EC 285 - 2026-09-10 - Lecture 01

``

EC 285 - Lecture 01 - 2026-09-10

Introductory Statistics · Zara Liaqat · SB 106

Week 1 · 16:00-17:11 · 71 min · Full transcript · Flashcards

Overview

This first lecture was mostly the course introduction. The lecturer explained what statistics is for: using a sample to make well-reasoned claims about a population you can't fully observe, and attaching a probability to those claims because the world is uncertain. She then introduced the core vocabulary: population vs. sample, parameter vs. statistic, descriptive vs. inferential statistics, and correlation vs. causation. The rest of the class covered admin, including grading, tutorials, assignments, exams and the course's AI Tutor. Sampling methods and the remaining Chapter 1 material were left for Tuesday.

What statistics is for

Decisions under uncertainty

Everyday decisions are made with incomplete information. Statistics is a set of tools to process (collect), summarize, analyze and interpret data so those decisions are better informed. The course's one-line summary: "how we can use a sample of individuals to make well-reasoned claims about an unobserved population."

Why: You can never be 100% sure about the real world, so the best you can do is use data (the more the better) and state how confident you are. Her examples were:

Probability

Probability is the likelihood or chance that something happens. It is a number between 0 and 1, used when you are not certain of the outcome. An example is a 30% chance of rain tomorrow.

Why: Because the environment is uncertain, the course uses probability distributions to put as accurate a probability as possible on events happening.

Hypothesis testing preview

The "well-reasoned claims" in the course summary are what the course calls hypothesis testing. You have a hypothesis about the population, and you test it with a representative sample. This is Chapter 9, and a lot of groundwork comes first.

Why: A finding is only credible, reliable and publishable if it makes sense and is statistically significant. She only gave a loose preview of that here: the probability attached to the result should be fairly high, for example estimates that are "90% statistically significant or 80% statistically significant." (Standard wording is different. A result is "significant at the 10% level" or "at the 5% level", which goes with 90% or 95% confidence. The significance level is the small number, not the big one. Use the Chapter 9 wording on tests.) In the lecturer's words, if "it's only a 50% probability that this will be true for the larger population, then it's not worth anybody's attention." The exact meaning of significance comes with hypothesis testing in Chapter 9.

Population and sample

Population and sample

Why study samples?: Surveying everyone is impossible, slow and expensive. Even Statistics Canada doesn't collect data on every single Canadian. A sample is faster and cheaper, and if it is representative it can give very good predictions about the whole population. The lecturer stressed that the sample size should be large: "the higher the better for your sample to be representative."

Parameter and statistic

A parameter is a characteristic of the population that can be estimated by teh statistic. A statistic is a characteristic of a sample, such as the average grade or average age in the sample.

Why: It is the same idea (for example, an average) computed on two different groups. The name tells you which group you mean. The population value is usually unknown, and the sample value is what you actually calculate.

Exam

"Parameter versus statistic, another potential multiple choice question." She added: "If I'm saying the parameter is equal to 50, it means I'm referring to the population. Statistic is for the sample." She said she'll keep calling out these "tiny differences in single terms," so you should know them.

Two branches of statistics

The course splits into these two parts. The smaller first part is descriptive, and most chapters are inferential.

Descriptive and inferential statistics

Why: A big spreadsheet of raw numbers means nothing to you on its own. So the first step is always descriptive: draw charts to spot patterns and trends and to catch problems in the data. Only then do you move on to forecasting, predictions and hypothesis testing.

Exam

"This, by the way, is a potential multiple choice question on the midterm, where I ask you the difference between descriptive and inferential statistics."

Charts for summarizing data

Charts turn a table of numbers into a picture, so you can see at a glance where the values are concentrated. The slides list these charts:

Variables are usually noted capital letters (X, Y) and are characteristics or measurements that can be determined for each member of the population.

A pie chart shows each category's share of the total as a slice.

Why: Even the five-row hospital table is easier to read as a pie chart. With a big spreadsheet of many rows and columns, you would not see any pattern without a chart.

percent of total=category counttotal count×100

The category count is the number of items in one category, and the total count is the number across all categories.
Not on any formula sheet, so memorise it.

Example: Hospital patients by unit (pie chart)

Patients per year: Cardiac Care 1,052; Emergency 2,245; Intensive Care 340; Maternity 552; Surgery 4,630.

  1. 1052+2245+340+552+4630=8819 -- total patients across all units
  2. 4630/8819×100=52.50% -- Surgery's share
  3. 2245/8819×100=25.46% -- Emergency's share
  4. 1052/8819×100=11.93% -- Cardiac Care's share
  5. 552/8819×100=6.26% -- Maternity's share
  6. 340/8819×100=3.86% -- Intensive Care's share

Answer: Rounded to whole percents, the slices are Surgery 53%, Emergency 25%, Cardiac Care 12%, Maternity 6% and Intensive Care 4%. The chart makes it obvious that most patients are in Surgery.

The software she named for making charts and analyzing data is Excel, R, EViews, Python and Stata. This course uses Stata, and Excel comes up in the last chapter.

Outliers

An outlier is a data point far away from the rest of the data.

Why: Plotting your data shows outliers right away. They matter because in a small sample a few outliers can bias the average, the median and the standard deviation. The average and the standard deviation are hit hardest. The median moves much less, because it only depends on the middle of the sorted data. In a huge class, two or three outliers barely move the average. An odd point might also be a data-entry mistake, so you go back and check it. You might then drop outliers, or report the mean both with and without them.

Made-up example: How one outlier moves an average

A small class has four midterm scores: 60, 65, 70, and one outlier at 30.

  1. (60+65+70+30)/4=225/4=56.25 -- the average including the outlier
  2. (60+65+70)/3=195/3=65 -- the average without the outlier

Answer: One score moves the average from 56.25 to 65, almost 9 points. This is why outliers matter much more in small samples.

Correlation and causation

Two variables are correlated when they move together, for example rising together in a straight-line pattern. Causation means one variable actually makes the other change. Correlation alone does not tell you anything about causation.

Why: Two things happening together doesn't mean one causes the other. Her example: on a hot, sunny day you eat ice cream at the beach and get a sunburn an hour later. Ice cream and sunburn may be correlated, but the ice cream didn't cause the sunburn. The heat and the missing sunblock did. The slide's scatter plot shows two variables that are positively correlated, but "we don't know anything about causation."

Exam

She said the distinction "is important and we talk about this quite a few times actually in the course."

Exam checklist

Admin & deadlines


Transcribed automatically from the lecture recording. Slides used: EC 285 - Course Introduction (slides).pdf, EC 285 - Ch. 01 - Sampling and Data (slides).pdf. Course info used: EC 285 - Syllabus - Fall 2026.pdf. Audio archived at /mnt/porsche/configs/lectures/archive/2026-09/EC 285 - 2026-09-10 - Lecture 01.opus.