A fatal flaw: Positive leadership style research creates causal illusions

This was a fascinating read, arguing and evidencing that positive leadership styles aren’t representations of leadership behaviours but rather “subjective evaluations of leaders”.

They empirically tested this idea over four studies.

Providing background:

·???????? While valid constructs are necessary in science – they need to be defined, measurable and causally linked to other constructs. However, “many behavioral constructs do not meet these criteria”

·???????? One systematic review of leadership and organisational behaviour research found that just 3% of variables “in these supposedly behavioral sciences capture the observable behaviors of individuals or groups” and rather, they capture “perceptions and evaluations”

·???????? They argue that “Ironically then, behavioral scientists rarely study behaviors, which is due to both improper conceptualization and incorrect measurement”

·???????? A problem of this popular stream of behavioural research is that they conflate leadership behaviour with evaluations of these behaviours, and hence “leading to causal illusions”

·???????? Further, they state that via their four empirical studies, they find support that “positive leadership styles lead to positive outcomes […] might be an artifact of conflation rather than a reflection of reality”

·???????? These validated constructs don’t validate effects of leader behaviours but “validate” a construct that conflates leader behaviors and evaluations of those behaviors”

·???????? Leadership styles are defined as patterns/characteristics/modes of behaviours, like authentic, ethical or servant. But the pattern of behaviours isn’t the same as behaviours per se, but a characterisation of a set of behaviours with a common theme

·???????? As such, leadership styles is a characterisation or judgement that carries an evaluative component; said differently, “Behaviors are objective, whereas evaluations are subjective”

·???????? They propose three sources of the evaluative component in leadership styles—

·???????? Positive valence – saying that a person leads authentically or ethically “carries an extremely positive valence”, types of moral judgements requiring an evaluation

·???????? Nonbehavioral leader features: Evaluations rest not just on observed behaviours but other observable leader factors, like some research showing that a leader’s facial appearance affecting observers evaluation of that leader. Knowledge about a leader’s previous performance also shapes evaluations of leadership style, and thus “evaluations of leader behaviors are not merely judgments about the observed behaviors ... [but] holistic assessments of the leader as a person”

·???????? Evaluator attributes: This facet is about evaluator-specific factors influencing their judgements of leaders

I can’t do this justice, so recommend you read the full paper. I’ve also skipped most of the explanatory background and all of the specific findings, and gone straight to the discussion.

Results

Key findings:

·???????? “positive leadership styles are outcomes that depend on non-behavioral, evaluative factors, such as information about a leader’s previous success or value alignment between leaders and followers”

·???????? Measures of leadership styles create “causal illusions by spuriously predicting objective outcomes, even when leader behaviors and other leader-specific factors are kept constant”

·???????? Their findings “cast serious doubts on previous research claiming that positive leadership styles cause positive outcomes”

·???????? And “positive leadership style research is not only wrong but also practically futile because its constructs and measures are amalgams that do not isolate concrete and learnable behaviors”.

Unpacking the findings, it’s highlighted that positive leadership styles are “conflated constructs that might be partially behavioral, but in large part represent positive summary evaluations of leader behaviors and other leader properties”.

Said differently, leadership styles are likely affected by leader behaviours, but whether leaders have an ethical, servant etc. style “is not an objective leader property but a subjective evaluation through the eye of the beholder”.

Further, positive leadership styles are likely to be “mixed leader-rated constructs”.

Next they argue that if leadership is in the eye of the beholder, then evaluating observer ratings is “simply the correct measured choice”. They disagree, because subjective evaluations of leadership styles are outcomes themselves and “using these styles as independent variables to predict other outcomes is a futile exercise that can only produce causal illusions”.

The causal illusions are said to stem from:

1. misinterpreting leadership styles as purely behavioural constructs

2. ignoring other variables, e.g. info about a leader’s past performance, leader-follower value alignment and more.

They argue that their findings are somewhat sobering, pointing to “fundamental conceptual and methodological weaknesses in past research” and “thus knocking past wisdom about positive leadership styles off its foundation”.

Perhaps more importantly, though, these findings provide a starting point for advancing future research in this area (which they discuss but I’ve skipped); but the short of it is: research needs to separate behaviours and their evaluations.

Finally they provide some other points, like:

1. Positive leadership styles like authentic, ethical or servant leadership “lack a solid empirical foundation, rendering these claims speculative”

2. “leadership styles are not leader behaviors per se but rather a mix of what leaders do and how followers evaluate leadership”

3. these findings “warn practitioners that lots of evidence on the effectiveness of positive leadership styles is likely driven by causal illusions and thus unwarranted”

As expected, several limitations are present in this work.

Authors: Fischer, T., Dietz, J., & Antonakis, J. (2024). A fatal flaw: Positive leadership style research creates causal illusions. The Leadership Quarterly, 101771.