PNB 3RM3 Lecture Notes - Lecture 12: Level Of Measurement, Intelligence Quotient, Data Analysis
Document Summary
Categorical: nominal (uses names to label levels of its variables, report percentages, types of category) Represents only qualitative differences; you can"t quantitate differences. Sex, species, major in university, political affiliation, religious affiliation. Cases on one level must be equivalent to others on that level; cannot overlap with cases on another level - levels are mutually exclusive. Any numbers used are arbitrary (doesn"t mean one group is better than the other) Tells you rank order - provide information on relative position. Travel website"s hotel rankings - you only know rank between hotels, but you don"t know by how much; the actual difference can"t be quantified. Prefer not to use ordinal scale; not lots of room for data analysis. Multiple conclusions can be drawn in statistics, but you can"t speak in ratios (i. e. It is twice as hot as it was yesterday) The value of zero truly means nothing; true zero exists.