Question

Define and give a example of each

Meta-analysis

Blobbogram

Funnel Plot

File drawer effect

Concurrent validity

Predictive validity

Answer 1

Meta-Analysis: A subset of systematic reviews; a method for systematically integrating specific qualitative and quantitative research data from a variety of selected studies to produce a single conclusion with higher statistical capacity. This conclusion is statistically stronger than the analysis of any single study, due to increased numbers of subjects, increased diversity among subjects, or cumulative results, and results.

For example, do individuals who wear sunscreen have less cases of melanoma than those who do not use sunscreens? A MEDLINE quest was performed using the words melanoma, sunscreening agents and zinc oxide, resulting in 8 randomized controlled trials, each of between 100 and 120 participants. Both research showed a beneficial association between wearing sunscreen and reducing the risk of melanoma. Participants from all eight studies (total: 860 participants) were pooled and statisticallyanalyzed to establish the impact of the relationship between sunscreen and melanoma. This meta-analysis showed a 50% reduction in the diagnosis of melanoma among sunscreen wearers.

Blobbogram: A blobbogram (sometimes called a forest plot) is a graph that contrasts many clinical or research studies that study the same subject. Initially designed for meta-analysis of randomized controlled trials, the forest plot is now often used in a number of qualitative studies. It's called a forest plot because of the line forest it creates.

For example, The blobbogram is from an popular medical examination; it reveals clinical trials of the use of corticosteroids to promote the growth of the lung in pregnancy where a baby is likely to be born prematurely. Even after there was ample evidence to prove that this procedure saved the lives of children, evidence was not well known and the procedure was not commonly used. After a thorough analysis of the evidence, care became more commonly used to prevent thousands of pre-term babies from dying of infant respiratory distress syndrome.

Funnel Plot: The funnel plot is a graph designed to test for publication bias; funnel plots are widely used in systematic reviews and meta-analysis. In the absence of a publication bias, it is expected that high-precision studies will be plotted close to the average, and low-precision studies will spread equally on both sides of the average, producing a approximately funnel-shaped distribution. Deviation from this shape can indicate publication bias.

For example,

An example of a funnel plot showing no bias in reporting. Every dot represents a test (e.g. calculating the effect of a certain drug); the y-axis reflects the precision of the test (e.g. standard error or number of experimental subjects) and the x-axis indicates the result of the research (e.g. the estimated average effect of the drug).

File drawer effect:

Publication bias is a kind of bias that exists in published academic work. This happens when the results of an experiment or a scientific study affects the decision to publish or otherwise distribute this.

For example,

Meta-analysis of the effect of stereotyping on girls' math scores showing asymmetry indicative of the bias of publishing.

Concurrent Validity: Concurrent validity is the type of Validity Criterion. If you construct some kind of check, you want to make sure it's valid: that it calculates what it's supposed to measure. Validity of the criterion is one way to do so. Concurrent validity tests how well a new test performs compared to a well-established test.

For example, If you are designing a new depression test, you should compare its results with previous depression tests (such as a 42-item depression level survey) that are highly credible. Concurrent means "as at the same time," and you can do both tests at around the same interval: you could measure the depression level on one day with your survey, and on the next day with the measure collection. A statistically significant outcome would mean that you have achieved concurrent validity. If the experiments are further apart (i.e. they are not performed concurrently), they will fall under the category of Predictive Validity instead of the criterion of validity.

Predictive Validity: Predictive validity is the degree to which the Scale or Test Score predicts the result for any criterion calculation.

For example, The validity of a cognitive work performance test is the association between test scores and, for example, the performance rating of the supervisor. Such a cognitive test would have predictive validity if the observed correlation was statistically significant.