#75461 - 4 Pages - Research project of computer gaming

SPSS LAB 2: Bivariate Analysis

1 | P a g e

OBJECTIVES

 Familiarizing yourself with standard bivariate analysis (cross-tabs)  Knowing which correlation test to use given level of measurement (PC, Ch-S., ANOVA)  Correlation doesn’t imply causation  Exploring data visualization strategy

PLESE NOTE

 We will NOT revise the statistical foundations of inferential statistics (t-test, p-value, etc)  We will NOT revise the four main level of measurement (nom., ordinal, interval, ratio)

DEADLINE (the good news)

 November 9th, Midnight  I will NOT be answering emails about Lab 2

SPSS LAB 2: Bivariate Analysis

2 | P a g e

What is correlation?

 Tells you if there is a relationship between two variables  It’s a measure of linear relationship between two variables

RELATIONSHIPS BETWEEN TWO CONTINUOUS VARIABLES (PEARSON_CORRELATION) EXERCISE 1 Variable 1: TOTAL_SKILL_ENDORSEMENT Variable 2: TOTAL_BIO_INFO_PROFILE

 Look at each variable in SPSS to familiarize yourself with what they measure.  Explain in a few words what they measure. (2 pts)  Use some of the univariate/descriptive skills we learned in LAB 1.  Explain in a few words what you looked at. (2 pts)  Tip = always first identify the level of measurement for each variable

Pearson’s Correlation  Measure of the strength and direction of association that exists between two continuous variables Three ways to run Pearson Correlation

1. Analyse  Correlate  Bivariate (mine doesn’t work)

2. Analyse  Descriptive Statistics  Crosstabs a. Click on Statistics  Check “Correlates”

3. SPSS SYNTAX

a. File  New  Syntax b. Copy Paste the Following Code and Run it

CORRELATIONS /VARIABLES = TOTAL_SKILL_ENDORSEMENT TOTAL_BIO_INFO_PROFILE /PRINT = TWOTAIL NOSIG /MISSING=PAIRWISE.

* Explain in a few words your interpretation of the Pearson Correlation results. (8 pts) EXERCISE 2 – RUNNING PEARSON COR. with CATEGORICAL VARIABLES Let’s look at the difference in the correlation between variable 1 and 2 for GENDER AND RACE.

 Analyse  Descriptive Statistics  Crosstabs  In Layer 1 of 1 Section add GENDER and RACE

* Explain in a few words your interpretation of the Pearson Correlation results. (8 pts)

SPSS LAB 2: Bivariate Analysis

3 | P a g e

EXERCISE 3 – RUNNING T-TEST (categorical and continuous)

T-Test: Statistical differences between the means of two groups

 Take a second to think about how T-Test is different from Pearson’s Correlation  Explain in a few words the differences between T-Test and PC. (2 pts)

Hand-on Example with Francois and Adam: Race and # of Language Run Two T-Tests:

(1) TOTAL_SKILL_ENDORSEMENT & GENDER (2) X_TITLE_WORK_1_DEGREE_CONNECTION & BACHELOR_DEGREE_RE2_DUMMY  Analyze  Compare Means  Independent-Sample T Test

o Test Variable = Dependent Variable o Grouping Variable = Independent Variable

*Explain in a few words your interpretation of the results. (8 pts) RELATIONSHIPS BETWEEN TWO CATEGORICAL VARIABLES (CHI_SQUARE)

 If you want to look at strengths of relationship between two categorical variables, you CANNOT use Pearson’s correlation.

 You must use Chi-Square Test  Example: LANGUAGES_RE2 AND BACHELOR_DEGREE_RE2_DUMMY

EXERCISE 4 Variable 1: GENDER Variable 2: BACHELOR_DEGREE_RE2_DUMMY

 “RE2” means that the variable has been recoded for a second time.  Explain in a few words how the variables were recoded the second time. (2 pts)

Before you run the Chi-Square, 1. let’s run a cross-tabs to look at the data (make sure to check off “correlations”):

2. VISUALIZATION: go to Graphs  Legacy Dialogs  Bar. Try to produce a meaningful

visual representation of the data in a way that helps you make sense of pattern(s)

Run Chi-Square: 1. Analyse  Correlate  Bivariate (mine doesn’t work)

SPSS LAB 2: Bivariate Analysis

4 | P a g e

2. Analyse  Descriptive Statistics  Crosstabs a. Click on Statistics  Check “Correlates” & “Chi-Square”

3. SPSS SYNTAX a. File  New  Syntax b. Copy Paste the Following Code and Run it

CROSSTABS /TABLES=LANGUAGES_RE2 BY BACHELOR_DEGREE_RE2_DUMMY /FORMAT=AVALUE TABLES /STATISTICS=CHISQ CORR /CELLS=COUNT /COUNT ROUND CELL.

* Explain in a few words your interpretation of the Pearson Correlation results. (8 pts) EXERCISE 5 Do you think it would be important to control for RACE when looking at language spoken? Put differently, do you think there is a statistical difference between the # of language spoken between white and non-white? (tips: use LANGUAGES_RE1, not “RE2”)  VISUALIZATION: go to Graphs  Legacy Dialogs  Bar and try to produce a meaningful visual

representation of the data in a way that helps you make sense of pattern(s) * Which test can you use to test that hypothesis of difference? (2 pts) * Explain in a few words your interpretation of the statistical test. (8 pts)

SPSS LAB 2: Bivariate Analysis

5 | P a g e

Frank and Adam’s Cheat Sheet Categorical variables are also known as discrete or qualitative variables. Categorical variables can be further categorized as either nominal, ordinal or dichotomous.

 Nominal: Religious Affiliation  Dichotomous: Gender  Ordinal: School Year

Continuous variables are also known as quantitative variables. Continuous variables can be further categorized as either interval or ratio variables.

 Interval scale: can be measure on a continuum  Ratio: Interval variable with real zero

Pearson’s The Pearson product-moment correlation coefficient, often shortened to Pearson correlation or Pearson's correlation, is a measure of the strength and direction of association that exists between two continuous variables.

 Between two independent variables Chi-Square

The Chi-Square Test of Independence is commonly used to test the following:

Statistical independence or association between two or more categorical variables.

The Chi-Square Test of Independence can only compare categorical variables. It cannot make comparisons between continuous variables or between categorical and continuous variables. Additionally, the Chi-Square Test of Independence only assesses associations between categorical variables, and can not provide any inferences about causation.

SPSS LAB 2: Bivariate Analysis

6 | P a g e

t-test

 Random Sampling  Independence of observations  DV: interval / Ratio scale  Normally Distributed Data

o Skew/Kurtosis  Homogeneity of Variance

o F-Max o Levene’s F-test

 No Need of equal sample size  Gender and degree

SPSS LAB 2: Bivariate Analysis

7 | P a g e

Statistically Significant Analysis (as a reference) CORRELATIONS /VARIABLES = TOTAL_SKILL_ENDORSEMENT TOTAL_BIO_INFO_PROFILE

 PEARSON CORRELATIONS (ONE TAIL) /VARIABLES = BACHELOR_DEGREE_RE2_DUMMY PERSONAL_INFO_SUMMARY_WORD_COUNT

 ANOVA CORRELATIONS (ONE TAIL) /VARIABLES = BACHELOR_DEGREE_RE2_DUMMY PERSONAL_INFO_SUMMARY_WORD_COUNT_CAT

 CHI-SQUARE CORRELATIONS /VARIABLES = GENDER DEPRESSION_SCALE

 ANOVA CORRELATIONS /VARIABLES = BACHELOR_DEGREE_RE1 DEPRESSION_SCALE

 ANOVA CORRELATIONS /VARIABLES = BACHELOR_DEGREE_RE1 TOP_TEN_SKILLS_RE1

 ANOVA CORRELATIONS /VARIABLES = TOTAL_SKILL_ENDORSEMENT LANGUAGES_RE2

 ANOVA CORRELATIONS /VARIABLES = LANGUAGES_RE2 BACHELOR_DEGREE_RE2_DUMMY

 CHI-SQUARE /TABLES=X_TITLE_WORK_1_DEGREE_CONNECTION BY BACHELOR_DEGREE_RE2_DUMMY /FORMAT=AVALUE TABLES /STATISTICS=CHISQ

 ANOVA

ONEWAY X_TITLE_WORK_1_DEGREE_CONNECTION BY BACHELOR_DEGREE_RE1 /STATISTICS DESCRIPTIVES /MISSING ANALYSIS.

 ANOVA

SPSS LAB 2: Bivariate Analysis

8 | P a g e