Session 7

Front 1

Systematization of data collection methods

Back 1

Front 2

Choice of methods:

Back 2

Front 3

Making choices about research methods

Back 3

–Each class of research methods comes with generic advantages and disadvantages

–McGrath: only one

of the three following goals can be optimized(intermediates like two ok, one bad are possible )
•Generalizability: reach as many people as you can
•Precision: have full control over study

•Realism: the degree to which the actual phenomenon is studied (vs. lab)

Front 4

Classifying research methods:

Back 4

–Descriptive research: snapshot of the status quo
–Correlational research: establishing a link between at least two variables

–Experimental research: understanding the cause of things

Front 5

Research question, conversation, and methods

Back 5

Front 6

Mixed­-methods approaches

Back 6

–impossible to design perfect single method study

–Two ways of fixing:

•Multiple contributions to the conversaton

•Mixed methods papers
–allow triangulation (E.g. large scale survey supplemented by interviews: survey gives you generalizability, interviews allow for realism and may also allow you to
make stronger claims with respect to causality)

Front 7

Sample and population

Back 7

–who you do research?
–sampling strategy must be fully alligned with research question, incorrect sampling prohibits correct solutions
–must be representative of the population
–anything can be the population, depends on the research question

Front 8

From the population to the sample

Back 8

–who do we need to work with? (pick cases where phenomenon of )

–normally only work with subgroups
–complete population are sampled very rarely (census)

Front 9

Probability sampling: (simple) random sampling

Back 9

– Each member of the population has an equal probability of being selected

– The use of random numbers applied to a list of the entire population

– Purely random samples are hard to achieve, but one can come close!

Front 10

Probability sampling: Stratified sampling

Back 10

The whole population is segmented into mutually exclusive subgroups/strata, and then units are randomly selected from each stratum

Front 11

Probability sampling: Cluster sampling

Back 11

A sample is selected in two or more hierarchical stages, with different units being selected at each stage

Front 12

Non-probability sampling

Back 12

Chance of being selected of each unit is unknown or predefined

Front 13

Non-probability sampling: Convenience sampling

Back 13

–The respondents are selected, in part or in whole, at the convenience of the researcher (with no or little effort to achieve an accurate representation)
–Generalising the result is difficult

–Can provide useful information, especially in a pilot study

Front 14

Non-probability sampling: Quota sampling

Back 14

–A convenience sample, with an effort made to ensure a certain distribution of demographic variables

Front 15

Non-probability sampling: Snowball(ing) sampling

Back 15

–Usually employed to access hard­to­reach populations and particular subgroups in the population

Front 16

Non-probability sampling: Judgement sampling

Back 16

–The researcher uses his/her judgement in selecting the units from the population for the study
–If the population to be studied is difficult to locate, or if some members are thought to be better/more knowledgeable/willing

–The determination is often made on the advice and with the assistance of theclient

Front 17

Non-probability sampling: Theoretical sampling (usually for qualitative work)

Back 17

–Selection of extreme case

–Selection of specific/typical cases

Front 18

Threats: sampling is strongly linked to validity

Back 18

–External (Generalizability)/internal (changing input factors during the study) validity must be ensured

–random error vs. systematic error