In Defense of Professional Failure

We don’t often discuss failure. We’re not supposed to fail, and we pretend that we don’t fail even when it may take multiple attempts to succeed at a task. When someone else admits to failing, it threatens this well-crafted lie that failure doesn’t happen. We do students and early-career professionals a disservice when we don’t discuss past failures and what is possible going forward. Mentees should know that not all interventions are successful. Not all courses work exactly as we planned. Not all teaching evaluations will be positive. Not all job applications produce a job offer. Not all manuscripts are accepted for publication. Not all grants are funded. This might be knowledge that everyone has, but it’s difficult not to be disappointed in the moment when these situations happen to us.

Many (graduate) students will at some point experience imposter syndrome, which occurs when typically high-achieving people doubt their own skills and success and are afraid of being discovered as fraud. According to Psychology Today, around 70% of adults may experience imposter syndrome. It is more common in women of color and people who struggle with self-efficacy, perfectionism, and neuroticism (refer to Flora, 2016, November 1). I remember feeling like an imposter in graduate school and hearing many other graduate students then and now express similar feelings. Most of us struggled with perfectionism (i.e., If you [aren’t] first, you’re last), and mentors would generally express bewilderment and quickly reassure us that we belonged. This, unfortunately, didn’t help much with imposter syndrome. Hendriksen (2017, August 8) described several ways in which imposter syndrome manifests and how to overcome it. We will focus on normalizing failure and supporting mentees/colleagues.

Fail Publicly

One way that people have acknowledged failure is to publish satirical papers about it. For example, many researchers experience the pressure to publish (or perish). It’s an open secret that many writers will encounter writer’s block and have difficulty starting a manuscript. Upper (1974) initially wrote about this frustration. Later, Didden et al. (2007) and Brodhead et al. (2019) expressed similar difficulties, and McLean and Thomas (2014) and Ampatzidis (2021) pointed out that this wasn’t a phenomenon specific to behavior analysis. With the exception of McLean and Thomas (2014), these publications are blank pages as a testament to the manuscripts that haven’t been written. These manuscripts may not be regularly cited, but they are shared widely on social media (e.g., Oransky, 2018, October 3). I think these articles have been so well-received because the authors are accomplished researchers, and it is rare that people who are successful are ever open about their failures with the larger community.

Stefan (2010) kept a record of her failures to acknowledge her work, accept her failures, and contextualize her successes. Curriculum vitae are complete histories of educational, teaching, and research experience. They are lists of successful achievements without any hint of the universities the author wasn’t admitted to, the courses they weren’t selected to teach, the jobs/fellowships they didn’t get, the publications/posters/talks that weren’t accepted, or the awards/scholarships/grants that weren’t awarded. Since her article was posted, several researchers have posted their cv of failure: Johannes Haushofer was a professor of psychology and public affairs at Princeton University and is now at Stockholm University and the National University of Singapore (see also Hrala, 2017, December 25), Veronika Cheplygina is an associate professor researching machine learning at the IT University of Copenhagen, David Smith is a professor of medicine at the University of California San Diego, and Kiran Tomlinson is a PhD candidate in computer science at Cornell University. Everyone should keep a cv or résumé of failures to understand that we’re not going to complete every task perfectly on the first attempt and that we can learn from these early attempts how to be successful (Herrera, 2019, February 3). We don’t necessarily need to share these documents with everyone, but the more that we accept that learning is a part of the process and tell others about it, the better off we will be.

Trial-and-Error Learning and Perseverance

We have previously written about how to support students in the classroom through trial-and-error learning (and errorless learning; refer to Swisher and Cihon, 2020, May 14), but these early attempts at completing professional tasks are part of trial-and-error learning. Shaping responses for errorless learning is easier when the task is simple (e.g., discriminate between a circle and a triangle; Schilmoeller et al., 1979) and encourages proficient performance early on, but job-related tasks are often complex. Shaping can be time-intensive for the teacher and the learner even with simple tasks, and it would not be an efficient use of supervisor or supervisee time to shape every response class. Learners will make more independent responses with error correction (or trial-and-error) procedures than with errorless learning procedures (e.g., to teach learners to match the names of comic book characters to their pictures; Leaf et al., 2020). While many of those responses will be accurate – once the task is learned, there will also be more errors than with errorless learning. When trainers are less tolerant of a trainee’s errors in the early stages of learning, participants at vocational schools indicated on a survey that they would feel bad and wouldn’t be motivated to try again (Zhao et al., 2018). Trainees’ affective-motivational responses were not affected if the trainer was more accepting of errors. At the very least, if supervisors use trial-and-error learning and aren’t normalizing errors, they should not punish early failed attempts.

Another way to conceptualize failure is when responses are not reinforced. In errorless learning procedures, every response is reinforced in acquisition; this is a continuous reinforcement schedule. In error correction (or trial-and-error) learning, only some responses are reinforced in acquisition; this a partial reinforcement schedule. The typical finding is that responding will persist longer in extinction after exposure to a partial reinforcement schedule than after exposure to a continuous reinforcement schedule (e.g., Thrailkill, 2023). Often, students who are described as having grit or resilience are praised for their ability to continue even though the task is hard/their efforts are executed under effectively extinction-based schedules. It’s not that these people (or rats; Harris et al., 2019) are less sensitive to failure, but they have learned that they’re still capable of succeeding (i.e., earning reinforcers) even when they’ve failed (e.g., rats with leaner partial reinforcement schedules persisted longer in extinction than rats with richer partial reinforcement schedules; Chan & Harris, 2019). That is, individuals have become unreasonably stubborn and keep trying until they succeed (or perceive that they won’t succeed).

Problems with Hiding Failures

Some of the issues regarding the replication crisis (or reproduction or reproducibility crisis; refer to Forbes et al., 2023) come from the file drawer effect (or publication bias; refer to Rosenthal, 1979). Journal editors are reluctant to publish manuscripts about failures to find an effect – even when the science is sound. Furthermore, journal editors are unlikely to publish replications of studies because they’re not considered original research – whether those replications confirm or disconfirm the original findings. When researchers find no effect, they often literally store the results in a filing cabinet or electronically in the laboratory. Very rarely do researchers publicly present these “failures” so that other researchers are aware of them. Some present talks or posters at conferences, but these records aren’t permanently available for researchers who want to know the actual effect size (e.g., Funder & Ozer, 2019) of some intervention for their own purposes or when conducting meta-analyses. Publishing preprints of manuscripts, publishing multiple studies (including those with no effect) in a mini meta-analysis, and publishing in journals that accept null results (i.e., no improvement after treatment) are two ways to correct for publication bias and make failures accessible. The All Results Journals accept manuscripts based on preregistration of methods and planned analyses regardless of the results. The following journals accept manuscripts with null results: Positively Negative, Journal of Articles in Support of the Null Hypothesis, Journal of Negative Results in BioMedicine, Journal of Trial & Error, and The Missing Pieces: A collection of Negative; Null and Inconclusive Results.

This isn’t to say that we shouldn’t acknowledge failure in other ways, but the current most public examples of normalizing failure are in publishing practices. If we have more honest conversations with students and colleagues, then perhaps we’ll support more professionals in their efforts to succeed. I know that I appreciate these conversations, and my undergraduate students are relieved to hear that they don’t need to be perfect from the start.

Image credit

[1] Image provided courtesy of fauxels under Pexels license

References

Ampatzidis, G. (2021, November 24). The unsuccessful self-treatment of a case of ‘writer’s block’: A replication in science education. Journal of Trial & Error, 2(1). https://doi.org/10.36850/e7 

Brodhead, M. T., Truckenmiller, A. J., Cox, D. J., Della Sala, M. R., Yough, M., & Hartzheim, D. U. (2019). A multidisciplinary replication of Upper’s (1974) unsuccessful self-treatment of writer’s block. Behavior Analysis in Practice, 12, 547. https://doi.org/10.1007/s40617-018-00290-w 

Chan, C. K. J., & Harris, J. A. (2019). The partial reinforcement extinction effect: The proportion of trials reinforced during conditioning predicts the number of trials to extinction. Journal of Experimental Psychology: Animal Learning and Cognition, 45(1), 43-58. http://dx.doi.org/10.1037/xan0000190 

Didden, R., Sigafoos, J., O’Reilly, M. F., Lancioni, G. E., & Sturmey, P. (2007). A multisite cross-cultural replication of Upper’s (1974) unsuccessful self-treatment of writer’s block. Journal of Applied Behavior Analysis, 40(4), 773. https://doi.org/10.1901/jaba.2007.773 

Forbes, H. J., Travers, J. C., & Johnson, J. V. (2023). Supporting the replication of your research. In D. J. Cox, N. Y. Syed, M. T. Brodhead, & S. P. Quigley (Eds.) Research ethics in behavior analysis: From laboratory to clinic and classroom (pp. 237-262). Academic Press. https://doi.org/10.1016/B978-0-323-90969-3.00003-7

Funder D. C., & Ozer, D. J. (2019). Evaluating effect size in psychological research: Sense and nonsense. Advances in Methods and Practices in Psychological Science, 2(2),156-168. https://doi.org/10.1177/2515245919847202 

Harris, J. A., Seet, M. S., & Kwok, D. W. S. (2019). The partial-reinforcement extinction effect does not result from reduced sensitivity to nonreinforcement. Journal of Experimental Psychology: Animal Learning and Cognition, 45(2), 185-202. http://dx.doi.org/10.1037/xan0000200 

Leaf, J. B., Cihon, J. H., Ferguson, J. L., Milne, C. M., Leaf, R., & McEachin, J. (2020). Comparing error correction to errorless learning: A randomized clinical trial. The Analysis of Verbal Behavior, 36, 1-20. https://doi.org/10.1007/s40616-019-00124-y 

McLean, D., & Thomas, B. R. (2014). Unsuccessful treatments of “writer’s block”: A meta-analysis. Psychological Reports, 115, 276-278. http://dx.doi.org/10.2466/28.PR0.115c12z0 

Rosenthal, R. (1979). The file drawer problem and tolerance for null results. Psychological Bulletin, 86(3), 638-641. https://psycnet.apa.org/doi/10.1037/0033-2909.86.3.638 

Schilmoeller, G. L., Schilmoeller, K. J., Etzel, B. C., & LeBlanc, J. M. (1979). Conditional discrimination after errorless and trial-and-error training. Journal of the Experimental Analysis of Behavior, 31, 405-420. https://doi.org/10.1901/jeab.1979.31-405 

Stefan, M. (2010). A CV of failures. Nature, 468, 467. https://doi.org/10.1038/nj7322-467a 

Thrailkill, E. A. (2023). Partial reinforcement extinction and omission effects in the elimination and recovery of discriminated operant behavior. Journal of Experimental Psychology: Animal Learning and Cognition, 49(3), 194-207. https://doi.org/10.1037/xan0000354 

Upper, D. (1974). The unsuccessful self-treatment of a case of “writer’s block.” Journal of Applied Behavior Analysis, 7(3), 497. https://doi.org/10.1901/jaba.1974.7-497a 

Zhao, B., Seifried, J., & Sieweke, J. (2018). Trainers’ responses to errors matters in trainees’ learning from errors: Evidence from two studies. Journal of Managerial Psychology, 33(3), 279-296. https://doi.org/10.1108/JMP-10-2017-0364

Blog post contributed by Melissa Swisher