Statistics bible
Pre-master Communications and Information
Sciences
*Some notes are in Dutch
, When to run which test
One-sample t-test
You compare the entire group’s scores (Words and Non-words) against some predetermined
cut off score (50%)
- Is the percentage of correct judgments for non-words more than 50%?
- Is the percentage of correct judgments for words more than 50%?
Independent t-test
Comparing the scores on a dependent variable (Drinks) of two different groups (Gender
boys vs. girls)
- Do boys drink more units of alcohol per week than girls? in dit geval is dit een
independent t-test, omdat het gaat om units of alcohol, dit kunnen 1, 2, 3, 4, 5, etc
units zijn…… oneindig dus.
LET OP!!! Wanneer het wel binnen een categorie valt (mobile phone (yes or no), gender
(men or woman)) dan gebruik je een chi-square
Bijv: Boys use their mobile phones more than girls
Paired samples t-test
Because the scores come from the same subjects, we are dealing with dependent scores and
so we need the Paired-samples T test.
- People’s correctness judgement of real English words is better than their correctness
judgement of non-words.
One-Way ANOVA
ANOVA is a statistical technique with which you can test the significance of the differences
among independent group means. The main difference with the Independent-Samples T test
is that ANOVA allows you to compare three (or more!) mean scores at once, rather than just
two.
- Younger people are less optimistic in life than older people (young and old people are
divided in three categories)
- The difference in how much stress respondents perceive depends on the
source of their stress.
o Respondents with stress about their health perceive more stress than
respondents who stress about their children or finances
o Respondents who experience stress about their children perceive more stress
than respondents with stress about their finances.
Factorial ANOVA
It is an extension of the one-way ANOVA, which can analyze the effects of only one
independent variable. You can use Factorial ANOVA to analyze the effect two factors X and
Y exert separately on some dependent variable A, but also to analyze their combined effect
on a (main effect & interaction effect)
- Text complexity will have an effect on appreciation. More in particular, the ‘simple’
version will be appreciated better than the other two versions
- Image type will have an effect on appreciation. More in particular, ‘supportive’
images will be appreciated better than ‘non-supportive’ images.
, - The ‘Complex’ version of the text will be appreciated better if it combined with the
supportive image than with the illustrative image.
- Ads with image animation score higher on attitude than ads without image animation
- Ads with a tagline score higher on attitude than ads without a tagline
- The effect of animation on the score on attitude (H1) will be stronger if the ad
contains a tagline
Chi-square
The Chi-square test (or χ2 test, as it is commonly written) is designed to analyze associations
between two categorical variables, i.e., variables that consist of frequencies or counts. In
addition, the Chi-square test allows you to analyze the frequency distribution of one
categorical variable against some expectation (AND CATEGORICAL)
Het verschil tussen een independent t-test en de chi-square is dat er bij de independent t-test
geen categorieën, maar continuous variabelen zijn. Bij de chi-square alleen maar ja/nee,
man/vrouw, A/B/C etc. vragen.
- Boys are more likely than girls to have their own mobile phone.
- Children from divorced parents are more likely to have a mobile phone than children
whose parents are not divorce.
- Comparing the preferences for three different brands of soda (A, B, C) among three
different age groups (18-25, 26-40, 41 and above).
- Checking whether the distribution of blood types in a sample population follows the
expected distribution based on the known frequencies in the general population
(Goodness of Fit test).
Correlation
Correlation analysis is a technique that allows examining the association between two
continuous variables, or between two ordinal variables
- Hypothesis 1: Students who drink less alcohol are more conscientious.
- Hypothesis 2: Students from a larger city drink more alcohol.
Voorbeeld van het verschil tussen continuous en ordinal variabelen:
Continue Variabele: De hoeveelheid tijd die studenten besteden aan studeren per week (in
uren). Dit kan elke waarde zijn, zoals 3, 5.5, 8, 12.3, enz.
Ordinale Variabele: De rangorde van de studenten in de klas. Stel je voor dat er 30
studenten zijn en ze zijn gerangschikt van 1 (de hoogste score) tot 30 (de laagste score).
Simple Linear Regression
Regression analysis is a statistical technique to quantify the relation between a predictor
variable and an outcome or response variable. The technique is used to examine the
hypothesis that the predictor significantly predicts the outcome.
- A person’s BIS score (behavioral inhibition) predicts his or her extraversion.
- A person’s score on the BAS_fun scale predicts how much alcohol s/he drinks.
- Interpersonal interest predicts how good people are in predicting the gift preference of
a person, with higher scores on interpersonal interest relating to a better gift selection.
Pre-master Communications and Information
Sciences
*Some notes are in Dutch
, When to run which test
One-sample t-test
You compare the entire group’s scores (Words and Non-words) against some predetermined
cut off score (50%)
- Is the percentage of correct judgments for non-words more than 50%?
- Is the percentage of correct judgments for words more than 50%?
Independent t-test
Comparing the scores on a dependent variable (Drinks) of two different groups (Gender
boys vs. girls)
- Do boys drink more units of alcohol per week than girls? in dit geval is dit een
independent t-test, omdat het gaat om units of alcohol, dit kunnen 1, 2, 3, 4, 5, etc
units zijn…… oneindig dus.
LET OP!!! Wanneer het wel binnen een categorie valt (mobile phone (yes or no), gender
(men or woman)) dan gebruik je een chi-square
Bijv: Boys use their mobile phones more than girls
Paired samples t-test
Because the scores come from the same subjects, we are dealing with dependent scores and
so we need the Paired-samples T test.
- People’s correctness judgement of real English words is better than their correctness
judgement of non-words.
One-Way ANOVA
ANOVA is a statistical technique with which you can test the significance of the differences
among independent group means. The main difference with the Independent-Samples T test
is that ANOVA allows you to compare three (or more!) mean scores at once, rather than just
two.
- Younger people are less optimistic in life than older people (young and old people are
divided in three categories)
- The difference in how much stress respondents perceive depends on the
source of their stress.
o Respondents with stress about their health perceive more stress than
respondents who stress about their children or finances
o Respondents who experience stress about their children perceive more stress
than respondents with stress about their finances.
Factorial ANOVA
It is an extension of the one-way ANOVA, which can analyze the effects of only one
independent variable. You can use Factorial ANOVA to analyze the effect two factors X and
Y exert separately on some dependent variable A, but also to analyze their combined effect
on a (main effect & interaction effect)
- Text complexity will have an effect on appreciation. More in particular, the ‘simple’
version will be appreciated better than the other two versions
- Image type will have an effect on appreciation. More in particular, ‘supportive’
images will be appreciated better than ‘non-supportive’ images.
, - The ‘Complex’ version of the text will be appreciated better if it combined with the
supportive image than with the illustrative image.
- Ads with image animation score higher on attitude than ads without image animation
- Ads with a tagline score higher on attitude than ads without a tagline
- The effect of animation on the score on attitude (H1) will be stronger if the ad
contains a tagline
Chi-square
The Chi-square test (or χ2 test, as it is commonly written) is designed to analyze associations
between two categorical variables, i.e., variables that consist of frequencies or counts. In
addition, the Chi-square test allows you to analyze the frequency distribution of one
categorical variable against some expectation (AND CATEGORICAL)
Het verschil tussen een independent t-test en de chi-square is dat er bij de independent t-test
geen categorieën, maar continuous variabelen zijn. Bij de chi-square alleen maar ja/nee,
man/vrouw, A/B/C etc. vragen.
- Boys are more likely than girls to have their own mobile phone.
- Children from divorced parents are more likely to have a mobile phone than children
whose parents are not divorce.
- Comparing the preferences for three different brands of soda (A, B, C) among three
different age groups (18-25, 26-40, 41 and above).
- Checking whether the distribution of blood types in a sample population follows the
expected distribution based on the known frequencies in the general population
(Goodness of Fit test).
Correlation
Correlation analysis is a technique that allows examining the association between two
continuous variables, or between two ordinal variables
- Hypothesis 1: Students who drink less alcohol are more conscientious.
- Hypothesis 2: Students from a larger city drink more alcohol.
Voorbeeld van het verschil tussen continuous en ordinal variabelen:
Continue Variabele: De hoeveelheid tijd die studenten besteden aan studeren per week (in
uren). Dit kan elke waarde zijn, zoals 3, 5.5, 8, 12.3, enz.
Ordinale Variabele: De rangorde van de studenten in de klas. Stel je voor dat er 30
studenten zijn en ze zijn gerangschikt van 1 (de hoogste score) tot 30 (de laagste score).
Simple Linear Regression
Regression analysis is a statistical technique to quantify the relation between a predictor
variable and an outcome or response variable. The technique is used to examine the
hypothesis that the predictor significantly predicts the outcome.
- A person’s BIS score (behavioral inhibition) predicts his or her extraversion.
- A person’s score on the BAS_fun scale predicts how much alcohol s/he drinks.
- Interpersonal interest predicts how good people are in predicting the gift preference of
a person, with higher scores on interpersonal interest relating to a better gift selection.
