The basic unit of information is a record. "Vital records" consist of three kinds of events: Births, Marriages, and Deaths. A 4th kind of record is sometimes available, a Banns record.
Accuracy varies. One way to check accuracy is to compare books in cases where duplicate data is available. Another way to check accuracy is to assemble all the records for one family and look for inconsistency of detail. The error details that are most obvious are house numbers, ages, maiden names, and given names of parents and grandparents. The general impression is that the sections of books with good penmanship tend to have fewer errors.
The amount of information in a record is not standard. The amount of information varies from place to place and from time to time.
Most of the Poland LDS microfilms are from the 19th Century. Records less than 100 years old are generally not filmed although some have been. Records more than 200 years old are less common because few record books survived.
A remarriage of a widow does not always mention the name of her previous husband. At least in south Poland before 1800, a widow's family name in a record is usually the family name of her prior husband. Watch out for widows with only one family name recorded. The format "Mary daughter of John Szymanski a widow" makes it clear that Szymanski is the maiden name. The format "Mary Szymanska a widow" is ambiguous.
The Banns are the official announcement that a couple plan to marry. Three readings of the banns, in church, were required back then. The priest was required to keep a record of which dates the banns were read. Few Banns books survive, but when available they are useful records, particularly when the actual marriage records are missing, since the banns record usually includes the marriage date. Sometimes the dates for reading of the banns are added to the Marriage record, too.
Nineteenth century Poles were not as aware of their age as we are. Where parent age and age of witnesses is recorded, there are examples of more than 20 records for one individual. The discrepancy of recorded age vs actual age is sometimes as large as 10 years. On the other hand, there are examples of other individuals whose age is always correct.
Nevertheless, recorded age, when available, can be useful for identification if there are 2 individuals in town with the same name but very different age.
Death ages of old people are notoriously exaggerated, as everywhere in the world. Some are accurate, but most seem to be recorded as much older than calculated from the birth record (when available and definite).
As expected, ages at marriage tend to fib a little, with the person recorded as older for people less than 20 and as younger for widows and widowers over 30.
Polish language books explain that the family name of Polish females is inflected to indicate if she is married or single. The books may mention that there are many different ways of doing this. It is very confusing. The books may not mention that even Poles get confused by name inflections. The "always married" inflections seem to be inconsistent, being sometimes used for unmarried young women. There are "always single" inflections, such as "-cionka". Check out the Translation Tips web sites. Some records avoid feminine inflections by using a format like "Mary daughter of John Szymanski".
Summary: When a marriage or death or witness record for a woman does not mention her status, one of the diminutive inflections tells you she is unmarried, even if she is old. On the other hand, the various adult inflections tell you nothing unless you can find a pattern for that particular record book by studying the book as a whole.
Polish spelling was not standard in the 19th century. Poles use various inflections, or variations for names. For example, in the 19th century records, Kolodziejczak family members are also recorded as Kolodziejczyk and Kolodziejski. That "l" is usually but not always crossed, and of course the crossed "l" is often indistinguishable from a "t".
Apparently, it was not very uncommon to use a first name "around the house" that is different than the name of record.
Alias family names were used. Some are recorded with the Latin word "alias". For example there was more than one Gwozdz family in a village that used the name Knap. To confuse the issue, there was at least one family that used the name Knap exclusively.
Sometimes, people changed their family names. That can be difficult to spot in the records. For example, one ancestral family (in Szelkow, north Poland) used the names Bajtko and Bajtczak before 1800. Most Poles would consider Bajtko and Bajtczak as valid inflections of the same name. After about 1800, the family switched to Banasiak, which is clearly a different name. It was not a clean switch. Not everyone in the family changed at the same time. Some did not change, and there were people not related with the same names. Here is a example convincing record: Apolonia Bajtko was married 25 Feb 1820. The first witnesses is Jozef Bajtczak, brother of the bride. The second witness is Jozef Banasiak, uncle of the bride on her father's side (the Polish word "styia" is used for this relationship). The priest who wrote the record felt no need to explain why they were all using different family names.
Many records indicate that the name of a person is unknown. Most commonly it is a maiden name, but you will see this also for given names. Rarely the name of the person of record is indicated as unknown.
There are several possible reasons: (1) the person has no name, (2) the name was unknown at the time of the record, and (3) the name was illegible at the time of copying of the record.
An example is the word "beznazwisko", which literally means "without family name". Certainly that word should be used for people who in fact do not have a family name. However, there is a significant example where the word "beznazwisko" was used to indicate that the name was unknown to the person making an index.
Another word is "niewiadomego" which means "unknown". Often seen is "niewiadomego nazwiska", which means "unknown family name". Again, it is usually ambiguous if the name was unknown at the time of record or just illegible at the time of copying.
In south Poland unknown names are usually indicated by "N.", which presumably is the abbreviation for niewiadomego.