The Criterion Choices Dr. Steve Training & Development INP6325 Why Evaluate Training? What can be gained from evaluating training? 1. Determine effectiveness of program. 2. Demonstrate benefits of training to top management and stakeholders 3. Demonstrate job-relatedness of training (legal implications) 4. Research value in aiding future training development 5. Ability to make personnel decisions (promotion, retention, etc.) Why not evaluate training? In a survey of 611 orgs, 92% claim to evaluate training programs, however the vast majority of these were simply trainee reaction, rather than learning or transfer What’s preventing them? 1. 2. 3. 4. Evaluation not often emphasized by management Training directors do not know how HR may not understand importance View that evaluation is expensive and risky Determining Criteria To determine criteria, must know purpose of evaluation: To predict job success, must evaluate training for relationship between training performance and job performance Must determine whether one training program is better than another or no formal training at all. Pessimistic view of criteria selection Guion (1961) 1. 2. 3. 4. 5. 6. 7. I/O psychologist has a hunch about a problem Read a vague, ambiguous job description Form a fuzzy concept of the ultimate criterion Develop a set of measures that can be combined to approximate ultimate criterion Judge relevance of measure: deficiency & contamination Data required for the measure are not available in company personnel files (and never will be) Select best available criterion Evaluating training effectiveness To assess effectiveness of training, must: 1. 2. 3. Develop criteria that assesses trainees’ learning of the KSAs necessary for the job Assess performance at the end of training (progress) Assess performance after period of time on job (transfer) Criterion Selection Ultimate Criterion – better time management B A D C A – KSAs not in Needs Assessment or Criterion B – Criterion Deficiency - KSAs in Needs Assessment, but not Criterion C – Criterion Contamination (error + bias) - KSAs in Criterion, but not Needs Assessment D – Criterion Relevance - KSAs in both Needs Assessment & Criterion Actual Criterion - files reports on time - Avoids overtime - Meets deadlines Criterion Deficiency Criterion Deficiency – training program intended to teach certain KSAs required for the job, but criteria used to evaluate training are missing KSAs - Sort Mail - Weigh pkg - K of pricing Example: Postal Clerk - mail sorting skill not part of the training Criterion Contamination Criterion Contamination – extraneous variables included in the criteria that were not part of training program Opportunity bias – some individuals might have a greater opportunity for successful job performance which had nothing to do with training salesmanship - KSAs Geographic location Example: Salesperson - performance affected by location (opportunity) as much as by training Criterion Reliability Criterion Reliability – consistency of criterion measure. Example: inter-rater reliability of supervisory ratings Negatively impacted by: Competence of judges Simplicity of behaviors Overtness of behaviors Operational definition of behaviors Two Views of Criterion Development Composite Criteria Mastery of needs assessment A+B+C=X Grade: A if X > 90% Problem: not very diagnostic May not know where one was successful or failed Multiple Criteria Mastery of needs assessment A Skill at writing task statements B Knowledge of task analysis techniques C Skill at presenting Results Problem: may meet some criteria, but not others. What constitutes success? Proximal vs. Distal Criteria Proximal – short term criteria Distal – Long term criteria Example: Political training Proximal: performance during campaign Distal: performance in office USA TODAY/CNN/Gallup Poll results Below are the results of a USA TODAY/CNN/Gallup Poll: 1. Question: Do you approve or disapprove of the way George W. Bush is handling his job as president? Poll of National Adults Approve Disapprove No Opinion 9/22/2001 90 8 2 9/22/2002 66 30 4 9/22/2004 48 49 2 9/22/2006 42 55 3 Levels of Criteria Proximal 1. Reaction – opinion of trainees (survey) 2. Learning – mastery of training material (test) 3. Distal Immediate knowledge Knowledge retention Behavior/Skill demonstration Behavioral – trainee job performance (ratings) 4. Affective reactions Utility judgments Transfer Results – org profit by training ($$$$) Reaction Criteria Guidelines for Reaction Criteria Development a. b. c. d. e. Questions based on information from needs assessment Q’naire includes quantifiable data (do not use ONLY open-ended questions) Q’naire should be anonymous Should include SOME open-ended questions Pilot test Q’naire for length and comprehension Benefit: provides info from all trainees, not just those with extreme opinions Disadvantage: may have nothing to do w/ eventual performance, but if training is perceived as poor it is less likely to be taken seriously or skills retained Reaction Criteria Example Strongly Disagree Disagree Neither Agree nor Disagree Agree Strongly Agree 1. The objectives of this program were clear 1 2 3 4 5 2. The instructor was helpful and contributed to the learning experience 1 2 3 4 5 3. There was an appropriate balance between lecture, participant involvement, and exercises in the program 1 2 3 4 5 4. The topics covered in this program were relevant to the things I do on my job 1 2 3 4 5 5. I can see myself performing more effectively after attending this program 1 2 3 4 5 6. The logistics for this program (e.g., arrangements, food, equipment) were satisfactory 1 2 3 4 5 Learning Criteria Test at end of training addressing material covered during training Example: Can receptionist trainee recall the steps to transfer a phone call on the company’s phone system? Pre-test/Post-test comparison Behavior Criteria Criteria should come directly from Task and KSA analyses. Use experimental methods to demonstrate improvements due to training. Assess whether performance goals are met. Example: Bus Driver performance: Stops and restarts without rolling back Tests brakes at tops of hills Uses mirrors to check traffic Signals following traffic Stops before crossing sidewalk when coming out of driveway Stops clear of pedestrian crosswalks Results Criteria Measure of training program in terms of meeting organizational goals Money saved = lower turnover, lower absenteeism, improved morale, improved productivity, etc. Utility – Whether training saves more money than it costs Ex: if no formal training in place, senior workers lose time showing junior workers what to do. Utility Analysis U = (T x N x dt x Sdy) – (N x C) U = $ value of training program (Value) – (Cost) T = # years duration of training effect on performance N = # trainees dt = effect size or true difference in pre-post perf = (Xc – Xe) /ryy Sdy = std deviation of performance in $ of untrained group C = cost per trainee Example: U = (2yrs x 100N x .5 x $5,000) – (100N x $200) = $500,000 - $20,000 = $480,000 over 2 yrs Utility Analysis Reasons for NOT using utility analysis Data may not be readily available or unreliable Seeking non-monetary benefits of training Other variables confound results Choosing Criteria Conclusions: Reaction information is important to know for whether trainees will accept training program, but it does not translate into effectiveness in terms of learning, transfer, or monetary savings Learning criteria is a good predictor of results Learning criteria is a modest predictor of transfer Behavior criteria is a modest predictor of results Other Criteria Concerns Can’t judge effectiveness of training strictly on outcome (summative), must also look at process (formative) Process measures tell us source of outcome changes Outcome alone is not diagnostic Importance variables affecting process include: differences between trainers, settings, student samples, motivation of groups, etc. Subjective vs. Objective Measures Subjective – ratings, opinions Problem of rater biases (halo, central tendency, leniency) Easy to use Was rater well-trained to make ratings? Subjective ratings may be improved by training the rater Objective – countable, observable measures such as production, absences, defects missed, etc. Problem of opportunity bias Objective measures should be used when possible, but must be aware of potential contamination and account for it