THE  A-Z OF TESTING

 

 

M

 

MAXIMUM PERFORMANCE

 

We can distinguish two types of test:

 

Typical Performance: which measure how you characteristically behave. These include measures of areas like personality and mood

 

Maximum Performance: which try to find out how well people can do things.

 

Contra received opinion, tests of maximum performance are administered and designed to encourage the candidates’ best efforts. Tests of ability and attainment fall into this area.

 

MEAN

 

The total of everyone’s scores on the test divided by the number of people who took it – the most common form of average.

 

MEASUREMENT ERROR

 

At the risk of repetition, all measurements include error. Using good tests has the advantage of allowing you to estimate the error and see how confident you can be in the score you’ve got. See also CLASSIC TEST THEORY and CONFIDENCE BANDS.

 

  

MOOD

 

More transient STATES ( as opposed to such relatively stable features of human beings as personality factors, often called TRAITS ) such as fatigue, concentration and sadness ( although obviously if these moods persist over a long period they may be pointing to other issues).

 

Moods can affect well-being and performance: for instance, tiredness affects, among other things, creativity and accuracy of checking. There are profile assessments of mood states which are characteristically used frequently to assess the way in which these states change.

 

MOTIVATION

 

In psychology, motivation is a rich area of study and looks at why people do things and why they chose a particular way to act out of the many possible options.

 

In its business applications, motivation has taken on considerable importance. We ask how we can motivate people to achieve individual and corporate objectives. For some time now the ‘management by objectives’ movement and the use of performance-related bonuses has rested on a very specific view of human motivation. More recently the popular field of sports psychology has influenced management of staff as research has identified low motivation as a major problem in UK industry.

 

Tests can measure people’s individual motivation and findings confirm that people are motivated in many more ways than has been thought; money does not come top of the list in motivating factors. Indeed, for some people offering performance related bonuses may have the opposite of the desired effect and, generally, older and younger people are motivated by very different factors.

 

Assessing motivation of existing staff can have a huge impact on the specifics of management.

 

MULTIPLE-CHOICE

 

Multiple choice tests ask questions or present statements then present a number of possible answers or reactions. Test takers chose their answers from the these options. The multiple choice format, apart from anything else, makes it much easier to automate and score tests.

  

N

 

NARRATIVE

 

Another term for report, interpretation or analysis, a narrative refers to a report, generated either by a computer system or an individual test expert which draws out the implications of test scores in prose. It may also contain statistical analyses and graphic representations of the scores.

 

Carefully translating the numerical output of a test into clear and accurate words is a major challenge for report / narrative writers.

 

You’ll occasionally come across the word NARRATIVE referring to a particular field of psychology, with some applications in business. NARRATIVE PSYCHOLOGY looks at the stories we tell ourselves. There’s some evidence that human beings are more comfortable coping with the world in terms of stories, rather than in terms of theories or analyses. Psychologists and consultants have been looking at both the images and narratives we tell ourselves about the organisations we work for and found that these are powerful tools for understanding how an organisation ticks, what its VALUES are and how it can be changed.

 

NEED

 

Perhaps best known in management circles via Abraham Maslow’s Hierarchy of Needs, needs are things you require and which you will be motivated to get. Typical needs are social approval, safety, having enough to eat. Maslow posited that individuals climb a hierarchy of needs from very basic ones ( being safe, warm and fed) to higher ones ( such as ‘self actualisation’) and some management theories use this framework to create differential reward systems

 

NORMAL DISTRIBUTION

 

The famous Bell curve. The frequency of many characteristics from height and weight of a population to the distribution of personality characteristics and the incidence of disabilities follow this shape.

 

The bell curve is used in test construction and has certain very specific mathematical characteristics which enable us to create norms for tests which are much more useful than, simply, RAW SCORES. The Normal Distribution allows us to report scores on a test in a way which allows us to compare them meaningfully with averages and other peoples’ scores.

 

You may see “ The Bell Curve” referred to in relation to a specific book which analysed intelligence test scores and argued that intelligence was largely genetically determined and then drew some highly controversial conclusions for American society.

 

We should stress that the bell curve is just a statistical tool used in a huge number of disciplines and that its use does not endorse the findings of that book, which have been roundly attacked as politically biased by researchers and academics in the field of intelligence studies.

 

NORMATIVE

 

On normative tests, an individual’s scores are compared with the scores of others, so that you can make statements like “ A scores better on numerical reasoning than 50% of UK male anaesthetists “ or “ B’s personality profile is similar to that of successful product managers in fast moving consumer goods companies.” See also IPSATIVE

 

NORMS

 

Norms are what you compare an individuals scores with when they take a normative test. They comprise tables  which enable you to turn an individual’s RAW SCORES into STANDARD SCORES which can then be compared with other people’s performance.

 

Norms are created when a large sample of individuals is asked to fill in a test and then their scores are analysed.

 

There are guidelines for the minimum number of people that should take a test to create norms in order for the test to be VALID. In fact this number varies considerably from many thousands for educational tests to a few tens for tests of clinical mental states.

 

What sorts of norm groups a manual contains tells you a lot about whether a test is relevant to your purpose.

 

A test may report general population norms, for instance. It has been tried out with a sample of people that reflects the make up of the UK population in terms of gender, age, ethnic origins, geographical area. It may offer specific norms ( successful accountants or aircraft maintenance staff or directors ). Expert test users can create their own norms relating to a particular defined group or within an organisation though there are some technical problems around interpreting scores based on very specific tightly defined norm groups that users should be aware of.

 

You should look at the description of the norm groups for any test to check you’re comparing it with a relevant sample.

 

Norm groups go out of date as jobs, training, education, society and the language in which ITEMS are written change. It’s easier to update norms on the internet than it is in printed versions of tests, where publishers often offer NORM SUPPLEMENTS; data gathered after the test was published but not yet integrated into the main manual.

 

Converting RAW SCORES into STANDARDISED SCORES used to be a complicated, eye-straining and mistake ridden task using tables in the back of a manual. Computer systems do this matching and conversion process for us, nowadays.

 

O

 

OPEN ENDED FORMAT

 

In a MULTIPLE CHOICE  test you chose your answer or reaction from a number of options offered to you. Open Ended Format tests don’t offer options: you provide your own responses. Exams which demand essays are open format assessments. So are tests which present you with a stimulus ( a picture for instance ) which you have to react to spontaneously.

 

OVER-INTERPRETATION

 

Human beings seem to place greater reliance on numbers than words. This brings dangers of over-interpretation.

 

If someone scores 6 on a test and someone else scores 7 we tend to see the difference as significant. We a difference in numbers as somehow more precise than the difference between the words “more” and “ less”.

 

Good psychometric tests use STANDARD SCORES, the NORMAL CURVE OF DISTRIBUTION, STANDARD OF ERROR OF MEASUREMENT and a range of sophisticated statistical techniques to highlight how much reliance you can put on the numbers testing produces.

 

This might seem a waste of time, given that you’re using tests to make important and often urgent decisions. Put this characteristic of testing another way and it becomes clearer why it’s a real benefit. Testing highlights what is significant about a person and prevents you from over-interpreting insignificant details.

 

Dealing with people is a complex business and anything that can help us focus on key issues will, in the end, help us make better decisions.

 

The issue of Over-Interpretation highlights another point. Since tests use numbers there’s a tendency to over-rely on them. Thomas International stresses that tests should never be used on their own. They provide very rich information on specific areas of human functioning but you need to use them with other techniques to get the rounded information you require to make a crucial decision.