Correlation

Correlation analysis is used when you want to analyze the relationship between scale variables.

As an example, if you want to see if there is a relation between personality and affection you can use a correlation analysis. Personality is a concept that, according to some authors, comprises 5 dimensions: extroversion, kindness, conscientiousness, neuroticism and openness to experience. As for affection, some authors refered 2 dimensions: positive and negative.

 Using the commands as shown in the following image, a new window will open (second image) where you can select the personality and affection variables .

 

correlation1

correlatio2

The output will show all the correlations between variables. The ones marked with ** are correlations significant at the 0.01 level  and the ones marked with * are significant at 0.05 level.

correlation3

 

 

Classifying variables

To set variables in SPSS, as of questions in a survey, it is not always easy. Although some are more difficult than others, if we are beginners in SPSS,   we tend to complicate or to devalue the correct completion of SPSS sheet.
Take for example the following cover page of a survey:

1. Age: _________2. Gender:

  1. Male
  2. Female

3. Marriage status:

  1. Single
  2. Married or living together
  3. Divorced or separated
  4. Widower

4. Educational background:

  1. Elementary school
  2. Middle School
  3. High school
  4. College

5. Family income (month):

  1. Up tp 500€
  2. 501€ e 1000€
  3. 1001€ e 2000€
  4. 2001€ e 3000
  5. 3001€ e 4000€
  6. 4001€ e 4500€
  7. More than 4500€

The previous questions in a survey, which are used to describe the sample of a study, are called sociodemographic variables. This variables will assume the following types and measures in SPSS.

Question

Type

SPSS

 

 

Type

Measure

Age

 Quantitative

Numeric

Scale

Gender

Qualitative

Numeric

Nominal

Marital status

Qualitative

Numeric

Nominal

Educational background

Qualitative

Numeric

Ordinal

Family income (month)

Qualitative

Numeric

Ordinal

Variables can be qualitative or quantitative:

  • Qualitative variables show a quality, present or absent, and each category is mutually exclusive and exhaustive. That is, if an individual belongs to a category he can’t belong to any other. The reason is that one of the categories comprehensively qualifies that individual. This type of variable can be in a nominal or ordinal scale.

An example of a nominal variable is gender: either you are male or female (the categories are mutually exclusive) and one of the categories comprehensively qualifies the individual. Numbers may be used to identify the categories of a measure. For this reason the variable takes a numeric type (Type = numeric): “0” for male and “1” for female.  As for marital status, categories can have the following values: “1” = single; “2” = married; “3” = divorced; “4” = widower. The use of numbers is an easier way to enter data in SPSS, saving time and effort.

For ordinal variables, besides being mutually exclusive and exhaustive , they show an order of magnitude. The variable educational background is a good example, with 4 categories. It can assume the following values: “1” = Elementary school; “2” = Middle School; “3” = High school; “4” = professional course; “5” = College. We know that an individual on category 2 presents a higher order than other category 1. This  does not mean, however, that the value 2 is twice the value of 1; nor does it mean that the difference between categories 2 and 3 is the same as the difference between categories 1 and 2.

  • Quantitative variables have the the same properties as qualitative variables do and additionally allow to measure the difference between values. That is, the difference between 8 and 10 is the same as the difference between values 100 and102. This difference is equal at any point of the scale. These variables can be presented in an interval or ratio scale. In either case, they are considered a scale measure and values are allways numeric.

An example of an interval variable is temperature in Celsius. A difference of five values is equal at any point of the scale, either between 16 and 21, or between 36 and 41. Although a zero value can exist , this does not mean the absence of heat and, as such, is not an absolute zero. In temperature, zero is an arbitrary value corresponding to the freezing point of water.

Ratio variables have an absolute zero, in addition to all the properties of an interval variable. This is true for income, where a zero corresponds to an absence of income.

 

 

 

 

Free statistics and reports

Statistics automated reports in PDF and DOCX

Better than SPSS or any other statistical package. Why?
STAT.COOL discusses the results and writes the reports automatically.

A website that calculates statistics and writes automatic reports, at no charge, was launched today.

The address is: https://stat.cool/

This site was developed by the team of SPSS Tips, and at this early stage offers the following statistics and free reports:

  • Frequencies
  • t test

The following statistics are also already available with a symbolic payment of 20 € per year, and the user can perform all the statistics he wants and get the respective reports with all analysis performed and explained.

  • Chi-square test of independence
  • ANOVA
  • Correlation
  • Linear regression
  • Factor analysis

Soon the following statistics and respective analyzes and automatic reports will be available.

  • Mann-Whitney
  • Kruskal-Wallis
  • Kolmogorov-Smirnov test for normality
  • Logit
  • Structural Equation Model

Join now!

Address: https://stat.cool/

SPSS tips and much more

Welcome to SPSS Tips, the site where you can ask SPSS questions, consult articles, and find application examples.

In addition to the use of SPSS, we also provide tips on how to build questionnaires, which statistics should be used in every case, and more.

See you soon