Do Current HRIS Meet the Requirements of HRM?:

An Empirical Evaluation using Logistic Regression

and Neural Network Analysis

Stefan Strohmeier

and Rüdiger Kabst

Chair of Management Information Systems

Saarland University, Am Stadtwald 61

66123 Saarbrücken, Germany

Chair of Human Resource Management

Justus-Liebig University Gießen, Licher Str. 66

35394 Gießen, Germany

Abstract. Our paper examines the question whether major features of current

HRIS actually meet requirements of HRM. To do so, we initially identify major

features of current HRIS, discuss why these features may or may not meet re-

quirements of HRM and derive corresponding hypotheses. Subsequently, we

employ an international large-scale survey to test these hypotheses by combin-

ing logistic regression and neural network analysis. Our results draw a rather

positive picture of HRIS: If equipped with the right functionality and delivery

features current HRIS are able to meet requirements of HRM.

1 Introduction

The usage of Information Technology (IT) in Human Resource Management (HRM)

dates back to at least the 1950s [47]. Since that time, numerous innovations have

considerably changed the features of the respective Human Resource Information

Systems (HRIS) [1, 4, 9, 18, 23, 34, 35]. At present, conceptions like virtual or elec-

tronic HRM denote the latest phase of HRIS usage [31, 39, 43].

However, describing current HRIS is a rather complex task. Likely, current HRIS

are best understood as configurations of different interacting systems that aim at the

generation and delivery of HR functionality in order to automate and informate

HRM. The configurational aspect emphasises that current HRIS usually comprise of

several different interacting systems. Roughly categorized, back-end systems aim at

the generation of HR functionality, while corresponding front-end systems undertake

the delivery of these functionalities to different categories of end-users [43].

In the interim, numerous back-end systems for HRM are offered by a bulk of spe-

cialised vendor organisations. With solitary HR applications, integrated HR applica-

tions (that provide a broader range of HR functionalities) and/or enterprise wide ap-

Strohmeier S. and Kabst R. (2007).

Do Current HRIS Meet the Requirements of HRM?: An Empirical Evaluation using Logistic Regression and Neural Network Analysis.

In Proceedings of the 1st International Workshop on Human Resource Information Systems, pages 31-44

DOI: 10.5220/0002418500310044

 SciTePress

plications (that integrates HR functionalities additionally with other organisational

functions such as production, sales or finance) there are three general varieties of

back-end systems [1, 11, 43].

Front-end systems then have to deliver the generated functionality of the back end-

systems to different categories of end-users. For this purpose increasingly web-

browsers can be used as general user interfaces of HR back-end systems [23, 30].

Combined with the delivery potentials of web-browsers, portal-technology addition-

ally enables a simple, unified and personalized access to all the heterogeneous back-

end systems and services any user group may need [44]. As a consequence, by com-

bining different back-end and front-end systems a multitude of different HRIS con-

figurations is imaginable. Each of these configurations then will offer a different set

of specific features that may (or may not) support HRM.

Abstracting from specific features, there evidently are some main features of cur-

rent HRIS that refer to functionality and delivery. Regarding the functionality pro-

vided by back-end systems, in contrast to mere operational systems of the past func-

tionality now can be considerably expanded into managerial areas and integrated with

other business functions [23, 43]. Regarding the delivery provided by web-based

front-end systems, new HRIS user groups such as managers, employees and appli-

cants can be opened up thereby offering completely new ways of organising HRM

[31, 43].

Acknowledging that these features represent major innovations with far reaching

implications for HRM, and additionally acknowledging that innovativeness of a tech-

nology does by no means imply its appropriateness, our paper explores to what extent

these features actually meet genuine requirements of HRM. Does the automation and

support of complex managerial HR functions such as selection and performance man-

agement really meet a requirement of HRM? Is the self-service based handing over of

HR functions to managers and employees (and hence to layman) really desirable and

required by HRM? Previous empirical research has described major features of HRIS

[3, 13, 15, 16, 20, 21, 22, 24, 25, 32, 33, 37, 46], however, has barely considered the

current developments in functionality and delivery, and especially, has not evaluated

theses features against actual requirements of HRM.

We hence try to fill this gap and provide some answers. To do so, we firstly ana-

lyse the HRIS features under debate, discuss why these features may or may not meet

requirements of HRM and derive corresponding hypotheses. Subsequently, we em-

ploy an international large-scale survey to test these hypotheses by combining logistic

regression and neural network analysis. After presenting and discussing our results,

we finally attempt to derive some conclusions for future work.

2 Features of HRIS and Requirements of HRM

2.1 Functionality

That it is functionality of a HRIS that centrally matters to HRM seems to be a truism.

However, according to previous empirical research the functionality of HRIS was

rather restricted. For a long time operational functions - repeatedly mentioned are

payroll processing, employee record keeping and attendance administration - consti-

tuted the main functional focus of HRIS [3, 13, 21, 25, 32, 33] while there is only few

evidence of non-operational HRIS in the recent past [3, 24, 46]. Owing to the ongo-

ing technological development, meanwhile the offered functionality of packaged

HRIS components has considerably expanded. Besides operational functions more

and more managerial applications, comprising also “softer” and hence harder pro-

grammable functions are offered. At present, there are extensive systems for all core

HR areas, such as recruiting and selection [e.g. 10], compensation and benefits [e.g.

14], training and development [e.g. 48], performance management [e.g. 36] and HR

planning [e.g. 23]. For some time past, there is no function in HRM, which cannot be

supported by corresponding systems [9, 23]. This gradually occurring transition to a

comprehensive managerial functionality constitutes a first alteration that - at least

optionally - characterises a main newer feature of HRIS.

Concerning the requirements of HRM, the identified previous concentration on

operational functions is consistently perceived as insufficient. Research manifested a

dichotomy of “unsophisticated vs. sophisticated” [33], “administrative vs. strategic”

[46], “administrative vs. analytical” [27], “transaction orientated vs. decision support

orientated” [7] or “non-strategic vs. strategic” [24] functionalities of HRIS. Though

slightly different in terminology and in main focus, these set of opposing pairs ac-

cordingly refer to the gap between actually offered operational functions and desir-

able, but - at least until a shorter time ago - not available managerial functions of

HRIS. The emphasis on managerial HRIS functions is directly derived from core

ideas of HRM: Rather than merely administering existing workforce, HRM crucially

aims at the proactive management of employees so as to actively contribute to busi-

ness strategy and organisational performance [19]. Hence managerial functions, in

particular recruitment and selection, training and development, compensation and

benefits, performance management and HR planning, are core activities of HRM [6].

As a consequence, managerial functions of HRIS are considered as particularly im-

portant. HRIS increasingly are seen as measures to support organisational strategy

and to contribute to organisational performance by not only automating managerial

HR functions, but in particular by creating relevant information. This information

shall enable comprehension and appropriate decisions leading to improved HR results

[7, 11, 23, 24, 27, 28, 30, 40]. Hence:

H1a: The implementation of managerial HR functions in HRIS will meet the re-

quirements of HRM.

Given this, the offering of operational functions constitutes an ordinary and by no

means innovative feature of HRIS. However, it would be precipitate to deem opera-

tional functions as obsolete. Though the contribution of operational HR functions to

organisational performance may be minor, they are nevertheless operationally neces-

sary. Payroll processing constitutes a plain example: Justifiably employees expect

their salaries and being lastingly incapable to ensure a correct and timely pay is a

worst case scenario to HRM. So as to avoid the harm as well as the costs of a manual

payroll, HRM of course needs operational payroll functionalities. Quite similar, op-

erational functionalities are frequently necessary to ensure compliance with national

legislation. For instance, a number of European countries require the online delivery

of income-tax relevant employee data for their revenue authorities, what is usually

accomplished again by payroll systems.

Hence, in comparison to managerial functions operational functions may be less

valued. However, HRM would badly miss operational functionalities if they were not

implemented. Hence:

H1b: The implementation of operational HR functions in HRIS will meet the re-

quirements of HRM.

As a further feature, current HRIS may be integrated with other organisational sys-

tems. This vision of an integrated Management Information System (MIS) is as old as

the organisational usage of IT. However, the complexities of its realisation have

caused numerous setbacks. Hence, rather than being integrated into the functions of a

general MIS, HRIS of the past tended to be isolated. In the interim however, there are

two major working approaches to actually integrate the functions of HRIS into a

general MIS [29]. Firstly, for more than a decade Enterprise Resource Planning Sys-

tems (ERP) offer possibilities. Based on a common integrated database usually nu-

merous modules for areas such as HRM, finance, production, sales and accounting,

with predefined interfaces are offered by a single ERP-vendor. HR systems, then, are

integrated into a general business concept of the corresponding system, which allows

cross functional data integration, business processes etc. [1, 11, 29, 40]. Additionally,

even if there are heterogeneous, apriori non-compatible systems, Enterprise Applica-

tion Integration (EAI) offers a second approach for integration. Based on inter-

application “middleware”, EAI applications serve as an integrative bridge between

the heterogeneous back-end systems of HR, other organisational systems and the

front-ends [29]. Besides specific EAI applications, portal systems increasingly dis-

pose of comprehensive EAI features as well [44]. Hence, external integration of func-

tions is a further, gradually obtainable feature offered by current HRIS.

This external integration directly corresponds with another core idea of HRM, i.e.

the coordination with other managerial functions in order to achieve a “fit” with cen-

tral organisational developments [2]. Integration features of HRIS allow the system-

atic coordination of operational as well as managerial HR functions with other organ-

isational functions. Concerning operational integration, e.g. HR cost data as results of

monthly payroll processing can directly be transferred to respective accounting sys-

tems. Concerning managerial integration, e.g. the production schedules of production

systems can be integrated with work scheduling in the HR.IS. The functional integra-

tion of the HRIS into a general MIS therefore offers a decisive measure for achieving

the often postulated “external fit” of HRM [2]. Hence:

H1c: The integration of HRIS functions with functions of a general MIS will meet

HRM requirements.

2.2 Delivery

Besides the mere functions especially novel kinds of web-based delivery of functions

have gained much attention [43]. Previously, functionalities were delivered to the

end-users via graphical or textual user interfaces. Since the advent of the World Wide

Web over a decade ago, HRIS can increasingly be accessed via web. Technically,

web servers enable HR back-end systems to deliver their functionality in terms of

HTML-sites that can be accessed by web-browsers as front-end systems. As a conse-

quence, web-browsers are increasingly used as the main user interface of HRIS [30].

Web-based delivery of functions is often associated with several advantages.

Firstly, browsers are common practice and end-users are accustomed to it. Browsers

offer graphical interfaces that are self-explanatory and easy to use. As its main advan-

tage however, access to back-end systems becomes ubiquitously possible. Functional-

ity can be delivered worldwide to any desired site of the organisation, but also to any

home office thereby enabling new possibilities of HR telecommuting. If combined

with portal technology, numerous and heterogeneous back-end systems and services

can simultaneously be accessed using a single sign on, based on a unified interface

and presenting personalised access. Users can easily switch between back-end sys-

tems without knowing the corresponding system, its navigation features etc. This

releases users from the burden of conventionally using numerous different systems

simultaneously and should lead to cost and time savings [23, 43]. These general de-

livery advantages as compared to conventional user interfaces should be valued by

HRIS users as well. Hence:

H2a: The delivery of implemented HR functions to HR professionals via World

Wide Web will meet the requirements of HRM.

However, really striking changes due to web-based delivery arise from the opening

up of new user groups, mainly line managers, employees and applicants. Clearly,

based on an elaborated user and rights administration, these groups could have gained

access to conventional HRIS as well. However, besides security considerations espe-

cially the handling of the systems proves to be uncomfortable for occasional users.

Hence, the user-friendly, ubiquitous delivery via web constituted the technological

breakthrough for inaugurating new actor groups to HRM.

Concerning the requirements of HRM, especially the connection of line managers

to HRIS corresponds with a conceptual characteristic of HRM: Denominated as “de-

centralization of HRM”, the participation of line management in HRM is discussed

since the 1990s. Such decentralization can be well supported by so called “manager

self-service”-features that enable managers to perform operational and managerial HR

functions related to their subordinated employees. Though initially not an explicit

demand of HRM, likewise employees themselves can be provided with web-based

access to HRIS. Besides rather operational tasks such as data maintenance or travel

authorizations requests also the composing of compensation options, choice of (web-

based) training measures etc. can be assigned to employees (“employee self-service”).

Restricted to recruiting and selection “applicant self-service” concepts are feasible

too.

Taken together, such delivery features have introduced novel ways of organising

HRM. HR functions can be assigned to a network of spatially segregated and hetero-

geneous actors that are supported and coordinated by IT [31, 43]. Especially this kind

of delivery has led to mentioned general notations like electronic or virtual HRM

[34]. These possibilities of organising HRM should comprise opportunities, but pos-

sibly also threats for HRM. This transfer of HR functions to external actors may save

time and costs, avoid redundant work and enhance cooperation between HR stake-

holder groups. Therefore, HRM may be “liberated” [40] for more important activities

that support business strategy and organisational performance. However, this also

puts pressure on HRM to truly act strategic and truly support to performance. Other-

wise e-HRM may result in an unwanted downsizing of the HR department so as to at

least realise the promised rationalisations. Though comprehensive e-HRM concep-

tions may entail challenges to HRM, the positive aspects should prevail. Hence:

H2b: The delivery of implemented HR functions to line managers, employees and

applicants via World Wide Web (“e-HRM-features”) will meet the require-

ments of HRM.

3 Method

3.1 Sample

The data employed in our study stem from the repeating Cranet Survey, which con-

tains data on HRM issues of private and public organisations in over 30 countries.

Data are ascertained by research partners in each of the countries. Information is

gathered from the senior HR person in the organisation via a postal survey using a

comprehensive address list. The data is broadly representative with respect to the

industrial sector (using European Union’s NACE categorization) in each country. In

the 2004 survey over 40,000 questionnaires were sent out worldwide. With a re-

sponse rate of approximately 17% more than 7,000 organisations participated [see for

details 8].

We restricted our analysis to European countries (5,771 organisations in 22 coun-

tries). Given this subset we additionally sorted out the organisations that did not dis-

pose of an internal HRIS (1,054 organisations/18.2%). This resulted in a sample of

4,717 organisations. As described below, we used a combination of logistic regres-

sion and neural network analysis to analyse the data. To assess performance, neural

network analysis (and optionally logistic regression analysis too) requires a holdout

or test sample of usually 20% of the total sample [42]. We followed this convention

and randomly selected a test sample of 957, while using the remaining set of 3,760

organisations for logistic regression and neural network analysis.

3.2 Measures

As the dependent variable, suitability of the HRIS was measured using a 4-point

Likert scale which referred to the extent that the HRIS actually met the requirements

of HRM. HRIS that met the requirements entirely or to a large extent were coded as

1, HRIS that met the requirements only to a small extent or not at all received a 0.

The provision of operational and managerial functions was measured by the num-

ber of implemented functions. In accordance with previous research [e.g. 3, 46] “pay-

roll processing”, “time and attendance administration”, “employee record keeping” as

well as “health and safety” were taken as operational functions. Managerial functions

referred to the “major four” HR functions (“recruitment and selection”, “compensa-

tion and benefits”, “training and development” and “performance management”) that

were completed with two HR planning functions (“career and succession planning”

and “work scheduling”). To ascertain cross-functional integration it was asked,

whether or not the HRIS was integrated into a general MIS.

Delivery was measured corresponding to the respective user groups. Concerning

the HR professionals it was asked whether the HRIS offered web-based access or not.

Concerning the novel user groups of e-HRM, the extent of web-based delivery was

measured on a 6-point Likert scale (from 0 for HRIS without any web-based delivery

to 5 for HRIS that allow highly complex web-based transactions).

3.3 Data Analysis

The kind of actual relationships between HRIS features and the requirements of HRM

are unknown at present. In order to map also more complex relationships such as

interaction- and threshold-effects, we decided to follow suggestions of the literature

and use a combination of the established logistic regression and the rather novel neu-

ral network analysis [5, p. 259].

Besides some further advantages, neural nets can identify and map highly com-

plex, non-linear (but also trivial and linear) relationships between independent and

dependent variables. For this general strength neural nets are increasingly used as

statistical tools in multivariate research [5, 17]. Neural nets are best understood as a

computational structure, which connects independent variables with dependent vari-

able(s). To accomplish this, neural nets arrange the independent variables in an input

layer and the dependent variable(s) in an output layer. Subsequently one or more

intermediate layers with several processing elements (“neurons”) are constituted to

mathematically map the probably highly complex, non linear connections between

independent and dependent variables as good as possible [see for a detailed descrip-

tion e.g. 5, 17, 38, 45].

As the most common approach, we used the multi layer perceptron (MLP) [5, 17,

41, 42]. Previously, the task of determining the adequate number of layers and com-

putational elements (“network topology”) and the adequate computational character-

istics of the elements were left to trial-and-error approaches of researchers [17, 5].

Meanwhile, this can be delegated to algorithms that systematically test different net-

work topologies and computational rules and select the individually most suitable

configuration [41]. The MLP that provided the best results showed a 5-2-2 topology

(5 input elements, 2 intermediate layers with 2 elements each).

To determine the relative influence of input variables on the output variable(s) of a

neural net sensitivity analysis is suggested. Sensitivity analysis systematically varies

the input values and measures the averaged variation of output values [41]. Sensitiv-

ity values, however do not allow determining direction and strength of the effects.

Hence, the supposed combination with regression may help to interpret findings.

The most meaningful measure of network performance is the accuracy of its clas-

sification and hence the generalizability of results. The accuracy of prediction is de-

termined on the basis of the test sample, i.e. data that were not used for the creation of

the net [38, p. 249ff]. We hence tested the classification results of the MLP and com-

pared it with the classification results of logistic regression.

4 Results

4.1 Descriptive and Bivariate Results

Means (M), standard deviations (SD) and correlations between variables of the study

are listed in Table 1.

Table 1. Descriptive Statistics and Correlations between Study Variables.

Variable M SD 1 2 3 4 5 6

1. Managerial Functions 2.14 1.52 -

2. Operational Functions 2.55 0.94 .31** -

3. Cross-functional Integra-

tion

0.38 0.48 .15** .12** -

4. Web Delivery (HRM) 0.19 0.39 .11** .02 .08** -

5. Web Delivery (e-HRM) 1.25 1.17 .22** .13** .12**

.22

6. Suitability 0.62 0.48 .32** .23** .12**

.06

.16

* p < .05 ** p < .01

n = 4,717

Concerning the managerial functions, 96% of HRIS offered at least one managerial

function. This basically confirms the notion that HRIS are increasingly managerial

orientated. However, this transition seems to be rather a beginning trend since only

3% of systems showed the full range of managerial functions. 38% of the HRIS were

integrated into a general MIS. Only 19% of HRIS offered web-based access for HR

end-user, while surprising 75% showed some e-HRM features. Overall 62% of HRIS

met the requirements of HRM and hence were seen as suitable. All examined HRIS

features were significantly positive correlated to its suitability.

4.2 Multivariate Results

Results of the logistic regression analysis are summarised in Table 2. Fit statistic first

shows that the selected features are able to explain suitability to a certain degree. As

expected, the results confirm that functionality actually matters. Especially, manage-

rial functions were valued, while operational functions as well showed clear rele-

vance. Additionally, the integration with general MIS functions proved to meet the

requirements of HRM. Therewith, all functionality related hypothesis could be con-

firmed. Delivery features showed mixed results. Web-based HRIS access for HR

professionals does obviously not constitute an essential requirement of HRM. Hence,

hypothesis 2a has to be rejected based on logistic regression. Conversely, web-based

delivery to new actor categories was an obvious predictor of suitability, therewith

substantiating the role of e-HRM.

Table 3 shows the results of the MLP.

Table 2. Results of Logistic Regression.

B Fit Statistic

1. Managerial Functions -.414**

2. Operational Functions -.331**

3. Cross-functional Integration -.296**

4. Web Delivery (HRM) -.074

5. Web Delivery (e-HRM) -.183**

-2 Log Likelihood 1248.68**

674.42**

Cox and Snell .133

Nagelkerke .182

McFadden .108

** p < .01 n = 3,760

Table 3. Results of the Multi Layer Perceptron.

Predictor Sensitivity Rank

1. Managerial Functions

0.443 (1)

2. Operational Functions

0.279 (2)

3. Cross-functional Integration

0.064 (5)

4. Web Delivery (HRM)

0.097 (4)

5. Web Delivery (e-HRM)

0.160 (3)

Estimated Accuracy: 70.36% n = 3,760

Interestingly, neural network analysis shows similar as well as variant results.

First, the importance of managerial functions could be affirmed. Managerial functions

showed the greatest influence for the classification results of the MLP found. Also,

operational functions definitively influenced the classification result. While being a

significant predictor in logistic regression, cross-functional integration of these func-

tions exerted only minor influence for the classification results of the MLP and was

actually the least important predictor of suitability. Given this, hypotheses 1c is de-

batable from an MLP perspective. Concerning delivery features, the MLP concurred

with logistic regression, since e-HRM features proved to have a perceptible influence

on suitability, while the influence of the web-based access for HR was rather unim-

portant.

Table 4 compares the classification accuracy of both methods based on the test

data.

Table 4. Comparison of Classification Results.

Neural Net Logistic Regression

Overall Correct Classified 71.16% 70.64%

Suitability Correct Classified 82.23% 84.27%

Unsuitability Correct Classified 49.22% 43.61%

n = 957

The comparison shows that neural net analysis was only marginally better in gener-

ally classifying suitable and unsuitable HRIS. While the MLP showed some advan-

tages in classifying unsuitable HRIS, logistic regression showed some advantages in

classifying unsuitable HRIS. Taken together, concerning the results commonalities

between both methods prevail, while the documented advantages of neural nets due to

there ability to map highly complex relationships [see the overview of comparative

studies in 16, p. 8ff] could not be exploited in our study.

5 Discussion

Our study aimed at the question whether main features of current HRIS actually cor-

respond with HRM requirements. The results generally underscore that offered fea-

tures are valued and mainly do meet the requirements of HRM.

At first, the long demanded transition from operational to managerial HRIS can be

confirmed: Managerial functions are actually valued by organisations. However,

though our study generally underlines the trend towards managerial HRIS the main

focus of current systems is operational now as before. Given the confirmed relevance

of managerial functions on the one hand and the only gradually growing usage of

managerial HRIS on the other hand, this seems to indicate a latent usage paradox:

Particularly the functions that are the best predictors of suitability are only hesitantly

utilized. Hence, concerning the numerous organisations with sheer or rather opera-

tional HRIS there is a simple message: Since managerial systems are both available

and, according to our results, suitable for meeting the requirements, it should be vali-

dated if and to what extent a transition to managerial HRIS will improve service de-

livery and decision quality in the respective activities.

Additionally, as hypothesised organisations attach importance to operational func-

tions as well. Therefore operational functions of HRIS are by no means obsolete. In

particular the time-consuming character of operational tasks is an ongoing driver

towards automation. Operational HRIS functions are thus the most important “libera-

tors” of HRM.

Taken together, viewing operational HRIS functions as “hygiene factors” and

managerial HRIS functions as “motivators” may provide an appropriate understand-

ing of the respective importance and the mutual interactions of both areas. Ordinary

operational functions of HRIS now as before are necessary and will cause substantial

dissatisfaction if they are missing. However, managerial functions will increasingly

constitute the more important and more interesting field of upcoming technological

support for HRM. This dichotomy gives also a glimpse of future competition on mar-

kets for integrated HRIS software. Coming competitive advantages will rather be

based on the offered managerial functions, while viable operational functionalities

will be seen as a matter of course. Hence offering extensive operational functions will

organisations not motivate to purchase software, but not offering them, may lead to

the refusal of a solution.

Concerning the integration of HRIS functions into a general MIS, there was mixed

evidence. This brings about some debate. Since the MLP showed at least few contri-

butions to satisfy the requirements too, possible effects of cross-functional integration

of HRIS should not be precipitately rejected but further examined. A possible expla-

nation of only minor contributions - should this be a lasting result - may refer to level

inadequacies [26]. The advantages of a fully integrated MIS may not, or at least not

mainly, refer to the departmental level and hence to HR and other departments, but to

the general organisational level. It hence may be simply not adequate to ask whether

this feature meets departmental requirements. In any case, the suitability effects of

cross-functional integration of HRIS need further investigation.

Referring to delivery features, web-based access for HR end-users obviously were

plainly overrated by hypothesis 2a. Given that in the interim usually all HRIS tend to

be graphically orientated, easy to use and self-explaining, the additional delivery

advantage that provides ubiquitous access does obviously not correspond with the

requirements of HRM.

Conversely and interestingly, e-HRM obviously satisfies a meanwhile frequent re-

quirement of HRM. This pivotally substantiates the status that e-HRM has achieved

in the interim. Concerning HRM requirements, e-HRM seems to represent an interest-

ing case: e-HRM seems to only partially arise from preceding requirements. Espe-

cially manager self-service may be seen as an answer to the preceding HRM discus-

sion and practical claims concerning a decentralisation. An empowerment of employ-

ees and applicants, however, to rather independently perform HR tasks and make

decisions for themselves was not broadly discussed and claimed before. Hence, larger

parts of e-HRM seem to be rather a “technology push”- than a “demand pull”- phe-

nomenon. This by no means devalues such a requirement, since a lot of technological

innovations, ranging from cars to television, were not broadly required prior to their

invention. Taken together, e-HRM features of HRIS evidently are the second major

innovation besides the managerial functions. Hence again, organisations without or

without elaborated e-HRM should evaluate, if the further utilisation and expansion of

e-HRM may result in improvements of their work. In the same way, HRIS software

vendors that have paved the way for e-HRM can be assured to have initiated a trend

that actually appeals to practical HRM and is of lasting importance.

6 Conclusions

This paper aimed at an evaluation of major features of current HRIS against the actual

requirements of HRM. Based on conceptual work, a large scale-survey and the paral-

lel application of two methodical approaches two major findings could be revealed:

First, managerial functions of HRIS (if based on adequate operational functions) are

especially valued by HRM. Second, in addition web-based delivery possibilities to

new user groups were particularly appreciated. Taken together, our study generally

draws a rather positive picture of contemporary HRIS. If equipped with the right

features HRIS are able to actually meet requirements of HRM.

Our study, however, also shows limitations. Firstly, our study is limited to five ma-

jor HRIS features, while there are numerous other detail features. A refinement and

enrichment of examined HRIS features hence would provide a deeper understanding

and probably explain more “variance”. Secondly, our study implies a universalistic

perspective [12] by suggesting that all organisations show more or less the same re-

quirements concerning HRIS. On the very general level of examined features this

seems to be admissible. However, at least when refining the analysis, contingencies

are to be expected. If for instance not functional categories but single HR functions

are examined, the suitability of HRIS will necessarily depend on whether or not the

corresponding function is actually performed. Hence a contingent perspective seems

to be a necessary future improvement. Thirdly, we used a perceptional summary

measure of suitability. Future studies hence may analytically distinguish several re-

quirements and - given the common criticism - should use objectives measures wher-

ever possible. Fourthly, our additional application of neural net analysis did not pro-

vide significantly better results and additionally yielded a deadlock situation concern-

ing cross-functional integration. However, accepting that each multivariate method

shows specific strengths as well as weaknesses a methodical enhancement of tradi-

tional methods with neural nets now as before makes sense if complex, non-linear

relationships are to be expected. After all, commonalities of both methods prevailed

and hence stand for methodical robust findings concerning the other features.

Despite of these considerations the technologisation of HRM forges ahead and re-

search is well advised to keenly keep up with it.

References

1. Ashbaugh, S., Miranda, R.: Technology for Human Resources: Seven Questions and An-

swers. Public Personnel Management, 31, 1 (2002) 7-20

2. Baird, L., Meshoulam, I.: Managing the Two Fits of Human Resource Management. Acad-

emy of Management Review, 13, 1 (1988) 116-128

3. Ball, K.S.: The Use of Human Resource Information Systems: A Survey. Personnel Re-

view, 30, 6, (2001) 677-693

4. Beckers, A.M., Bsat, M.Z.: A DSS Classification Model for Research in Human Resource

Information Systems. Information Systems Management, 19, 3 (2002) 41-51

5. Bell DeTienne, K., DeTienne, D.H., Shirish, A.J.: Neural Networks as Statistical Tools for

Business Researchers. Organisational Research Methods, 6, 2 (2003) 236-265

6. Boselie, P., Dietz, G., Boon, C.: Commonalities and Contradictions in HRM and Perform-

ance Research. Human Resource Management Journal, 15, 3 (2005) 67-94

7. Broderick, R., Boudreau, J.W.: Human Resource Management Information Technology

and the Competitive Edge. Academy of Management Executive, 6, 2 (1992) 7-17

8. Brewster, C., Mayrhofer, W., Morley, M. (eds.) Human Resource Management in Europe:

Evidence of Convergence? (2004) Butterworth Heinemann, London

9. Bussler, L., Davis, E.: Information Systems: The Quiet Revolution in Human Resource

Management. Journal of Computer Information Systems, 42, 2 (2002) 17-20

10. Chapman, D.S., Webster, J.: The Use of Technologies in the Recruiting, Screening, and

Selection Processes for Job Candidates. International Journal of Selection and Assessment,

11, 2/3 (2003) 113-120

11. Dery, K., Wailes, N.: Necessary but Not Sufficient: ERPs and Strategic HRM. Strategic

Change, 14, 5 (2005) 265-272

12. Delery, J.E., Doty, H.D.: Modes of Theorizing in Strategic Human Resource Management:

Test of Universalistic, Contingency and Configurational Performance Predictions. Acad-

emy of Management Journal, 39, 4 (1996) 802-836

13. DeSanctis, G.: Human Resource Information Systems: A Current Assessment. MIS Quar-

terly, 10, 1 (1986) 14-28

14. Dulebohn, J.H., Marler, J.H.: e-Compensation: The Potential to Transform Practice? In:

Gueutal, H.G., Stone, D.L. (eds): Brave New World of e-HR. Human Resources Manage-

ment in the Digital Age, Jossey Bass, San Francisco (2005) 166-189

15. Elliott, R.H., Tevavichulada, S.: Computer Literacy and Human Resource Management: A

Public/Private Sector Comparison. Public Personnel Management, 28, 2 (1999) 259-274

16. Florkowski, G.W., Olivas-Luján, M.R.: The Diffusion of Human Resource Information

Technology in US and Non US Firms. Personnel Review, 35, 6 (2006) 684-710

17. Garson, G.D.: Neural Networks. An Introductory Guide for Social Scientists. Sage Publica-

tions, London Thousand Oaks New Delhi (1998)

18. Groe, G.M., Pyle, W., Jamrog, J.J.: Information Technology and HR. Human Resource

Planning, 19, 1 (1996) 56-61

19. Guest, D.E.: Personnel and HRM: Can You Tell the Difference? Personnel Management, 1,

21 (1989) 48-51

20. Haines, V.Y., Petit, A.: Conditions for Successful Human Resource Information Systems.

Human Resource Management, 36, 2 (1997) 261-275

21. Hall, L., Torrington, D.: Why Not Use the Computer? The Use and Lack of Use of Com-

puters in Personnel. Personnel Review, 15, 1 (1986) 3-7

22. Hannon, J., Jelf, G., Brandes, D.: Human Resource Information Systems: Operational

Issues and Strategic Considerations in a Global Environment. International Journal of Hu-

man Resource Management, 7, 1 (1996) 245-269

23.

Hendrickson, A.R.: Human Resource Information Systems: Backbone Technology of

Contemporary Human Resources. Journal of Labor Research, 24, 3 (2003) 381-394

24. Hussain, Z., Wallace, J., Cornelius, N.E.: The Use and Impact of Human Resource Infor-

mation Systems on Human Resource Management Professionals. Information and Man-

agement, 44, 1 (2007) 74-89

25. Kinnie, N.J., Arthurs, A.J.: Personnel Specialist’s Advanced Use of Information Technol-

ogy. Personnel Review, 25, 3 (1996)

26. Klein, K.J., Dansereau, F., Hall, R.: Level Issues in Theory Development, Data Collection,

and Analysis. Academy of Management Review, 18, 2 (1994) 195-222

27. Kovach, K.A., Cathcart, C.E.: Human Resource Information Systems (HRIS): Providing

Business with rapid Data Access, Information Exchange and Strategic Advantage. Public

Personnel Management, 28, 2 (1999) 275-282

28. Lawler, E.E., Mohrman, S.A.: HR as a Strategic Partner: What Does It Take to Make It

Happen? Human Resource Planning, 26 (2003) 15-29

29. Lee, J., Siau, K., Hong, K.: Enterprise Integration with ERP and EAI. Communications of

the ACM, 46, 2 (2003) 54-60

30. Lengnick-Hall, M.L., Moritz, S.: The Impact of e-HR on the Human Resource Manage-

ment Function. Journal of Labor Research, 24, 3 (2003) 365-379

31. Lepak, D.P., Snell, S.A.: Virtual HR: Strategic Human Resource Management in the 21

Century. Human Resource Management Review, 8, 3 (1998) 215-234

32. Lin, C.Y.Y.: Human Resource Management in Taiwan: A future Perspective. International

Journal of Human Resource Management. 8, 1 (1997) 29-43

33. Martinsons, M.G.: Benchmarking Human Resource Information Systems in Canada and

Hongkong. Information and Management, 26, 6 (1994) 305-316

34. Mayfield, M., Mayfield, J., Lunce, S.: Human Resource Information Systems: A Review

and a Model Development. Advances in Competitiveness Research, 11, 1 (2003) 1349-152

35. McLeod, R.Jr., DeSanctis, G.: A Resource-Flow Model of the Human Resource Informa-

tion System. Journal of Information Technology Management, 6, 3 (1995) 1-15

36. Miller, J.: High Tech and High Performance: Managing Appraisal in the Information Age.

Journal of Labor Research, 24, 3 (2003) 409-424

37. Ngai, E.W., Wat, F.K.: Human Resource Information Systems: A Review and Empirical

Analysis. Personnel Review, 35, 3 (2006) 297-314

38. Ripley, B.D.: Pattern Recognition and Neural Networks. Cambridge: University Press

(1996)

39. Ruël, H.J.M., Bondarouk, T., Looise, J.C.: E-HRM: Innovation or irritation. An Explor-

ative Empirical Study in Five Large Companies on Web-based HRM. Management Revue,

15, 3, (2004) 364-381

40. Shrivastava, S., Shaw, J.B.: Liberating HR through Technology. Human Resource Man-

agement, 42, 3 (2003) 201-222

41. SPSS, Inc.: Clementine 10.1 Algorithms Guide. Chicago (2006)

42. SPSS, Inc.: Clementine 10.1 Node Reference. Chicago (2006)

43. Strohmeier, S.: Research in e-HRM: Review and Implications. Paper accepted for publica-

tion in Human Resource Management Review 1, (2007)

44. Strohmeier, S.: (HR- and Employee-) Portals. Article accepted for publication in Torres, T.,

Arias-Oliva, M. (eds.): Encyclopedia of HRIS, Idea Group, Hershey (2007)

45. Swingler, K.: Applying Neural Networks: A Practical Guide. Academic Press, New York

(1996)

46. Teo, T.S.H., Soon, L.G., Fedric, S.A.: Adoption and Impact of Human Resource Informa-

tion Systems (HRIS). Research and Practice in Human Resource Management, 9, 1 (2001)

101-117

47. Walker, A.J.: A Brief History of the Computer in Personnel. Personnel Journal, 16 (1980)

32-36

48. Welsh, E.T., Wanberg, C.R., Brown, K.G., Simmering, M.J.: E-learning: Emerging Uses,

Empirical Results and Future Directions. International Journal of Training & Development,

7, 4 (2003) 245-258