Levels of Analysis
Reality can be seen as organized on a number of levels. Each level incorporates the ones "below." Visually, the idea might be depicted with a number of concentric circles or layers of an onion. For example, Miller (in Living Systems) sees the world organized much as the following.

Regarding the levels or units, we are trying to answer two questions:
1. How are the phenomena on any level (or within any unit of organization) interrelated? How do they interact?

2. How do the lower levels (or smaller units) constitute the higher levels, and how do the higher levels serve as contexts that help to organize the lower levels? For example, what arrangements of interpersonal exchanges equal or produce a certain form of school organization? And how does a form of school organization (once it is institutionalized) guide the interpersonal exchanges of members?

Degrees of Cut-up-ability, or Scales of Measurement
The phenomena that are incorporated by concepts at each level can be seen as "having" various degrees of "cut-up-ability." At the same time, different forms of instrumentation (measurement) impose different degrees of cut-up-ability on the phenomena being measured. The different degrees of cut-up-ability are called "scales" or "levels of measurement." These are as follows.

1. Ratio scale. Implies a true zero (e.g., zero episodes of aggression occurred; zero income). In addition, there are equal intervals between quantities; e.g., the difference between 0 and 1, 1 and 2, etc., is 1). Examples of phenomena and concepts (that identify phenomena) that exist at a ratio level include behaviors and other events that can be counted. Examples of instrumentation or measurements that impose a ratio scale on whatever is being measured include: a) frequency counts (/////); b) magnitude estimations ("Let's see. If my anxiety level last week could be given a rating of 200, my anxiety level this week is down to about 100."); and c) questions that ask for a real number ("What was your income last year, in dollars?" or "How many people would you say you can depend on for emotional support?").

2. Interval scale. There are equal intervals between quantitative degrees, but no true zero. Intelligence and temperature are examples of interval-level phenomena, concepts, or variables. "But," someone asks, "what about the expression, 'It's zero degrees outside'"? The answer is: The zero is an arbitrary visual or verbal symbol for a certain spot on the temperature-measuring continuum. It does not mean no temperature. Examples of measurement items that place the quantity of a variable on an interval scale include the following:

What was your total income last year?
(1) between zero and $10,000.
(2) between $10,001 and $20,000.
(3) between $20,001 and $30,000.
(4) between $30,001 and $40,000.
(5) more than $40,000.

How much do you like yourself? Like |__1_|__2_|__3_|__4_|_5__| Dislike

3. Ordinal Scale. An ordinal-scale variable or measure implies degrees, but not equal intervals between degrees. In addition, no true zero exists or is imposed. Probably most attitudes, perceptions and feelings are in reality ordinal-level. So, too, might be concepts such as "stage of civilization or social development" (implying more or less of certain attributes that define the concept); "stage of cognitive development" (again, implying more or less of certain attributes), or "degree of social cohesion." Examples of items by which one might measure (represent) where a phenomenon is on an ordinal scale include the following.

Stages of civilization or social development [Note: stages that do not imply degrees of "goodness," but perhaps degrees of energy production, population size, or energy use.]
(1) Hunting and gathering
(2) Village gardening
(3) Pastoral
(4) Industrial

Cultural Stages of Knowledge Acquisition and Understanding [Each stage is defined by: degrees of openness and adaptation to new and possibly contradictory information; freedom of thought.]
(1) Theological
(2) Philosophical
(3) Scientific
(4) ?

How confident are you that you will be able to master this material? [Note: The four levels look like levels of an equal-interval scale, but are the differences between 4 and 3, 3 and 2, 2 and 1 really equal?
(1) Quite confident
(2) Fairly confident
(3) Fairly doubtful
(4) Quite doubtful

4. Nominal Scale. A nominal scale implies qualitative not quantitative differences. It refers to kinds or types of things. Examples include marital status, occupation, and ethnic identity. "But," someone might say, "different occupations have different degrees of prestige, income, and control over the work process. So, why isn't occupation an ordinal, interval, or ratio variable?" Good question. But we have to be very clear about the phenomenon or concept that we are talking about. If we were talking about "occupational prestige," that would be a phenomenon that is perhaps on (and could be measured by) an ordinal scale. And if we were talking about the average income for occupations, that would be a ratio-level phenomenon. But if we are talking about "occupation" per se (steam fitter, salad chef, or track coach), occupation is simply a nominal-level phenomenon.

What about "marital status"; that is, married, separated, divorced, widowed, not married? Aren't these statuses somehow quantitatively different from each other? Answer: Do each of the statuses imply degrees of anything that is essential to the definition of the status? Or is each one merely a kind of status with which there might be associated some quantitative phenomena? Only you can decide! Because things do not come with labels on them.

In other words, a woman is either pregnant or she is not; child-bearing status is a nominal scale variable. However, a woman can be 1-9 months pregnant; pregnant with a tiny to a huge fetus; pregnant with one or more children, etc. Similarly, you could be a high jumper or a shot putter. One is not more or less than the other. But, as a shot putter, your performance could be measured on a ratio scale, for example.

A few cautionary comments
1. You can use a lower-level scale for measuring a phenomenon that is on a higher level. For example, you can measure yearly income (which is a ratio-level phenomenon) with an interval-level item (see example, above, for interval scales). Or, you could measure the amount of mutual reward exchanges in a family (which is ratio-level) with an ordinal-level item. For example: Rate the amount of observed mutual reward interaction among family members. (a) Not much mutual reward;
(b) A fair amount of mutual reward; (c) A lot of mutual reward

However, you lose information when you do this. You could have learned exactly how many dollars a person earned, or you could have learned exactly how many episodes of mutual reward were observed. (I did not say "how many occurred" because there is no way to know that.) In other words, if you use a lower level of measurement than you could haveused, you lose precision.

2. However, you CANNOT (!!!) use a higher-level scale to measure a phenomenon that is really on a lower scale. For example, you cannot place a nominal-level variable on an interval scale. Following is a truly bizarre rating of civilizations. Each number is not merely a name or code (e.g., for entering the civilization into a computer); it is seen as implying degrees of something.

American =5 Latino =4 Asian =3 Slovik =2 Germanic =1

As you can see, the above rating scale makes no sense. What if a person were a combination of Latino and American ethnicities? Would she be scored as 4 + 5/ 2 = 4.5? Does Slovik have one more unit of something than Germanic, but one less unit of it that Asian? Nor should you (although many people do, including me) measure ordinal-level phenomena (such as attitudes and perceptions) on an interval-level or ratio-level scale. Why shouldn't you? Because you are imposing more orderliness on reality than there might be. [Does shame, for example, really come in different equal-interval packages?] And because the measurements may not be valid. [How accurately can a person give a precise numerical rating to a less-than-precise feeling?]