Statistical data such as vital statistics is useful to grasp local characteristic and problems.
We use, and how should we evaluate number of statistical data? Here, we want to give viewpoint and precaution when we use statistical data.
What do you want to know?
We make purpose clear what you want to know using statistical data.
If purpose becomes clear, we collect what kind of statistical data and can map out a course how we should evaluate.
We see whether you see in city in ward, or is it Codium in area?
When we use statistical data, we decide whether you evaluate at unity (unit) of which scale.
Do you see whether you see whether you see in credit in city by unit of ward by more finely local unit? Please decide unit of evaluation to match purpose using statistical data.
We look for statistical data
We look for statistical data which number according to sex age-grade appears in.
Average life expectancy has a longer woman than man, and the rough death rate rises so as to become old. In this way, number of statistics is often different from sex in tendency by age.
About the death and birth used well, we can obtain from vital statistics.
We obtain population according to sex age-grade of unit (city, ward, area) to evaluate.
If there is much local population, people who may suffer from disease increase. Therefore it is necessary to think about influence on number of statistics by difference in population.
About population, national census population and population (thing which modified birth, death, transference cases to national census population) by estimate are used well.
We can download material of government figures such as vital statistics from WEB page. (collection of links to obtain statistical data)
You take real number, or do you take health index?
(1) We use real number
At first, let's look at real number.
Is it best that evaluation of statistical data uses health indexes such as "the rough death rate" or "the age adjustment death rate?"
Health index has good appearance, and, with index considering influence such as population or age, there is advantage such as being easy to compare in local interval and annual.
However, may it be said that the area does not have any problem because value of health index is low?
For example, in A city of a population of 1 million people and B city of 100,000 people, we do as the number of the suicides was 200 A wards, B ward 50. In this case, as for "the rough death rate" of suicide (a population of 100,000 pairs), B ward becomes 50.0 in A ward 20.0. Although it is low, in A ward, phenomenon that the number of the suicides has much gets angry at frequency of suicide from B ward. Is there not that there is much number of the suicides itself in A ward in problem?
It depends on way of thinking of side using statistics "how much do assume problem?" but, at first, what we consider number itself of statistical data slowly and carefully and evaluate is important.
Viewpoint of evaluation: Sex age-grades (graph), secular variation (graph)
We see evaluation of real number according to sex age-grades and are possible by seeing secular variation. We graph, and it will be good to make plain visually.
(2) We use health index
We decide health index to seek what you finally want to know by (purpose).
There is health index a lot. We extract index along purpose two or three and demanded and there is few and finishes labor than we seek health index of limit to hit on at random and is easy to evaluate.
It is important to understand the characteristic (good point, bad point) to choose the most suitable health index.
Let's check whether there is not usable existing health index.
In health statistical data, there is thing which seeks health index such as vital statistics special report and life table according to municipality beforehand, and is released. There is difficult point to take time before it is released, but we use such an information, and what we evaluate will be good.
（We obtain existing health index）
When we use existing health index, it is necessary to confirm what kind of expression it was demanded in using number of which statistical data. When this is not health index that we made same how to find when we compare to see secular variation between areas, this is because it cannot compare obediently.
When we seek health index, let's think about accidental error.
When we use time to seek health index about phenomenon showing infrequent outbreak and local statistical data with a little population, it is necessary to think about error with numerical chance.
Method to lower error accidentally may adopt statistical technique, but method to average number in the plural number year generally has stolen.
Depending on statistical data to use, let's think about systematic error.
When the people's life basics investigation or national census use statistical data made based on questionary survey, it is necessary to think even if there is systematic error.
This is because answer may be twisted depending on having marriage or not and contents of question such as occupations. In addition, you may think that the number of the death and the number of the birth of vital statistics used well, national census population are hardly angry at systematic error.
Viewpoint of evaluation: Sex age-grades (graph), secular variation (graph), other comparisons (we separate map by color) with area
We see health index that we sought according to sex age-grades and see secular variation and we compare with other areas and can use. It will be good we graph, and to do by classification of map clearly visually.
Depending on characteristic of index, let's think about method of evaluation.