The Complexity Turn in Human Resources Theory and Research

Abstract

This study proposes and tests empirically a configural asymmetric theory of the antecedents to hospitality employee happiness-at-work and managers’ assessments of employees’ quality of work-performance. The study confirms and goes beyond prior statistical findings of small-to-medium effect sizes of happiness-performance relationships. The study merges data from surveys of employees (n = 247) and surveys completed by their managers (n = 43) and by using qualitative comparative analysis via the software program, fsQCA.com. The study analyzes data from Janfusan Fancyworld, the largest (in revenues and number of employees) tourism business group in Taiwan; Janfusan Fancyworld includes tourist hotels, amusement parks, restaurants and additional firms in related service sectors. The findings support the four principles of configural analysis and theory construction: recognize equifinality of different solutions for the same outcome; test for asymmetric solutions; test for causal asymmetric outcomes for very high versus very low happiness and work performance; and embrace complexity. Additional research in other firms and additional countries is necessary to confirm the usefulness of examining algorithms for predicting very high (low) happiness and very high (low) quality of work performance. The implications are substantial that configural theory and research will resolve perplexing happiness-performance conundrums. The study provides algorithms involving employees’ demographic characteristics and their assessments of work facet-specifics which are useful for explaining very high happiness-at-work and high quality-of-work performance (as assessed by managers)—as well as algorithms explaining very low happiness and very low quality-of-work performance. The study is the first to propose and test the principles of configural theory in the contest of hospitality frontline service employees’ happiness-at-work and managers’ assessments of these employees quality of work performances.

Relationships between variables can be non-linear with abrupt switches occurring, so the same “cause” can, in specific circumstances, produce different effects.

(“The Complexity Turn”, Urry 2005 ).

Appendix: Examples of Computing Scores for Complex Antecedent Conditions

Consider the following descriptions of five employees. Bob is very young with little education, he is unmarried with no children, he works part-time, he is a new employee, he is very happy-at-work; Bob’s manager rates Bob’s job performance to be very low.

Edwina is very young with little education, unmarried, children at home, she works full-time, three years of working in the firm, she is very unhappy-at-work; Edwina’s manager rates Edwina’s job performance to be very high.

Helen is 54 years old, married, grown children, 18 years working in the firm, working full-time, very little education, working full-time, very happy at work; Helen’s manager rates her job performance to be very high.

Linda is new to the firm, 24 years old, university graduate, married, not children, working full-time, very happy-at-work; Linda’s manager rates her performance to be acceptable but not high, “she has a long way to go but she shows promise.”

Consider the following complex antecedent conditions:

• Model D: ~age●~education●~married●~children●gender

• Model R: ~age●~education●~married●children●~gender

• Model V: ~age●education●married●~children●~gender

• With ~age = the negation of age (i.e., high score means very young);

• ~education = very low education score

• ~married = not married

• ~children = no children

• ~gender = female (thus, gender = male).

Using the logical “AND” in Boolean algebra, the membership score for the complex statement is equal to the lowest score among the scores for the simple antecedents in the complex statement. Computing the complex antecedent scores for models D, R, and V, for the four employees:

Case	Age	Education	Married	Children	Gender	D	R	V
Bob	0.01	0.01	0.01	0.01	0.99	0.99	0.01	0.01
Edwina	0.01	0.01	0.01	0.99	0.01	0.01	0.99	0.01
Helen	0.98	0.01	0.99	0.99	0.01	0.01	0.01	0.01
Linda	0.06	0.82	0.99	0.01	0.01	0.01	0.01	0.82

Bob’s score for ~age = .0.99; his score for ~education = 0.99; his score for ~married = 0.99; his score for ~children = 0.99; his score for gender = 0.99. Thus, Bob’s score for model D equals 0.99—the lowest score among the five simple antecedent conditions. Here are Linda’s scores for the simple antecedent conditions in Model V: ~age = .94; education = 0.82; married = 0.99; ~children = 0.99; ~gender =  0.99; Linda’s score for model V is equal to the lowest score among the five values (i.e., .82).

Construction of XY plots by hand is possible with the each set of scores for models D, R, and V on the X-axis and the scores for full-time, happiness, and job performance on the Y-axes. Note that full-time equals 0.00 and part-time equals 0.01; “very happy” equals 0.99 and very unhappy equals 0.01; very high performance equals 0.00 and very low performance equals 0.01. With five demographic antecedent conditions, all possible combinations include 32 models for the complex combinations