Developing and Defining Constructs
Construct: specified in a theory in order to describe, explain, and predict a wide range of
related empirical findings.
Constructs studied in psychology include: intelligence; personality; memory;
These are not directly observable material entities. They can also not be totally
captured just by their titled.
Operational Definitions of Constructs
A construct needs to be translated into an operational definition. It is not limited to how
to measure the construct. In laboratory studies, it spells out the conditions in which the
construct is supposed to occur.
It identifies the independent variable
The structure of a scientific theory starts with abstract constructs, which then become
independent/dependent variables. In many cases, these variables will be linked by
intervening variables, hypothetical internal states.
Positive psychology: seeks to learn about human thriving, optimism, resilience, joy, and
Provides a theoretical framework to study attitudes, actions, and feelings.
Mental health and mental illness were separate: Theorized that mental health was an
enduring sense of well-being and happiness that a person regularly experiences in life.
He used the DSM to explain mental illness. The DSM has two component required for
diagnosing a mental illness:
Presence of a particular set of symptoms as described in the DSM
That the illness is pervasive and interferes with normal functioning
He was able to match the constructs to the behaviours exhibited by people. This was
considered valid research because:
The extent to which the various variables converged
The extent to which the different variables discriminated between mental health
and mental illness. This study was meant to prove that mental health and mental illness were separate and
that the lack of mental illness didn’t actually mean that are person was totally mentally
Measurement of Psychological Constructs
Measurement: assignment of numerals to objects or events, based on some objective
The objects or events are variables, selected as indicators to measure a given
The variable will in turn specify a form of measurement
Advantages of Multiple Measurements
Essential in order to develop a complete understanding of a phenomenon.
Levels of Measurement
Numbers are used to represent a particular scale of measurement: refers to levels of
measurement of a variable
Nominal Level of Measurement
Def.: identifies variables whose values have no mathematical interpretation; vary in kind
or quality but not amount.
Also known as categorical or qualitative level
Classifies data according to category only
While attributes of categorical variables do not have a mathematical meaning, they must
be assigned to cases carefully. The attributes must be:
Mutually exclusive: every case must have only one attribute
Exhaustive: every case can be classified into one of the categories
These allows for every case to correspond to one attribute.
Special Case of Dichotomies
Dichotomy: variables having only two values
This can work for a statistical procedure requires that variables be quantitative a
dichotomy can be perfectly acceptable.
Example: gender. There is no number for gender, it’s just either/or.
Ordinal Level of Measurement
Def.: numbers are assigned to response choices, to allow for ‘greater than’ and ‘less
than’ distinctions. Allows for a ranking of responses
An example of this is the Likert Scale: often involves a question about attitude or
opinions that asks for a numerical rating of the extent of agreement or disagreement.
The different values of variable must be mutually exclusive and exhaustive. They must
cover the range of observed values and allow for each case to be assigned to only one
A limitation of the ordinal level is that you cannot sume that the respondents perceive
the differences between response scale points as equidistant.
Interval Level of Measurement
Def.: a blend of nominal and ordinal level of measurement. It gives a name or category
for each observation, with a number serving as a code for a level.
Like an ordinal scale, responses are numerically ordered or ranked from lowest
to highest on some particular characteristic.
In this case, the distances between any two points are of known size. The
numbers represent fixed measurement units but have no absolute zero point.
The key thing is that because the interval scale has no true zero points, ratio
measurements make no sense.
Numbers can be added and subtracted nut rations are not meaningful.
In psychology, standardized measures of intelligence, personality, and aptitude use an
interval level of measurements. They have an arbitrary zero point because a person
cannot actually have no personality or intelligence.
Ratio Level of Measurement
Def.: represent fixed measuring units and an absolute zero point
In this case, ratio numbers can be multiplied and divided, as they start at an absolute
Comparison of Levels of Measurement
All four levels of measurement allow researchers to assign different values to different
Scores obtained on an ordinal, interval, or ratio scales all allow for mathematical
operations and are often referred to as quantitative variables.
A nominal scale is used for qualitative variables
Nonparametric statistics: test hypotheses for variables that use either a nominal or an
ordinal scale of measurement. Parametric statistics: used with variables that are measured on either an interval or ratio
Psychometric tests and Measurements
Psychometrics: school of psychology that studies the application of psychological tests
as objective measures of the mind and mental processes.
Intelligence is only partially captured by IQ test scores and shouldn’t be confused with
actual wisdom or virtue.
The 15 point IQ gap
A gap exists between White Americans and African-Americans, with the White
Americans scoring higher. There two important points to consider about this gap:
All caste-like minorities show a similar 15 point gap between them and the
nonstigmatized members of their societies
Race-based cognitive inequalities are not hereditary
Really, this gap reflects outcome bias: when a test consistently produces lower scores
for an identifiable class of people defi