lecture 1-4 research methdology
concept
general decription of a social phenomenon
hypothesis
an expected, but unconfirmed relationship between variables
variable
characteristics of units that can vary (measurable/quantifiable)
hypothesis characteristics:
- can come from various sources
- provides a prediction/expectation of the results of the research
- written as a statement
- true or false?
measurement levels of variables
categorical variables = categories
continious variables = numbers
formulating descriptive hypothesis:
- must state expectation about the value taken by a variable of interest
formulating explanatory questions
- state the relationship between dependent and independent variable
1 continious statement = increases in one variable, leads to increases/decreases from another variable
2 difference statements = one continious and one categorical variable
3 conditional statements= if-then
types of continious variables
- postive relationship
- negative relationship
bivarate relationship
expected relationship between two variables
two independent variables relationship
more than two variables, with one dependent variable
mediation relationship
when an independent variable affects a dependent variable X3 -> X4 -> Y
- full mediation = 'the more time spend outdoor, the higher the change of contracting malaria'
- partial mediation
moderation relationship
two variables that are related
- when a moderator is continious, the relation between the individual variable and depoendent variable will be intensified or buffered with the values of the moderating variable
'the higher the heat, the higher the change of being hospitalized. this effect is itensified by age.'
- when a moderator is discrete, the relation is simply different between the two
'the higher the heat, the higher the chance of being hospitalized. this change is higher for people over the age of 50.'
identifying causal relationships
1. association
2. establish the direction of the influence
3. eliminate rival explanations
spurious relationships
a third variable that creates a apparent relationship.
'sunny weather causes higher ice cream consumption and higher sunburn sensitivity'
unit of analysis
about whom or what statements are made in the research
aggregated data
using combined responses across smaller units to describe a characteristic of a bigger unit
nested data
small units that are embedded within various levels and how this relates to outcomes
units of analysis: fallacy
mistaken or delusory belief or idea, founded on unsound reasoning
incorrect to generalize from smaller to bigger units/vice versa
1. ecological fallacy = group or area characteristics used to infer inidvidual characteristic
2. individualistic fallacy = inidvidual characteristics used to infer group characteristics
how to measure concepts
1. using ecisting measure insturments
2. develop measures/instruments
3. conceptualizing to operationalizing
conceptualizing =
describing/defining of concepts that one aims at measuring/observing.
latent constructs
abstract concepts that cannot be observed directly.
operationalizing
making concepts observable
measurement levels: categorical and continious
categorical =
1. nominal = variables with two or more categories, without intrinsic order 'are you a student yes or no'
2. ordinal = variables with various categories that can be ordered or ranked 'level of english levels'
continious =
1. interval = variables that can be measured along a continuum, with anumerical value 'temperature'
2. ratio = interval variables with the measurement 0 being none
reliability
- stability, consistency, precision
- random error
validity
- are you measuring what you intend to measure with this Definition?
- systematic error
improving reliability
1. test-retest
2. split-half method
3. intercoder reliability
improve validity
first:
face validity (intended to measure)
content validiy (intstrument covers all aspects (dimensions) of the concepts definition
1. Empirical validity
- predictive/criterion validity = assesing results against future external measures en comparing measurement instruments outcomes with the external outcomes.
- concurrent validity = compare other measurement instruments made at the same time
2. construct validity (relating measurement to general theoretical framework)
- correlates with related variables
- differences between known groups
- convergent validity (different ways of measuring the same concept)
- divergent/discriminant validity (a measurement instrument of a construct differentiates from other constructs from which it is supposed to differentiate)