INFO 370 - L08 - October 25, 2004 Notes By: Fortier, Egaas, Yaptinchay, Prins! * Assignment 1 - will get back in a week * The parts for the Team Proposal are due on Wednesdays for the sake of feedback Research Design and Observation Deduction vs. Induction (Review from Chapter 1) > Difference is often about having a hypothesis or not > Deduction - Research guided by an initial assumption about the state of affairs * Evidence leads to a conclusion * In social research: constructing logical argument which you then test > Induction - Research aimed at examining a phenomenon w/o considering any existing assumptions * Observe and draw conclusions Examples > What do Word users think about the XP version of the product? (Inductive) > Does having a compute in every room in the dorms result in higher grades? (Deductive) > Experience in developed countries has shown that wide use of the Internet accelerates the economy. Does it do the same in developing countries? (Deductive) > How well do elementary school students understand what the Web is? (Inductive) > Types of Research * Exploratory research is always inductive * Descriptive research can be inductive deductive. * Explanatory (cause and effect) research is mostly deductive * Evaluation is always deductive Deductive Evaluation > The initial assumption that guides the evaluation is formulated as a _hypothesis_ > Hypothesis: A tentative statement about the relationship between two or more variables *that will be tested in the evaluation project* The Null Hypothesis H₀ If you disprove the null hypothesis, the opposite must be true Example: ① 'bright colors increase speed' ② 'bright colors do not increase speed' If one of those is disproved, the only possible reality is the other. "Exception disproves the rule." Examples of Hypotheses > Interfaces that use bright colors increase the speed of interaction > Placing computers in community centers will increase computer use in members of the community > The greater the use of computers, the smaller the rate of errors in typing > The rate of job-finding Variables > Characteristic that can take different values > Independent variable: hypothesized to cause variation in another variable > Dependent variable: hypothesized to vary depending on the influence of another variable What are the independent and dependent variables? > Interfaces that use bright colors increase the speed of interaction. - Independent: bright colors - Dependent: speed of interaction > Placing computers in community centers will increase computer use in members of the community - Independent: computers in community - Dependent: computer use in community > The greater the use of computer, the smaller the rate of errors in typing. - Independent: computer use - Dependent: typing errors > The rate of job-finding is higher among those who do not make typing errors - Independent: typing errors - Dependent: job-finding Path Analysis: Multiple variables Inductive Research > Begins with specific data, which are then used to develop (induce) a general explanation or assumption. > Examples: - Let's see what happens when people use computers everyday. - Let's find out why some people like bright interfaces while others don't. Research Design > To describe or to explain? > Inductive or Deductive > Naturalistic or Controlled (experimental) > Participant or Non-participant > Qualitative or Quantitative data Useful Databases > ACM Digital Library > Portal - Greater Western Library Alliance > INSPEC > IEEE Explorer > Science Direct Research Design: Definition > Framework which guides data collection > Determined by the research question > Also specifies - Variables to be measured - Population and sampling method - Method of observation - Method of analysis MOST IMPORTANT SLIDE OF TODAY'S TALK!!!!! Choice of Research Design > Cross-sectional - Case study - What is happening? * small group or one subject (small sample) * descriptive type of study - Comparison - How are A and B different at this one point in time? * sample divided into sub-groups > Longitudinal study Panel - Has there been a change in A (the Panel) over time? > Longitudinal comparison - Are A and B different over time? (repeated cross-sectional) > Experiment - Is the difference between A and B due to a change in the independent variable? Case Study > Descriptive > Collecting information from one group or one person at one point in time Comparison study > cross-sectional design > collect observations between two different groups > To what extent do the two groups differ on the dependent variable > Select variable related to the concept under study > Measure the same variable in the same way in two or more groups at the same or nearly the same time Longitudinal Study > 2+ case studies of the same group searated by an interval of time > Measurement of variable at 2 points in time > Intervention / Maturation Longitudinal comparison > Quasi-panel design > Different groups of people are studied at the two points of time > Avoids problem of tracking subjects > Problems - can’t draw causal inferences - spurious relationships Experimental Design > A relationship between the variables needs to be established > All other reasons for the relationship must be eliminated > Two groups: Experimental and Control > Random Assignment of units of analysis to control experimental groups > Active intervention (treatment) for the experimental group - change the conditions of the independent variable (X) > After intervention, measurement of dependent variable (Y) for both groups > Double-blind - Neither subject nor experimenter know whether the subject is i the experimental or control group. > Within- vs. between-subjects variation Planning an observation > Entry > Setting > Task or situation being observed - Created by you? > Means of recording available Non-Participant Observation > Observation > Think-aloud > Surveys > In-depth interviews > Logs - computer files > Diaries / Journals Participant > Focus groups > Oral history > Ethnography # END #