pubhlth 206a week 6: sources of error

study designs

• each offer diff advantages & disadvantages for valid group comparisons

• trade-offs b/w feasibility (cost, ethics, time, etc.), & validity impacts what study design is best used

internal validity

do the observed results accurately reflect the true association?

external validity (generalizability)

• to whom can results be applied

• requires internal validity

**NOTES**

• if a study lacks internal validity, external validity is irrelevant

• we don’t compromise internal validity in an effort to achieve external validity (generalizability)

type of sources of errors

random & systematic

systematic error

• selection bias

• informaiton bias

• confounding

error due to chance

• random error

• dots are spread out

error due to recognizable source

• systematic error

• dots are clustered

bias definition

systematic error in design or conduct of study that leads to incorrect estimate of association

bias sources

• can be caused by investigator or study participants during design or conduct of study

• can occur in experimental, cohort, case-control, cross-sectional, & ecological studies

• not inherent characteristic of the study population

• few studies have no bias or errors

bias effects

• create appearance of an association when there is none, or mask an association that really exists

• selection bias & information bias cannot be fixed in the analysis; confounding can be fixed to a point

bias solution

• limit in study design & study conduct

• critically evaluate after study has been completed, & discuss its effect

  • discuss:

    • bias type & sources

    • direction

    • magnitude

    • impact on study results & impact on study interpretation

direction of bias: towards the null

• positive & preventive association is biased TOWARDS the null

• true association is UNDERESTIMATED

direction of bias: away from the null

• positive & preventive association is biased AWAY from null

• true association is OVERESTIMATED

feasibility

cost, ethics, time, etc.

fundamental goal of study designs

• estimate the truth

• study samples allow us to estimate the truth among everyone

random and systematic error combo?

possible to have both source of errors

selection & information bias combo?

possible to have both types of bias

types of bias

• selection bias

• information bias

selection bias

WHO is in the 2×2 table

information bias

WHERE in the 2×2 table

summary

• validity & feasibility determine what type of study design to
  pursue

• internal validity: do the observed results accurately reflect the true
  association?

• sources of error in estimating the truth: random & systematic error

• 3 types of systematic error or bias: confounding, selection bias, & information bias

• direction & relative extent of bias (over- or under-estimate true association