Introduction to

Statistics
AN INTUITIVE GUIDE FOR ANALYZING DATA
AND UNLOCKING DISCOVERIES

Jim Frost
Copyright © 2019 by Jim Frost.

All rights reserved. No part of this publication may be reproduced, distributed

or transmitted in any form or by any means, including photocopying, record-
ing, or other electronic or mechanical methods, without the prior written per-
mission of the publisher, except in the case of brief quotations embodied in
critical reviews and certain other noncommercial uses permitted by copyright
law.

To contact the author, please email: jim@statisticsbyjim.com.

Visit the author’s website at statisticsbyjim.com.

Ordering Information:

Quantity sales. Special discounts are available on quantity purchases by edu-

cators. For details, contact the email address above.

Introduction to Statistics: An Intuitive Guide / Jim Frost. —1st ed.

3
 Contents

Prepare for an Adventure! .................................................... 7

The Importance of Statistics ............................................... 10
Draw Valid Conclusions ................................................... 11

Avoid Common Pitfalls .................................................... 12

Make an Impact in Your Field ......................................... 13

Protect Yourself with Statistics ....................................... 14

Statistics versus Anecdotal Evidence ............................. 14

Organization of this Book ................................................ 20

The Chapters/Sections below are only in the full ebook.

i
iii
v
 INTRODUCTION

Prepare for an
Adventure!

NOTE: This sample contains only the introduction and first chapter.
Please buy the full ebook for all the content listed in the Table of Con-
tents. You can buy it in My Store.

I love statistics and analyzing data! I also love talking and writing
about it. I was a researcher at a major university. Then, I spent over a
decade working at a major statistical software company. During my
time at the software company, I learned how to present statistics in a
manner that makes it more intuitive. I want you to understand the es-
sential concepts, practices, and knowledge for statistics so you can an-
alyze your data confidently. That’s the goal of my book.

Over the years, I've literally received thousands of requests from as-
piring data scientists for statistical guidance. This book is my answer -
years of knowledge and thousands of hours of hard work distilled into
a thorough, practical guide for understanding statistics.

I hope this book helps you see statistics through my eyes―as a key that
can unlock discoveries trapped in your data. Discovery is a bumpy
road. There might be statistical challenges en route, but they can be

7
 Jim Frost

exciting and rewarding to resolve. Sometimes it is the perplexing mys-

tery in your data that prompts you to play detective and leads you to
surprising discoveries!

I hope you’ll view this book as a voyage. We start with the practical
nuts and bolts about the field of statistics and data analysis. Then we
move to the broader context of the process and challenges of using
statistics to expand human knowledge. Now, that’s an adventure in
statistics!

In this book, you'll learn about the many facets of the field of statistics.
We’ll start with an overview of statistics and why it is so important.
Then, we’ll move on to the necessary skills and knowledge about the
different types of data, and how to summarize data both graphically
and numerically. Importantly, I’ll show you how to discover relation-
ships between your variables. After that, we go to the next level and
learn about populations and how to estimate their properties using
samples.

Finally, we look at how statistics are part of the scientific method and
help expand the scope of human knowledge. This process includes de-
signing experiments that identify causation, rather than mere correla-
tion, by accounting for challenges that all studies face. The book closes
with criteria you can use to evaluate the quality of both the data and
the design of the experiment itself. These criteria are useful for cri-
tiquing your research or other researchers' experiments.

Reading this book will allow you to take your skills to the next level.
You’ll be able to proceed onto more advanced statistical studies such
as hypothesis testing, regression analysis, and experimental design.
Alternatively, you’ll have a solid foundation for a career in data sci-
ence and machine learning.

You’ll notice that there are not many equations in this book. After all,
you should let your statistical software handle the calculations while

8
 Introduction to Statistics: An Intuitive Guide

you instead focus on understanding your results. Consequently, I fo-

cus on the concepts and practices that you’ll need to know to perform
the analysis and interpret the results correctly. I’ll use more graphs
than equations! If you need the equations, you’ll find them in most
textbooks.

Please note that throughout this book I use Minitab statistical soft-
ware. However, this book is not about teaching particular software but
rather how to perform basic graphing and analysis. All common sta-
tistical software packages should be able to perform the analyses that
I show. There is nothing in here that is unique to Minitab.

9
 CHAPTER 1

The Importance of
Statistics

Statistics are everywhere today. You’ll run across them in your every-
day life. Four of five dentists prefer this toothpaste! On the news. This
scientific poll used random sampling and has a margin of error of 3%.
In school. Your grades were averages of scores. Perhaps you were
graded on a curve? Then, you took the SAT, ACT, or GRE and received
your results with percentiles. You’ll find statistics in almost every as-
pect of life: Nielsen TV ratings, surveys, political polls, insurance
rates, and so on. And then, of course, there are the ubiquitous sports
statistics, batting averages, third-down conversions, and countless
more! Politicians and policy wonks use statistics to make their cases.
Companies use statistics to use data to make decisions and increase
their profits.

The field of statistics is crucial for expanding scientific knowledge.

Statistical analyses are present in virtually every scientific study. In-
deed, these analyses determine whether the results of the study are
significant and worthy of being published. It’s powerful stuff. But
what is the field of statistics exactly?

10
 Introduction to Statistics: An Intuitive Guide

The field of statistics is the science of learning from data. Statistical

knowledge helps you use the proper methods to collect the data, em-
ploy the correct analyses, and effectively present the results. Statistics
is a crucial process behind how we make discoveries in science, make
decisions based on data, and make predictions. Statistics allows you to
understand a subject much more deeply. Surprisingly, the field isn’t
only about numeric results. It also involves a wide range of practices,
decisions, and methodologies for both collecting data and analyzing
them in a manner that produces valid findings and sound conclusions.

Let’s start by covering two main reasons why studying the field of sta-
tistics is crucial in modern society. First, statisticians are guides for
learning from data and navigating common problems that can lead you
to incorrect conclusions. Second, given the growing importance of de-
cisions and opinions based on data, it’s crucial that you can critically
assess the quality of analyses that others present to you.

In my view, statistics is an exciting field about the thrill of discovery,

learning, and challenging your assumptions. Statistics facilitates the
creation of new knowledge. Bit by bit, we push back the frontier of
what is known.

Draw Valid Conclusions

Statistics are not just numbers and facts. You know, things like 4 out
of 5 dentists prefer a specific toothpaste. Instead, it’s an array of
knowledge and procedures that allow you to learn from data reliably.
Statistics enable you to evaluate claims based on quantitative evidence
and help you differentiate between reasonable and dubious conclu-
sions. That aspect is particularly vital these days because data are so
plentiful along with interpretations presented by people with an
agenda.

Statisticians offer critical guidance in producing trustworthy analyses

and predictions. Along the way, statisticians can help investigators
avoid a wide variety of analytical traps.

11
 Jim Frost

When analysts use statistical procedures correctly, they tend to pro-

duce accurate results. In fact, statistical analyses account for uncer-
tainty and error in the results. Statisticians ensure that all aspects of a
study follow the appropriate methods to produce reliable results.

These methods include:

• Collecting reliable data.

• Analyzing the data appropriately.
• Drawing reasonable conclusions.

Avoid Common Pitfalls

Please notice that statistical practices begin before the analysis phase!
Using statistical analyses to produce findings for a study is the culmi-
nation of a long process. This process includes constructing the study
design, selecting and measuring the variables, devising the sampling
technique and sample size, cleaning the data, and determining the
analysis methodology among numerous other issues. The overall qual-
ity of the results depends on the entire chain of events. A single weak
link might produce unreliable results. The following list provides a
small taste of potential problems and analytical errors that can affect
a study.

Biased samples: An incorrectly drawn sample can bias the conclu-

sions from the start. For example, if a study uses human subjects, the
subjects might be different than non-subjects in a way that affects the
results.

Overgeneralization: Findings from one population might not apply to

another population. Statistical inferences are always limited, and you
must understand the limitations.

Causality: How do you determine when X causes a change in Y? Stat-

isticians need tight standards to assume causality, whereas others ac-
cept causal relationships more easily. When A precedes B, and A

12
 Introduction to Statistics: An Intuitive Guide

correlates with B, many mistakenly believe it is a causal connection!

However, you’ll need to use an experimental design that includes ran-
dom assignment to assume confidently that the results represent cau-
sality.

Incorrect analysis: Are you analyzing a multivariate study area with

only one variable? Or, using an inadequate set of variables? Perhaps
you’re assessing the mean when the median might be better? You can
use a wide range of analytical tools, but not all of them are correct for
a specific situation.

Violating the assumptions for an analysis: Most statistical analyses

have assumptions. These assumptions often involve properties of the
sample, variables, data, and the model. Adding to the complexity, you
can waive some assumptions under specific conditions—sometimes
thanks to the central limit theorem. When you violate an essential as-
sumption, you risk producing misleading results.

Data mining: Even when analysts do everything else correctly, they

can produce falsely significant results by investigating a dataset for
too long. When analysts conduct many tests, some will be statistically
significant due to chance patterns in the data. Fastidious statisticians
track the number of tests performed during a study and place the re-
sults in the proper context.

Numerous considerations must be correct to produce trustworthy

conclusions. Unfortunately, there are many ways to mess up analyses
and produce misleading results. Statisticians can guide others through
this swamp! We’ll cover everything, and more, on that list above.

Make an Impact in Your Field

Researchers us statistical analyses in almost all fields to make sense of
the vast amount of available data. Even if the field of statistics is not
your primary field of study, it can help you make an impact in your
chosen field. Chances are very high that you’ll need working

13
 Jim Frost

knowledge of statistical methodologies both to produce new findings

in your area and to understand the work of others.

Conversely, as a statistician, there is a high demand for your skills in

a wide variety of areas: universities, research labs, government, indus-
try, etc. Furthermore, statistical careers often pay quite well. One of
my favorite quotes about statistics is the following by John Tukey:

“The best thing about being a statistician is that you get to play in eve-
ryone else’s backyard.”

My interests are quite broad, and statistical knowledge provides the

tools to understand all of them.

Protect Yourself with Statistics

I’m sure you’re familiar with the expression about damned lies and
statistics, which was spread by Mark Twain among others. Is it true?

Unscrupulous analysts can use incorrect methodologies to draw un-

warranted conclusions. That long list of accidental pitfalls can quickly
become a source of techniques to produce misleading analyses inten-
tionally. But how do you know? If you’re not familiar with statistics,
these manipulations can be hard to detect. Statistical knowledge is the
solution to this problem. Use it to protect yourself from manipulation
and to react to information intelligently.

The world today produces more data and more analyses designed to
influence you than ever before. Are you ready for it?

Statistics versus Anecdotal Evidence

One way to appreciate the power of statistical analysis is to compare
it to an alternative method for learning. Statistics and anecdotal evi-
dence are almost directly opposite methodologies. Let’s see how they
compare!

14
 Introduction to Statistics: An Intuitive Guide

Anecdotal evidence is a story told by individuals. It comes in many

forms that can range from product testimonials to word of mouth. It’s
often testimony, or a short account, about the truth or effectiveness
of a claim. Typically, anecdotal evidence focuses on individual results,
is driven by emotion, and presented by individuals who are not sub-
ject area experts.

The following are examples of anecdotal evidence:

• Wow! I took this supplement and lost a lot of weight! This pill
must work!
• I know someone who smoked for decades, and it never pro-
duced any significant illness. Those claims about smoking are
exaggerated!
• This anti-aging cream took years off. It must be the best!

I’m sure you’ve heard that you can’t trust anecdotal evidence. How-
ever, we still ask our friends for recommendations about restaurants,
travel destinations, auto mechanics, and so on. The tricky thing about
anecdotal evidence is that even when an individual story is true, it can
still be entirely misleading. How does that work?

The table below shows how statistical methodology is the opposite of

anecdotal evidence.

15
 Jim Frost

Statistics Anecdotal Evidence

Samples are large and representa- Small, biased samples are not
tive. Using proper methodologies, generalizable.
they are generalizable outside the
sample.
Scientists take precise measure- Unplanned observations are
ments in controlled environments described orally or in writing.
with calibrated equipment.
Other relevant factors are meas- Pertinent factors are ignored.
ured and controlled.
Strict requirements for identify- Anecdotes assume causal rela-
ing causal connections. tionships as a matter of fact.

A quick look at the table should be enough to convince you that anec-
dotal evidence is not trustworthy! However, it’s even worse thanks to
psychological factors that prime us for believing these stories.

Humans are more likely to tell and remember dramatic, extraordinary

personal stories. Throw in some emotion, and you’re more likely to
believe the story. In psychological terms, statistical analysis of data
that scientists collect carefully from well-designed experiments lacks
that emotional kick. Sad but true.

Furthermore, if B follows A, our brains are wired to assume that A

causes B.

Finally, anecdotal evidence cherry-picks the best stories. You don’t

hear about all of the unsuccessful cases because people are less likely
to talk about them.

When Fred tells an emotional story about taking a supplement and

losing a lot of weight, we’ll remember it and assume the supplement
caused the weight loss. Unfortunately, we don’t hear from the other

16
 Introduction to Statistics: An Intuitive Guide

ten people who took the supplement and didn’t lose weight. We also
don’t know what else Fred might be doing to lose weight.

Collectively, these factors bias conclusions drawn from anecdotal ev-

idence towards abnormal outcomes and unjustified causal connec-
tions.

Next, I’ll illustrate the problems graphically and explain how statistics
and the scientific method deals with them.

The graph below displays the results from anecdotal stories of people
who took a hypothetical weight loss supplement. Think of this chart
as a summary of the results presented in a TV commercial. We’ll even
assume that these people are telling the complete truth. The supple-
ment looks effective, right? They’ve lost a lot of weight! When you
see the individuals and hear their emotional stories about weight loss,
we want to believe that the supplement worked.

Regrettably, the graph doesn’t provide the full story. Remember, an-
ecdotal evidence uses small non-random samples that aren’t general-
izable beyond the sample. The individuals might have been cherry-
picked for their narratives, or perhaps they presented the tales on

17
 Jim Frost

their own. Either way, it is a sample based on having a dramatic and

emotionally compelling story. As the fine print says, their results are
not typical!

Unfortunately, our minds are wired to believe this type of evidence.

We place more weight on dramatic, personal stories.

A scientific study of the weight loss supplement

Now, let’s imagine that we conduct a scientific experiment using a
more substantial, random sample that represents the broader popula-
tion. We’ll also include a treatment and control group for comparison.
We must go beyond a few compelling stories and get the bigger pic-
ture that scientific studies can provide.

In this graph, blue dots represent the supplement takers, and red dots
represent those who didn’t take the supplement.

These results are not as impressive as the other graph. Some who took
the supplement lost the weight shown in the TV ad, but many more
lost much less weight. Those people didn’t come forward with their
less exciting stories!

18
 Introduction to Statistics: An Intuitive Guide

Furthermore, those who did not take the supplement fit the same pat-
tern as those who did. Collectively, taking the supplement didn’t pro-
duce more significant weight loss than the control group.

Because this study uses a random sample rather than a self-selected

sample, we have reason to believe that we can apply these results out-
side the sample.

How Statistics Beats Anecdotal Evidence

In statistics, there are two basic methods for determining whether a
dietary supplement causes weight loss: observational studies and ran-
domized controlled trials (RCTs).

In an observational study, scientists measure all pertinent variables in

a representative sample and then generate a statistical model that de-
scribes the role of each variable. For each subject, you measure varia-
bles such as basal metabolic rate, exercise, diet, health, etc., and the
consumption of dietary supplements. After you factor in the role of
all other relevant variables, you can determine whether the supple-
ment correlates with weight loss. Anecdotal evidence provides none
of this critical, contextual information.

Randomized controlled trials (RCTs) is the other method. RCTs are

the gold standard because they allow you to draw causal conclusions
about the treatment effect. After all, we want to determine whether
the supplement causes weight loss. RCTs assign subjects to treatment
and control groups randomly. This process helps ensure that the
groups are comparable when treatment begins. Consequently, treat-
ment effects are the most likely cause for differences between groups
at the end of the study.

Don’t worry. We’ll cover all of this information about experiments in

much greater detail starting in chapter 7. For now, understand that
using data to identify causal relationships (the supplement causes the

19
 Jim Frost

weight loss) and being able to generalize the results beyond the origi-
nal sample requires the researchers to use various data collecting pro-
cedures and experimental designs.

Making decisions based on anecdotal evidence might not always be

harmful. For example, if you ask a friend for a restaurant recommen-
dation, the risk is low, especially if you know his/her tastes. However,
if you’re making important decisions about things like finances,
healthcare, and fitness, don’t base them anecdotal evidence. Look at
scientific data and expert analysis even though they’re not as flashy as
emotionally charged stories presented by relatable people!

If you find yourself being won over by anecdotal evidence, remind

yourself that the results are not typical!

Organization of this Book

Unless you are involved data analysis, you’re probably more familiar
with the graphical and numeric results that someone else produced.
However, the field of statistics covers much more than that, including
a wide variety of processes and methodologies for producing trust-
worthy results.

Consequently, I’ve split this book into two major parts.

The first portion covers the essential tools of the trade, including data
types, summarizing the data, and identifying relationships between
variables. I’ll describe the types of data and how to present them
graphically. Learn how summary statistics represent an entire dataset
and describe where an observation falls within it. These statistics in-
clude measures of central tendency, measures of variability, percen-
tiles, and correlation. Then, we’ll move onto probability distributions.
Probability distributions help you understand the distribution of val-
ues and calculate probabilities. We’ll pay extra attention to the crucial
Normal Distribution.

20
 Introduction to Statistics: An Intuitive Guide

Collectively, this knowledge allows you to understand the basics of

the different types of data, how to summarize a dataset, identify rela-
tionships between different types of variables, and use probabilities to
know how the values are distributed. These skills will allow you to
summarize a dataset and explain relationships between variables to
others.

The second portion covers the practices and procedures for inferen-
tial statistics and using statistics in the scientific process. Inferential
statistics allow you to use a relatively small sample to learn about an
entire population. However, making this leap from a sample to the
population requires additional procedures and methodologies. I’ll also
cover how the field of statistics fits in with the scientific method along
with the essentials of designing experiments to answer questions.

END OF FREE SAMPLE

NOTE: This sample contains only the introduction and first chapter.
Please buy the full ebook for all the content listed in the Table of Con-
tents. You can buy it in My Store.

21