Please enable JavaScript.
Coggle requires JavaScript to display documents.
Module 3: psychological scaling (measurement is the assigning of numbers…
Module 3: psychological scaling
measurement is the assigning of numbers to individuals in a systematic way as a means of representing properties of the individuals
"measurement" consists of rules for assigning symbols to objects so as to do one of the following
represent quantities of attributes numerically
(scaling)
or
define whether the objects fall in the same or different categories with respect to a given attribute
(classification)
no matter what definition we use we need to do the following
we need to be able to quantify the attribute of interest
we must be able to quantify our attribute of interest in a consistent and systematic way
(standardization)
we must remember that we are measuring attributes of individuals or (objects), not the individuals per se
Levels of measurement
Nominal measurements
: used to assign numbers to categories in order to keep track of them and keep them in order
this is a
qualitative
form of measurement
Ordinal measurement
: used to rank things in order from high to low or low to high. here the numbers do have some quantitative meaning
Interval scale
: i.e. celsius thermometer. zero does not mean zero it is just assigned zero on the scale
Ratio scale
: Kelvin thermometer. zero actual does mean zero or represent zero
this is
quantitative
data
can compute mean, median, and inter-quartile range
when we have interval level data, we can calculate statistics such as means, standard deviations, variances, and various statistics of shapes (skew and kurtosis)
Unidimensional scaling models
how much of an attribute a given person has
allows us to measure the
inter-individual
and
intra-individual
differences of participants
what this allows us to do is hold one attribute constant, scale across another (compute the score) and then measure the differences of the third attribute
we want to first test the survey on SMEs (
subject matter experts
)
with participants we want to get their attitudes or semantics
with SMEs we want them to keep their semantics neutral so that way we can adequately measure if each question is correctly measuring the construct
the SMEs can help us to scale the stimuli, or rule out which questions are not going to tap into the construct
in
ordinal
testing we can have the SMEs rank order the questions to help us pick the best ones for the current construct we are trying to measure
in
interval level measurement
the SMEs not only rank order but also provide the range of difference between each question in regards to how close it is to measuring the construct
Multidimensional scaling models
subjects are asked to give their general impression or broad rating of similarities or differences among stimuli
subjects might be asked to look at a variety of products and rank which ones are similar and which ones are different
MDS or
multi dimensional scaling
is particularly useful when participants are sure they like a product or category more than another but are unable to articulate why