summarize the article Of the five, the second item, the dependent variable in MIS research, is a particu$-$

larly important issue. If information systems research is to make a contribution to the

world of practice, a well$-$defined outcome measure $($or measures$)$ is essential. It does

little good to measure various independent or input variables, such as the extent of

user participation or the level of $\frac{\text{I}}{S}$ investment, if the dependent or output variable

I$/$S success or MIS effectiveness$-$cannot be measured with a similar degree of

accuracy.

The importance of defining the $\frac{\text{I}}{S}$ dependent variable cannot be overemphasized.

The evaluation of I$/$S practice, policies, and procedures requires an I$/$S success mea$-$

sure against which various strategies can be tested. Without a well$-$defined dependent

variable, much of $\frac{\text{I}}{S}$ research is purely speculative.

In recognition of this importance, this paper explores the research that has been

done involving MIS success since Keen first issued his challenge to the field and

attempts to synthesize this research into a more coherent body of knowledge. It

covers the formative period $1981-87$ and reviews all those empirical studies that have

attempted to measure some aspects of "MIS success" and which have appeared in

one of the seven leading publications in the I$/$S field. In addition, a number of other

articles are included, some dating back to $1949,$ that make a theoretical or conceptual

contribution even though they may not contain any empirical data. Taken together,

these $180$ references provide a representative review of the work that has been done

and provide the basis for formulating a more comprehensive model of $\frac{\text{I}}{S}$ success

than has been attempted in the past.

A Taxonomy of Information Systems Success

Unfortunately, in searching for an $\frac{\text{I}}{S}$ success measure, rather than finding none,

there are nearly as many measures as there are studies. The reason for this is under$-$

standable when one considers that "information," as the output of an information

system or the message in a communication system, can be measured at different

levels, including the technical level, the semantic level, and the effectiveness level. In

their pioneering work on communications, Shannon and Weaver $(1949)$ defined the

technical level as the accuracy and efficiency of the system which produces the infor$-$

mation, the semantic level as the success of the information in conveying the in$-$

tended meaning, and the effectiveness level as the effect of the information on the

receiver.

Building on this, Mason $(1978)$ relabeled "effectiveness" as "influence" and de$-$

fined the influence level of information to be a "hierarchy of events which take place

at the receiving end of an information system which may be used to identify the

various approaches that might be used to measure output at the influence level"

$($Mason $1978,$ p$.227).$ This series of influence events includes the receipt of the

information, an evaluation of the information, and the application of the informa$-$

tion, leading to a change in recipient behavior and a change in system performance.

The concept of levels of output from communication theory demonstrates the

serial nature of information $($i$.$e$.,$ a form of communication$).$ The information system

creates information which is communicated to the recipient who is then influenced

$($or not!$)$ by the information. In this sense, information flows through a series of stages

from its production through its use or consumption to its influence on individual

and$/$or organizational performance. Mason's adaptation of communication theory