Lecture 10: Working With Quantitative Data

Introduction

Welcome to Week 10 of the social science research methods course at the University of Maine at Augusta.  In this week’s lecture, we’ll talk about working with quantitative data and the connection between a data matrix and the skills we developed earlier in the semester, namely variable conceptualization and variable operationalization.  To operationalize a variable is to describe in exact detail how a the value of a variable will be measured in an observation. The construction of a data matrix is when the quality of an operationalization is tested; in research, data entry is usually the point when problems in operationalization are most clearly exposed.  Working with a data matrix is not just a technical exercise, then, but also a test of the clarity of your thought.

Three aspects of work should be completed by you this week: 1) a poll at the end of the lecture, 2) a problem set, and 3) an assignment in which you construct your own data matrix regarding the recently completed 2016 elections.

The Data Matrix Brings Clarity to Quantitative Analysis

As we’ve discussed earlier in the semester, the contrast between “qualitative” and “quantitative” research unfortunately suggests that quantitative research must only measure counts (quantities) and that uncountable aspects (qualities) of observed objects cannot be captured in quantitative analysis.  This is an inaccurate characterization of quantitative research.

What quantitative research does insist upon is that quantities or qualities be expressed as variables with values that that either express some form of number or indicate some kind of category.  As you read in Chapter 12, measurements of these numbers or categories are recorded in a particular form called a data matrix, with information being located in particular places to convey particular information to a person who knows how to read the matrix.

Let’s review the structure of a data matrix. Imagine you have three space aliens named Zog, Xana and Blzp.  Everyone say hello to Zog, Xana and Blzp:

Three Aliens: Zog, Xana and Blzp

 

How would you describe these aliens?  If you were a journalist, you might write a newspaper story.  If you were a visual artist, you might represent them in an oil painting.  As a quantitative social scientist, you would characterize them in a data matrix.

The first question a social scientist might ask is, “what’s my unit of analysis?”  A unit of analysis is an explicit description of the type of entity being studied.  In this case, we might say that the alien is the unit of analysis.  Every particular alien is called a case — an individual unit being studied.  Every observable characteristic that can take on different values is a variable.  A data matrix is a table in which every variable has its own vertical column, every case has its own horizontal row, and every combination of case and variable is a cell in which the observation of a variable for a case is entered.

Here is a data matrix for the aliens:

ID # Name Name Length # Eyes # Teeth Dominant Color # Lobes Symmetric?
1 Zog 3 4 0 Yellow 5 0
2 Xana 4 2 0 Red 11 0
3 Blzp 4 0 8 Blue 5 1

This data matrix involves 8 variables and 3 cases. But how did I know what to place inside the cells?  There is some implicit operationalization going on here — that is to say, I’ve clearly made some decisions about what sort of information I’ll write down and what that information will represent.  Can you figure out explicitly what those decisions have been?  Can you declare in exact language what the operationalization of each variable is?  Give it a shot by clicking and writing in the Padlet box below.

Research Methods Operationalization Prompt Made with Padlet

One of these variables is an extension and alteration of another variable. Name Length takes the value of the variable Name and types in a new variable based on the “length” of Name (what is “length,” anyway? Operationalize it!). Taking the value of one variable and altering that value in a specific way to obtain a value for a new variable is called recoding.

Activity: Whip Up Your Own Data Matrix For Maine Election 2016

Just in case you’ve been hiding in a steel-reinforced cage somewhere in Upper Silesia, I should probably let you know that we’ve had an election in the United States in the past week.  Yes, it’s true! 😉 When most people mention “the election,” they’re referring to the race for President of the United States, but if you turned out to vote this past Tuesday, you’ll have noticed that your ballot was long, taking up many pages and asking you to make elections for many ballot questions and offices.  Among these in the state of Maine was an election for members of the Maine State Senate.

The great thing about elections is that they leave a trail of all kinds of variables with many cases behind them.  For this week’s graded activity, you’ll be looking into history that’s less than a week old, creating a data set that you could use for research purposes.  To complete this week’s DIY Activity, which is due by the end of the day on November 19, you should complete the following steps:

1.  Visit the Maine Secretary of State’s “Upcoming Elections” web page at http://www.state.me.us/sos/cec/elec/upcoming/index.html and download the list of general election candidates at http://www.state.me.us/sos/cec/elec/upcoming/gecandidates1116.xlsx.

2.  Open the gecandidates1116.xlsx file in either Microsoft Word or Google Sheets (online at http://sheets.google.com) to review the file.

3.  In this file, locate all Maine State Senate (Office “SS”) candidates for 2016.

4.  Create a proper data matrix as described in Chapter 12 of Dixon et al, featuring at least three variables and all Maine State Senate candidates as cases.

5.  Visit the Kennebec Journal’s Election 2016 results webpage at http://www.centralmaine.com/interactive/2016-election-results/ and using the page’s “Legislature” tab, find election results for the 35 Maine State Senate races.

6.  Add a “votes” variable and an “outcome” variable for each candidate in your data matrix based on what you find in Step 5. Code appropriately.

7.  Save your data matrix in a single file and upload that file in the “DIY Activity #8” section of our course Blackboard page.

Next Step In the Class Research Project: Vote For Your Favorite Campaign Sign Hypothesis

Thank you for sending in hypotheses regarding campaign signs. If I indicated to you in feedback that you’re having trouble with the format of a hypothesis, please be sure to check back with readings and lectures on hypotheses to nail this skill down; it’s important.

The next step in our class research project is to select hypotheses for our research work.  Please vote for your favorite below; we’ll move forward soon with the top three: