STA10003 Supplement B - Reporting Information about Single Variables Classifying a variable according to its level of measurement: When reporting information about single variables, we need to take into account the level of measurement before making decisions about which procedure is required, and the statistics that are the most appropriate to report. Data can be classified under the two broad headings: Categorical or Metric. Categorical data has a nominal or ordinal level of measurement, whereas metric data has an interval or ratio measure. Levels of Measurement: Categorical: Nominal & Ordinal Nominal: used to label the group eg 1 = Male 2 = Female Ordinal: used to label AND order st nd rd th eg 1 , 2 , 3 , 4 in a race, Chapter 1, Chapter 2, Chapter 3 in Text Metric: Interval & Ratio Interval: numbers are used to label and order AND the intervals between the numbers are equal 0 0 eg Temperature in Cor F Ratio: numbers are used to label and order and the intervals between the numbers are equal AND ZERO means a complete absence of something eg number of correct answers in a test In an existing SPSS data file, categorical data will usually have named categories, as numbers are used to either represent different named categories only [nominal data] or to represent different named categories which also have some form of ranking [ordinal data]. By contrast, when a variable recorded in an SPSS file is metric, it occurs naturally as numbers which represent a reasonably accurate quantitative measurement, so these numbers will not have name labels assigned to them. Producing the analysis and writing the report: The report is a description of the information / findings from the sample. FOR CATEGORICAL VARIABLES: SPSS FREQUENCIES procedure: Frequencies Table [default] Percentage Bar Chart [requested] Page 1 of 11 .. STA10003 Supplement B Information .. ..
STA10003 Supplement B - Reporting Information about Single Variables REPORT SHOULD CONTAIN: An introductory sentence about the variable of interest, a description of the sample [size and who / what] and reference to the Figure representing graphical information about the distribution. Information relating to the 'most typical' [mode] response category - including percentages. Any other information that will provide an 'overall' picture of the data. For example, was there one response category that was quite different from the others? FOR METRIC VARIABLES: SPSS EXPLORE procedure: Descriptives Table [default] Histogram [requested] and Boxplot [default] Percentiles Table [requested] REPORT SHOULD CONTAIN: An introductory sentence about the variable of interest, a description of the sample [size and who / what] and refer to the Figure representing graphical information of the distribution. The shape of the distribution [symmetrical or skewed - if skewed the direction of the skew] and interpretation of the 'measure of centre' [mean / sd for symmetrical or median for skewed data] An interpretation of the 'middle 50%' [25th to 75th percentiles - which provides an indication of how spread out the data is Any other information that will provide