SAS Lesson 4 - Refining your research question by selecting rows

Loading...

来自 Wesleyan University 的课程

数据管理与可视化

666 个评分

Explore Related Videos

At Coursera, you will find the best lectures in the world. Here are some of our personalized recommendations for you

Wesleyan University

数据管理与可视化

666 个评分

课程 1（共 5 门，Specialization Data Analysis and Interpretation）

无论是定制百万流量的网页广告，还是提高小餐馆的库存效率，数据的作用变得越来越显而易见。曾几何时，我们不知道如何通过数据挖掘找提升业务效率的办法。但在这个课程中，你会了解到到什么是数据，什么是能够通过数据挖掘解决的问题 — 当然也可能你从来没有研究过这些。你会了解到到如何通过现有的数据进行一个问题的研究，找到各种变量、变量之间的关系、及基础的统计学和如何清晰得展示你的研究结果。在完成这门课程之后，你将能够运用强大的数据分析工具－类似SAS或Python －来管理和可视化你的数据，以及如何应对数据丢失、组合的变化和图表处理等问题。同时，通过这门课程，你可以和大家一起分享你的所学并获得反馈，也能够了解你的同伴们是如何运用数据解决他们的问题。

从本节课中

Writing your first program - SAS or Python

In this session, we will discuss how to write a basic program that allows you to load a data set and examine frequency distributions. Your task will be to write a program that helps you to explore the variables you have selected for your own research question. You may choose either Python or SAS. Both are made freely available, and we have created a helpful guide to support you in making the decision. Once you have selected your platform, just follow the instructions in the appropriate "GETTING STARTED...." file, and then check out this week's video lessons aimed at helping you write and run your first program. You need only view the lessons for one of the statistical platforms (SAS or Python).

SAS Lesson 1 - Defining exploratory data analysis3:00

SAS Lesson 2 - SAS coding conventions8:30

SAS Lesson 3 - Running your program and examining frequency distribution9:58

SAS Lesson 4 - Refining your research question by selecting rows5:33

与讲师见面

Lisa Dierker
Professor
Psychology

Once you've examined frequency tables for your variables,

the next decision you need to make is whether you want to subset the data, and

ask a question based on specific observations.

Or whether you want to ask the question of the entire sample.

In other words are there particular rows in the data set?

That is individual observations that you want to keep and

others that you want to exclude?

For example, we're most interested in the association between smoking and nicotine

dependence in young adults who might still be establishing their smoking patterns.

Therefore, we decide to examine only those adults who are 25 or

younger and exclude all adults older than the age of 25.

We also decide that we want to focus on recent smokers,

rather than individuals who've smoked more than a year ago.

So the original research question, for example, is further clarified

to include the association between smoking and nicotine dependence.

But only among young adults who have smoked in the past twelve months.

The research question is not changed.

Rather, the question will be asked based only a subset of the observations that

are available in a specific data set that will help answer the research question.

To implement these decisions, we'll go back to our program and

add two additional statements just before the end of our data step.

These additional statements are logic statements.

Specifically, you need to add logic statements that tell the program to

include only those observations that will help you answer your research question.

Logic statements require one or more operations.

Here's a table of common operations used in SASS.

These include equal to, not equal to, greater than, less than.

Greater than or equal to and less than or equal to.

The following logic statements will need to be added to this program

in order to call in the specified subset of the observations

now to be used as the basis of the research question.

IF CHECK321 = 1; which is the variable for

cigarette smoking status within the past 12 months.

And IF AGE LE 25; indicating selection of only those

individuals less than or equal to 25 years old.

These logic statements go at the end of the data step

before the PROC SORT statement.

The variable AGE should also be added to the table statement, so

that distribution can be examined.

When this program is saved and rerun, you can see frequency distributions for

each of the variables.

But only for the 1706 individuals age 25 and

younger who reported smoking in the past 12 months.

>> I can be confident that I am now working with this sample.

Because as you can see, all of the 1,706 observation has a 1,

meaning yes for having smoked in the last 12 months.

And if I scroll down to the distribution of age,

I can see that only 18 to 25 year olds are included in the sample.

If, for your own research question,

you are completely unsure at this point whether or

not you want to limit your question to a subset of observations, no worries.

In that case, there would be no need for

you to add any additional logic statements to your program.

This is always a decision and an example that you can revisit in the future.

Here's a way to add notes to your program to help you remember what specific lines

of code do.

As you keep adding to your code,

it might become difficult to remember why you wrote what you wrote.

You can comment out your notes so

that SASS doesn't read the notes as actual program statements.

The method for doing this, is to include your notes in between a /* and a */.

Here's a note added to our example code.

It's a note to remind ourselves that the data is being subset

to include only past 12 month smokers age 18 to 25.

The note is written in green within the program.

You can be confident that it will not be read as SASS syntax.

You should definitely use this commenting out technique early and

often as you build your own program.

These notes can be very important, because they will help you keep track of decisions

you've made and actions you've taken in writing your code.

At this point, and

whenever you've added something to your code, you should save it.

探索我们的目录

免费加入并获得个性化推荐、更新和优惠。

Coursera 致力于普及全世界最好的教育，它与全球一流大学和机构合作提供在线课程。

© 2019 Coursera Inc. 保留所有权利。

通过 App Store 下载

通过 Google Play 获取