Reliability Definition: The stability or consistency of a test. Assumption: True score = obtained score +/- error. ~ Domain Sampling Model ~ Item Domain Test Test-retest method [error is due to changes occurring due to the passage of time] Test 1 ------------- r Test 1 ------------- Issues: • Length of time between test administrations if crucial (generally, the longer the interval, the lower the reliability) • Memory • Stability of the construct being assessed • Speed tests, sensory discrimination, psychomotor tests (possible fatigue factor) Parallel/Alternate Forms [error due to test content and perhaps passage of time] Test 1 ------------- r Test 2 ------------- Issues: • Need same number & type of items on each test • Item difficulty must be the same on each test • Variability of scores must be the same on each test Two types: 1) Immediate (back-to-back administrations) 2) Delayed (a time interval between administrations) KR-20 and Coefficient Alpha [error due to item similarity] • KR-20 is used with scales that have right & wrong responses (e.g., achievement tests) • Alpha is used for scales that have a range of response options where there are no right or wrong responses (e.g., 7-point Likert-type scales KR-20 Rtt = k k–1 % of those getting the item correct Σ pi (1 – pi) σy 2 # of items variance of test scores variance of scores on each item Alpha ά = k 1 – Σ σ i2 k–1 # of items σ y2 variance of test scores Factors Affecting Reliability 1) Variability of scores (generally, the more variability, the higher the reliability) 2) Number of items (the more questions, the higher the reliability) 3) Item difficulty (moderately difficult items lead to higher reliability, e.g., p-value of .40 to .60) 4) Homogeneity/similarity of item content (e.g., item x total score correlation; the more homogeneity, the higher the reliability) 5) Scale format/number of response options (the more options, the higher the reliability) Types ofof Validity Types Validity Content Validity [the extent to which test items represent the domain] a) Subject Matter Expert Opinions (e.g., CVR statistic) b) Internal consistency reliability c) Correlation with other similar tests Types of Validity (cont.) Criterion-Related Validity [correlation between test scores and job performance scores Predictive – Concurrent -- [Correlation between test scores of applicants and their performance scores when some time interval has passed after they are hired] [Correlation between test scores and performance scores of current employees] • Range restriction issue on performance scores • Time, cost, & pragmatic concerns • Job experience factor • Motivation level • Guessing, Faking • Range restriction issue on performance scores Types of Selection Tests • Application Blanks (Weighted Application Blanks, Biographical Information Blanks) • Honesty (Integrity) Tests, Drug Tests • Interviews (e.g., Situational) • Aptitude Tests (e.g., Mechanical, Clerical) • Work Samples • Assessment Centers (Situational Exercises) Application Blanks • Content of items (use of job analysis) • Number of application blanks (one for each position or job category) • Legal issues • Image of organization (e.g., format, recruitment issue, perceived fairness of questions) • Accuracy of data Education (e.g., 21%), salary (22%), job title (24%), years worked (29%) College students willing to include one lie on AB (95%); 45% had done so Yahoo CEO Scott Thompson caught padding his resume Yahoo director Patti Hart, who was in charge of the company's CEO search, will not seek re-election at the next shareholder meeting. An exact date for the meeting has not been set, but last year's event was held in June. Hart, who joined Yahoo's board in 2010, is the CEO of International Game Technology Statements from both Yahoo and IGT said that IGT's board asked her not to seek re-election at Yahoo. Thompson stepping down as Yahoo CEO (May 13th, Washington Post) Yahoo chief executive Scott Thompson, hired from eBay in January to turn around the struggling Web portal, is stepping down after it was discovered that his academic credentials were misrepresented, according to a person with knowledge of the matter In a series of published biographical statements stretching back for years – including his bio on Yahoo’s website ---- Thompson has said that he "holds a Bachelor's degree in accounting and computer science" from Stonehill College. But his degree is actually only in accounting … Yahoo sent out a statement saying references to Thompson earning a computer science degree were an "inadvertent error.” … It's an error Thompson made repeatedly. References to Thompson's nonexistent computer science degree are featured in his bios on sites for PayPal, the eBay subsidiary where he previously served as president. Frequency of Common Inappropriate Application Blank Questions Worded Appropriately Not Asked Item Not Appropriate Past salary 98.9% 0% 1.1% Minimum acceptable salary 72.7% 0% 27.2% Reference source 59.1% 0% 40.9% Age 54.5% 37.5% 8.0% relatives 50.0% 10.2% 39.8% Conviction records 43.2% 28.4% 28.4% Health 40.9% 2.3% 56.8% Military service 30.7% 30.7% 38.6% Marital status 27.3% 0% 72.7% 25.0% 43.2% 31.8% residence 23.9% 0% 76.1% Physical description, photo 19.3% 0% 80.7% Rent or own car or home 18.2% 0% 81.8% Handicap 17.0% 6.8% 76.2% Organizational membership 15.9% 21.6% 12.5% Work schedule 13.6% 63.6% 22.7% Information about Who to notify in case of emergency Length of time in Years of experience and previous salary are the strongest predictors of starting salary, and starting salary is the greatest predictor of current salary. --- Mickey Silberman, Jackson Lewis, LLP Industry Liason Conference (2011) Other possible concerns include: personal email, personal web page, and/or Facebook accounts Effect of Name on Resumes and Interview Rates Name type Resume Quality Low High “White” sounding name “Black” sounding name 50% less chance of being invited for an interview versus “Whites” with high qualifications (Bertrand & Mullainathan; 2004) Social Media and Selection Frequency of Use --•18% indicated that they have used social networking websites to screen applicants, while 11% planned on using such sites in the future (survey of over 400 organizations by the Society for Human Resource Management in 2011) • 45% of employers used social networking sites to investigate job applicants (survey of over 2,600 hiring managers conducted by Harris Interactive for CareerBuilder.com) Consequences? • 35% of organizations in Harris survey said they did not hire candidates Protection Act (2012) due to content available on social networking sites.Password Most common examples • of negative information included Provocative attire Images of drug or alcohol use, Complaints about previous employers Password Protection Act (2012) • Prohibit an employer from forcing prospective or current employees to provide access to their own private account as a condition of employment • Prohibit employers from discriminating or retaliating against a prospective or current employee because that employee refuses to provide access to a password-protected account. ~ Honesty Testing ~ Two Types 1) Overt integrity tests 2) Covert (Honesty test questions imbedded within personality measures, e.g., among items assessing drug use/attitudes, tendency toward violence) ~ Honesty Testing ~ Types of Question (Overt): 1) Frequency and extent of theft (e.g., What percentage of people take more than $1.00 per week from their employer?) 2) Punitiveness toward theft (e.g., should a person be fired if caught stealing $5.00?) 3) Thoughts about theft (e.g., Have you ever thought about taking company merchandise without actually taking any?) 4) Perceived ease of theft (e.g., How easy would it be for a dishonest person to steal from an employer?) 5) Likelihood of detection (e.g., What percent of employee thieves are ever caught? Other types of questions: • Do you think you are too honest to take something that is not yours? • Do you believe most employers take advantage of the people who work for them? • How often do you tell the truth? • I have frequently associated with fellow employees who admitted they were stealing merchandise from the company. Validity Issues: a) Correlations with polygraph results, b) Future behavior (e.g., # days with cash shortage, discharges), c) Admissions of past theft, d) Shrinkage reduction, Contrasted groups (e.g., scores by criminals vs. general population scores) Honesty Testing --- Types of Questions (cont.) Covert Tests (tests about one’s attitudes thought to relate to honesty) • It would be better if almost all laws were thrown away [Conventionality, traditionalism] • Sometimes I am strongly attracted by the personal articles of others [Self-discipline] • I like to take chances [Sensation seeking] • On average, how often during the week do you go to parties? [Sensation seeking] • Does it bother you when people ask stupid questions? [Forgiveness] • How often do you make your bed? [Tidiness] Drug Testing • Approximately 57% of companies use pre-employment drug testing (Society for Human Resource Management survey (2011) • What does a positive drug test score indicate? Some contend that drug use does not negatively affect performance in the majority of jobs (Macdonald, 1997) Other research indicates that drug use is related to poor job performance (e.g., increased injuries and involuntary turnover) (Bass et al., 1996; Normand, Salyards, & Mahoney, 1990). Drug Testing Some Issues: • Those being tested Applicants, employees or both • Testing procedure Random or for cause • Type of company Public or private • Type of test Sensitivity, Cross-reactivity (TLC, EIA/RIA, GCMS) • Type of job Safety concerns or not ~ Reference Checks ~ (Exceptionally common technique; e.g., 95% usage by organizations) In-Person (e.g., interview) • Costly, time consuming • Used in jobs that involve the concern for risks (e.g., security, $) • Can elicit different types of information (differences between in-person and written reference information) Mail (or e-mail) • Low return rate using “snail” mail (e.g., 56 – 64%) • Standardized questions, format • Written record of responses • Ensure confidentiality of responses (signed statement by applicant) ~ Telephone Checks ~ (More frequently used than written references) • Allows follow-up or clarification of answers given • Less resistance to giving certain types of information can be collected • Quick process • Important data can be gleaned from various verbal cues (e.g., pauses, hesitations, voice inflections, voice level, intonations) • Relatively high return rate • Better responsiveness, more interactive nature of the method • More confidence in the identity of responder Sample Biographical Information Blank Items • During high school, how many times did you make the honor roll? • How much freedom or independence did your parents allow you in grade school? • How important did your favorite high school teachers stress discipline in the classroom? • How many times did you change schools before you were sixteen years old? • Compared to other people in high school, how many friends did you have? • How old were you when you spent your first week (or more) away from your parents? • How bothered are you if you a job is left undone? • How often do you read craft and mechanics magazines? • How quickly do you normally work? • How well do you feel you can understand the feelings of others? • How well do you tolerate performing routine tasks? Classification of Biographical Items ~ Employment Interview ~ • Frequently used to make selection decisions (over 90%) • Social exchange (interpersonal) process • Search for information COMMON PROBLEMS WITH THE “TRADITIONAL” INTERVIEW • Variety of Interviewer Biases * 1st Impressions * Expectancy Effect * Contrast Effect * Stereotype Matching • Different Questions Asked to Applicants (Lack of standardization) • Disagreement on the Desirability of Interview Responses • Little Formal Interviewer Training • Subjective (or no) Scoring System • Interview Conducted and Scored by One Person • Poor Reliability, Validity, and Job Relevancy (Open to Legal Challenge) ~ Summary of Situational Interview Process ~ • Perform a Job Analysis Using the Critical Incident Technique • Place Critical Incidents into Relevant Job Dimensions (e.g., Safety, Responsibility, Interpersonal Skills) • Reword Critical Incidents Into Question Form • Decide on the desirability of responses [Think of how good, average, and mediocre workers would have answered such a question] • Conduct interviews in groups of two or more. Each interviewer scores applicant independently. A single score is given after group discussion Situational Interview Process (cont.) Differential Aptitude Test (DAT) Abstract Reasoning PROBLEM FIGURES ANSWER FIGURES A B C D E Differential Aptitude Test (DAT) Abstract Reasoning (cont.) PROBLEM FIGURES ANSWER FIGURES A B C D E Differential Aptitude Test (DAT) Mechanical Reasoning A B Which weighs more? (If equal, mark C.) Differential Aptitude Test (DAT) Space Relations A B C D ~ Work Sample Tests ~ (performing a piece, or sample task, of the job) • Good content validity • Less/no adverse impact ~ Assessment Center Process ~ Candidates Participate in Situational Exercises Sample Group Exercises Leaderless Group Discussion Business Game Sample Individual Exercises Interview Simulation Scheduling Exercise In-Basket Observed, discussed, and scored on various dimensions (e.g., communication, decision making, planning/organizational skills) by trained raters • Overall scores computed and ranked for personnel decisions • Indiviudaul dimension scores used for developmental purposes Types of Selection Tests (cont). Ability Tests Sensory (e.g., hearing, vision) Motor (e.g., dexterity, strength, agility) ADA concerns • Reasonable accommodation • Essential job duties Cognitive (e.g., Intelligence) Wonderlic Personnel Test Cognitive Ability Tests --- Pros & Cons Pros --• Significant predictors of managerial performance (Hunter & Hunter, 1984), especially scores on verbal and numerical ability (Grimsley & Jarrett, 1973, 1975). • Data from of meta-analytic studies indicate they are some of the best predictors of performance available across an array of jobs (Bobko, Roth, & Potosky 1999; Schmidt & Hunter, 1998). Cons --- • Potential adverse impact (minorities score significantly lower these tests (Roth, Bevier, Bobko, Switzer, & Tyler, 2001) • They do not assess other types of intelligence (e.g., analytical, creative and practical intelligence (Sternberg, 1985, 2000). ~ “Big 5” Personality Factors ~ • Extraversion --- Outgoing, sociable • Neuroticism (Emotional Stability): Depressed, anxious, worrisome, etc. • Agreeableness: Flexible, forgiving • Conscientiousness: Careful, thorough, persevering • Openness to Experience: Curious, imaginative Overall, conscientiousness and extraversion are best predictors of managerial performance across jobs Personality measures add to prediction above and beyond other commonly used measure such as cognitive ability Self-Report Inventories ~ Myers-Briggs Type Indicator ~ Introversion Extraversion (Internally motivated; enjoys spending time prefers one-to-one communication) (Action-oriented; motivated enjoys alone; by outside world and social interactions) Sensing Intuitive (Desires concrete information; practical; in orientation) (Imaginative, creative factual and detailed improvises comfortable with ambiguity; focuses on contexts and connections) Thinking (Searches for facts and logic in a decision situation Focus on tasks and work to be accomplished Easily able to provide an objective and critical analysis) Judging Feeling (Sensitive to other’s needs, Seeks consensus, does not like conflict) Perceiving Plans before acting, breaks tasks down Is comfortable to acts without planning, into subsets, uses deadlines to accomplish tasks Enjoys working on multiple tasks at the same time, enjoys flexibility in work. Sample Myers-Briggs Items • • • • When you go somewhere for a day, would you rather: a) plan what you will do and when, or b) just go In a large group, do you more often: a) introduce yourself, or b) get introduced Is it harder for you to adapt to: a) routine, or b) constant change Do you think it is a worse fault to be a) unsympathetic b) unreasonable Literal ______________________ Figurative Forgive _____________________ Tolerate Impulse _____________________ Decision Thinking --- Feeling Very Clear Clear 30 25 Moderate Slight 20 15 10 5 T Slight Moderate 5 10 Clear 15 20 Very Clear 25 30 F Holly (7) Steve (10) Paul (2) Bill (5) Sally (2) Frank (1) Betty (19) Thinking • Looks at the logical consequences of actions; • Examines pros and cons of approaches objectively; • Energized by critique and analysis; • Focuses on tasks • Desires to find a standard to apply in all situations; reasonable and fair Feeling • Focuses on how others may be affected by decisions by mentally placing themselves into situations; • Guided by personal values; • Energized by appreciating and supporting others; • Focuses on interactions • Strives for harmony and positive interactions