#75461 - 4 Pages - Research project of computer gaming

Lab 1 – SPSS univariate statistics

1 | P a g e

Lab Assignment 1: Descriptive Analysis

What does the trajectory of UBC graduates look like on the job market? Institutional and provincial survey initiatives all assure us that graduated students are doing great! They are mostly all employed and extremely satisfied with what their post-secondary education brought them. What kinds of jobs are they holding? Survey’s data-points are highly aggregate and tend to answer this question with concepts like “public sector”, “private sector”, or “service industry”? How many jobs can a recent graduate go through in their first few years on the job market? Do UBC domestic graduates stay in the province or do they tend to be mobile?

To start answering labor-market questions with a finer level of granularity, Dr. Bartolic’s students built a dataset using UBC graduate LinkedIn profiles. Using the UBC 2016 graduation convocation guide, they created a dataset that included all bachelor-level 2016 spring graduates (n=5437) and collected LinkedIn data for an initial sample of 130 cases. Preliminary analysis indicates that 23% (n=1177) of all UBC 2016 spring bachelor graduates have a LinkedIn account.

In today’s lab, we will start exploring this unique dataset using SPSS. Let’s figure out what your future might looks like together!

SPSS WORKFLOW AND DATA MANAGEMENT

1. Open SPSS

2. Load Dataset File (ubc_grad_2016_SPSS.sav)

 File  Open  Data [click menu at the top of the screen]

3. Today we are working with AN EXISTING DATASET, but for your project, your first step in SPSS would be:

 File  New  Data (if entering raw data) or you would download your data file from Qualtrics

4. Spend a few minutes to explore the dataset. Look at the bottom left:

 Data View Structure: Column [variable]; row [case]

 Variable View Structure: Row [variable], Column [variable details]

Fig. 1: Data View and Variable View

Lab 1 – SPSS univariate statistics

2 | P a g e

Exercise 1 – ANSWER SOME BASIC QUESTIONS NOW (10 pts)

Now that you are a bit more familiar with the dataset, please answer the four following questions:

1. What is the # of cases in the dataset (2 pts):

2. What does each “case” represent? (2 pts)

3. What are the # of variables in the dataset (2 pts):

4. Name two variables from the dataset and explain what they are measuring (4 pts):

Exercise 2 – CREATE A NEW VARIABLE: RELIGION (10 pts)

Pay attention students, this is one of the first steps you will have to do after you have collected your data. You will open SPSSS, create a new dataset (step 3 below), and create new variables to populate your dataset (all those independent and dependent variables we have been bugging you about for 1 whole month now!)

PAY ATTENTION DEAR STUDENTS!!!!!

Two ways to create a new variable, both from DATA or VARIABLE VIEW

The best way is the following:

1. Click on Variable View (see Fig. 1 for help)

2. Say you want to add the variable RELIGION after RACE, right click on the number associated with the variable after RACE and select Insert Variable:

 Right Click  Insert Variable

3. Start with Column “Name” and Name the Variable: RELIGION (no space)

Lab 1 – SPSS univariate statistics

3 | P a g e

4. Go to Column “Label” and describe in a few words what the variable is

5. Go to Column “Values” and click on the small three dots. This should open a pop-up.

Fig. 2: CREATE A NEW VARIABLE (variable view)

Column “Values” (continued):

Here you have some thinking to do. How do you want to code the variable?

 Dummy Variable [binary/dichotomous]

o Christian code as “1” and non-Christian as “0”

 But there might be a lot of religious diversity that you wish to capture

 Nominal Variable

o Christian; “0”, Muslim; “1”, Jewish; “2”; Other; “3”, Missing; “888”

o Sounds good? Well, let’s code. The results should be:

Fig. 3: Value Label for New Variable

Lab 1 – SPSS univariate statistics

4 | P a g e

Are we done creating the new variable? No!

Next Two Steps:

6. Go to Column “Missing” and click on three small dots. A pop-up should open.

 Select “Discrete Missing Values” and Enter “888”. Press OK.

7. Go to Column “Measure” and select the appropriate type of variable. See you’re putting your hard-bookish knowledge into practice here.

8. SAVE YOUR FILE OFTEN TO NOT LOOSE YOUR PRECIOUS WORK!

9. When you save, SPSS will open a pop-up “output file”. You don’t need to save it. Just close it and don’t think about it.

Next Final Steps:

10. Switch to “Data View” (see Fig. 1 for help)

11. Find your new variable RELIGION.

12. Manually input an attribute for each respondent. (Obviously here, this is a make-believe variable, but with your own data you would input real data for each case.)

13. SAVE YOUR WORK.

Lab 1 – SPSS univariate statistics

5 | P a g e

Exercice 3 – RECODING A VARIABLE: DEPRESSION ITEMS (10 pts)

1. In the dataset (either Data or Variable View Mode), find the 11 depression-related variables and indicate their name below (2 pts).

2. Reeeeeee-coding time. Why do we need to recode data?

a. Imagine for the Religion Variable, you didn’t assign numerical values to each attribute and instead your variable Religion in your dataset was RAW. Raw means that instead of number as attributes, each respondent’s response would be a “string”: e.g. Muslim, Jewish…. Why is that a problem? (1 BONUS pt)

Find Coding inconsistencies

3. The depression questionnaire included 11 items. The general idea is that the higher a respondent’s score on the scale the more severe/likely is their depression.

a. Carefully read each item and find inconsistencies with the idea that a “high score” indicates “depression” and “lower score” indicates “non-depression”.

b. How many did you find? Indicate question [#] here. (2 pts)

DEPRESSION QUESTIONNAIRE

[1] In the past week, I felt depressed

[2] In the past week, I felt that everything I did was an effort

[3] In the past week, my sleep was restless

[4] In the past week, I was happy

[5] In the past week, I felt lonely

[6] In the past week, people were unfriendly

[7] In the past week, I enjoyed life

[8] In the past week, I did not feel like eating. My appetite was poor

[9] In the past week, I felt sad

[10] In the past week, I felt people dislike me

[11] In the past week, I could not get going

Answer choice:

(1) Hardly ever

(2) Some of the time

(3) Most of the time

Lab 1 – SPSS univariate statistics

6 | P a g e

4. You need to reverse the numerical attributes for the item(s). (6 pts)

 Bad Solution: is to do it manually. Go into Data View and then manually recode each “1” into “3”, etc.

 Why is this a bad practice?

 Good solution:

o Transform  Recode into Same Variables

o Select a variable you wish to RECODE

o Press “Old and New Values”

o Enter Old value and New value, press add, and repeat

o Press Continue

o Make sure everything is fine

o Press ok

o Go check and make sure the variable(s) were recoded (how to do this?) Compare OLD variable to the RECODED variable – they should be reversed.

 VOILA!

Lab 1 – SPSS univariate statistics

7 | P a g e

Exercise 4 – CREATING A NEW (SCALE) VARIABLE: FROM INDIVIDUAL DEPRESSION ITEMS to a DEPRESSION (SCALE) VARIABLE (5 pts)

 The goal is to create a scale (here a continuous variable) for depression, a global variable.

 Try to think of what is involved in terms of SPSS “mechanics”:

o Creating a new variable. What kind? A RECODE

o A RECODE of what? A sum of the score of all depression items

o To create a Depression Variable.

 Normally with the full SPSS package, computing a variable:

o Transform  Compute Variable

o But your Student Version is lame!

o How can you do it? Time to “pivot”, to think on your feet and solve the problem

 How did you decide to go about and create a Depression Variable? Explain your steps. (2 pts)

Lab 1 – SPSS univariate statistics

8 | P a g e

Exercise 5 – DESCRIPTIVE/UNIVARIATE ANALYSIS (10 pts)

 Run descriptive statistics for GENDER, RELIGION, LANGUAGE_RE1, and JOB

 Use measure of central tendency (mean, median, mode when appropriate). CHOOSE the correct one! (based on level of measurement)

 Use measure of dispersion (range, outliers when appropriate)

1. First use Analyse  Descriptive Statistics  Frequencies

Fig. 4 – Running Frequency Analysis

Fig. 5 – Advanced Options for Frequencies Statistics

Lab 1 – SPSS univariate statistics

9 | P a g e

2. Click on “Statistics” on the right-hand side to select measures of cent. and disp.

3. Click “Continue”

4. SPSS creates an OUTPUT file with the results. Don’t close it.

5. Frequencies Results are two-fold:

o Statistics Table

o Individual Frequency Table for each variable

6. NOTE: DON’T BE ALARMED BY A MISSING NUMBER.

Table 1 – MISSING DATA

Gender of

Graduate

COUNTRY

RECODE 1

# of Languages

(fixing missing)

# of Job(s)

currently hold by

graduate

N Valid 130 111 130 130

Missing 65405 65424 65405 65405

7. BONUS POINT ALERT. Can you explain these huge numbers for missing data? (1 pt)

8. Interpret your results. Use a succinct paragraph to describe the variables (10 pts).

o Tip: what stands-out to you? What do these basic stats tell you about the respondent?

o Tip: Focus on Valid Percent and Cumulative Percent

Lab 1 – SPSS univariate statistics

10 | P a g e

ANSWER SHEET

MAKE SURE YOU ANSWER ALL THE QUESTIONS

Exercise 1 – GET SOME BASIC THINGS NOW (10 pts)

1. # of cases in the dataset (2 pts):

2. What each “case” represents? What are they? (2 pts)

3. # of variables in the dataset (2 pts):

4. Name two variables from the dataset and explain what they are measuring (4 pts):

Exercise 2 – CREATE A NEW VARIABLE: RELIGION (10 pts)

 Snippet or according screen-shot of RELIGION VARIABLE (Data View) (5 pts)

 Snippet or according screen-shot of RELIGION VARIABLE (Variable View) (5 pts)

Exercice 3 – RECODING A VARIABLE: DEPRESSION ITEM (5 pts)

 Snippet or according screen-shot of RECODED VARIABLE(S) (View) (2 pts)

 2. A  Bonus Question (1 pt): Why need to recode string attributes with numerical values?

 3. B  Which question’s answers need to be reversed? Indicate question #. (2 pts)

Exercise 4 – COMPUTING A VARIABLE: FROM DEPRESSION ITEMS to VARIABLE (5 pts)

 Snippet or according screen-shot of COMPUTED VARIABLE(S) (Data View)

 How did you decide to go about and create a Depression Variable? Explain your steps. (2 pts)

Lab 1 – SPSS univariate statistics

11 | P a g e

Exercise 5 – DESCRIPTIVE/UNIVARIATE ANALYSIS (10 pts)

 7  Bonus Question (1 pt): Can you explain these huge numbers for missing data?

 Short paragraph describing univariate analysis. Not everything. What makes sense to discuss, what gives us a feel and an overview of the data.

o Tip: what stands-out to you? What do these basic stats tell you about the respondent?

o Tip: Focus on Valid Percent and Cumulative Percent