Study Guides (238,237)
Canada (115,036)
York University (9,811)
Sociology (180)
SOCI 2030 (26)


11 Pages
Unlock Document

York University
SOCI 2030
Tracy Supruniuk

Chapter 9 & July 12 ‘12 APPROPRIATE RESEARCH QUESTIONS  Experimental research allows researcher to focus sharply on causality (changing one variable, causes another variable to change)  There are 3 criteria for causality: 1) There must be an association between variables. Ex. Dropping pen & weather getting hotter = no logical association 2) Need to establish a temporal order (one have to occur before the other) Ex. Take drug before losing weight 3) Need to rule out rival causal factors – even though there is an association & comes before, must consider if another variable is causing the change  Experimental research is good for focusing on small groups & set variables & suited for when you can manipulate the independent variable TERMS USED IN EXPERIMENT 1. Random Assignment  Dividing subjects into groups at the beginning of experimental research using a random process, so the experimenter can treat the groups as equivalent 2. Subjects  Cases or people used and on whom variables are measured 3. Experimental group/treatment  This is the group that experiences the modification or stimulus (aka treatment) 4. Control Group  The group that does not receive modification or stimulus (treatment) – served as comparison group – usually don’t know about the other group. Deception is needed at times due to Hawthorne Effect – whenever doing an experiment must measure the Dependent variable 5. Dependent Variable  Outcomes in an experiment that change in response to a treatment 6. Post-test  Measurement of the dependent variable in experiment after the treatment. When doing an experimental/control group, you want to randomly select people and do a pre/post test. Goal: to have both groups similar, need similar characteristics 7. Matching  To avoid selection bias – randomly select people based on similar characteristics into each group – split up big group into subgroups and match. Ex. 240 women and 260 men, choose 15 women and 15 men for control group and then 15 women and 15 men for experimental group 8. Experimental Design  Arranging the parts of an experiment and putting them together TYPES OF EXPERIMENTAL DESIGNS (and ‘design notation’) 1. True Experiments  Has 3 general characteristics: random assignment, experimental & control group and involves pre/post test A) – Classical Experiment E = equivalence (randomly selected) O = Observations X = Treatment # By O = time observation has been done E O 1 O - 2xperimental group E O 1 O2- Control group - Each line represents a different group, observed at same test, pre-test, post test: one of the best for establishing causality design B) – Post-test only control group design - same as classical but no pre-test & used when no bias result - difficult to rule out rival causal factors – randomly selected E X O 2 E O 2 C) – Solomon Four Group Design -subjects are randomly assigned to 2 control groups & 2 experimental groups - only one control group & one experimental group receives pre-test - all four receives post-test – combines both groups to find out strengths & Weaknesses E O 1 O 2 E O 1 O 2 E X O 2 E O 2 2. Pre-experimental Designs (Non-random)  Lacks random assignment & weaker than classical  Used when cannot use all features of classical but have weaker internal validity – avoid unless exploratory A) – One Shot Case Study - an experiment with only experimental group & post-test X O 2 B) – One-group pre-test – post-test design - has one group, a pre-test, a treatment, and post test but no control group O 1 O 2 C) – Static Group Comparison X O 2 - two groups, post test O 2 3. Non-equivalent control quasi design  Stronger than pre-experimental designs, used in special situations or when experimenter has limited control over independent variable A) – Time Series Design - Several repeated pretests, posttests & treatment for one group over a period of Time O 1 O2X O3O O 4 5 6 B) – Multiple Time Series Design - Same as time-series design by looks at experimental AND control group O O O X O O O 1 2 3 4 5 6 O 1 O2 3 O4O O5 6 C) – Non-Equivalent Control Quasi Design - 2 groups involved, more than one experimental group Group 1 - O X O 1 2 Group 2 - O 1 O 2 Control Group O O1 2 DIFFERENT ISSUES INVOLVED WITH EXPERIMENTS Internal Validity – ability to eliminate alternative explanations of dependent variables 1. Selection Bias: When groups in an experiment are not equivalent (randomly assigned) – matching can help – this may also affect external validity 2. History: Something that occurs and effect the dependent variable during an experiment, but is unplanned and outside the control of the experimenter. Outside even unrelated to the experiment affects outside variable – Ex. Rainy day causes depression – less of an issue when it affects both groups 3. Local History: only affects one of the outside group 4. Maturation: Natural process of growth, boredom, that occur to subjects during the experiment and affect the dependent variables – change of dependent variable due to age on time process not treatment, usually happens in longitudinal studies 5. Experimental Mortality: Subjects failing to participate through entire experiment 6. Testing Effect: Pretest measure that affects the experiment. More than the treat alone affects the dependent variable. 7. Instrumentation: Reliability threatened due to a change in how a concept is measured during experiment 8. Diffusion of Treatment: Treatment spills over from experimental group to control group and they begin to modify their behavior because of the learn of treatment – contaminate 9. Compensatory Behavior: Might need to compensate to nor have the treatment (control group) 10. Experimenter Effect: Researcher indirectly communicated desired findings to subjects ex. Body language – to avoid this do a double-blind experiment – neither the subjects or the person directly deals with the subjects know the specifics of the experiment – give a placebo (false treatment or sugar pill which has no effect on experiment) 11. Statistical Regression: Tendency for random errors to move group results toward the average External Validity – ability to generalize research findings in your study 1. Reactivity: This can also affect internal – occurs because subjects are aware of experiment and being studied – Hawthorne Effect 2. Field Experiments: less control over conditions 3. Testing Effects: Pretest effect dependent variable rather than treatment alone 4. Selection Bias: When groups in experiment are not equivalent ETHICAL ISSUES  Intrusive – can put people in physical danger or cause embarrassment – harm  Deception – placebo, Hawthorne effect  Informed Consent CAN AVOID THIS THROUGH: planning & pilot tests, instructions to subjects and post- experiment interviews Chapter 8 & July 17 ‘12 APPROPRIATE TOPICS FOR SURVEY RESEARCH  Good for self-reported beliefs or behaviors. When asking about past behavior: recollective questions  Can ask about many things at one time, measure many variables, test several hypothesis  Can ask about: 1. Self Reported Behavior 2. Attitudes/beliefs/opinions 3. Expectations (ask what someone would do in a specific situation and see if they did so) 4. Self Classification – are you a patient person? 5. Characteristics – demographic characteristics (age, sex etc) 6. Knowledge – useful in accessing people’s knowledge TYPES OF SURVEYS  Mailed surveys  Phone surveys  In-person surveys  Hand-out surveys  Electronic surveys MAILED SURVEYS Advantages: cheapest, can be conducted by a single researcher, can cover a wide Geographical area, offer anonymity Disadvantages: low response rate, cannot control the conditions, no one available to Clarify or probe. WEB SURVEYS Advantages: fast & cheap, flexible Disadvantages: coverage, privacy, verification, design issues TELEPHONE SURVEYS Advantages: Wide population can be reached, control sequence of questions Disadvantages: high cost, limited length, some people unreachable INPERSON INTERVIEWS Advantages: highest response rate, long time, extensive probe Disadvantages: high cost, can be biased SURVEY CONSTRUCTION (QUESTION WORDING)  Can begin with jotting down questions but need to follow guidelines 1. Be cautious of your audience o Know your audience and word questions in a way they will understand & will not be offended/cater to their expertise – don’t risk alienating respondents 2. Avoid Confusing/Vague Wording o Make questions as short as possible; avoid double negatives 3. Avoid Biased Questions o Ex. Do you still use crack? (Implying that they have used it before) researchers ask
More Less

Related notes for SOCI 2030

Log In


Don't have an account?

Join OneClass

Access over 10 million pages of study
documents for 1.3 million courses.

Sign up

Join to view


By registering, I agree to the Terms and Privacy Policies
Already have an account?
Just a few more details

So we can recommend you notes for your school.

Reset Password

Please enter below the email address you registered with and we will send you a link to reset your password.

Add your courses

Get notes from the top students in your class.