How do you get your class size in probability and statistics?

Answer 1

One determinant is what the data will be used for, whether they need to be combined with other data for further analyses. If, for example, you want to study human mortality, you may want to use the same class sizes as for population statistics - which are usually 5-year age bands. That will enable you to calculate deaths rates. However, you may additionally wish to have a band for "less than 1 year" and "1 to 5 years" to allow you to study infant mortality in greater detail. At the other end, you may wish to group together "90 and above".

If there are no such constraints, you should divide up the range of values into 6-10 classes. Ideally, no class should have very few (<5%) observations. If there are, then combine adjoining classes - there is no requirement for the class size to be the same across the whole range.

One determinant is what the data will be used for, whether they need to be combined with other data for further analyses. If, for example, you want to study human mortality, you may want to use the same class sizes as for population statistics - which are usually 5-year age bands. That will enable you to calculate deaths rates. However, you may additionally wish to have a band for "less than 1 year" and "1 to 5 years" to allow you to study infant mortality in greater detail. At the other end, you may wish to group together "90 and above".

If there are no such constraints, you should divide up the range of values into 6-10 classes. Ideally, no class should have very few (<5%) observations. If there are, then combine adjoining classes - there is no requirement for the class size to be the same across the whole range.

One determinant is what the data will be used for, whether they need to be combined with other data for further analyses. If, for example, you want to study human mortality, you may want to use the same class sizes as for population statistics - which are usually 5-year age bands. That will enable you to calculate deaths rates. However, you may additionally wish to have a band for "less than 1 year" and "1 to 5 years" to allow you to study infant mortality in greater detail. At the other end, you may wish to group together "90 and above".

If there are no such constraints, you should divide up the range of values into 6-10 classes. Ideally, no class should have very few (<5%) observations. If there are, then combine adjoining classes - there is no requirement for the class size to be the same across the whole range.

One determinant is what the data will be used for, whether they need to be combined with other data for further analyses. If, for example, you want to study human mortality, you may want to use the same class sizes as for population statistics - which are usually 5-year age bands. That will enable you to calculate deaths rates. However, you may additionally wish to have a band for "less than 1 year" and "1 to 5 years" to allow you to study infant mortality in greater detail. At the other end, you may wish to group together "90 and above".

If there are no such constraints, you should divide up the range of values into 6-10 classes. Ideally, no class should have very few (<5%) observations. If there are, then combine adjoining classes - there is no requirement for the class size to be the same across the whole range.

Answer 2

One determinant is what the data will be used for, whether they need to be combined with other data for further analyses. If, for example, you want to study human mortality, you may want to use the same class sizes as for population statistics - which are usually 5-year age bands. That will enable you to calculate deaths rates. However, you may additionally wish to have a band for "less than 1 year" and "1 to 5 years" to allow you to study infant mortality in greater detail. At the other end, you may wish to group together "90 and above".

If there are no such constraints, you should divide up the range of values into 6-10 classes. Ideally, no class should have very few (<5%) observations. If there are, then combine adjoining classes - there is no requirement for the class size to be the same across the whole range.