Lecture 2 Notes: Data Coding & Entry, Levels of Measurement, and Visual Plots

Vocabulary flashcards covering data coding, levels of measurement, SPSS data entry, survey types, data plots, and related concepts from Lecture 2.

Quantitative data

Numerical data produced when researchers measure something and assign a number value.

Discrete data

• Quantitative data with finite values or categories; the values can only be counted and cannot be mathematically manipulated.

• also called categorical data - this will be nominal

Continuous data

• Quantitative data that can take an infinite number of values; values are interval or ratio and can be mathematically manipulated.

• eg blood pressure, height, weight, body temp

sources of quantitative data

• surveys

• observations

• secondary databases - data collected by somebody other than you

types of surveys

• questionnaires - paper or electronic instruments w questions to collect data from individuals - self-administered surveys

• interview-type surveys — face-to-face or over the phone/online between researcher and individual — interviewer-administered

types of survey questions

• close ended — predetermined answers, easiest to analyze

• open ended — answer in their own words, harder to analyze but can be done

• closed and open ended — still restrictions, but then later ask for open ended

Nominal level of measurement

• Assign numerical values to categories with no inherent order; values cannot be ranked; categories are mutually exclusive and exhaustive. these cannot be manipulated.

• dichotomous is only two options, but falls under dichotomous

Ordinal level of measurement

• Categories can be ordered or ranked;

• intervals between categories are not equal;

• categories are mutually exclusive and exhaustive.

• subjective ratins will have ordinal measurement properties

Scale (Normal) level of measurement

• Also called interval/ratio or normal scale; infinite possible values, equal intervals, true zero; suitable for parametric statistics.

• normal distribution, continuous

• this is where we can do mathematical manipulation

• allows for parametric tests, which have more power or ability to detect relationships and/or differences. non normal have to use nonparametric tests, and have less power to determine relationships and/or differences compared to parametric

Dichotomous variable

A nominal variable with two levels (dummy variable); examples: 1=Male, 2=Female.

Data coding

Assign numerical codes to categories or attributes to prepare data for SPSS.

Levels of measurement

• Concept used to determine data type and appropriate statistical tests; main levels: nominal, ordinal, scale.

• assignment of numerical values to attributes of variables according to some rules

Data dictionary / data codebook

Documentation describing variables, coding, and measurement to guide data entry and analysis.

SPSS Variable View

SPSS window where you define variable properties: names, labels, values, and measure.

SPSS Data View

SPSS window showing the actual data; rows are cases, columns are variables.

Variable name

Meaningful, concise identifier for a variable in SPSS.

Variable label

Descriptive text describing a variable to aid interpretation.

Variable values

Numeric codes used to represent categories within a variable.

Value labels

Text labels attached to numeric codes in SPSS to describe categories.

Missing values

Codes or blanks indicating missing data; SPSS can use codes to distinguish reasons for missingness.

detecting data entry problems

• eyeball it when entering the data - double and triple check

• generate and inspect visual plots

• generate and inspect descriptive statistics

  • like mean, minimum, maximum, range

  • mean will be really useful because it will be altered by outliers. if the mean is too big or small, it might suggest that there is incorrect data for variables

Frequency distribution

• A table showing counts (or percentages) of each value for a variable. — how many times each score or value occurs for that variable. this will use a table

• frequency plot is a graphical representation of values on a variable and can be used to represent all types of data

normal distribution

most scores are for middle values, with a small number of scores in the low or high values

negatively skewed distribution

extreme scores, or the tail of the curve are on the low end, or the left sidea

Bar chart

Best for nominal, dichotomous, and ordinal data; not ideal for normal/scale data.

Histogram

Best for scale/normal data; not appropriate for nominal or dichotomous data.

Frequency polygon

Line-connecting plot for distributions; best for scale/normal data, not nominal.

Box and whisker plot

Plot useful for ordinal and normal/scale data; shows quartiles and spread.

mean

• average - takes into account all available info in computing central tendency of a frequency distribution

• add all raw scores and divide by number of scores

median

• good measure of central tendency for ordinal level raw data

• the middle value

• especially good when you have outliers

mode

• most common category or score

• least precise info on central tendency

• this would be the tallest bar in bar graph or histogram

measures of variability

• if all scores in distribution are the same, there is no variability

• if scores are all different and far apart, variability will be high

range

• a measure of variability

• highest minus lowest score

standard deviation

• measure of variability

• based on deviation of each score from the mean of all scores

interquartile range

• in box plot, distance between top and bottom of box

• the whiskers are the expected range

Pie chart

Alternative to bar chart for nominal data; bar charts are usually easier to read.

Parametric statistics

Statistical tests with higher power used with normal/scale data.

Nonparametric statistics

Statistical tests used for non-normal data; generally less powerful.

SPSS Data Editor concepts

Interface for entering and managing data, including Data View and Variable View.

Two views in SPSS

Data View shows data; Variable View shows variable properties and measurement.

Mutually exclusive and exhaustive

Categories do not overlap (mutually exclusive) and cover all possibilities (exhaustive).