Research in psychology 1 3 Experimental methods Variables

Research in psychology 1. 3 Experimental methods

Variables � A variable is any characteristics or factor that can vary, e. g: gender, age, grade points, stress, motivation, etc. � Independent variable - the IV produce a change in another variable - deliberately manipulated by the researcher - all other variables are kept constant - Example: new antidepressant medication � Dependent variable - measured after alteration of the IV - is it influenced by the IV? - Example: Did the medication affect the depression?

Operationalized �Operational definition translates an abstract term (variable) into something observable and measureable. �The operational definition gives the variable meaning within a particular study. ’ - In precise terms, what is being measured? - Example: aggression vs. how many times the participant will kick the doll during one hour

Hypothesis � Experimental hypothesis is a prediction of how the IV will affect the DV Example: the new therapy will decrease the participants anxiety more than the old one � A null hypothesis is prediction that there will be no change Example: The new therapy will have no effect on the participants anxiety compared to the old one Most often two conditions: � Experimental (treatment) condition - Situation where a variable is being manipulated � Control condition - Situation where a variable is not being manipulated � Is there a significant difference between the two?

Be a thinker p. 27 �Identify the Independent variable and dependent variable in each of the following experimental hypothesis.

Placebo �People who receiving a treatment show a change in behaviour because of their expectations, not because the treatment itself had any specific benefit

Case �Study on the effect of the new antidepressant drug �One group receives the new antidepressant drug and told they receive it – experimental condition (treatment group) �One group receives a placebo pill but told they receive the new antidepressant drug – control condition (group) �Does the antidepressant work better than the placebo?

Experiments �Laboratory experiments + easy to replicate + easy to hold variables constant - artificial environment - low ecological validity �Field experiments + Ecological validity - hard to hold variables constant �Natural experiment + Unique situations - No control over variables

Experiments �Laboratory experiments Example: Study on the effect of the new antidepressant drug �Field experiments Example: Piliavin and Rodin (1969) in the New York subway – investigated helping behaviour regarding sober or drunk person �Natural experiment Example: aggression before and after TV came, stroke victims

Confounding variables (undesirable variables that influence the IV and DV) �Demand Characteristics (aka Hawthorne effect, taken from the Hawthorne Works plant of Western Electric in the US) - Participants act differently because they are in a study and trying to guess what the researcher is after - To counteract: Use single blind control: participants are not told the aim

Confounding variables �Researcher bias (observer bias) - When expectations of the researcher affects the findings, often in subtle and unintentional ways - To counteract: Use double blind control in which both participant and experimenter are unaware if the participant is in the control group or the experimental group

Confounding variables �Participant variability �When characteristics of the sample affect the dependent varible �To counteract: use random sampling

Correlation studies – an experiment cannot be carried out but data are collected which show a relationship �Data is gathered that relates to the IV and the DV �If one variable change the other change as well �Positive correlation: - Same change in both variables - both in increase or both decrease - Example: Life expectancy and hours of exercise + 1 = perfect positive correlation �Negative correlation: - When one variable increase the other decrease - Education and time in jail - minus 1 = perfect negative correlation

Correlation studies �Example: 1. Researcher measures one variable (wealth) 2. Researcher measures a second variable (happiness) 3. The researcher statistically determines whether wealth and happiness are related. +

Bidirectional ambiguity �Cause-and-effect? �Example: Better social relationships and greater happiness are correlated �But, which causes which? = Better social relationships = greater happiness or Greater happiness = better social relationships or is there that another variable responsible for the behaviour? �Correlation between eating ice cream and drowning? ?