Week 3 Discussion Responses - Analytics

Discussion Response

By M,F

So in any military branch recruiting is a pretty big deal.  Without people coming into the ranks the natural attrition of the current population will dwindle down to dangerous levels where we would be in an undesirable position.  My current job falls right into the middle of this.  As an ROTC instructor part of our job is to recruit and retain in the goal of meeting the assigned goal numbers of our students who commission as officers annually.  These numbers will greatly fluctuate for each school when compared.  All ROTC programs are designed the same in essence, but differ in size due to the location and size of the school as well as the opinion on the US military in the local demographic.

Ideally we want to have 100% of all students who start Army ROTC as Freshman, but we know that the probably of this is unlikely (less than 1) due to variable of people dropping out, transferring to other schools, or being medically or even academically disqualified.  Although, highly unlikely, that all who start will finish and commission, it isn’t impossible. 

Through the skill of utilizing probability distributions we can help predict patterns or even certain variables of students that may hinder or prevent them from completing all four years. Utilizing previous years of data we can lean on empirical probability or relative frequency to determine the general odds of returning students and their success.  We can use this format with students of particular degree programs (engineering, nursing..both difficult and time consuming or political science or communication…not as intensive as the aforementioned).  By identifying such variables and their associated probability of success or failure we can adjust as instructors as necessary to know which students will need close guidance and some additional support to complete the program as well as their degree and commission as an officer.  

Along with this, on a simpler level, we know that based on particular class size as freshman we can determine the probability of reaching our goal of commissioned officers for that class at the end of four years.   

Discussion response 2

By P,T

According to Lind, A probability distribution gives the entire range of values that can occur based on an experiment. Probability is used to detect a pattern in certain data. The probability of weather accuracy is used by everyone on a daily basis. It can tell an individual whether to go on vacation if it's good weather, or go to the beach and also whether to bring an umbrella to work. These are just some of the examples of how we rely on weather accuracy. When considering the probability of weather all variables such a temperature, wind and precipitation are accounted for. 

Looking at the weather of one day forecast is more accurate than looking at 7 or 14-day forecast. According to Robbins Comparing predicted highs to actuals, the same day forecast averages ~2 degrees off, the 1-4 day forecasts average 2.5 - 3 degrees off, and the 9-day averages nearly 6 degrees off.  The table below shows the error rates for precip, the numbers represent prediction vs actual

The graph below shows the mean SD, the further the days out. In this chart again we see that the variability of weather and accuracy decrease over days out

Robbins, R. (2015, January 29). How Accurate Are Weather Forecasts? Retrieved July 12, 2017, from http://www.huffingtonpost.com/richard-robbins/how-accurate-are-weather-_b_6558770.html

Lind, D. A., Marchal, W. G., & Wathen, S. A. (2015). Statistical techniques in business & economics(16th ed.). New York, NY: McGraw-Hill Education