SA/D Topic 4 -- Data Gathering
- the process of communicating actions, facts, feelings, interpretations,
and plans (schedules, activities, etc.) regarding a system.
- collecting information necessary to understand and model the existing
system and the future system
- heavily involves users, and therefore requires a great deal of personal
contacts and communication
- purpose: to gain knowledge
- requires: interpersonal skills
- careful plan to minimize problems and time
- 30% of the analysis time
Data Gathering Techniques
- Reading and studying written materials.
- Observing is best suited to understanding human or machine processes that
currently exist (passive participation).
- Participating is an advanced form of observing in which the data gatherer
actually performs the activity (active participation).
- Surveying and questionnaires are best for written responses, and are used
in situations where a small amount of data is needed from a large number of people.
- Interviewing provides a verbal response, and is a good method for
collecting data on feelings and interpretations. It is the most widely used, and often
most poorly used, technique.
- Brainstorming is used in group interviews as a means of arriving at new
ideas. It is good in undefined situations in which no solution is evident.
Sources for Data Gathering
- Internal Sources
- External Sources
- Professional journals
- External documents (government regulations, etc.)
Considerations Affecting Data Gathering
-- breadth vs. depth of collection
- usually breadth and then depth of collection
- breadth refers to a general idea of the situation
- depth focuses on a specific area
-- open vs. closed questioning
- open-ended questions result in more information
- open-ended questions are good in an undefined environment
- closed question are not useful in analysis
-- direct vs. indirect questioning
- direct questions result in specifics
- indirect questions are good for sensitive questions
-- structured vs. unstructured questioning
- structured approach is more systematic and planned
- structured approach is good if scenario is well-defined
- unstructured provides overall objectives and guidelines
- unstructured approach is good if there are many unknowns
Problems of Data Gathering
- Approaching the wrong people
- Must find the right people from management directions
- Biasing the system or person by being there
- Your presence affects observation of the existing system.
- In interviews people say what they think they should.
- People telling you what they want you to hear
- People often give a biased view of their own, rather than a realistic assessment.
- Not sensing all the stimuli that are presented
- We perceive things selectively, and might miss some stimuli.
- Not interpreting correctly the things you sense
- Do you understand the jargon being used?
- Did you understand all that was being said?
- Not uncovering feelings or intentions
- User's feelings can influence the use or acceptance of a system.
Prerequisites to Data Gathering
- Sanctioned by the group where you collect data?
- Management, unions, etc. must understand your purpose.
- Participants aware of your process and intentions?
- People have to know why you are there.
- Familiar with the subject?
- Read up on the subject and prepare -- prior preparation prevents poor performance
- Selected data gathering techniques?
- What's best for this situation and environment.
Questionnaires -- used to
ascertain attitudes, beliefs, behaviors, and characteristics of several key people in an
organization who may be affected by the current and proposed systems.
- attitudes -- what people say they want in a new system
- beliefs -- what people think is actually true
- behavior -- what people do
- characteristics -- properties of people or things
Reasons to Use Questionnaires
- people to be questioned are widely dispersed
- large number of people are involved
- important to determine proportion that supports or opposes change
- trying to gauge overall opinion before undertaking project
- questionnaires require extensive planning
- questions must be transparently clear
- flow of questions must be presented in a way that brings out pertinent
- respondents' questions must be anticipated and accounted for
- administration of the questionnaire must be planned in detail
- those that leave all possible response options open to the respondent
- well-suited to situations in which it is necessary to solicit opinions
about some aspect of the system
- good when it is impossible to list effectively all possible responses to
- useful in exploratory situations in which analyst is not able to
determine exactly what problems exist with a system--responses can be used to focus on
cited problems more narrowly via interviews
- require that you anticipate the type of response that you will get
because you must be able to correctly interpret responses
- must be narrow enough to guide respondents to answer in specific way
- example: phrase questions in context of satisfaction vs. dissatisfaction
with current system
- example: suggest some system features to prompt respondents to recall
features of interest
- those that limit the response options available
- useful when it is possible to list effectively all possible responses to
questions and when all responses listed are mutually exclusive (choosing one precludes all
- useful when sampling large groups of people (easier to interpret)
|Speed of Completion
|Breadth and Depth
|Ease of Preparation
|Ease of Analysis
Guidelines for Choosing Language
- Use the language of respondents whenever possible. Keep wording simple.
- Be specific rather than vague in wording. Also avoid overly specific
- Keep questions short.
- Do not talk down to respondents through low-level language choices.
- Avoid bias and objectionable questions.
- Target questions to the right respondents.
- Include only technically accurate questions.
Review topics on scales and questionnaire design (Ch. 6).
Observation -- witnessing
firsthand the activities of users and decision makers.
- provides insights into what is actually done, not just what is documented
- reveals the relationships between decision makers and others.
- helps to confirm or negate conclusions resulting from other data
- allows SA to see how decision makers gather, process, share, and use
information to get the work done.
- decide which activities are to be observed
- decide the level of detail necessary
- create categories that adequately capture key activities
- prepare appropriate scales, checklists, or other materials for
- decide when to observe
- Time Sampling -- allows the analyst to set up specific intervals at which
to observe activities
- Event Sampling -- allows the sampling of entire events rather than random
time periods, and provides for observation of an integral behavior in its natural context.
- best to combine both
|| Time Sampling
|| Event Sampling
- reduces bias when observations are randomized
- allows a representative view of frequent activities
- allows observation of actual behavior as it occurs
- allows observation of important events
- data gathering is fragmented with no time for decision to
- misses infrequent but important decisions
- requires a great deal of time
- misses a representative sample of frequent decisions