Article in journal

[noetsi]

I am reading articles on statistics at work a lot and I found this one - which rather bemused me. Because it dealt with medical journals where you would expect high level expertise.

P value, odds ratio,and 95% confidence interval for coefficients in MLR was reported in 75.2% and sufficient cases (>10) percovariate of limiting sample size were reported in the 58.7% of the articles. No article reported the test for conformity of linear gradient for continuous covariates. Total score was not significantly different across the journals.

I am not familiar with the later statistic, but on a whole that seems a rather low floor to get over. Remarkable that these were not reported in the other quarter of the articles.

However, involvement of statistician or epidemiologist as a co‑author improved the average quality score significantly (P=0.014).

It is encouraging to know when they asked professional statisticians to participate the statistics got better.:yup:

Conclusions: Reporting of MLR in many Indian journals is incomplete. Only one
article managed to score 8 out of 10 among 109 articles under review. All others scored less.

Its a bit scary that the quality was that low in medical journals in India in 2011 (more likely the previous decade). I have not yet found the scale they used to measure quality, perhaps it was very high. But previous findings in US medical journals for logistic regression generated similar concerning findings.

A recent study conducted on 196 original articles published in two Indian pharmacology journals between 2002 and 2010 found that only
22% reported appropriate descriptive statistics. Only two out of 196 articles mentioned about the assumptions related to the
statistical method applied, and 69.3% articles used appropriatestatistical methods.[1]

Which means 30 percent did not use appropriate methods....in journals where you would think the statisics would be rather precise

 

[noetsi]

These are the critier if any are interested for quality (they don't seem very high).

These 10 criteria are (i) sufficient cases (>10) per covariate, (ii) conformity of linear gradient for continuous variables, (iii)
testing of interactions, (iv) testing of collinearity, (v) validation of model, (vi) P value, odds ratio, and 95% confidence interval of odds ratio, (vii) selection of potential covariates, (viii) coding of covariates, (ix) goodness‑of‑fit or classification summary, and (x) model fitting procedure

I wonder if "conformity of linear gradient for continuous variables" means they were tested for linearity. Notice there is no requirement that they be tested for outliers or highly leveraged points.

 

[rogojel]

if you read German, the book " Der Hund der Eier legt" has plenty of examples from european medical journals, from the 90-ties. No reason to expect an improvement unfortunately.

The title means "the egg-laying dog" BTW.

regards
rogojel

 

[Karabiner]

These 10 criteria are (i) sufficient cases (>10) per covariate,

But rules of thumb for logistic regression do not refer to number of cases,
but to number of cases in the smaller outcome group?!

With kind regards

K.