1. Introduction to Statistics

Intro to Collecting Data

1. Introduction to Statistics

Intro to Collecting Data: Videos & Practice Problems

Topic summary

Data collection methods are crucial in studies, classified into experiments and observational studies. Experiments involve applying a treatment to measure its effects, allowing for causation assumptions. In contrast, observational studies collect data without interference, meaning causation cannot be assumed. For example, testing a medication with a placebo is an experiment, while surveying students about sleep habits is observational. Understanding these distinctions is essential for interpreting results accurately and applying statistical principles like the Central Limit Theorem and confidence intervals in research.

concept

Introduction to Collecting Data

Video duration:

Introduction to Collecting Data Video Summary

Data collection is fundamental in research, and the method chosen can significantly influence the conclusions drawn. There are two primary methods for collecting data: experiments and observational studies, each with distinct characteristics and implications for causation.

In an experiment, researchers implement a change, often referred to as a treatment, and measure its effects. This method allows for the assumption of causation, meaning that any observed differences in results can be attributed to the treatment applied. For instance, if a study tests a medication by administering it to one group while giving a placebo to another, the differences in outcomes can be directly linked to the medication's effects.

Conversely, an observational study involves observing and measuring characteristics without any intervention. Researchers collect data on the current state of affairs without influencing the subjects or conditions. For example, surveying college students about their sleep habits and grades provides insights into their behaviors but does not allow for conclusions about causation. Trends may be observed, but the researcher cannot definitively state that one factor affects the other.

To further illustrate these concepts, consider the following examples:

1. **Testing a Medication**: In this scenario, 15 subjects receive a placebo while another 15 receive the actual medication. This setup is an experiment, as it involves a controlled change (the medication) and allows for the assumption of causation regarding any differences in symptoms observed.

2. **Surveying College Students**: Here, researchers ask 30 students about their sleep habits and grades. This is an observational study since it merely collects data without manipulation, and thus, causation cannot be assumed even if trends are noted.

3. **Rolling Dice**: In an experiment where a fair die and a loaded die are rolled 10 times each, the comparison of results indicates a change in conditions. This is also an experiment, allowing for the assumption of causation based on the differences observed in the outcomes of the rolls.

Understanding these methods is crucial for interpreting research findings accurately. By recognizing whether a study is experimental or observational, one can better assess the validity of the conclusions drawn regarding causation.

Problem

A regional store manager wants to test whether increasing store hours increases profits, so they randomly select half of their locations to stay open an extra hour later in the evenings and compare profits between stores at the end of the month. They notice that stores open later saw higher profits on average. Is this an experiment or an observational study? Can they determine the extra hours caused the increase in sales?

Observational study; no

Experiment; yes

Observational study; yes

Experiment; no

Problem

A store surveys its target demographic and learns that 86% of people would purchase the product they’ve been heavily advertising. Is this an experiment or an observational study? Can they conclude their current advertising strategy caused the high percentage of interest?

Experiment; yes

Experiment; no

Observational study; yes

Observational study; no

Problem

An office manager wants to determine how employees feel about their personal growth and professional achievement in the last quarter. Should they run an observational study or experiment?

Experiment

Observational study

Neither

Both

Problem

A software development company created a new app for fitness, and they want to determine if using the app can lead to weight loss and increased strength in customers. Should they run an observational study or experiment?

Observational study

Experiment

Survey

Case study

Here’s what students ask on this topic:

What is the difference between an experiment and an observational study in data collection?

An experiment involves applying a treatment or change to a group and measuring its effects, allowing researchers to assume causation. For example, testing a medication by giving one group a placebo and another the actual medication is an experiment. In contrast, an observational study involves observing and measuring characteristics without manipulating the situation, meaning causation cannot be assumed. For instance, surveying students about their sleep habits and grades is observational. Understanding these distinctions is crucial for interpreting results accurately and applying statistical principles.

Why can causation be assumed in experiments but not in observational studies?

Causation can be assumed in experiments because researchers control and manipulate one variable (the treatment) while keeping other factors constant. This allows them to isolate the effect of the treatment on the outcome. For example, comparing the effects of a medication versus a placebo ensures that any observed differences are due to the medication. In observational studies, no manipulation occurs, so other uncontrolled factors may influence the results, making it impossible to establish a direct cause-and-effect relationship.

Can you provide an example of an experiment and explain why it qualifies as one?

Sure! Testing a new fertilizer by applying it to one field and leaving another field untreated is an example of an experiment. This qualifies as an experiment because a treatment (the fertilizer) is applied to one group, and its effects are measured by comparing the treated field to the untreated one. The controlled application of the fertilizer allows researchers to assume causation if differences in crop yield are observed.

What are some common methods used to collect data in observational studies?

Common methods for collecting data in observational studies include surveys, interviews, and direct observation. For example, researchers might survey individuals about their eating habits, interview employees about job satisfaction, or observe traffic patterns at an intersection. These methods focus on gathering information about the current state of things without manipulating or influencing the variables being studied.

How do you determine whether a study is experimental or observational?

To determine whether a study is experimental or observational, ask whether the researchers applied a treatment or change to the subjects. If they did, it’s an experiment. For example, testing a medication with a placebo involves applying a treatment, making it experimental. If no treatment was applied and the researchers simply observed or measured existing conditions, it’s observational. For instance, surveying students about their sleep habits involves no manipulation, so it’s observational.

Your Statistics for Business tutors

Patrick Ford

Physics and Math Lead Instructor

Colleen Daly