Methods I Flashcards
What type of study is the following: Fuel consumption of pizza delivery motorbike
Descriptive
What type of study is the following: Vote intention for the next elections
Descriptive
What type of study is the following: Sales rate by sex.
Correlational
What type of study is the following: Reading comprehension by age.
Correlational
What type of study is the following: Test scores improved by new educational program?
Causal
What type of study is the following: Ad click rate increased by new interface design.
Causal
What is meant with the study type ‘descriptive’?
Statistics on that what exists: factual gathered information.
What is meant with the study type ‘Correlational’?
Studying the relationship between multiple things.
What is meant with the study type ‘causal’?
Determine if one thing causes (an effect on) the other.
What are independent variables (IV)?
Variables that are manipulated by the researcher or nature.
What are dependent variables (DV) ?
Variables that are presumably affected by the Independent Variables.
In the process chart, how do we classify IV’s?
As (un)controllable factors on the process.
In the process chart, how do we classify DV’s?
As a response of the factors (IV’s).
What kind of variables are controllable factors?
IV’s
What kind of variables are uncontrollable factors?
IV’s
What kind of relationship do we expect from numerical vs numerical?
positive vs negative
What kind of relationship do we expect from categorial vs categorial?
count differences
What kind of relationship do we expect from numerical vs categorial (or vise versa) ?
differences across groups
What is sought after with causality (or causation)?
When we can say that one aspact causes change in another aspect -> we find the IV and DV respectively.
What do require to establish causation (causality) ?
Correlation between IV and DV, The change in IV happens before the change in DV, No other explaination possible.
Does correlation mean there is also causallity?
No, relation does not mean one causes the other.
Does causality mean there is correlation?
Yes, this is required
What is the main reason that it is hard to establish a single explaination in causality?
Unknown (and often uncontrollable) factors.
What is the third variable problem?
That there are (un/)known variables at play that influence our data without our knowledge.
What are lurking variables?
A type of third variable; Unkown to us, they can cause changes in DV’s and IV’s
What is often the cause for lurking variables in your data?
Poor planning
What are confounding variables?
Type of third variable; Known to us, They cause change in only DV’s. But their effect cannot be destinguished from that of any other IV.
What is often the cause for confounding variables in your data?
Poor design of research
What are mediating variables?
Third variable problem: Variables that facilitate the relationship between two other variables.
What are moderating variables?
Third variable problem: They have an effect on the relationship between two other variables.
What is the description of sampson’s paradox?
When the effect between two groups dissapears/reverses when looking at the subgroups.
What does sampson’s paradox imply?
That there is a third variable to group by, either (un)controllable and (un)known.
What is the description of Inspection Paradox?
The perceived observations or results are skewed from the truth due bias. (e.g. arrival time between trains feels longer than factually true.)
What paradox is this? : The average classroom size according to students is larger than it actually is, because students are more likely to get (randomly) placed in larger classrooms.
Inspection
What are quantriative variables?
Numerical variables such as age or testscores: measured quantity
What are qualitative variables?
Not numerical variables or the numerical value is meaningless.
What is meant with: levels of a variable should be exhaustive and mutually exlusive?
Exhaustive: Should contain all possible levels, Mutually exclusive: There should be a clear destinction between each level, and only one possible at a time.
What are the different types of levels of measurements?
Nominal, ordinal, interval, ratio
What level of measuremnt are variables of the type nominal?
Qualitative: sex, nationality, company (usually categorial variables)
What level of measuremnt are variables of the type ordinal?
Qualitative, There is order, but no sense of magnitude: difficulty (none < some < much), agreement (totally > partially > slightly).
What level of measuremnt are variables of the type interval?
Quantitative, There are magnitudes (equally spaced) but there is no zero or ground: Dates, age range, or temperatire (IN C OR F !)
What level of measuremnt are variables of the type ratio?
Quantitative, sense of magnitude where the zero has a meaning: mass, temperature in kelvin, length, duration.
Is it sensible to compute the median of ordinal variables?
Yes, ordinal variables are categorial.
Is it sensible to compute the mean of ordinal variables?
No, the mean is meaningless as ordinal variables have no sense of magnitude or scale. (categorial)
What is the main difference between ratio and interval variables?
Ratio variables have a (meaningfull) zero, where intervals are just magnitudes without a sens of scale.
What is the difference between thurstone and lickert?
Thurstone is agree/disagree without neutral, where lickert offers multiple magnitudes and neutral options.
What are indexes?
Quantitative (numerical) scores that combine multiple variables and reflect a single, general construct: e.g. an economical index may include dept, residence and purchasing power.
