Editing Personality test (section)

== Topics ==

===Norms===
The meaning of ''personality test'' scores are difficult to interpret in a direct sense. For this reason substantial effort is made by producers of personality tests to produce norms to provide a comparative basis for interpreting a respondent's test scores. Common formats for these norms include [[percentile]] ranks, [[standard score|z scores]], [[sten scores]], and other forms of standardized scores.

===Test development===
A substantial amount of research and thinking has gone into the topic of personality test development. Development of personality tests tends to be an iterative process whereby a test is progressively refined. Test development can proceed on theoretical or statistical grounds. There are three commonly used general strategies: Inductive, Deductive, and Empirical.<ref>{{cite journal|last=Burisch|first=Matthias|title=Approaches to personality inventory construction: A comparison of merits|journal=American Psychologist|date=March 1984|volume=39|issue=3|pages=214–227|doi=10.1037/0003-066X.39.3.214}}</ref> Scales created today will often incorporate elements of all three methods.

Deductive assessment construction begins by selecting a domain or construct to measure.<ref name="Burisch, M. 1984">{{cite journal | last1 = Burisch | first1 = M | year = 1984 | title = Approaches to personality inventory construction: A comparison of merits | journal = American Psychologist | volume = 39 | issue = 3| pages = 214–227 | doi=10.1037/0003-066x.39.3.214}}</ref> The construct is thoroughly defined by experts and items are created which fully represent all the attributes of the construct definition.<ref name="Burisch, M. 1984"/> Test items are then selected or eliminated based upon which will result in the strongest internal validity for the scale. Measures created through deductive methodology are equally valid and take significantly less time to construct compared to inductive and empirical measures. The clearly defined and face valid questions that result from this process make them easy for the person taking the assessment to understand. Although subtle items can be created through the deductive process,<ref>{{cite journal | last1 = Jackson | first1 = D. N. | year = 1971 | title = The dynamics of structured personality tests: 1971 | journal = Psychological Review | volume = 78 | issue = 3| pages = 229–248 | doi=10.1037/h0030852}}</ref> these measure often are not as capable of detecting lying as other methods of personality assessment construction.<ref name="Burisch, M. 1984"/>

Inductive assessment construction begins with the creation of a multitude of diverse items. The items created for an inductive measure to not intended to represent any theory or construct in particular. Once the items have been created they are administered to a large group of participants. This allows researchers to analyze natural relationships among the questions and label components of the scale based upon how the questions group together. Several statistical techniques can be used to determine the constructs assessed by the measure. Exploratory Factor Analysis and Confirmatory Factor Analysis are two of the most common data reduction techniques that allow researchers to create scales from responses on the initial items.{{citation needed|date=April 2015}}

The [[Five Factor Model]] of personality was developed using this method.<ref>{{cite journal|last=McCrae|first=Robert|author2=Oliver John|title=An Introduction to the Five-Factor Model and Its Applications|journal=Journal of Personality|year=1992|volume=60|issue=2|pages=175–215|doi=10.1111/j.1467-6494.1992.tb00970.x|pmid=1635039|citeseerx=10.1.1.470.4858|s2cid=10596836 }}</ref> Advanced statistical methods include the opportunity to discover previously unidentified or unexpected relationships between items or constructs. It also may allow for the development of subtle items that prevent test takers from knowing what is being measured and may represent the actual structure of a construct better than a pre-developed theory.<ref>{{cite journal|last=Smith|first=Greggory|author2=Sarah Fischer |author3=Suzannah Fister |title=Incremental Validity Principles in Test Construction|journal=Psychological Assessment|date=December 2003|volume=15|issue=4|pages=467–477|doi=10.1037/1040-3590.15.4.467|pmid=14692843}}</ref> Criticisms include a vulnerability to finding item relationships that do not apply to a broader population, difficulty identifying what may be measured in each component because of confusing item relationships, or constructs that were not fully addressed by the originally created questions.<ref>{{cite book|last=Ryan Joseph|title=Understanding Psychological Assessment: Perspective on Individual Differences|year=2001|publisher=Springer|pages=1–15|edition=1|author-link=Understanding Test Construction|author2=Shane Lopez |author3=Scott Sumerall |editor=William Dorfman, Michel Hersen}}</ref>

Empirically derived personality assessments require statistical techniques. One of the central goals of empirical personality assessment is to create a test that validly discriminates between two distinct dimensions of personality. Empirical tests can take a great deal of time to construct. In order to ensure that the test is measuring what it is purported to measure, psychologists first collect data through self- or observer reports, ideally from a large number of participants.
{{Further|topic=the matched series of timed cognitive aptitude tests|Morrisby Profile}}

=== Self- vs. observer-reports ===
A personality test can be administered directly to the person being evaluated or to an observer. In a self-report, the individual responds to personality items as they pertain to the person himself/herself. Self-reports are commonly used. In an observer-report, a person responds to the personality items as those items pertain to someone else. To produce the most accurate results, the observer needs to know the individual being evaluated. Combining the scores of a self-report and an observer report can reduce error, providing a more accurate depiction of the person being evaluated. Self- and observer-reports tend to yield similar results, supporting their validity.<ref name=":0">{{Cite book|title=Individual Differences and Personality|last=C.|first=Ashton, Michael|isbn=9780128098455|edition=3rd|oclc=987583452|date = 2017-06-13|publisher=Elsevier Science }}</ref>

===Direct observation reports===
Direct observation involves a second party directly observing and evaluating someone else. The second party observes how the target of the observation behaves in certain situations (e.g., how a child behaves in a schoolyard during recess). The observations can take place in a natural (e.g., a schoolyard) or artificial setting (social psychology laboratory). Direct observation can help identify job applicants (e.g., work samples<ref>{{Cite web |url=http://hr.ucdavis.edu/recruitment/selection/work-samples.html |title=Interview Work Sample |website=Human Resources |publisher=University of California, Davis |access-date=2018-04-08 |archive-url=https://web.archive.org/web/20180409043719/http://hr.ucdavis.edu/recruitment/selection/work-samples.html |archive-date=2018-04-09 |url-status=dead }}</ref>) who are likely to be successful or maternal attachment in young children (e.g., [[Mary Ainsworth]]'s [[strange situation]]). The object of the method is to directly observe ''genuine'' behaviors in the target. A limitation of direct observation is that the target persons may change their behavior because they know that they are being observed.<ref>Schonfeld, I.S., & Mazzola, J.J. (2013). Strengths and limitations of qualitative approaches to research in occupational health psychology. In R. Sinclair, M. Wang, & L. Tetrick (Eds.), ''Research methods in occupational health psychology: State of the art in measurement, design, and data analysis'' (pp. 268-289). New York: Routledge.</ref> A second limitation is that some behavioral traits are more difficult to observe (e.g., sincerity) than others (e.g., sociability). A third limitation is that direct observation is more expensive and time-consuming than a number of other methods (e.g., [[self-report study|self-report]]).<ref name=":0" />

=== Personality tests in the workplace ===
Though personality tests date back to the early 20th century, it was not until 1988 when it became illegal in the United States for employers to use polygraphs that they began to more broadly utilize personality tests.<ref>{{cite journal |last1=Stabile |first1=Susan J. |title=The Use of Personality Tests as a Hiring Tool: Is the Benefit Worth the Cost |journal=University of Pennsylvania Journal of Labor and Employment |date=2001 |volume=4 |page=279}}</ref> The idea behind these personality tests is that employers can reduce their turnover rates and prevent economic losses in the form of people prone to thievery, drug abuse, emotional disorders or violence in the workplace. There is a chance that an applicant may fake responses to personality test items in order to make the applicant appear more attractive to the employing organization than the individual actually is.<ref>{{cite journal | last=Ones | first=D.S. | title=Personality at Work: Raising Awareness and Correcting Misconceptions | journal=Human Performance | publisher=Informa UK Limited | volume=18 | issue=4 | year=2005 | issn=0895-9285 | doi=10.1207/s15327043hup1804_5 | pages=389–404| s2cid=36707701 }}</ref>

Personality tests are often part of [[management consulting]] services, as having a certification to conduct a particular test is a way for a consultant to offer an additional service and demonstrate their qualifications. The tests are used in narrowing down potential job applicants, as well as which employees are more suitable for promotion.<ref name=":1">{{Cite news|last=Cunningham|first=Lillian|date=December 14, 2012|title=Myers-Briggs: Does it pay to know your type?|url=https://www.washingtonpost.com/national/on-leadership/myers-briggs-does-it-pay-to-know-your-type/2012/12/14/eaed51ae-3fcc-11e2-bca3-aadc9b7e29c5_story.html|newspaper=Washington Post}}</ref> The United States federal government is a notable customer of personality test services outside the private sector with approximately 200 federal agencies, including the military, using personality assessment services.<ref name=":1"/>

Despite evidence showing personality tests as one of the least reliable metrics in assessing job applicants,<ref>{{Cite news|last=Martin|first=Whitney|date=2014-08-27|title=The Problem with Using Personality Tests for Hiring|work=Harvard Business Review|url=https://hbr.org/2014/08/the-problem-with-using-personality-tests-for-hiring|access-date=2021-12-17|issn=0017-8012}}</ref> they remain popular as a way to screen candidates.

===Test evaluation===
There are several criteria for evaluating a ''personality test''. For a test to be successful, users need to be sure that (a) test results are replicable and (b) the test measures what its creators purport it to measure. Fundamentally, a ''personality test'' is expected to demonstrate [[Reliability (statistics)|reliability]] and [[Test validity|validity]]. Reliability refers to the extent to which test scores, if a test were administered to a sample twice within a short period of time, would be similar in both administrations. Test validity refers to evidence that a test measures the [[Construct (philosophy)|construct]] (e.g., neuroticism) that it is supposed to measure.<ref>{{cite book|last1=Urbina|first1=Susana|title=Essentials of Psychological Testing|date=2014-06-30|publisher=John Wiley & Sons, Incorporated|location=Hoboken. New Jersey|isbn=978-1-118-70725-8|pages=127–128, 165–167|edition=Second|url=https://public.ebookcentral.proquest.com/choice/publicfullrecord.aspx?p=1727717&ppg=1|access-date=4 May 2018}}</ref>

===Analysis===
A respondent's response is used to compute the analysis. Analysis of data is a long process. Two major theories are used here: classical test theory (CTT), used for the observed score;<ref>See: {{cite book |last1=Lord |first1=F.M. |last2=Novick |first2=M.R. |year=1968 |title=Statistical theories of mental test scores |place=Reading, MA |publisher=Addison-Wesley}}</ref> and item response theory (IRT), "a family of models for persons' responses to items".<ref>{{cite book |last=Mellenbergh |first=G.J. |author-link=Gideon J. Mellenbergh |chapter=Chapter 11 - Tests and questionnaires: Analysis |page=[https://books.google.com/books?id=LCnOj4ZFyjkC&pg=PA244 244] |chapter-url=https://books.google.com/books?id=LCnOj4ZFyjkC&pg=PA235 |editor-last=Adèr | editor-first=H.J. |editor-link=Herman J. Adèr | editor-last2=Mellenbergh | editor-first2=G.J. | title=Advising on Research Methods: A Consultant's Companion | publisher=Johannes Van Kessel Publishing | year=2008 | isbn=978-90-79418-01-5 }}</ref><ref>For a full summary of IRT, see: {{citation | last1=Hambleton | first1=R.K. | last2=Swaminathan | first2=H. |date=April 1985 | title=A Look at Psychometrics in the Netherlands | url=https://files.eric.ed.gov/fulltext/ED273665.pdf | id={{ERIC|ED273665}} }}</ref> The two theories focus upon different 'levels' of responses and researchers are implored to use both in order to fully appreciate their results.

===Non-response===
Firstly, item non-response needs to be addressed. Non-response can either be ''unit'', where a person gave no response for any of the ''n'' items, or ''item'', i.e., individual question. Unit non-response is generally dealt with exclusion.<ref name="Mellenbergh, 2008">{{cite book |last=Mellenbergh |first=G.J. |author-link=Gideon J. Mellenbergh |chapter=Chapter 11 - Tests and questionnaires: Analysis |pages=235–70 |chapter-url=https://books.google.com/books?id=LCnOj4ZFyjkC&pg=PA235 |editor-last=Adèr | editor-first=H.J. |editor-link=Herman J. Adèr | editor-last2=Mellenbergh | editor-first2=G.J. | title=Advising on Research Methods: A Consultant's Companion | publisher=Johannes Van Kessel Publishing | year=2008 | isbn=978-90-79418-01-5 }}</ref> Item non-response should be handled by [[Imputation (statistics)|imputation]] – the method used can vary between test and questionnaire items.

===Scoring===
The conventional method of scoring items is to assign '0' for an incorrect answer and '1' for a correct answer. When tests have more response options (e.g. multiple choice items) '0' when incorrect, '1' for being partly correct and '2' for being correct.<ref name="Mellenbergh, 2008"/> Personality tests can also be scored using a dimensional (normative) or a typological (ipsative) approach. Dimensional approaches such as the Big 5 describe personality as a set of continuous dimensions on which individuals differ. From the item scores, an 'observed' score is computed. This is generally found by summing the un-weighted item scores.