|This death certificate is only a primary source for the death|
info. His parents names are actually wrong, and his birth year
is in conflict with records from when he was alive, suggesting
the informant, his son, may have gotten it wrong.
- Birth records are the only primary source for a birth. This may include a birth announcement in the newspaper, but the further back you go, the less likely this becomes. Equally, the further back you go, the less likely that civil vital records were kept. Delayed birth certificates aren't a primary source, but may be the only record of a birth in existence. Also keep in mind that some places would fine individuals for reporting a birth too late, which means they may have lied about the birth date to avoid being fined.
- Baptism records are only a primary source for the baptism, not the birth. However, if the baptism occurred only a few days after the birth, that's pretty much as good as a primary source for the birth too (if it recorded the birth date - do not assume the baptism and birth date are the same if both aren't recorded). Especially if there's no birth record in existence, a baptism record is likely as good as it's going to get. However, if the baptism took place years after the birth, maybe even months, that is not a primary source for the birth because enough time has passed for the actual birth details to have been remembered incorrectly.
- Marriage records are only a primary source for the marriage. Particularly if the parents of the bride or groom were deceased, you can't be sure their names are correct. Be careful not to mistake a marriage bann, engagement announcement, or marriage license for the actual marriage.
- Death records are only a primary source for the death. If it includes an address where the deceased was living at the time of death, then it's also a primary source for that. But it's a secondary source for the birth date and location, both because the document is normally recorded many years after the birth (unless it's an infant death), and because the informant for the death record is often someone who wasn't even alive or present at the time of the deceased's birth. It's also not a primary source for the parent's names or birth locations; it's very common for those details to be incorrect. Death records are usually a good source for the burial location, even though they are recorded before the burial takes place, and therefore that info theoretically could change before it happens.
- Obituaries are generally considered a type of death record and therefore can be considered a primary source for the death if they are published within a few days of the death, as is typical. Excepting potential printing errors, of course (i.e. the informant may have provided the correct death information, but the typist misprinted it).
- Gravestones aren't really a primary source for anything! At the most, they may be a primary source for the location of the burial, but I have seen gravestones erected for people before their death, who then actually wind up buried elsewhere. However, when this happens, there's usually a lack of at least a death date on the gravestone. It's also not a primary source for the date of the burial, since gravestones don't normally have the burial date listed on them. You might think it's a primary source for the death date, but gravestones often aren't created for weeks or even months after the death, plenty of time for people to remember the exact date incorrectly.
- Cemetery/burial records are only a primary source for the burial information. Unlike gravestones, these usually include the interment date and wouldn't exist unless the deceased was actually buried there.
- Census records are only a primary source for data that was current at the time the census was taken, such as: residence, occupation, citizenship, literacy, etc. All other data that occurred prior to the census - birth/age, marriage, immigration, etc is secondary. Additionally, even things like the occupational data may be subject to the knowledge of the informant and could be incorrect. Also don't forget that in the US, pre-1880 censuses did not record relationships to the head of the household. While you can often surmise relationships based on the order in which people are listed, ages, and names, you can't be sure about them without other supporting documents to confirm.
- Family bibles may or may not be a primary source for any or all of the data within, depending on when each item of information was recorded and who recorded it. Unfortunately, there's generally no way to know for sure when the data was recorded, or who by. You can sometimes get an idea based on the handwriting and/or different types of pens used at different times. For example, you might note the birth info was recorded at a different time from the death info. But this still doesn't assure they were recorded at the time of those events. They could have each been recorded years after the fact, whenever the author (and we may not even know who the author was) got around to it.
- Wills and Probates can contain a lot of valuable and reliable information, like the names of someone's children, the details of their estate/property, etc. But even though they are related to the death fact, they typically don't contain a date or location of the death, let alone a cause of death. Don't mistake the will or probate dates for the death date, but you can usually get a time frame for the death date - sometime after the will date, and before it was probated.
- Lineage books are a secondary source for everything in them, since they are written after all the events took place. However, many lineage books use primary sources for at least some of their information. That doesn't necessarily mean the entire book is reliable, but the particular data coming from primary sources should be. Not all authors note their sources, but many do.
Judging which secondary source is more reliable for what type of conflicting data can be difficult, and we have to weigh when, how, and by who the data was recorded/provided. You may think it makes the most sense to go with a birth year that you find on most of the records for an individual, but what if all those records are from later in his life, or even after his death? A record from his childhood, closer to when he was born, and when his parents, who were there for the birth, were still alive and one of them may have been the informant may actually prove to be the more reliable source. Of course you can never know for sure, so it's also best to put all recorded facts in your tree as alternate data, but you still have to pick a default/preferred one. Hopefully, this has given you some things to consider when choosing a default/preferred fact to go with, and given you a good understanding of primary and secondary documents.