fr md3 lecture 7

What is the science–practice gap in personnel selection?

The gap between scientific evidence about effective selection and actual selection practices.

What is the difference between information collection and information combination?

Collection obtains predictor scores; combination determines how those scores are weighted together.

What is holistic combination?

Combining information through human judgment and intuition.

What is mechanical combination?

Combining predictor scores using a predefined algorithm or rule.

Why is mechanical prediction often more accurate than holistic judgment?

Algorithms combine information consistently, whereas human judgments can be inconsistent.

What is unit weighting?

Giving every predictor the same weight when combining scores.

Why can simple algorithms work well?

Consistency often matters more than finding perfectly optimized predictor weights.

Why is evidence-based selection sometimes unpopular?

Reasons include limited knowledge, perceived costs, applicant reactions, poor communication, and reduced autonomy.

What is the perceived autonomy problem in algorithmic selection?

Decision makers may resist algorithms because they feel these reduce their control over decisions.

How can decision makers retain autonomy while using algorithms?

They can design their own algorithm or adjust its predictions through clinical synthesis.

What is clinical synthesis?

Using an algorithm's prediction as advice that a human can override.

What is mechanical synthesis?

Combining an algorithm's prediction with a human judgment using predefined weights.

How does interview training affect interview structure?

It is associated with more consistent questions, more note-taking, and more standardized evaluations.

What is the Binomial Effect Size Display (BESD)?

A table showing the change in success and failure rates associated with a predictor's validity.

In the lecture's BESD example, what success rate corresponds to r = .30?

65% for the high-predictor group and 35% for the low-predictor group.

What was the main finding of the mechanical synthesis study?

Clinical synthesis outperformed holistic prediction, and mechanical synthesis outperformed clinical synthesis.

What was the effect size for clinical synthesis versus holistic prediction?

Cohen's d = 0.50.

What was the effect size for mechanical synthesis versus clinical synthesis?

Cohen's d = 0.83.

How did retaining human judgment affect user perceptions?

It increased perceived autonomy and intentions to use the method.

Did information about optimal predictor weights improve judgment consistency?

No. It changed cue weighting but did not significantly improve consistency or validity.

What is predictive validity?

The extent to which a predictor accurately predicts a relevant future outcome.

What was the selection context in the Dutch court judge example?

Selecting candidates for the Dutch judiciary's training programme.

What is a predictive applicant design?

Predictors are measured during selection and criterion performance is measured later.

How was performance measured in the judge-selection validation study?

Trainers rated 69 items covering 11 competencies at the end of four years of training.

What was the correlation between the assessment consultant's hiring advice and training performance?

r = .02.

Why was the consultant's hiring advice weakly related to training performance?

The process relied on multiple stages of holistic judgment, introducing inconsistency and noise.

Which cognitive test had the highest reported validity in the judge-selection study?

The verbal test, with r = .33.

What was the validity of average interview competency ratings in the judge-selection study?

r = .38.

What was the validity of the sum of six interviewers' decisions?

r = .21.

What was the validity of the best LTP competency judgment?

Analytical ability and judgment, r = .22.

What is a major problem with unstructured interviews?

They lack standardized questions, rating scales, and predefined evaluation criteria.

Why can more selection predictors reduce decision quality?

Weak or redundant predictors can dilute validity and increase confidence without improving accuracy.

What is ordinary least squares (OLS) regression?

A method that estimates predictor weights to minimize squared prediction errors.

Why does adding predictors not always increase validity in practice?

With suboptimal weights, extra predictors can reduce validity when they add little unique criterion-related information.

Why must predictor scores be rescaled before combining them?

Predictors measured on different scales cannot be meaningfully added without appropriate rescaling.

How can predictor scores be standardized when sufficient data are available?

Convert scores to a common standardized scale, such as z-scores.

What can be done when samples are small and norm data are unavailable?

Transform predictors to a common scale with identical endpoints.

What is a limitation of converting scores to a common ordinal scale?

It simplifies comparison but discards information about score differences.

What was the result of the lecture's simple trichotomized predictor rule?

The combined score correlated r = .24 with performance, compared with .22 for unaided holistic judgment.

What is the overall practical lesson of Lecture 7?

Even simple, consistently applied algorithms can improve selection accuracy over unaided holistic judgment.