Soc 220

1
Soc 220

• January 21, 2009
• Prof. Brines
• MGH 231

2
Levels of Measurement

• Nominal
• Ordinal
• Interval
• Ratio

3
Examples Ordinal, Interval, or Ratio?

• SAT verbal test scores (these range from 300 to
  800)
• Q1 What was your familys income last year?
• 0 Less than 10,000
• 1 10,000-24,999
• 2 25,000-49,999
• 3 50,000-74,999
• 4 75,000 or more

4
From the Panel Study of Income Dynamics

• Q2. About how much time did you spend last week
  doing work around the house? I mean cooking,
  cleaning, and other household tasks.
• ________________ hours (fill in the blank).

5

• Variables Measures
• Type of Variation Level of Measurement
• Discrete (categorical choppy) Nominal
• Ordinal
• Interval
• Continuous Ratio

6
Measurement Decisions II

• Checks for
• Validity
• Reliability
• Precision

7
Measurement Validity

• Does the measure get at what I intend to measure,
  or is it really measuring something else?
• Invalid measures introduce error into
  research
• A) Idiosyncratic error
• B) Generic error
• C) Method factors

8
Idiosyncratic Error

• Problems w/a measure that affect respondents
  answers in unique ways. These tend to affect only
  small numbers of respondents.
• Items that respondents dont understand, that are
  worded in ways that trigger unintended response,
  or that are sensitive due to recent events.

9
Idiosyncratic error

• Q1 How much do you agree or disagree with this
  statement?
• It is better to get married than to go through
  life being single.
• 1 2 3 4 5
• strongly strongly
• disagree agree

10
Generic Error

• Problems with measures where the responses of
  groups dont reflect what true nature of the
  opinion, behavior, etc. being measured.
• E.g., socially desirable responses

11
Generic error Social Desirability

• Q3 How much do you agree or disagree with this
  statement?
• Adult children should provide financial support
  to aged parents if they are in need.
• 1 2 3 4 5
• Strongly Strongly
• Disagree Agree

12
Method-based Error

• Problems with measures that affect most/all
  respondents answers.
• Unclear questions unbalanced response choices

13
Method Factors

• Q4 Do you support the recent transportation
  decision made by the Seattle City Council?
• ? No
• ? Yes

14
Method Factors

• Q5 How wrong do you think it is for someone who
  is not a hunter to own a gun?
• Very wrong Very wrong
• Wrong Wrong
• Slightly wrong Neither right nor wrong
• Not at all wrong Right
• Very right

15
Face Validity

• Is the measure valid on its face (to all
  appearances)?
• Variable Measure
• Physical aggression of times has hit,
• shoved or slapped others
• Creativity Persons ranking of favorite
  movies

16
Reliability

• Is the measure consistent in its measurement of
  the variable?
• Consider a fixed attribute musical aptitude
• Test scores for Joe Smith reliability
• T1 33 33
• T2 42 31
• T3 32 34
• T4 17 32

17
Precision

• How exact or finely-measured is the variable
  when I use this measure?
• Variable Measure
• What is your age (in years)
• Age ______________ ?
• vs.
• What is your birthdate?
• ______________ ?

18
Beware of measures that invite false precision!

• Q2. About how much time did you spend last week
  doing work around the house? I mean cooking,
  cleaning, and other household tasks.
• ________________ hours (fill in the blank).

19
Variables and Measures Checklist

• Is the measure consistent with the definition for
  the variable?
• Does the level of measurement align with the
  choppiness or continuousness of the variable?
• Hows the face validity of my measure?
• Am I trying to force too much precision?
• Am I trying to measure something about a
  sensitive subject that is likely to elicit a
  biased report?