Teach me STATISTICS in half an hour! Seriously.

THE CHALLENGE: "teach me statistics in half an hour with no mathematical formula"

The RESULT: an intuitive overview of distributions, sampling and hypothesis testing that (almost) scrapes in under 30 minutes :)

Whether you're studying statistics for the first time or are just curious as to what the study of statistics is all about, this video is a great introduction to several core concepts.

Oh, and all the examples I use come from the 2018-19 NBA season.

Steph's 3 point percentage is what is called a 'ratio', which is numeric. It can be argued to be discrete (as for a given number of attempts, there are only so many possible percentages that can be achieved), or continuous (if we allow the number of attempts to change). Heated arguments either way here are semantic, in my opinion :)

0:00 Introduction
1:18 Data Types
8:04 Distributions
13:51 Sampling and Estimation
22:30 Hypothesis testing
31:28 p-values
35:35 BONUS SECTION: p-hacking

[jennifercathro]

This is a fantastic overview of statistics! I feel like I'm listening to a great story, and cant wait to hear all the other chapters. As a college student enrolled in stats, I am so glad I stumbled across this video! Your narration manages to take the painfully mundane content of statistics and turn it into riveting story. Thanks to you, the probability of me passing the class just skyrocketed. :)

[andrewcornell]

"perfect for those just enrolling in a statistics course"

my final is in 2 hours

[heatherwoods]

I spent an entire semester crying my way through trying to understand Statistics. Literally sobbing as I was studying. 4 years later and I'm retaking the course (someplace new) and found this video. Suddenly all of the previous class makes sense and I can see what the Prof was trying to do... he just couldn't present it in a way that made any sense. Always in the weeds of details without an overarching idea of what I was looking at. Thank you so so much!

[DaarinaAC]

"The idea here is for you to develop your intuition around statistics..." I'm all ears. I also have an exam tomorrow.

[BruceNitroxpro]

As a former college instructor of statistics, I must complement the narrator of this series on the value of the short course for the average viewer.

[chriscockrell]

0:00 Introduction
- 30 minute summary of statistics

1:18 Data Types
- numerical (continuous and discrete) or categorical (nominal and ordinal)
Steve Curry example, nominal data. Is a proportion considered discrete or continuous numerical variables?

8:04 Distributions - Probability density function - P.D.F. Normal (bell curve), uniform (bi-modal), left and right skew -shapes. PDF is unique to an event and based on data. Probability of choosing data (a player) of a certain attributes (conditions on dimensional attributes - a certain height). Sampling, averaging many events, can be viewed as having a steeping PDF and a lower variance.

13:51 Sampling and Estimation - more data provides better confident interval. Sampling
mean - average, must be numeric (mu) Makes sense that none numerical data uses mode and median x (with bar hat)
standard deviation (row) - difference from the average, on average s
proportion (pi) - is a measure of data, whether nominal or numerical. p
correlation (row) - the connection between dimensions r
gradient (beta) - change between dimensions. Regression b
(theta) - general variable. (and the greek with statistic and english for probability of sampling)

22:30 Hypothesis testing - binomial distribution (PDF) Since 50% is null, 7.5 is the likely outcome, middle of the bell. less than 5% would reject the null. In this case, we'd not reject. We don't accept or prove the alternative, we reject or don't reject the null. Legal system, innocent until proven guilty.
31:28 p-values - our test statistic is 30.4% (area under 9).
35:35 BONUS SECTION: p-hacking

I find it interesting that all data is conceptual, or also referred to as nominal (I don’t mean or in the set theory this OR that categorization way, just as a different name for the same concept way). That includes data that is numerical. Just because we assign a character "3, " doesn't make 3 not a name. Numerical is a name that can be used as a measure. Measures are about numerical comparison. Both the concept of numerical and measure are tied to counting, while nominal is not. When we say nominal data, we literally mean, a name of a concept, but we also mean "not a measure” or “Lacking ordering of elements.”

(Ordering is always a numerical property of “distance” but distance is not always a property of ordering, as nominal data can be given structure and ordinality.)

The measure (or what you are calling a proportion of data - where we use numerical values as a comparative evaluation) of a dimension of all or any data (to my understanding - correct me if you disagree with my narrative, I'm only starting to learn about data) can have the attribute (a specific entry of a dimension) of "numerical" (As opposed to "nominal, " but if we are being honest, "non-numerical" is more descriptive).

I'd argue that reality is continuous. Discrete is the way that we process or model reality. The same way that we process continuous functions that represent physical signals in reality (currents, voltages, energy, ...). We discretize to understand and find meaning. The ship of Theseus is a good example of a continuous reality. I'd even go as far as to say you can't process data without discretization. Calculus and differential equations attempt to approach this continuous reality, but show that the reality is right out of our reach for understanding as we discretize our understanding. Continuity is a form or experience and reality, and, as such can't be presented unless in a discrete way. Other narratives around data types is: 1. Nominal, 2. Ordinal, 3. Interval/Ratio.

Both measures (proportions), produced from the given data set, are example discrete dimensions of data. I'd argue that counting is not continuous. Saying numerical data has an attribute of "continuous, " isn't illogical. If it is, I'd challenge you to provide an example. To say that data has an attribute of nominal (non-numeric) or numerical, is logical. This topic deserves a whole essay to do it justice.

It is worth noting that there is value to comparing the concept of continuous (changes don't occur instantaneously) and discrete, but I suspect we are a little too liberal in our use of saying numerical data is either continuous or discrete. I don't think it is an accurate statement, but it is part of how we communicate meaning. We should change that.

Another observations are that the numerical is also ordinal data and that ordinal is a subset of nominal and numerical is a subset of ordinal.

The key with math and legal and politics and life, is understanding what and how you should look at something. That is the toughest part about math. Identifying what you are looking at and how you should conceptualize it. Example. It is a function, an equation, and/or a relationship. I suspect it is much the same with Stats and Prob. Great Philosophical video. I'll have to check out some of your others.

[Prosegrrl]

I almost quit my statistics course because I was so lost. This and your other videos are definitely helping me!

[stephenday]

I love the calming voice and the lack of a music track in the background, it makes the video so much easier to listen to. So many of these videos are nearly impossible to listen to due to invasive music.

[geoffbakerhytch]

I'm a recently retired British (high-) secondary school teacher of mathematics; your approach appeals to me very much, and is similar to way I have taught the subject. I find your explanations really engaging, relevant and crstal clear! Well done, and thank you for helping to spread a better understanding of mathematical ideas and techniques.

[ShailendraSingh-exyj]

Me, 45 minutes before a 1.5-hour university exam on statistics.

♫ Where is my mind... ♫

[russelltreadway]

I became obsessed with NBA statistics when I was 12 after buying my first basketball cards. There was a problem, some cards didn't include averages, only totals. There were cards that included less seasons, but included career totals, etc. So I to applied the basic statistics I knew to get the missing info (91 pre-internet living in Mexico btw) and I got most of the concepts from doing that, but without the terminology of course. So now I'm trying to learn more so I can fill in the gaps, mainly because I have an interest in risk theory, probability and all things related. Your videos are great and easy to understand, thank you so much!

[aakuthotaharibabu]

It feels like listening to a story rather than a class. Great work. thank you

[varunpratapsingh]

Goddamn, i spent 4 hrs struggling to know the exact and basic meaning of hypothesis theory and p value, thank god i found this today!
Thanks a lot!

[therealforestelf]

do more of these, I love watching speedrun explanation videos the day before the exam, feels good and you reabsorb so many things just when you hear it from a different source again! friendly greetings from Germany!

[nethravathibg]

Zed, you are an amazing teacher/ professor/ trainer, I must say. You took even challenge so humbly, yet you didn't miss your purpose, i.e., to teach us :) Kudos to you! Im happy I found a guru like you! Now, to answer your 1st question in data for proportion, I think it's continuous data.

[aravindssingapore]

Even though this video is just an inch longer than 30 mins, it is still a huge, gigantic feat - you have done the impossible once again, Justin. This is a Guiness World Record-worthy effort. Cheers to you 🍻

[imamfahrurrozi]

I've zero understanding of what statistics really is before watching this video.
All beginer should watch this!! Best statistics overview ever!

[SandeepEllawala]

I am on the verge of making a change in my job role that involves learning statistics, among a whole bunch of other things, after a decade-long career in ITSM. I was uncertain as to whether I would enjoy spending time going deeper into statistics again. This video is what I needed to get started. Thank you for sharing your knowledge. I have subscribed to your channel, and looking forward to seeing more content in the future. 🙏

[misskay]

I'm reading all these comments about everyone watching right before an exam and I'm here watching 3 weeks before my statistics class even starts🙃 I'm trying to get all the advantages i can!

[maichka]

Thank you. I'm a BA in psychology, but it's been a while and now getting ready for an MA entrance exam this is a really good intro before all the equations come into the picture.