Research methods

what is operationism?

a past doctrine that taught researchers that a concept is identical to its operationalization

what are “converging operations”?

using multiple ways to measure a concept

what is “theory dependence of measurement”?

how you measure something depends on how you define it and what youre trying to find out

what is “reactivity of meaurement”?

measuring something causes it to behave differently

what are the 4 ways of measuring something?

• self report

• observation

• quantitative

• qualitative

what is the “conservation of inference law”?

• the less you infer in the measurement phase the more you will have to infer in the evaluation phase

• when you infer a lot in the beginning you wont have to infer later when trying to make sense of the results

what are two fundamental distinctions of study designs?

experimental and non experimental

what type of design would you choose when you want to know something about causal influence of external factors ?

experimental deisgn

what design would you choose if you want to learn about the variation between people?

non-experimental

what are the two different non-experimental designs?

descriptive and correlational

what is the difference between descriptive and correlational designs

in descriptive you only look at one variable while in correlational you look at two or more and want to learn how they are related

what is the difference between cross sectional and longitudinal?

cross sectional = measurements are done at one point in time

longitudinal = measurements are done at different points in time to look at development of variable

what are the four conditions for causality ?

• covariation

• precedence

• exclusion of alternative explanation

• logical mechanism

what are the four different type of validity as suggested by Cook?

• internal validity

• external validity

• construct validity

• statistical conclusion validity

what is internal validity?

whether there is actual causal relationship

what is external validity?

how generalizable are the results (do they actually apply in the real world)

what is construct validity?

is the correct thing being measured, what is the underlying construct that is causing the change

what are the 5 different types of causation?

• causation

• reverse causation

• mediator variable

• moderator variable

• conceptual confound

what is conceptual confound?

the two concepts youre measuring are so similar and overlap so that its hard to find out, whether one is causing the other or whether you might just be measuring the same thing

what is statistical conclusion validity?

• similar to internal validity, because also about internal

• its about the accuracy of the conclusions drawn from statistical analyses in a study

• (its not about the actual causality in the relatioship but about whether you correctly detect it with your statistical analysis)

what are quasi-experiments?

non randomized experimental studies

how can ypu categorize experimental designs?

randomized and non-randomized

what are the different non-randomized designs?

• one-group Post test only design

• one group pre test & post test design

• nonequivalent groups post test design

• nonequivalent groups pre test & post test design

• Interrupted time series design

what are the different control groups?

• no treatment controls

• wait list controls

• placebo controls

• comparative treatment groups

• non inferiority trial

what is the one group post test only design?

you only look at how one group looks like after the intervention

what is the one group pre test &post test design?

• you measure the group before and after the intervention

what is the gross and net effect in the one group pre test &post test design?

gross effect = difference between measurement 1 and measurement 2

net effect = gross effect minus confounding variables

what design is an open clinical trial or an uncontrolled trial?

one group pre test & post test design

what are threats to internal validity?

• endogenous change = spontaneous remission

• maturation trends = change is causec by maturation

• reactivity to measurement = change is caused by measurement

• secular drifts = change is caused by long term social trends

• interfering events= eg change is caused by rising prices

• regression of the mean= participants are selected because of their extreme mean scores

what are 3 threats to construct validity?

• confounding variables

• expectancy effects

• Hawthorne - effect = research itself causes the positive effect

what is Nonequivalent groups post test only design?

• you measure two groups, one who had the intervention and the other one who didnt have it

• but you didnt assign anyone to intervention or no intervention

what is another name for non equivalent groups post test only design?

natural experiment

what is the non equivalent groups pre test & post test design

one group gets the intervention the other one doesnt and then you measure them before and after

what is the biggest threat to internal validity in non equivalent groups designs?

uncontrolled selection

→ the groups might systematically differ in age, gender etc

what is the interrupted time series design?

• multiple obsrvations before and after intervention

what is a threat to validity in interrupted time series design?

interfering events

what is the randomized groups pre test & post test design?

• participants are randomly assigend to either intervention or no intervention

• you measure before and after intervention

what are multifactorial designs?

designs when there are multiple between groups factors

→ mroe things that make the groups different

what are blocking factors?

variables that are different in the participants (eg the kind of illness)

what are blocked designs?

first you allocate participants to the blocking factor and then you seperate them intp groups?

what is analysis of covariance and when is it used?

you control for variables that are different in participants so that these differences wont influence your results (you use it when the blocking variable has a linear relationship with the outcome variable)

what are dismantling studies?

try to find out the effective component of a treatmemt (comparing the same treatment but switching put some variables to look at their effects)

what are constructive or additive designs?

investigating how effectivenes of treatment changes when you add some components

what are some problems of randomized designs?

• ethical problems

• can be unneccessary to randomize

• doesnt account for treatment choice (since people are just assigned)

• randomization cant control for researcher allegiance

what is researcher allegiance?

bias of researcher which leads to him finding positive results for the treatment he favours

what are some practical issues with randomized designs?

• randomization cant ensure that groups will be equal

• attrition => drop outs reduce equivalence of groups even more

• leakage => people with treatement sharing it with control group

• funding problems

  costly

what is chronometry?

measuring duration of processes to find out information processing

what is Hick-Hyman law?

reaction time increases linearly related to amount of information extracted from stimulus

what is Donders approach to reaction time measurement?

measured the reaction time for three different tasks and then subtracted them from each involved process

what are the three types of tasks that donders uses?

Task A= simple RT (stimulus detection)

Taks B= Choice RT (detection, recognition and identification)

Task C= Go/no go RT (detection, identitifcation)