Ali - it is difficult. There is no 'rule of thumb'. It's the responsibility of the author to decide the extent of 'replication'. If it occurs too much, then the author may be accused of 'salami-sclicing' - but what this means in its truest sense or exactly what constitutes it is a subjective thing. I've a number of published articles that are 'similar' - but they are not the same. Unless an article is mostly a direct copy of another - it is difficult for editors to accuse authors of direct copy. If I take PhD students, especially because they will not have experienced salami-slicing, I will steer them away from the dilemma. I will suggest several publications (all on the same topic of course - but one will perhaps be national and another international, one will be contextual and based on the literature review, and another will be methodological based on the findings. If it is say a mixed methods study - then separate components of the study can be published separately - and another might report the 'whole'.

It depends on the publishing house's rules. Usually you have to ask for permission if you take a table, a figure or often, even a paragraph (e.g. 100 words in a row) from another published article, also if you take lots of text (e.g. 700 words in total, it can be taken from and used in different places) from another article: in these cases, it is not enough to cite these parts, you have to ask for permission to publish these. It doesn't matter if you originally created these tables, figures or paragraphs or not: you still have to ask for permission.

If you need that I think you should ask the chief editor in the old article. if they accept you can re-published it and change some things such as the title and add new data and information.

Ali - it is difficult. There is no 'rule of thumb'. It's the responsibility of the author to decide the extent of 'replication'. If it occurs too much, then the author may be accused of 'salami-sclicing' - but what this means in its truest sense or exactly what constitutes it is a subjective thing. I've a number of published articles that are 'similar' - but they are not the same. Unless an article is mostly a direct copy of another - it is difficult for editors to accuse authors of direct copy. If I take PhD students, especially because they will not have experienced salami-slicing, I will steer them away from the dilemma. I will suggest several publications (all on the same topic of course - but one will perhaps be national and another international, one will be contextual and based on the literature review, and another will be methodological based on the findings. If it is say a mixed methods study - then separate components of the study can be published separately - and another might report the 'whole'.

My initial thought is that it's more about how much new data is included than how much old data. There has to be enough that is completely new (data, interpretations, etc...) to warrant a stand-alone publication. At that point, how much 'old' information is included isn't too important.

You should only include the information you need to explain or provide context for the new results, and all of the previously published results need to be cited (so that it is clear that it is not new information). As previous comments have noted, you need permission and/or proper acknowledgement for any work that is reproduced directly (or almost directly) from previous work.

John - a good response - but not all publications are data-based, so the rules may well change when that is the case.