The sample must:
1. Have all the same relevant characteristics
2. Have tem in same proportion as the target.
3. Selective Attention Biased sampling for every
Stereotypical characteristics there are many non-stereotypical ones Enumerative induction
Particulars -> General
Parts -> whole
Form; X percent of observed members of group A have property p. X percent of all members of group A have property P.
Target Group/ Population = the whole Group
Sample/sample members observed members
Relevant property/ property
In question = Property / Characteristic
We are interested in.
Samples are weakened by:
1. Being too small -. Hasty generalization
2. Not being representative
-The more homogenous a group the smaller the sample can be.
-Less homogenous needs a bigger sample
-Social Psychological, Cultural properties - very diverse large samples Opinion Polls Should:
1. Be strong
2. Have T premises
3. Use a large enough sample
To accurately represent the target in relevant ways.
Question phrasing, order, restricted
Choices can represent bias
Random sampling selected samples at random
Will not yield accurate results
1. Agreement if two or more occurrence of a phenomenon has only one factor in common that must be the cause. 2. Difference the relevant dacoit that is present when the phenomenon occurs and absent when it doesn’t must be the cause 3 joint method of agreement and difference A identify relevant factors common to the occurrences.
Increased probability that the conclusion is T.
Controlled trials use this method
4. correlation/Concomitant Variation - Used he the cause of an occurrence seems to vary as the occurrence varies Shows possible causal relation but not a clear cause & effect relation Increase
1. Misidentifying relevant factors
2. Mishandling multiple factors
3. Being misled by...