Experimental and non-experimental correlational studies Flashcards
considerations in causal effects and inference considerations
- establish a cause-effect linkage between specific variables: if X then Y (so changes in A leads to changes in B)
- identify targets that can be controlled or manipulated (changes in outcome)
- internal validity
Independent variable / predictor variable / treatment variable
Dependent variable / ultimate / outcome variable
- manipulated by the researcher to see if it affects the DV
- measured to see if affected by IV
3 necessary bases for inferring causation
- correlation
- temporal precedence (ordering - so variation in IV should occur before changes in outcome)
- ruling out extraneous factors
EXPERIMENT
- mani_ulation
- random_zation
- standar_ization
- deliberately varying the predictor conditions (levels of treatment, control/experimental groups)
- randomly assign ppl to different groups and each one has a different treatment
- formalized procedures, enviro. factors constant, lab setting
What are the two EXPERIMENTal design with treatment and control groups?
Independent measures (between subject design): randomly assigned to Condition 1 or Condition 2
Repeat measures (within-subjects design): experience all levels of the IV (their own controls)
+ when few participants are available
+ conduct the experiment more efficiently
- order effects
- fatigue effect / reduced motivation
=> counteract with counterbalancing
Disadvantages of a lab EXPERIMENT
- artificiality decreases the ecological and external validity
- decreases generalizability of the results)
- no. of variables manipulated is limited (focus on single IV of interest)
Define correlational research
- researcher does not have as much control as they do in an experiment
=> emphasizes ind. differences and effort to establish relationships among those differences on various personality characteristics
What do correlational research use?
- statistical measures to establish association or correation
CORRELATIONAL
what are the factors shaping the relationships bet. 2 variables?
- reliability (2 variables can only correlate to the extent each measure is reliable)
=> if we use a poor measure of low quality, that decreases the reliability and leads to inaccurate data and correlation - third variable issue (al. explanations)
Is A the cause of B or is there C?
-restricted variation of scores (maximize ind. differences)
=> size of correlation bet. 2 variables
CORRELATIONAL
Longitudinal studies
- follow the same group over time
- can A predict B?
- measured more than once, measure same variables (ideally)
*can explore temporal ordering or sequence of these variables
CORRELATIONAL
cross-sectional studies
- measured once
Types of correlation coefficients
s_nchronous
au_o / te_t-re_est / lag_ed
cro_s-la_ged
synchronous: cross sectional
- correlation bet. X and Y measured at the same point in time
auto: same variable measured at different. times
Variable X measuredat time 1 and variable X measured at time 2
cross-lagged:
- X at time 1 and Y at time 2
Mediator vs Moderator variable
Mediator: “accounts” for X and Y relationship
Moderator: different levels of moderator variables show different sizes of the relationship