Psych 211-25A (Week three)

Variables and Measurement

• Variable - any event, situation or behaviour or individual characteristic that can have more than one possible value

• Levels - values that a variable may possibly take

• A variable must have at least two levels

Levels:

• conceptual variables - abstract ideas that form the basis of a research question

  • cannot be directly measured

  • usually what we are most interested in

  • variables on a theoretical level

• operational variables - the definition of an abstract concept used by a researcher to measure or manipulate the concept in a research study

Reliable and Valid measurement scales

• reliability - consistency

• validity - accuracy

• Reliability of measurement scales - the extent to which a measure is consistent

• Validity of measurement scales - the extent to which a measurement is accurate

Assessing reliability:

• Correlation Coefficients - measure the direction and strength of a relationship between variables

  • can vary from -1 to 1 (or 0-1)

  • 0 = no relationship

  • The closer the correlation coefficient is to -1 or 1, the more reliable the measure is

• 3 ways to test for reliability:

  • 1) Test - Retest reliability - reliability measured by looking at the correlation between scores on a measure given at one time with scores on the same measure given at a later time

  • 2) Internal consistency reliability - reliability measured by assessing the correlation between items on a scale collected at one point

  • 3) Inter - rater reliability - the extent to which the ratings of two or more judges or observers correlate with each other

• The four validities

  • 1) construct validity - the extent to which a measured (OPERATIONAL) variable actually measures the conceptual variable it is designed to assess (MEASURE OF ACCURACY)

  • How well our operational definition measures our conceptual variable

    • we never have access to conceptual variables due to all being made up of internalised factors (anxiety, intelligence etc)

    • Context matters when discussing

  • 2) Statistical validity

  • 3) Internal validity

  • 4) External validity

• Validity - the extent to which a claim, conclusion, finding or decision is accurate

• must always test for reliability first to get to validity

Umbrella of Construct Validity:

• Face validity - the extent to which the measure looks like it is measuring what is supposed to measure

  • based on face value (worth or implication of something)

• Content validity - the degree to which the operationalised variable adequately captures the conceptual variable of interest

  • coverage of the content

  • maximising the overlap

• Criterion validity - the extent to which a measure accurately predicts peoples behaviour (or the thing of interest)

  • Concurrent validity = present behaviour

  • Predictive validity = future behaviour

• Convergent validity - the degree to which scores on a measure are related to scores on other measures of the same or similar construct

  • correlation with other established measures of anxiety (does my new measure correlate?)

  • does it relate to things within the concept (anxiety)

  • Don’t always want peak correlation, otherwise new methods can be deemed pointless

• Discriminant validity - the extent to which the scores on a measure are not related to scores on conceptually unrelated measures

  • Don’t want the scale to relate to things that have nothing to do with anxiety

Summary:

• Studying conceptual variable = OPERATIONALISE first

• afterwards = measure test to see if it is RELIABLE

• afterwards = test if your measure is VALID

Scales of measurement

• Qualitative variables - variables that describe a trait, category, or group

  • NOT OKAY to do MATHEMATICAL operations on these variables

  • examples: gender, biological sex

• Quantitative variables - variables that describe difference in magnitude or amount

  • OKAY to do MATHEMATICAL operations on these variables

  • examples: distance in meters, time in seconds

Four main scales of measurement:

• Nominal scale - a scale of data measurement that involves non ordered categories i.e gender, political party

  • type: Qualitative data

  • operation: Logical only (is, isn’t)

• Ordinal scale - a level of measurement that involves ordered categorical responses i.e grade in class

  • type: Qualitative

  • operation: Logical (is, isn’t, greater, less)

  • i.e 1st, 2nd, 3rd place

• Interval scale - a scale of measurement that uses numerical values that are equally spaced out i.e consistent interval between numbers

  • type: Quantitative

  • operation: Logical (is, isn’t, greater, less) AND mathematical (+,-)

  • i.e temperature (Celcius and Farenheit)

  • Zero point is always arbitrary

• Ratio scale - a scale of measurement that uses numerical values that are true ratios of each other; they have a true zero point i.e time in seconds, length in inches

  • type: Quantitative

  • operation: Logical (is, isn’t, greater, less) AND mathematical (+,-,*,/)

  • i.e time in seconds, temperature in Kelvin