Term
|
Definition
| deviation from true values |
|
|
Term
|
Definition
due to sample or chance
-increase sample size |
|
|
Term
systematic error
-leads to?
-due to?
-affects?
-soltuion?
- 3 types |
|
Definition
-leads to bias
- due to study design flaws
- affects validity
- solution:
careful study desgin
use of appropriate statistical techniques
3 types of systematic error:
- selection
- information
-confounding |
|
|
Term
|
Definition
| SYSTEMATIC ERROR in design, conduct or analysis of study that results in a MISTAKEN ESTIMATE of an exposure's effect on risk of disease |
|
|
Term
| flow chart of research process highlighting steps of study and where the three types of bias occur. |
|
Definition
|
|
Term
|
Definition
systematic difference between those particpating in study and those who are not
- 3 types:
- nonresponse
-loss to follow up
- healthy worker |
|
|
Term
| what type of validity issues occur during the selection of subjects and selection bias |
|
Definition
selection of subjects = external validity issues
selection bias = internal validity issues |
|
|
Term
|
Definition
| bias generated by certain types of people not responding to survey or dropping out of study |
|
|
Term
|
Definition
| participants who withdraw from a study are systematically different from those who remain |
|
|
Term
|
Definition
| participants involved in study misrepresent target pop because they can participate (ie are healthier) |
|
|
Term
| methods of evaluating possible selection bias and addressing possible selection bias. |
|
Definition
evaluate:
- scrutinize sampling method
- explore non-responders
to adress possible selection bias:
- careful study design
- obtain randomized sample of target pop
- reduce likelihood of non-repsonse bias, loss to follow up, ect. |
|
|
Term
information bias
- 3 types |
|
Definition
determination of exposure and/or outcomes status from some participants are incorrect
3 types:
- misclassification
- recall
-reporting |
|
|
Term
Misclassification bias
- 2 types (describe) |
|
Definition
participants incorrectly assigned to exposure group (+/-) or outcome group (+/-)
2 types:
1. differential:
- no difference between groups/participants in their likelihood to be accurately assigned. issue inherent to data collection methods. bias towards null.
2. non-differential:
- one group more likely to be inaccurately assigned than another group. bias can be towards or away from null. can find associations that don't exist, or no associations when they do. |
|
|
Term
[image]
which numbers = pos correlation and neg correlation and null? |
|
Definition
neg correlation < 1
pos correlation > 1
null = 1 |
|
|
Term
|
Definition
when an individual may not remember correctly, or accurately, something that occurred in the past.
- to solve:
- reduce retrospective time period |
|
|
Term
|
Definition
when an individ may be reluctant to report undesirable behaviour.
to solve:
- reduce stigma/socially desirable answers |
|
|
Term
confounding bias
- 3 elements that it must contain |
|
Definition
observed association between exposure and outcome is affected by third factor.
3 elements:
1. associated with exposure
2. associated with outcome
3. not a result/caused by exposure |
|
|
Term
| Ways of solving confounding variables. |
|
Definition
· Address in design
o Randomization
§ During RCT, all potential CFVs are distributed equally between groups
o Restriction/exclusion
§ Only one level of CFV eligible to participate
o Matching
§ Select individuals purposively to ensure even distribution of CFVs between groups
· Address in analyses
Must first know what confounding variables are!
o Stratification
§ Stratify groups into different levels of the confounder
§ Examine association between E and O at each level
o Multivariate modeling
§ Statistically control for potential CFVs
|
|
|
