Part 1: Levels of Measurement (The "NOIR" Scale)
Before you can analyze data, you must understand what "type" of data you have. The level of measurement (scales of measurement) determines which mathematical operations (like addition or averaging) are allowed. We classify data into four levels, often remembered by the acronym NOIR. Going from lowest to highest, the 4 levels of measurement are cumulative. This means that they each take on the properties of lower levels and add new properties.
1. Nominal Scale Level (The "Naming" Level)
The word "Nominal" comes from the Latin nomen, meaning "name." At this level, numbers or words are used solely as identifiers or categories. There is no mathematical value to the labels themselves.
- Key Characteristics: Data is qualitative and mutually exclusive (you belong to one category or another). You cannot say one category is "more" or "less" than another.
- The "Number" Trap: Sometimes we use numbers for nominal data, like Zip Codes or Jersey Numbers. You can’t add two zip codes together to get a "better" location; the number is just a shortcut for a name.
- Examples: Eye color, gender, political party, types of flooring, or "Yes/No" survey responses.
- Mathematical Limit: You can only calculate the Mode (the most common category).
The "Ordinal" level introduces rank. It tells you the position of data points relative to each other, but it doesn't tell you how much better or bigger one is than the other.
- Key Characteristics: There is a logical sequence or "natural order." However, the intervals between the ranks are unknown or inconsistent.
- The "Gap" Problem: If you come in 1st, 2nd, and 3rd in a race, the ordinal scale tells us the order of finish. It does not tell us if the 1st place runner beat 2nd place by one second or ten minutes.
- Examples: Likert scales (Strongly Disagree to Strongly Agree), class rank (Valedictorian, Salutatorian), or "Small, Medium, Large" drink sizes.
- Mathematical Limit: You can find the Mode and the Median (the middle rank), but you cannot calculate a meaningful Mean (average).
The "Interval" level gives us order and tells us that the distance between each point is exactly the same. However, it lacks a "true zero."
- Key Characteristics: The difference between 70° and 80° is exactly the same as the difference between 30° and 40°. Because the intervals are equal, addition and subtraction become possible.
- The "Zero" Problem: Zero on an interval scale is just another point on the line; it does not mean "nothing." For example, $0^\circ\text{C}$ doesn't mean there is "no temperature"—it’s just the freezing point of water. Because there is no starting point, you cannot make "twice as much" statements.
- Examples: Temperature (Fahrenheit/Celsius), IQ scores, and Years (the year 0 is an arbitrary point in time, not the "beginning of time").
- Mathematical Limit: You can calculate the Mean, Median, and Mode. You can add and subtract, but you cannot multiply or divide (no ratios).
4. Ratio Scale Level (The "True Zero" Level)
This is the "gold standard" of measurement. It has all the properties of the previous three, but adds a True Zero point. Zero actually means "the absence of the thing being measured."
- Key Characteristics: Because there is an absolute zero, you can finally perform multiplication and division. You can meaningfully say one value is "double" or "half" of another.
- The "Ratio" Advantage: If you weigh 200 lbs and your friend weighs 100 lbs, you are exactly twice as heavy. This is only possible because 0 lbs means "no weight."
- Examples: Weight, height, distance, time duration (e.g., "it took 5 minutes"), and money ($0 means you have no money).
- Mathematical Limit: All statistical operations are allowed. This is the level required for the most advanced types of data analysis.
Key Definitions
1. Frequency (f): The number of times a specific value occurs.
2. Relative Frequency: The proportion (or percent) of the total data that a value represents.
Formula: Frequency/Total Number of Data Points
3. Cumulative Relative Frequency: The running total of relative frequencies. It shows the percentage of data that falls at or below a certain value.
Example:
Suppose 20 students reported the hours they worked yesterday. Their responses were as follows:
5, 6, 3, 3, 2, 4, 7, 5, 2, 3, 5, 6, 5, 4, 4, 3, 5, 2, 5, 3The following table lists the different data values in ascending order and their frequencies.
A frequency is the number of times a value of the data occurs. According to the table, there are three students who work two hours, five students who work three hours, and so on. The sum of the values in the frequency column, 20, represents the total number of students included in the sample.
Part 3: Rounding
When calculating the frequency, you may need to round your answers so that they are as precise as possible. A simple way to round off answers is to carry your final answer one more decimal place than was present in the original data. Round off only the final answer. Do not round off any intermediate results, if possible. If it becomes necessary to round off intermediate results, carry them to at least twice as many decimal places as the final answer.
For example, the average of the three quiz scores four, six, and nine is 6.3333.... Since the data are whole numbers, we would round this to 6.3.
Reference
Gemini AI
No comments:
Post a Comment