Error in Research

Question 1

What is the actual goal of research?

Answer 1

To conduct a study that gives us an estimate of the XY association that is as close as possible to the truth

Question 2

Why is it important to examine study accuracy?

Answer 2

To assess how precise our study was and how valid the results were; tells us whether or not we have a reputable study

Question 3

What are the 2 components of study accuracy?

Answer 3

Precision + Validity

Question 4

What is the difference between systematic & random error?

Answer 4

Systematic error is concerned with bias (validity) whereas random error is concerned with precision

Question 5

What are the 3 types of systematic error in studies?

Answer 5

1. Selection Bias
2. Measurement Bias
3. Confounding

Question 6

Name the 4 types of selection bias

Answer 6

1. Response Rate
2. Volunteer Bias
3. Healthy Worker effect
4. Loss to follow-up

Question 7

Name the 4 types of measurement bias

Answer 7

1. Recall bias
2. Social desirability bias
3. Interviewer bias
4. Observer bias

Question 8

What are the 2 types of random error in studies?

Answer 8

1. Sampling error
2. Measurement error

Question 9

Why can we not determine the true effect of X on Y within research?

Answer 9

Because no study is completely error-free (so the true relationship between X and Y cannot be known)

Question 10

What is validity in terms of research?

Answer 10

Validity is how much we should believe the findings in a study to be close to the truth

Question 11

How do we determine whether a study has validity?

Answer 11

by assessing the amount of error/bias a study has

Question 12

If a study has error should we automatically ignore the findings?

Answer 12

No, because all studies have some degree of error

Question 13

What is internal validity?

Answer 13

the extent that study results are representative of the sample

Question 14

What is external validity?

Answer 14

The generalizability of study results to the population

Question 15

Can we have external validity without internal validity?

Answer 15

No. The study findings aren’t accurate to the sample so it cannot be generalized to other populations.

Question 16

When darts are clustered together in the same place on a dart board but not on the bullseye, what does this tell us about study accuracy?

Answer 16

The study is precise but lacks validity

Question 17

What is selection bias?

Answer 17

systematic error due to selection & maintenance of sample

Question 18

Explain response rate

Answer 18

the requirement that 70% or more of a random sample agree to a study-otherwise creating response bias

Question 19

What studies are Response bias applicable to?

Answer 19

One’s that use random sampling

Question 20

What is volunteer bias?

Answer 20

Idea that those who are healthier are often more likely to volunteer for a study

Question 21

What studies can volunteer bias be a concern?

Answer 21

studies that use a convenience sample

Question 22

What is the healthy worker effect?

Answer 22

The idea that selecting a sample that are all individuals belonging to an organized group may be healthier on average than the rest of the population, creating selection bias

Question 23

If we have a study with membership bias do we automatically disregard the findings?

Answer 23

No. We must keep in mind that the study results will have low external validity unless generalized to a similar sample population/membership.

Question 24

How can the internal validity of a study with healthy worker effect be compromised/reduced?

Answer 24

When the comparison group is from the general population and not the same source population (organization) as the cases

Question 25

How can we avoid volunteer bias?

Answer 25

By using a random sample

Question 26

What is loss to follow-up concerned with?

Answer 26

The maintenance of a sample over time; we need to maintain at minimum 70% of the sample to avoid bias

Question 27

What studies can loss to follow-up be a concern for?

Answer 27

Any studies that have more than 1 data collection point (follow-up on samples)

Question 28

What is measurement bias?

Answer 28

systematic error due to the way data was gathered in a study

Question 29

What is observer bias?

Answer 29

When the ascertainment of Y in participants is not made independent of the researcher’s knowledge of the X status.

Question 30

When is observer bias relevant?

Answer 30

When Y is determined by the researcher using signs rather than a biological test

Question 31

What do we use to identify whether random error is present?

Answer 31

The CI: large CI that crosses effect sizes suggests random error is present

Question 32

How do we reduce sampling error?

Answer 32

By using a large sample

Question 33

What is measurement error?

Answer 33

Random error in instruments used to measure X or Y because they were not calibrated correctly or researcher was not trained how to use the instrument properly

Question 34

What can we say about a study that has high systematic error but low random error?

Answer 34

It is precise but lacks validity