Section 8A: Systematic error in research II: information bias

Question 1

What are the 2 main types of bias in research?

Answer 1

1. Selection bias
2. information bias

Question 2

What is selection bias?

Answer 2

Systematic error arising from mistakes conducted during the selection of study sample

Question 3

What is information bias?

Answer 3

Systematic error arising from mistakes conducted during measurement of key study variables (exp & out)

Question 4

How does information bias arise?

Answer 4

From wrong/inaccurate assessment of either exp or outcome variables

Question 5

From who can information bias arise?

Answer 5

• from researchers (unintentionally)
• from participants (un or intentionally)

Question 6

How can info bias arise from researcher actions?

Answer 6

• wrong/inacc diagnosis coz of a clinical error
• wrong/inacc measurements due to faulty instrument/machine
• wrong/inacc measurements due to poor training of assessor
• mistakes during recording of data
• mistakes during transferring data from paper - electronic

Question 7

How can info bias arise from participant actions (misreporting)?

Answer 7

Wrong/inaccurate answers from participants due to:

• misinterpretation of a question
• sensitivie issue relating to question
• poor recall (recall bias)
• • intentionally

Question 8

What are the types of information bias?

Answer 8

• misclassification bias
• hawthorne effect
• surviellance bias

Question 9

What is misclassification bias?

Answer 9

errors made in classifying either disase or exp status (or any other covariate). E.g.

• cases incorrectly classified as control or vice versa
• Exposed incorrectly classified as non-exposed or vice versa

Question 10

What are the types of misclassification bias?

Answer 10

• Recall bias
• Interviewer bias
• Observers bias

Question 11

What is recall bias?

Answer 11

• participants with particular outcome or exp may remember events more clearly or amplify their recollections.
• Very common in case-control studies

Question 12

What are some common examples of recall bias?

Answer 12

• pregnancy & birth defects - mthrs recall exp (alc, drugs, smoking) more carefully
• cancer & smoking - cancer patients may overreport & healthy may underreport
• food poisoning - sick ppl remember meals in detail, healthy ppl forget
• Occupational exp - disease workeers recall exp (asbestos, chemcials) more than healthy controls
• MH - depressed/ptsd recall childhood adversitiies more readily

Question 13

What is interviewer bias?

Answer 13

expectations or opinions of interviewer interferes with the judgment of the interviewee

Question 14

What is observers bias?

Answer 14

Researchers’ expectations/knowledge of exposure status influence how outcomes are assessed (especially if they aren’t blinded to exposures or medical history)

Question 15

Whaat is the Hawthorne effect?

Answer 15

participants act differently knwoing they’re being watched

Question 16

What are examples of the Hawthorne effect?

Answer 16

• Workers in factory improve productivity coz know they’re being observed, not coz of any intervention
• Patients in weight-loss study exercise more or eat healthier during the study because they’re being monitored.

Question 17

What is surveillance bias?

Answer 17

Group with known exposure or outcome may be followed more closely or longer than the comparison group (researcher’s bias).

Question 18

What is an example of surveillance bias?

Answer 18

Patients on new drug monitored more closely than those on standard care, so side effects are reported more often

Question 19

What is validity?

Answer 19

extent to which assessment tool (e.g. questionnaire, instrument, etc) measure accurately what it is intended to measure

Question 20

What is the most common type of validity used in med research?

Answer 20

Criterion validity

Question 21

What is criterion validity?

Answer 21

Results from assessment tool usd are compared with those of an established (known as gold standard) assessment tool

Question 22

What is reliability?

Answer 22

• How consistently a measure produces the same results under the same conditions.
• Also called reproducibility or repeatability.

Question 23

What are the 2 main types of reliability?

Answer 23

1. Inter-obsever reliability
2. Intra-observer reliability

Question 24

What is interobserver reliability?

Answer 24

• degree of agreement between results when 2(+) researchers administer assessment tool on same ppl under same conditions
• examines consistency across ppl/assessors

Question 25

What is **intra**observer reliability?

Answer 25

* Describes agreement between results when assessment tool is used by same researcher on 2(+) occasions under same conditions & in same population * Examines consistency within same person over time

Question 26

What is an example of a measure that is reliable & valid?

Answer 26

calibrated thermometer consistenly gives true body temp every time it's used

Question 27

What is an example of a measure that is reliable but **not** valid?

Answer 27

* bathroom scale that always shows 5kg too high. * consistent (reliable) but not accurate (valid)

Question 28

What is an example of a measure that is valid but **not** reliable?

Answer 28

* BP measurement reflects true BP but reading vary a lot to due stress, posutre, "white coat" effect * why many studies take multiple BP & average to improve reliability

Question 29

What is an example of a measure that is **not** reliable & **not** valid?

Answer 29

broken thermometer gives random, incorrect readings that are neither **consistent** nor **accurate**

Question 30

How can we minimize information bias?

Answer 30

* blinding * standardized protocols * validated instruments * training & quality control * multiple sources/repeated measurements * equal intensity of follow up

Question 31

How can blinding minimize info bias?

Answer 31

prevents interviewer/observer bias

Question 32

How can standardized protocols minimize info bias?

Answer 32

reduces interviewer bias & recall bias

Question 33

How can validated instruments minimize info bias?

Answer 33

reduces measurement error

Question 34

How can training & quality control minimize info bias?

Answer 34

* minimize observer bias * improves inter/intra observer reliability

Question 35

How can multiple sources/repeated measurements minimize info bias?

Answer 35

reduces recall & improves reliability

Question 36

How can equal intensity of follow up minimize info bias?

Answer 36

prevent surveillance bias