Research roundup: Political skill, emotions, and...hobos?

Time to catch up on the research:

First, Blickle, et al. show that political skill ("the capacity to understand others in working life effectively, and to apply such knowledge to induce others to act in ways that add to one's personal or organizational goals") can predict job performance beyond GMA and personality factors.

Fast on its heels comes a study from Bing et al. that also shows support for the concept of political skill (what is it, election season?). This time the researchers use meta-analysis to show a positive relationship between political skill and both task and contextual performance, with a stronger link to the latter. The relationship is also stronger as interpersonal and social requirements of the job increase.

Next, Giordano, et al. provide evidence that can help organizations detect deception in interviews (hint: reviewers out-performed interviewers, and giving a warning helps).

Fukuda, et al. show that the factor structure of two emotional intelligence measures remained consistent when translated into Japanese.

Hjemdal et al. continue gathering support for a measure of resilience (essentially the ability to successfully deal with stressors); this time in a French-speaking Belgian sample.

Interested in the Bookmark method of setting cut scores? Then you'll want to take a look at this study by Davis-Becker et al. and its slightly disturbing results.

The measurement of reading comprehension is common, and it's important to understand the construct(s) being measured. This piece by Svetina, et al. helps us parse out the concept.

The discussion about the March issue of the Journal of Personality and Social Psychology was dominated by Bem's study of precognition, but a hidden gem is there to be found: Hopwood et al.'s study of how personality changes over one's lifespan. Looks like it's largely due to environmental influences and the most change occurs early on.

Interested in PO fit and its relationship with attitudinal outcomes? Then you'll want to check out Leung & Chaturvedi's study.

Last but not least, Woo analyzes a sample of U.S. workers and finds support for "Ghiselli's hobo syndrome" among a small group: basically these folks frequently change jobs and have positive attitudes about it. Not surprisingly, these hobos reported being less satisfied with their current jobs. I guess there's just no pleasing some people!

Side note: I just ran across the Questionmark blog. A lot of really cool stuff that focuses on education and learning applications, but a lot of the content is directly applicable to personnel assessment. They're a vendor, so there are some sales pitches, but it's worth weeding through.

Final note: I apologize ahead of time if some of these links don't work; I've tried my best to link to non-session based abstracts but it can be hard to tell.

New issues of Personnel Psych and IJSA: Muslim applicants, selection perceptions, and more

Two of the big journals have come out with new issues, so let's take a look at the highlights:

First, in the Winter 2010 issue of Personnel Psychology:

King and Ahmad describe the results of several experiments that found interviewers and raters altered their behavior for confederates exhibiting obvious Muslim-identified behavior (e.g., clothing) depending on whether applicants exhibited stereotype-inconsistent behavior. For those that didn't, reactions were shorter and more negative. On the other hand, no difference was found in offers between those dressed in Muslim-identified clothing and those that weren't. So behavior--specifically its stereotypicality--and not simply something obvious like dress, may be key in predicting/preventing discriminatory behavior.

Do you consistently read the "Limitations" section of journal articles? Brutus et al. did, for three major I/O journals from 1995 to 2008 and found that threats to internal validity were the most commonly reported limitation. Interestingly, they also found that the nature of limitations reported changed over time (e.g., more sampling issues due to volunteers, variance issues). You can see an in press version here.

Next up, Henderson reports impressive criterion-related validity (combined operational = .86) for a test battery consisting of a g-saturated exam and a strength/endurance exam after following a firefighter academy class for 23 years. He suggests that employers have considerable latitude in choosing exams as long as they are highly loaded on these two factors, and also suggests approximately equal weighting.

Struggling to communicate the utility of sound assessment? Last but not least, Winkler et al. describe the results of sharing utility information with a sample of managers and found that using a casual chain analysis--rather than simply a single attribute--increased understanding, perceived usefulness, and intent to use.


Let's switch now to the December 2010 issue of the International Journal of Selection and Assessment (IJSA). There's a lot of great content in this issue, so take a deep breath:

Interested in unproctored internet testing? You'll want to check out Guo and Drasgow's piece on verification testing. The authors recommend using a Z-test over the likelihood ratio test for detecting cheating, although both did very well.

Walsh et al. discuss the moderating effect that cultural practices (performance orientation, uncertainty avoidance) have on selection fairness perceptions.

Speaking of selection perceptions, Oostrom et al. found that individual differences play a role in determining how applicants respond--particularly openness to experience. The authors recommend considering the nature of your applicant pool before implementing programs to improve perceptions of the assessment process.

Those of you interested in cut scores (and hey, who isn't) should check out Hoffman et al.'s piece on using a difficulty-anchored rating scale and the impact it has on SME judgments.

Back to perceptions for a second, Furnham and Chamorro-Premuzic asked a sample of students to rate seventeen different assessment methods for their accuracy and fairness. Not surprisingly, panel interviews and references came out on top in terms of fairness, while those that looked the most like a traditional test (e.g., drug, job knowledge, intelligence) were judged least accurate and fair. Interestingly, self-assessed intelligence moderated the perceptions (hey, if I think I'm smart I might not mind intelligence tests!).

And now for something completely different (those of you that get that reference click here for a trip down memory lane). A study by Garcia-Izquierdo et al. of information contained in online job application forms from a sample of companies found on the Spanish Stock Exchange. A surprisingly high percentage of firms asked for information on their applications that at best would be off-putting, at worst could lead to lawsuits, such as age/DOB, nationality, and marital status. The authors suggest this area of e-recruitment is ripe for scientist-practitioner collaboration.

Last but not least, a piece that ties the major topics of this post together: selection perceptions and recruitment. Schreurs et al. gathered data from 340 entry-level applicants to a large financial services firm and found that applicant perceptions, particularly of warmth/respect, mediated the relationship between expectations and attraction/pursuit intentions. This reinforces other research that has underlined the importance of making sure organizational representatives put your best foot/face forward.

Lewis case emphasizes need for valid tests

On Monday, May 24, the U.S. Supreme Court ruled in Lewis v. City of Chicago that plaintiffs filing an adverse impact discrimination claim under Title VII of the Civil Rights Act have 300 days to file from each time the test is used to fill a position--not just 300 days from when the test was administered.

For the most part, this mainly impacts employers that run an exam and use the results to fill positions for several years. In this case it was a firefighter entry-level exam, and my guess is it will mostly be public sector agencies and large employers that should pay particular attention to the ruling.

Why? Well for one it means more potential adverse impact lawsuits. If you were counting on being safe after 300 days from the exam, that is no longer the case. Second, it emphasizes the need to follow professional guidelines when developing an exam. Employers can successfully defend against an adverse impact case by showing that the selection practice is "job related for the position in question and consistent with business necessity..." (and that no alternatives with similar validity with less adverse impact were available). This means your exams need to be developed and interpreted by people who know what they're doing.

The City claimed that evidence related to an employer's business necessity defense might be unavailable by the time the lawsuit is brought--the court was not swayed. This means you'll want to hang on to your exam development records for at least a year beyond the last time you use the results to fill a position.

Another point worth noting: this case boils down to the validity of a cut score used by the City, which they themselves admitted wasn't supportable. Proper exam development and interpretation includes setting a job-related pass point based on subject matter input and/or statistical evidence that it is linked to job performance.

The Lewis ruling doesn't fundamentally change what we should be doing. It just emphasizes that we need to do it right.

Click here for a quick overview of the facts of the case.

Lessons from NYC Fire case - part 2

Part 2 of 2

Last time I discussed five important lessons we can take away from recent rulings in the Vulcan v. City of New York case. In this post I'll review the remaining lessons and also discuss the relief order.

----

6) The city failed to provide sufficient evidence that the exam(s) tested for a sufficient number of the critical KSAs. They also failed to explain why they chose not to measure several KSAs identified as critical.

Lesson: the courts do not require employers to measure every single critical KSA. But there is an expectation that employers attempt to measure a sufficient number that represent a significant portion of the job requirements. In this case, that included non-cognitive abilities such as resistance to stress, teamwork, and conscientiousness, that were not measured.

7) The city failed to adequately consider how to measure a significant number of essential KSAs. While some of their concerns were valid (e.g., structured interviews for all applicants would be an operational nightmare), there are many different forms of testing that should have been considered, including situational judgment tests (SJTs) and biodata, which can be used to measure non-cognitive components.

Lesson: triers of fact expect employers to be up on the various assessment methods available and be able to explain why they chose not to use certain ones. This includes tests that are relatively easy to develop (e.g., SJTs) as well as ones that require substantial resources and statistical expertise (e.g., biodata).

8) The city failed to conduct a reading level analysis on the exams to ensure that it was not "pointlessly high." The plaintiff introduced evidence suggesting the reading level was above 12th grade; in addition, it appeared to exceed the reading level of materials at the academy.

Lesson: never forget that every assessment method is in some sense measuring additional KSAs beyond those you intend. For written exams, reading comprehension is always a requirement (barring accommodation). It's quite easy to conduct a reading level analysis (MS Word has it built in) to ensure that the level is reasonable and matches other job-related material.

9) The city failed to show that the cutoff scores (pass points) established for the exams were based on adequate rationale, namely "the necessary qualifications for the job of entry-level firefighter." Instead, the cutoff scores were based on operational need (the number of job openings expected). This is particularly important in multiple-hurdle selection processes such as in this case, where a failure on one exam component precludes an applicant from participating in the rest of the (potentially compensatory) assessment process.

Lesson: ultimately applicants have to pass the test(s) to be considered for employment. Cutoff scores should be established using the expertise of both SMEs and test developers and should be based on the minimum competency levels required upon entry to the job. At a minimum (and I would not rely solely upon this), the scores should be analyzed to identify any logical "break-points."

----

After ruling for the plaintiffs on both the adverse impact and disparate treatment claims, the judge issued a relief order on 1/10/10. In it, he imposes several things, including the following:

1) The city must develop a new testing procedure for entry-level firefighter in conjunction with the relevant parties. Following the development of the test, there will be a hearing to determine if this test should be used rather than the current test (developed in 2007 and not at issue in this litigation).

2) The court shall develop a process by which the approximately 7,400 applicants covered by this case can file a claim for monetary relief.

3) The city will identify 293 black candidates on the eligibility list and offer them priority hiring. (No quotas are being imposed, although the judge leaves this possibility open)

4) Retroactive seniority for those hired.

In addition, several other issues are up for debate, including the appointment of a special master or monitor, standards that will be relied upon in constructing the new exam, and the need for additional relief.

---

So what did we learn from all this? If you follow--fairly closely--best practices when developing and administering exams, you will be on solid ground defending them. If you don't, and your exam has a discriminatory effect, you may be called on it--and it's not a pleasant process. I'll leave you with this quote from the January ruling on disparate treatment:

"The history of the City's efforts to remedy its discriminatory firefighter hiring policies can be summarized as follows: 34 years of intransigence and deliberate indifference, bookeneded by identical judicial declarations that the City's hiring policies are illegal."

Setting cutoff scores on personality tests

What's the best way to set a cutoff score for a personality test, knowing that some candidates inflate their score? It all depends on your goal. Are you trying to maximize validity or minimize the impact of inflation?

According to a research study by Berry & Sackett published in the Winter '09 issue of Personnel Psychology, if your goal is to maximize validity, your best bet is to wait until applicants have taken the exam, then set your cut-score (e.g., the top two-thirds); this was particularly true when selection ratios are small (i.e., organization is very selective).

If your goal is to minimize the number of deserving applicants who are displaced by "fakers", you're better off establishing the cut point ahead of time, by using a non-applicant derived sample (e.g., job incumbents, research group). The results were generated using a Monte Carlo simulation.

Interestingly, the authors also replicated the work of other researchers who have shown that the impact of faking on the criterion-related validity of personality measures is relatively low. There are a few other very good points made in this article:

- Expert judgment methods of establishing pass points (e.g., Angoff method) may be difficult to use for personality tests since experts may find it difficult to judge individual items. Methods used to select a certain number of applicants or methods based on a criterion-related validity study (both used as variables in this study) are more appropriate for personality tests.

- There is no consensus of how prevalent faking on personality exams is; estimates range from 5-71%. It likely depends on the situation and how motivated test takers are to engage in impression management.

- Some recommend setting a very low cutoff score for personality tests, which would exclude only those likely not suitable for the position (and not faking), while others prefer a more stringent cutoff to maximize utility.

- A reasonable range of d-values for score inflation on personality inventories is .5-1.0 (used in this study).

- There exists very little research on the skewness of faking score increases. A positively-skewed distribution (meaning most people faked a small amount) was used in this study. (I would think this would also vary on the situation)

So bottom line: where--and how--you set your cutoff score on personality inventories depends on whether you want to maximize the predictive validity or minimize the number of deserving applicants that get left out of the process.

Other good reads in this issue:

- Police officer applicants reactions to promotional assessment methods

- The impact of diversity climate on retail store sales

- The construct validity of multisource performance ratings

- Labor market influences on CEO compensation