursing profession who enthusiastically did a last minute push through over their extended social media groups for data and the respondents who kindly took out time for the survey.
Another thought is about the distribution of hours worked.
Wven if random, it "should be" "centered on" certain hours a day* number of days, with deviations from centre penalised, while picking a sample..
The observations 38 appear many times for example, however without an explainable reason (we are talking of work-distribution among nursing staff sample)
So do "primes" "47, 37, 29"
It is not to argue that they "shouldn't occur", but there has to be some reason for their being so significant/vibrant.
At this stage we may conclude that most of the respondents may not have been under full-time nursing employments in strict sense of the term. 42, 48,72,60, 50,40 appearing more often would give us less variation but more regularity in the data. Since we haven't tried stratification, we do not know "how often they should occur". We thus do not re-draw observations.
This data set has 5 favorites. Sign in to add it to your favorites!
Adding a data set to your favorites makes it easier to come back to in the future!