This summary of How do you measure personality? - Chapter 4 - Barelds - 2016 is written in 2020

When you want to measure personality you have to make sure to do this the right way. That's also why in psychology researchers always have to think about the reliability and validity of a test.

What is reliability?

When it comes to measuring personality traits, it is of course important that this is done well and responsibly, because you are playing with an important aspect of people's life. The most important psychometric criteria (relating t the technical characteristics of a test, which can be examined statistically) will be discussed. These concern reliability, validity and standardization.

If you want to determine how tall a person is, we can use a tape measure. If we measure a person's height several times using the same tape measure, we will probably get the same measurement every time. This is what we call a reliable measurement. If you want to measure psychological characteristics, there is no such thing as a measurement tape to measure the characteristic. So how do you know when psychological characteristics are measured the right way with a reliable measure. A benefit of psychological characteristics and personality traits it that someone's personality is often stable over time. So if you test someone's personality multiple times the test-hertest reliability is often very strong. If the correlation is perfect it will show a r=1, this also means that the test is perfectly reliable. If you think about other psychological characteristics, such as emotions, psychological measurement is difficult, because emotions are not stable over time. 

You can also look at the relationship between the items in the test. If the questionnaire has 20 items you can look at the correlation between the twenty items and this can give you an indication of the reliability of the test. You will get an estimate of the correlation between the twenty items. The most commonly used estimate of reliability is the Cronbach's alpha coefficient. If there is a perfect reliability, the Cronbach's alpha is equal to 1.

What is validity?

So the reliability relates to the question of whether the score a person obtains is a good representation of reality, validity mainly relates to the question whether the test does what it should do. The question of whether a psychological test measures what it should measure is what people call validity. There are two types of validity people often use: "Does the test measure what it should measure?", what is known as construct validity and "Does the test predict what it should predict?" what is known as criterion validity. 

What is construct validity?

The construct validity is about if the test is capable of measuring a particular contruct. You can answer this question by means of an empirical research. The three mains ways are to investigate (1) differences between groups, (2) links with other tests, and (3) the internal structure of the test. 

(1) Can the test make a distinction between people with and without a narcissistic personality disorder? If the difference between the two people/groups can be made, then this supports the construct validity of a test. 

(2) You can examine the links between this test and other tests. This can help to clarify what the test actually measures. The obvious way would be to examine the relationship between a test and other tests that aim to measure the same construct. If strong links are found, this is an indication that the instrument measures what it should measure. It is difficult to find similar tests that measure exactly the same construct, so a researcher often has to settle for tests that measure roughly the same construct or tests. 

(3) You can also look at the internal structure. This can be done by looking at the correlations between the different parts of the test or by means of a factor analysis. When factor analysis is used to examine the internal structure of a test, it analyses whether the test contains as many factors as you would expect. 

What is criterion validity?

Criterion validity is also known as predictive validity. It looks at how well a test is capable of predicting certain behavior, also known as the criterion. There is no precise definition of the term criterion, but you can describe it as what you want to predict. 

What are other types of validity?

Other well known types of validity are face validity and content validity. Face validity refers to the subjective impression you might have about what the test measures or predicts. Content validity refers to if the content of a test corresponds to the construct to be measured. 

Is there a connection between reliability and validity?

Correlations are often used to determine validity. A correlation is theoretically between -1 and 1, where 1 implies a perfect positive connection, and -1 implies a perfect negative relationship, and 0 implies there is no connection whatsoever. This also means that when the test has a good reliability the test automatically also has a good validity. But, when the test has a poor validity, this doesn't automatically mean that the test is also unreliable. An example; it can be completely unclear what the test measures, but the test does measure this construct in a reliable way.

What is standardization?

If you take a test and get your raw scores back on that test, that score is meaningless if no comparison is possible. There are two types of norm scores you can use: absolute norm scores and relative norm scores. If a norm group is used to compare the scores, it is important that the norm group meets a number of conditions. 

• The norm group must be relevant to the purpose of the test
• It is essential that a norm group is representative
• The size of the group is also relevant (COTAN says at least 400 people in a standard norm group)
• The norm group also needs to be up to date

What is COTAN?

The Dutch Committee on Tests and Testing (or COTAN) is a committee set up by the Netherlands Institution of Psychologists (NIP) to assess the quality of all psychological tests. There are seven basic criteria by which they assess tests:

1. Principles of test construction
2. Quality of test materials
3. Quality of the manual
4. Norms
5. Reliability
6. Construct validity
7. Criterion validity

All these criteria are assessed on a three-point scale; unsatisfactory (o), satisfactory (v) and good (g). 

What are projective tests?

A projective test is a test in which the subject is presented with an unstructured task. The idea is that the subject will tackle these tasks in their own way without any help, to reveal something of their own personality traits, preferences etc. There are many different projective techniques:

• Association methods: the subject is confronted with a stimulus and has to say the first thing that comes to mind in response to that stimulus
• Constructive methods: the subject's task is to construct something on the basis of something that is offered to him, usually a story
• Completion methods: the subject is given an incomplete task to complete
• Choice or ordering method: in this case, the subject can choose from a number of alternatives, or the subject is required to classify a number of stimuli.
• Expressive methods: the subject is instructed to draw a tree, a person or a house, for example.

A well known projective test is the Rorschach inkblot test, the TAT (thematic apperception test) and the sentence-completion test. 

What is observation?

Observation is used as a method for determining personality traits when it is difficult to obtain this information directly from the subject. Observations can take place in the subject's natural surroundings, in a standardized laboratory situation or in a situation somewhere in between. The problem that can occur in all these situations is that the person being observed usually knows that this is happening and therefore may behave differently. (Think about a 'Big Brother' television program.) 

What are the most commonly used personality questionnaires?

The following personality questionnaires are the most commonly used in the Netherlands: MBTI, MMPI-2, NEO-PI-R, the FFPI-II and the NPV. 

What is the MBTI: Myers-Briggs Type Indicator?

The MBTI is a questionnaire that is used in human resources management. This questionnaire is based upon the thinking of Carl Gustav Jung (1921) and was developed by mother and daughter couple Katherina Briggs and Isabel Myers. The MBTI measures four traits and people score a high or low for each trait.

1. Extroverted (E) versus Introverted (I): do you mainly get energy from the things around you or are you more focused on your inner world?
2. Sensing (S) versus Intuitive (I): do you prefer factual information or more intuitive and general information?
3. Thinking (T) versus Feeling (F): do you make decisions mainly on the basis of logic or more on the basis of feeling?
4. Judging (J) versus Perceiving (P): do you like your life to be ordered and structured or more flexible and spontaneous? 

The questions in MBTI are what are known as forced choice items, such as:

I prefer to work: (a) alone, (b) in a team

The original MBTI consists of 93 questions. If someone, for example, scores high on Extroverted scale they get the letter E, so each person who takes the MBTI will get 4 letters for the four scales. There are sixteen possible letter combinations in total. These sixteen combinations of letters correspond to sixteen types of MBTI. 

What is the MMPI-2: Minnesota Multiphasic Personality Inventory 2?

The MMPI, and MMPI-2 are internationally the most widely used personality questionnaires in the healthcare sector. The MMPI is used to measure personality and psychopathology. 

The MMPI-2 is quite complex in terms of scales contained in this test. The MMPI-2 has seven validity scales, ten clinical scales, fifteen content scales, and fifteen supplementary scales. 

One problem with the MMPI-2 is that it contains a relatively large number of poorly formulated questions. Another problem is the large item overlap between the scales. This is also why the COTAN assessed the MMPI-2 quite negatively, with 'unsatisfactory' ratings for the norms, construct validity and criterion validity. 

What is the NEO-PI-R?

The NEO-PI-R is a questionnaire developed by Costa and McCrae (1992) to measure their Five Factor Model of Personality. It is one of the most widely used internationally for measuring personality traits and is certainly the most widely used instrument for measuring the factors from the popular Five Factor Model. 

It consists of 240 items, distributed over the five dimensions (48 questions per scale): Neuroticism, Extroversion, Conscientiousness, Openess to experience and Agreeableness. 

• Neuroticism: anxiety, angry hostility, depression, self-consciousness, impulsiveness and vulnerability
• Extraversion: warmth, gregariousness, assertiveness, activity, excitement-seeking and positive emotions
• Openness to experience: fantasy, aesthetics, feelings, actions, ideas and values
• Agreeableness: trust, straightforwardness, altruism, compliance, modesty and tender-mindedness
• Conscientiousness: competence, order, dutifulness, achievement striving, self-discipline and deliberation. 

What is the FFPI-11: Five-Factor Personality Inventory I?

The FFPI-II is the revised version of the FFPI and is based on the Dutch psycho-lexical Big Five structure. The FFPI consists of a hundred short phrases which can be answered both by the person themselves or by someone who knows the person well ('likes a chat', 'neglects tasks'). A unique aspect of the FFPI-II is that the questionnaire has no less than ninety facet scores. How a person scores on the facets can be calculated on the basis of the scores obtained on the five scales of the FFPI-II thus revealing the person's most characteristic facets. 

What is the NPV: The Dutch Personality Questionnaire?

The Dutch Personality Questionnaire is a personality questionnaire that was developed to measure a wide range of personality constructs that are relevant to different practice areas. The seven scales have much in common with the Big Five Factors. The NPV is a test that is often used in various practices areas in the Netherlands.

Are there any problems with personality assessment?

Various methods that can be used to map personality traits have been mentioned in the previous sections. One well-known problem concerns the tendency of some people to give answers that are not, or not only, related to the content of the question. These are known as responsive tendencies. Some people tend to want to agree to question, this is called acquiescence. It could also be what is known as careless responding, in that case the subject is answering questions at random. One way to detect this kind of behavior is to include an infrequency scale in the questionnaire.

Another response tendency is social desirability. This has to do with answering questions in such a way that you come across as someone who is socially desirable.