Recipes for Science Chp2
- The role of experiments in testing hypotheses
- The main features of an experiment in an example of scientific research
- Extraneous variables and why these must be controlled in an experiment
- the problem of underdetermination and how scientists deal with it
- Identify three other uses of experiments in science besides hypothesis-testing
- Hypotheses are used to generate expectations
- Expectations are compared with observations
- That comparison is used to develop, confirm, reject, or refine a hypothesis
Experiments is used to produce and collect useful data that is used to refine a hypothesis
Comments: Cause and effect. How to measure the dependency of variable? How to tell the dependent variable ? How to tell the independent variable? Independent to what extend? Quantum struggling.
Extraneous variable : other variables that may influence the value of the dependent variable
Comments: which introduces the least number of extraneous variables?
One typical confounding variable
Observer bias (Hawthorne effect)
- Place and period of time
- Individual or collaborative experiments
- Background conditions
The choice of method for data collection depends on one’s research interests, the hypothesis, the variable of interest and the available instruments.
- Data collection involves one or more specialized instruments.
- scientists don’t always come to agreement with the choice of instruments.
- Measuring instruments can produce measurement error and require calibration.
Qualitative data can be measured by in quantitative form.
An Experiment that decisively adjudicates between two hypotheses, settling once and for all which is true.
Defintion: the underdetermination of hypotheses by data is the case that evidence not sufficient to determine which of multiple hypotheses is true.
=> there is always be some untested hypotheses that weren’t tested
=> for this reason, few experiment can be crucial experiments that once and for all settle the hypotheses
One hypothesis involves a number of auxiliary assumptions
(Assumptions that need to be true in order for the data to have the intended relationship to the hypothesis under investigation)
Three sources of uncertainty about what an experiment shows:
- Extraneous variables
- Unanticipated hypotheses
- Auxiliary assumptions
Replication - to minimize uncertainty ::how?::
Comments: replications in which experiment produces the same result doesn’t produce more information. Unsuccessful replications did produces new information ::(?)::
- Confirming and disconfirming evidence
- Evaluating the functioning of scientific instrument
- Determining the physical constants
- Exploratory experiment
- Identify the features of a perfectly controlled experiment and characterize the importance of each
- Describe the difference between direct and indirect variable control
- Describe the steps to conducting a perfectly controlled experiment of a given hypothesis
Expectations should be clearly and precisely defined.
It offers a measurable way to proceed testing
Intervention: a direct manipulation of the value of a variable
Intervention is the center of perfectly controlled experiment
However, some intervention can not be conducted due to ethical reasons i.e. syphilis and practical reasons i.e. distance from Moon to Earth
- Direct variable control : all extraneous variables are held at constant values during an intervention
- Indirect variable control: allow extraneous variables to vary in a way that s independent from the intervention
Double blind experiment
- Distinguish between lad and field experiments
- Define external validity and internal validity
- Describe the main types of non-experimental design
- High degree of internal experimental validity
Def: scientists can correctly infer conclusions about the relationship between the independent and dependent variables with great certainty
- Experimental setup and data analysis are easily to be followed and repeated
External experimental validity
The extent to which experimental results generalize from the experimental conditions to other conditions
1. Population validity
2. Ecological validity