Home / Summaries / Applied Research Methods: Development and Mental Health / imputation-values-missingness

Missing data techniques and low response rates

Q: When do you use matching instead of randomisation in psychotherapy outcome research?

A subject is 'matched' or 'paired' with a similar subject to reduce the chance that other variables obscure the primary comparison of interest. Is used if you want to use data from other samples as controls, which thus can’t be randomized. It can also be used if you have a small sample size and randomization is not feasible. Matching then increases validity of conclusions, as participants are more resembling each other.

10 important questions on Missing data techniques and low response rates

Is it more important to identify the number of variables with missingness OR number of participants with missingness

Number of participants

What are missingness mechanisms?

Reason why data is missin

Which missingness mechanisms are ignorable? Which are non-ignorable?

Ignorable
MCAR (Missing Completely at Random) = does not depend on the observed or missing values
MAR (Missing at Random) = partly depends on the observed values, but not on the missing values lead to lower statistical power!

Non-ignorable
MNAR (Missing Not at Random) = depends on the missing values themselves

What are three proactive strategies for minimizing missingness?

Proactive: the best thing is to not have any missing data Prevent it with advanced warnings, personalized surveys, follow-up reminders, and monetary incentives (increases participants motivation) Reactive: Numerous analytic strategies: Unwise deletion Single imputation Multiple imputation

How does listwise delection handle missing values?

Deleting cases with any missing values (however, this violates a fundamental principle of missing data analysis since real data are away)

How does pairwise deletion handle missing values

Uses all available data if missingness is minimal and randomly distributed across cases and variables

What is the advantage of using multiple imputation compared to single imputation

The multiple imputation does not provide a deterministic idea of what the missing value should be, it allows to have a range of scores. So it provides a much more accurate estimation in terms of the parameter estimates

What is the best way of handling missing values

Mean item imputation: compute mean score based on the items if scale is missing, consider multiple imputation o Item(s): mean item imputation o Scale(s): multiple imputation (if not MCAR) o Entire survey(s): nothing to do

When do you use matching instead of randomisation in psychotherapy outcome research?

A subject is "matched" or "paired" with a similar subject to reduce the chance that other variables obscure the primary comparison of interest. Is used if you want to use data from other samples as controls, which thus can’t be randomized. It can also be used if you have a small sample size and randomization is not feasible. Matching then increases validity of conclusions, as participants are more resembling each other.

Are these disadvantages when the effects of an experimental treatment condition are compared to those of list or treatment as usual condition? If yes, which disadvantages?

Comparison to no treatment or uncontrolled treatment > easy to show that experimental condition is effective but less overall power

The question on the page originate from the summary of the following study material:

Applied Research Methods: Development and Mental Health

View summary

A unique study and practice tool
Never study anything twice again
Get the grades you hope for
100% sure, 100% understanding

Remember faster, study better. Scientifically proven.

Missing data techniques and low response rates

10 important questions on Missing data techniques and low response rates

Is it more important to identify the number of variables with missingness OR number of participants with missingness

What are missingness mechanisms?

Which missingness mechanisms are ignorable? Which are non-ignorable?

What are three proactive strategies for minimizing missingness?

How does listwise delection handle missing values?

How does pairwise deletion handle missing values

What is the advantage of using multiple imputation compared to single imputation

What is the best way of handling missing values

When do you use matching instead of randomisation in psychotherapy outcome research?

Are these disadvantages when the effects of an experimental treatment condition are compared to those of list or treatment as usual condition? If yes, which disadvantages?

Summaries related to An overview of clinical research; the lay of the land

Applied Research Methods: Development and Me…

Development and Mental Health 2

Development of Primary Relationships samenva…

Behavior and Environment 2 boek samenvatting

Theoretical psychology

Prosocial and antisocial behavior

Brein en Cognitie 3

Affective Science and Psychopathology

Early Development

Adolescence: Chances and risks

Adulthood and ageing

Athena samenvatting