Exploring necessary conditions in HRM research: Fundamental issues and methodological implications

• Author: Henk Rhee,Jan Dul,Marco Guerci,Sven Hauff
• Published date: 01 January 2021
• Date: 01 January 2021
• DOI: http://doi.org/10.1111/1748-8583.12231

METHODOLOGY REVIEW

Exploring necessary conditions in HRM research:

Fundamental issues and methodological

implications

Sven Hauff

1

|Marco Guerci

2

|Jan Dul

3

|Henk van Rhee

3

1

Department of Humanities and Social

Sciences, Helmut Schmidt University,

Hamburg, Germany

2

Department of Social and Political Sciences,

Università degli Studi di Milano, Milan, Italy

3

Rotterdam School of Management, Erasmus

University, Rotterdam, The Netherlands

Correspondence

Prof. Sven Hauff, Helmut Schmidt University,

PO Box 700822, 22008 Hamburg, Germany.

Email: hauff@hsu‐hh.de

Abstract

Although the notion that HRM activities are necessary con-

ditions for achieving certain outcomes is widespread in

HRM research, most empirical studies follow an additive

sufficiency logic. That is, they analyse whether HRM activi-

ties increase an outcome, assuming that they can compen-

sate for one another. However, this does not correspond

to the idea of necessity where single HRM activities are

required for an outcome to occur and cannot be compen-

sated when absent. We discuss the differences between

sufficiency and necessity logics in terms of theory, method-

ology, and practical relevance. Also, we suggest using a new

approach and analysis technique in HRM research: neces-

sary condition analysis. We illustrate the application of the

method by analysing data on the relationship between

ability‐, motivation‐, and opportunity‐enhancing HRM prac-

tices (i.e., high‐performance work practices) and employee

performance. This illustrative example highlights that neces-

sary conditions require particular theoretical attention and

suitable empirical methods. The paper concludes by show-

ing the value of studying necessary conditions, because

such analyses allow straightforward recommendations to

be given of high relevance for HRM practice, which differ

from and add to recommendations based on additive suffi-

ciency logic and analyses.

Received: 30 June 2017 Revised: 23 November 2018 Accepted: 26 January 2019

DOI: 10.1111/1748-8583.12231

Hum Resour Manag J. 2019;1–19. © 2019 John Wiley & Sons Ltdwileyonlinelibrary.com/journal/hrmj 1

18© 2019 John Wiley & Sons Ltd Hum Resour Manag J.2021;31:18–36.wileyonlinelibrary.com/journal/hrmj

KEYWORDS

AMO framework, HPWPs, NCA, necessary conditionanalysis,

necessity, sufficiency

1|INTRODUCTION

Human resource management (HRM) research usually analyses how different HRM activities influence individual or

organisational outcomes. The respective relationships can be theorised following two different logics. In a sufficiency

logic, single HRM activities are sufficient but not necessary to increase an outcome; if one HRM activity is not in

place, the outcome may be reduced, but other HRM activities can be still effective, and may compensate for the miss-

ing activity. In contrast, a necessity logic implies that an outcome can only be achieved if a specific HRM activity is in

place. Thus, the necessary HRM activity represents a constraint, a bottleneck, or a critical factor that must be imple-

mented at the right level to allow a desired outcome level to exist. Indeed, if this necessary HRM activity is not in

place, the outcome cannot be achieved, and investments in other related HRM activities are pointless.

The notion that HRM activities are necessary for achieving specific outcomes is present in different theoretical

elaborations of HRM scholars. For example, researchers state (intentionally or unintentionally) that financial incen-

tives are necessary to motivate senior executives (Pepper, Gore, & Crossman, 2013), commitment is necessary for

low employee absenteeism (Hausknecht, Hiller, & Vance, 2008), specific knowledge and skills are necessary for inno-

vation (De Winne & Sels, 2010), fit between HRM and competitive strategy is necessary for organisational success

(Paauwe & Boselie, 2007), or more generally, that an adequate set of human resources is a necessary condition for

business survival (Boxall, 2007).

However, to date, we have very limited empirical evidence to support such necessity. This lack of knowledge

relates to the analytical strategies employed in HRM research. Indeed, regression‐based methods that are usually

applied to analyse HRM's effects refer to sufficiency logic, in terms of Y = a + b

1

·X

1

+b

2

·X

2

+b

3

·X

3

+…+ϵ.

1

In this

logic, each input (e.g., a specific HRM activity) is sufficient to increase the outcome, but not necessary: A lack of an

input reduces the outcome, but it will not prevent the outcome if other inputs (e.g., other HRM activities) compen-

sate for it. This does not correspond to necessity logic, according to which the outcome will not exist and will not be

impacted by the levels of other inputs if a single necessary input is missing or below a certain level(Goertz, 2003b).

Thus, regression‐based methodologies do not allow one to test for necessary conditions.

To date, the distinction between the sufficiency and necessity logics seems not widely acknowledged in HRM

research, which has typically treated them as interchangeable, both from a theoretical and an empirical perspective.

This is problematic, because the two logics have fundamentally different theoretical reasonings, and should be empir-

ically tested using different methods. Thus, we need more clarity on the differences between sufficiency and neces-

sity logics. Notably, this may not only help scholars in very different HRM research fields, it is also of high practical

value. Indeed, exploring sufficient relationships points out which HRM activities on average help to increase an out-

come of interest. In contrast, exploring necessary conditions can accommodate practitioners' interests in terms of

which (levels of) activities are required for achieving a specific outcome (level).

We seek to highlight that necessity logic differs from sufficiency logic and should be adequately treated in HRM

research. Thus, we first elaborate the particularities of sufficiency and necessity logics. Then, we suggest using a new

approach and analysis technique, that is, necessary condition analysis (NCA; Dul, 2016; Dul, van der Laan, & Kuik,

2018), which is able to identify necessary conditions in empirical data. We will demonstrate the differences between

the sufficiency and necessity logics with an illustrative example based on data on the relationship between high‐

performance work practices (HPWPs; Appelbaum, Bailey, Berg, & Kalleberg, 2000) and employee performance.

The illustrative example shows that NCA‐based results (which captures whether HPWPs are necessary for a certain

level of employee performance) differ from regression‐based results (which capture whether HPWPs are sufficient for

2HAUFF ET AL.

HAUFF ET AL.19