Help me with Standard Deviation

the question is , do you have to show that you understood standard deviation or so you have to prove that the standard deviation of the given values of R&D spending has an effect on operating profit and its standard deviation?

in my opinion, as others ahve stated before, that is hightly unlikely to being able to correlate, first, there is no point in standard deviation telling anything about it.


broken down,;

1. out of all the R&D spending values you get the median value .
2. you average minimum values and the average maximum values define the positive and the negative halves of the total standard deviation.
3. here it becomes tricky: even though you can get the same values on operating profits (median and SDEV) i see absolutely no correlation for these factors:
a. market delay: R&D spending in 2019 does not mean profits in 2020. it's not how it works. depending on the industry, time to market ranges from 5-10 years.
b. types of spending and market trends: "betting on the wrong horse" or markets floating away skew the correlation possibility as well, due to new trends coming in faster than the aove time frames
c. efficiency: some companies can spend little in R&D and get huge profits (example Pharma , lifestyle (trends) or biotech) , others have to invest hard for their profits...


do you have the original text of what you need to accomplish and show?

 

I have run other numbers to show correlation (or lack thereof), the Standard Deviation is a red herring that is there for me to show understanding, which requires that I provide an insight from the standard deviation. I have divided the SD by the Mean and arrived at 0.67, which indicates something, but I don't know that something. But again, the SD has effectively no relevance on my exercise.

I found that there is little correlation in the data to indicate that spending more on R&D equates to higher EBIT. The five companies I compared all have very similar ratios of R&D to EBIT, which I was able to show via numerical ratios and pretty charts. But as the SD has effectively no bearing on my project, I got stumped on how to provide an insight for an irrelevant statistic.

 

This also gets to the very important point of causation versus correlation. Just because two things are correlated doesn't mean one caused the other. Healthy companies with strong market position and added value might be able to afford a higher R&D budget but the EBIT could be driven by some other fundamentals. Over a large sample size though, you might see a pattern of what % investment drives a higher return ignoring all other factors. But if you wanted to include other factors, then you need multi-variable regression with correlation factors calculated for each factor.

Also, a sample of 5 data points will not give a meaningful standard deviation calculation. You need a large enough sample size to be statistically significant. There is another calculation called confidence intervals that shows what the spread of a number is based on the sample size. With a sample size of 5, the confidence interval for the standard deviation will pretty much equal the total range of the numbers.... 15 years since I took business statistics so that's about all the brain has left in it on this subject...

 

^ exactly.

a high ebit can be caused by operational excellence, lean processes, products with high margins.
does not essentially have to be correlated.

it's like saying women who ride horses are generally healthier. yes they are, but that might originate from better social position, better diet, less stress and horse-riding is only a by-product.

 

Maybe the answer is that you report that the data supports no conclusion...and here's why...

 

Bingo. That was my conclusion, that the companies spend roughly the same on R&D as a percentage of EBIT, which may be indicative of the entire sector, but further analysis is required. Therefore, the evaluation is inconclusive.

But I need to calculate SD as part of the rubric and be able to provide insights to the statistic.

 

You understand enough about SD to do that. Get to work.

Standard deviation is nothing more or less than predictability. In business decisions, financial decisions, that matters a great deal. How confident are we that this or that decision is the way to go? With the data you presented, not that we saw all of it, there was a bunch of I don't know, could be.

It was an interesting question because it was questioning how well you understood the concepts of SD and confidence intervals. You have to be confident (haahaaa) to say the data as yet supports no conclusion. So explain that as a function of SD, move one and two and three SDs away from the middle, and what bearing that has on your confidence in your opinion as to the correlation of X versus Y.

 

I think Monsterdood just got you an A!

 

Just actually reading this - I think the teacher just wants you to calculate r^2 by hand and demonstrate that you know that r^2 and standard deviation are inversely related. That prolly gets you a B/B+ and (IMO) is the primary "ask".

If you wanna get an A - lag the EBIT(DA) against the R&D expenditures and figure out what lag amount has the highest R^2 (i.e., "a 3 year lag is best to explain the impact that R&D expenditures have on future profitability"). If you wanna dig into their 10Ks (I'm sure they're all public since you know their R&D expenditures), you could even break it out by division/product line.

If you wanna get an A+, do all of the above but blow the teacher's mind by using current year EBITDA as the independent variable and current year R&D expenditures as the dependent variable showing how much current year profitability (as a proxy for cash flow) dictates what that year's R&D budget/expenditures are. You could do a similar exercise with prior year cash and equivalents (from the balance sheet, obviously) regressed against current year R&D spending.

 

Likely.

No.

 

I’m going through six sigma green belt training and studied SD a few weeks ago.

gave me flashbacks to my college stats class.

 

Okay, I read everything posted here, read some other junk, retyped my project, and submitted it last night. Eff it. Here's the feedback that I received:
Meets Specifications

Hooray for me! :|

Seriously, here's a sample of specific feedback from my project:

• Excellent project, you have a good understanding of using the correct formulas and assumptions to create a great financial model where required and creating good slides supported by visualisations with clear descriptions to explain the analysis using the measures of centre and spread to support the claims.
• Student uses standard deviation and range to generate insights.
  • Fantastic work on the calculation and insights generated for types of ranges and standard deviation. This helps the audience understand the values you have given and why they matter in your analysis. Well done!
• Communication Phase

• • Your analysis does not state or imply facts and changes caused by correlation. Well done!
• Business Metrics
  • You have provided appropriate assumptions based on the gross margin, revenue growth and operating margins correctly for the specific financial model scenario. This shows your ability to correctly use the required formulas to predict these outcomes, well done.

This feels like a podium speech, but thanks to everyone who contributed to the discussion and encouraged my progress. As this is a pass/fail course, I'm not going to post up my report card (it ain't a prominent university). I'll simply say "done" at the end and have the satisfaction of finishing something outside of my strengths.

Anyone looking to hire a Business Analyst?

 

Lol! Well done you! I don't care about this or that equation... when you learn the concept, that is yours to keep. Forever.

 

Based on your hobbies and where you spend time online, I've got some serious character concerns.

In all seriousness though, congrats on getting the assignment done successfully

 

Pot/Kettle. Does this mean that I get to question your judgement as well? Which infers a double negative, so I'm good!

 

Implies, implies... Take the win!

 

I'm sticking with infer as there is a conclusion in the statement, not a hint.

 

A battle for another day. Not today.