Deck 8

Question 1

longitudinal study design

Answer 1

doing research over time
trend: different sample from same population, repeated cross-sectional design
panel: one sample of the population is measured repeatedly over time
cohort: cohorts are groups of units that have one specific characteristics in common –> can be both combined with trend & panel studies

Question 2

trend study

Answer 2

same population, drawing different samples each time. good for following general development. all limitations of cross-sectional research still apply (spurious relationships, validity, reliability)

Question 3

panel study

Answer 3

one sample from the same population is measured repeatedly over time. a lot of dropout, get replaced with new research units (= replacement sample), and/or the initial sample is enlarged. drop out is rarely random. refreshment sample does mostly do not match drop-out

Question 4

problems with panel studies

Answer 4

drop out: not random! replacing drop-out units with other units –> imperfect matching.

new units cannot be strictly compared to the rest, because they have not yet been conditioned by testing

conditioning: by repeated measures, values can change (people start to think about the subject, know what already happens.
change of maeasurement validity over time –> respondents learn to anticipate follow up questions after testing.

Question 5

cohort study

Answer 5

trend or panel study of a population that shares a characteristic occuring in a particular time period. either the same or a different sample from this specific population is measured at intervals (panel vs trend). adding a cohort is an enrichment of your panel or trend study

Question 6

case control study

Answer 6

you study the (historical) past to build a story about the past that helps you understand the present, looking at cases and controls. (e.g. old people with (case) and without (controls) lung cancer) and then asking them questions about their risk behaviour in the past (smoking)

Question 7

retrospective cohort study

Answer 7

identifying units who were exposed to someting in the past (e.g. born during great depression) and compare their current status

Question 8

prospective cohort study

Answer 8

identify a cohort, measure current risk factors, and follow them over time to see development

Question 9

clinical trial

Answer 9

experiment: expose units to conditions and have control groups and see what happens

Question 10

single and multiple cohort design

Answer 10

a single cohort is studied over an extended period, e.g. 14/15 MSc students. with multiple cohort, multiple cohorts are studied over an extended period, e.g. 14/15 MSc students, 14/15 BSc students, 14/15 PhD

Question 11

three types of time related effects to consider in cohorts

Answer 11

age effect (maturation): changes over time in the variable of interest due to the aging process of the research unit
period effects (history): changes over time in the variable of interest due to an event that effects units of all ages
cohort effects (generation): differences between cohort groups

Question 12

reliability in longitudinal design

Answer 12

same as in cross sectional (precision of measurement at a given measurement occasion) +
precision/random error over repeated measurement occasions

Question 13

measurement validity in longitudinal design

Answer 13

coverage of concept, accuracy of measurement (same as cross sectional)
+
equivalent coverage and accuracy over time. and changes in systematic error of the same instrument over time

Question 14

internal validity in longitudinal design

Answer 14

confounding variables & design and model misspecification (same as cross)
+
dealing with changing confounders over time
design and model mis-specifications over time
conditioning

Question 15

external validity in longitudinal design