Chapter 10 Design of experimental research in communication
Exploring the consumer- product relationship How do Lux affect the senses? When, why and how do you recognize Lux? What words are associated with Lux? How are Lux remembered? How are expectations and preferences acquired? What makes you change? When does Lux seem similar to other soap? When are they different? What emotions does Lux evoke?
Exploring the consumer- product relationship How do Lux affect the senses? (Sensation) When, why and how do consumers recognize Lux? (Perception and Attention) What words are associated with Lux? (Language) How are Lux remembered? (Memory) How are expectations and preferences acquired? What makes you change? (Learning and Preferences) When do Lux seem similar to others? When are they different? (Categorisation) What emotions does Lux evoke? (Emotion)
Centre for Experimental Consumer Psychology University of Wales, Bangor The Centre has three main foci of research effort. –The perceptual, attentional and linguistic processes involved in recognition, choice and categorisation of low-involvement consumer goods. This work (supported by Unilever and ESRC) is targeted both at establishing an understanding of perceptual, attentional and memory processes to inform better product design and advertisement and at the testing of theories of object recognition, attentional capacity and categorisation. –The children’s acquisition of taste preferences (basic principles of categorisation and learning) –The establishing how well-being, mood, and emotion determine responses to commercially produced stimuli (eg food, marketing images). (affective aspects of cognition and lead to projects linking emotion, perception and action.) –
Why experiment? Behavior is not random but a kind of pattern that influences by causal relationship – พฤติกรรมไม่ได้เกิดขึ้นแบบสุ่มสี่สุ่มห้า แต่เป็น รูปแบบที่เกิดขึ้นจากความสัมพันธ์ระหว่างเหตุ และผล Identifying cause-effect relationship among variables that researcher suspect exist while controlling other influences (variables)
Notion of experiment Manipulating independent variables. - when, how, how much subjects receive treatment – บังคับตัวแปรว่าผู้ที่ถูกทดลองควรจะได้รับ treatment เมื่อไหร่ อย่างไร และขนาดไหน Control variables (age, gender, education) - Avoid intervening variables ควบคุมตัวแปร เพื่อหลีกเลี่ยงตัวแปร สอดแทรก Attempt to find causal relationships - long-term cause-effect relationships หา ความสัมพันธ์ระหว่างตัวแปร
Control Remove or hold constant the effects of intervening variables ( ตัวแปรสอดแทรก ) –Elimination and removal intervening variables ( ตัวแปรสอดแทรก ) - Temperature –Holding constant - population, กำหนด ค่าคงที่ –Matching จับคู่เปรียบเทียบ –Blocking intervening variables กันตัวแปร สอดแทรก –Randomization of sampling and assigning samples to experimental or control groups สุ่มเลือกและจับกลุ่มตัวอย่างเข้ากลุ่มทดลอง –Statistical control (ANCOVA, Covariate - intervening variable)
Additional control issues Halo effects - Strong +/- impressions ความประทับใจเชิงบวก / เชิงลบ Placebo effects - fake treatment ให้ Treatment แต่จริงๆ treatment ที่ให้ไม่มีผล อะไร
Experimental designs Treatment/intervention - Experimental variable X (Indonesian Advertisement) Test, measure - Observation 0 (Dependent variables/result /like?) Randomization (sampling)
Experimental designs Full/True experimental designs Preexperimental designs Quasi experimental designs Factorial designs
Full/True experimental designs Conditions Control and experimental groups Ramdomization Intervention
Full/True experimental designs Pretest-Posttest control-group design Solomon four-group design Post-test only control group design
Pretest-Posttest control-group design Compare two groups Group 1: Test ---> Treatment ---> Test (R O --> X -- -> O) Group 2: Test ---> No Treatment ---> Test (R O > O) Group 1 = Experimental group Group 2 = Control group
Solomon four-group design Compares four groups Group 1: Test ---> Treatment ---> Test (R O --> X ---> O) Group 2: Test ---> No Treatment ---> Test (R O > O) Group 3: No Test ---> Treatment ---> Test (R X -- -> O) Group 4: No Test ---> No Treatment ---> Test (R O)
Post-test only control group design Compares 2 groups Group 1: No Test ---> Treatment ---> Test (R X ---> O) Group 2: No Test ---> No Treatment ---> Test (R O)
Preexperimental designs Treatment ---> Test (X ---> O) –One-shot case study Test ---> Treatment ---> Test (O 1 --> X - --> O 2 ) –One group pretest-posttest (before- intervention-after) Comparing two groups that one group receives treatment and the other does not –X 1 ---> O 1 = X 0 ---> O 2 ? –Static-group comparison No internal validity
Quasi Experiments Almost experiments! Exploring a research field for potential hypotheses for subsequent studies ทดลองในสนามจริง เพื่อหาสมมติฐานใน การศึกษา Not keen to generalize to the population from the sample group
Factorial Designs Factors - more than one independent variable, many independent variables that cause certain result! Could be Full, Quasi, or Preexperimental!
Good experiment Pilot study ทดสอบก่อนกับกลุ่มเล็กๆ Manipulation checks - secondary variable to check whether an experimental variable is operating ตรวจสอบตัวแปรที่ต้องการวัด ว่าใช้ได้จริง หรือไม่ Care in interpretation --->
Care in interpretation Experiments are short term เป็นการ ทดสอบระยะสั้น Any nonlinear relationship? มี ความสัมพันธ์ที่ไม่เป็นเส้นตรง Multiple dependent variables มีตัว แปรตามหลายตัว
When to use what? Time เวลา Budget งบประมาณ Samples กลุ่มตัวอย่างที่หาได้เป็น ตัวแทนประชากร ? Variables controlled ตัวแปรที่สามารถ ควบคุมได้