SPSS | Basics & Principles
Rationale of statistics
● looking for meaningful measures of centrality as a reference & basis of prediction
● comparison of what you have found to a predicted situation
Measure of central tendency
● Mode: highest frequency (nominal variables)
● Median: middle score (ordinal variables)
● Mean: sum of all scores divided by number of scores (scale variables)
Variables
● Nominal (tells us only that values are different)
o binominal: hold one value (e.g. gender)
o multinominal: hold more than two values that cannot be ordered logically (e.g.
country of origin)
● Ordinal (tells us values are different & how they should be ordered, e.g. class)
● Scale (tells us that values are different, how they should be ordered and exactly how big
the gaps are between values, e.g income)
Likert scale
● Strongly agree, agree, partly agree, neither disagree nor agree, partly disagree, disagree,
strongly disagree,...
● nuances in between, in total / 7 different options possible
Semantic differential
● e.g. slow ↔ fast, good ↔ bad, small ↔ big etc (scale e.g. -3 ↔ +3)
Sample
● represents a larger population
● includes standard errors, never perfect
● confidence interval: the certainty percentage that the numbers represent reality
➔ the bigger the sample, the more reliable the data (the bigger the confidence)
● +/- = standard deviation of 68% (average distance of all scores towards the mean)
➔ is the finding significant?
Simple interpretations
● valid percentage excludes missing values
● choose most efficient way of presenting as possible
Rationale of statistics
● looking for meaningful measures of centrality as a reference & basis of prediction
● comparison of what you have found to a predicted situation
Measure of central tendency
● Mode: highest frequency (nominal variables)
● Median: middle score (ordinal variables)
● Mean: sum of all scores divided by number of scores (scale variables)
Variables
● Nominal (tells us only that values are different)
o binominal: hold one value (e.g. gender)
o multinominal: hold more than two values that cannot be ordered logically (e.g.
country of origin)
● Ordinal (tells us values are different & how they should be ordered, e.g. class)
● Scale (tells us that values are different, how they should be ordered and exactly how big
the gaps are between values, e.g income)
Likert scale
● Strongly agree, agree, partly agree, neither disagree nor agree, partly disagree, disagree,
strongly disagree,...
● nuances in between, in total / 7 different options possible
Semantic differential
● e.g. slow ↔ fast, good ↔ bad, small ↔ big etc (scale e.g. -3 ↔ +3)
Sample
● represents a larger population
● includes standard errors, never perfect
● confidence interval: the certainty percentage that the numbers represent reality
➔ the bigger the sample, the more reliable the data (the bigger the confidence)
● +/- = standard deviation of 68% (average distance of all scores towards the mean)
➔ is the finding significant?
Simple interpretations
● valid percentage excludes missing values
● choose most efficient way of presenting as possible
