Four%20Paradigms%20of%20IS%20Development

1
Four Paradigms of IS Development

• Survey Design and Hypothesis Testing
• (Proposal)
• by Jie (Jennifer) Xu

2
Four Paradigms

• Functionalism developer-as-systems
  expert
• Social Relativismdeveloper-as-systems-facilitator
  
• Radical--StructuralismDeveloper-as-labor-partisan
  
• Neohumanismdeveloper-as-emancipator-or-social-the
  rapist.

3
Order
Functionalism Developer-as-system-expert
Social Relativismdeveloper-as-systems-facilitator
Subjective
Objective
Radical Structuralism Developer-as-labor-partisan
Neohumanism developer-as-emancipator-or-social-th
erapist
Conflict
4
Y ß1F ß2S ß3R ß4N e

• where
• y the indicator of the approach used
  
• Dummy variables
• F indicator of Functionalism,
• F 1 if one believes
  Functionalism, F0 otherwise
• R indicator of Radical Structuralism,
  
• R 1 if one believes Radical
  Structuralism, R0 otherwise
• S indicator of Social Relativism,
• S1 if one believes Social
  Relativism, S0 otherwise
• N indicator of Neohumanism,
• N1 if one believes Neohumanism,
  N0 otherwise
• e error term.

5
Assumptions

• (1) Survey takers are rational and answer
  questions honestly
• (2) No autocorrelation
• ( This assumption is reasonable since
  the survey deals with people at one time period )
  
• (3) Heteroscidasticity is allowed.
• ( var(u) can varies with age, education
  level, sex, etc.)

6
Data
Hypothetical Data y F
S R N 1
1 0
0 0 4
0 0
0 1 3 0
0 1
0 3 0 1
0 0 2
0 1 0
0 . .
. .
. . . .
. .
7
Estimation

• Approach
• OLS ( without heteroscidasticity)
• WLS ( with heteroscidasticity)

8
Hypothesis Testing

• Use ----
• F test to test the joint significance of the
  coefficients.
• (If the F value is greater then a critical
  value, we reject the null hypothesis and conclude
  that there are more than one paradigms.)
• T test to test the significance of a specific
  coefficient.

9
Survey Design

• (1) Sample Selection
• Sample Size gt 30
• Participator
• managers
• IS developers

10
(2) Question Set a) Assumption Part These
questions ask mainly about the assumptions that
the survey taker may believe. These questions are
grouped into 4 sets. Each set consists of equal
number, say 5, of questions. Group
1--Functionalism Group 2--Social Relativism
Group 3--Radial Structuralism Group
4--Neohumanism. All the questions are
yes/no questions. The survey taker is supposed
to answer "yes" if he believes the assertion
provided or "no" otherwise. A typical
question may look like this " The system
developer play a neutral and objective role in
system development. A. Yes B.
No."
11
E.g. Assumption question set
12

• For example, the number of "yes" that
  participator A gives for each set is given
• Number of "Yes"
  Paradigm
• 5
  Functionalism
• 2
  Social Relativism
• 1
  Radial Structuralism
• 0
  Neohumanism
• The this participator uses paradigm
  of Functionalism and the values of the dummy
  variables (F,S,R,N) are set to 1, 0,0,0,
  respectively.

13
a) Application part

• Choice value of the
  indicator of paradigms
• A-Functionalism 1
• B-Social Relativeism 2
• C-Radical Structuralism 3
• D-Neohumanism 4

14
Problems with Methodology

• (1) Data problem
• Sometimes it is hard to classify the
  survey takers just by the number of "yes".
  Suppose a participator has two 4's for set 2 and
  set 3 and 0 for the
• other two sets. Then we can not tell whether
  the person believes Social Relativism or Radical
  Structuralism. If we set both S and R to 1, then
  this may
• lead to multicollinearity. One alternative
  way is to allow some weights to each question and
  allow higher weigh to the question about the role
  of system
• developer, since the answer to this question
  is largely determines the paradigm people use.

15

• (2) Sample problem
• The sample is mainly consists of
  developers and managers in mediate or large
  companies. Thus it may not be random. If most
  the survey takers are from the large companies
  which usually hold one specific paradigm, the
  result of the hypothesis test will tend to favor
  the null hypothesis.

16

• (3) Survey design problem
• Questions provided cannot well describe the
  properties of these different paradigms.
• If the survey taker is not familiar with context
  of the empirical case provided in the application
  part, he/she may end up with an "wrong" choice.
• (For instance, if the UTOPIA example is
  given, and the survey taker are not familiar with
  the technique in typesetting and the relationship
  between editor and typesetter, he/she may have no
  idea about which approach to choose.)