grobman, k. h. "confirmation bias." teaching about. developmentalpsychology.org, 2003....

Grobman, K. H. "Confirmation Bias." Teaching about. Developmentalpsychology.org, 2003. Web. 16 Sept. 2012. <http://www.devpsy.org/teaching/method/confirmation_bias.html>.

SequenceFits the instructor's Rule?

Guess the instructor's Rule

How Sure?

2,4,6 :-)______________________

_______%

__,__,__ ______________________

_______%

__,__,________________________

_______%

__,__,________________________

_______%

__,__,________________________

_______%

__,__,________________________

_______%



How Sure?

2,4,6 :-) counting up by two's 50%






• Grobman, K. H. "Confirmation Bias." Teaching about. Developmentalpsychology.org, 2003. Web. 16 Sept. 2012. <http://www.devpsy.org/teaching/method/confirmation_bias.html>.



How Sure?


10,20,30 :-) counting up by multiples 60%

100,500,894 :-)counting up with all even numbers

70%

1,9,20 :-) counting up 80%

27,13,4 :-( counting up 90%

55,2,999 :-( counting up 100%

• Grobman, K. H. "Confirmation Bias." Teaching about. Developmentalpsychology.org, 2003. Web. 16 Sept. 2012. <http://www.devpsy.org/teaching/method/confirmation_bias.html>.

Methods of Good Scientific Practice

Any study done carefully and based on observation is scientific.

Science must follow certain rules.

The rules of science make the scientific process as objective as is possible.

Objective = Not influenced by feelings, interests and prejudices; UNBIASED

vs.\Subjective = Influenced by feelings, interests and prejudices; BIASED

Science can be done poorly.

• Science can be done poorly, just like any other human endeavor.

• Quality control mechanisms in science increase the reliability of its product.

Anything done scientifically can be relied upon to be accurate and reliable.

Different scientists may get different solutions to the same problem.

• Results can be influenced by the • race, • gender, • nationality, • religion,• Politics,• economic interests of the scientist.

• 2 types of bias can result in different solutions to the same problem:• Sampling Bias• Measurement Bias

People need to be able to evaluate scientific information in order to make informed decisions about:

• Health care• Environmental issues• Technological advances• Public health issues

Knowledge of what science is, what it can and cannot do, and how it works, is important for all people.

Objectivity is the key to good science.

To be objective, experiments need to be designed and conducted in a way that does not introduce bias into the

study.

What is good science?

• A prejudiced presentation of material

• A consistent error in estimating a value

Two main types of bias:

1. Sampling bias

2. Measurement Bias

Bias =

Sampling BiasSample = A group of units selected to be “measured” from a larger group (the population).

Sampling bias is introduced when the sample used is not representative of the population or inappropriate for the question asked.

1. Use a RANDOM SAMPLE = every individual has an equal likelihood of being chosen.

2. Limit the question asked to the specific group sampled.

SAMPLE SIZE: Is the sample big enough to get a good average value?

SELECTION OF SAMPLE: Does the composition of the sample reflect the composition of the population?

Factors that contribute to sampling bias

Factors such as location, age, gender, ethnicity, nationality and living environment can affect the data gathered.

How to minimize sample selection bias:

Measurement Bias

Is the method of data collection chosen in such a way that data collected will best match reality

Evaluate the technique:

• - Were measurements taken accurately?

• - Were there any additions to the environment that will influence results?

• - Was the experiment designed to isolate the effect of a single factor?

Good science depends on a well-designed experiment that minimizes bias by using the appropriate:

•Sample size•Sample selection •Measurement techniques

***for the question being investigated

Summary

Bias in Clinical Trials• Clinical trials-

Studies to test the effectiveness of medications and treatments. The subjects are observed for the effects.

• Sample size –

the larger the sample size the better

• Placebo effect-

A beneficial effect produced by a placebo drug or. treatment that can not be attributed to the properties of the placebo itself. Therefore is due to the patients belief in the treatment

Tunnel vision- When a scientists goes into the experiment with preconceived notions that can effect how he/she interprets the results. While looking FOR certain things, they may miss other things.

Blind studies- group. When the subject does not know if they are part of the treatment group or the control

Double blind studies- When neither the subject nor the scientists know who is receiving the treatment and who is the control.

Independent duplication = Two or more scientists from different institutions investigate the same question separately and get similar results.

Peer-reviewed journal = A journal that publishes articles only after they have been checked for quality by several expert, objective scientists from different institutions.

The scientific community engages in certain quality control measures to eliminate bias.

Results are verified by independent duplication and publication in a peer-reviewed journal.

Identifying good science: Look for signs of bias!

• Language

• Appropriate data reported to back conclusions

• Data source

Language“Scientifically-proven”

* Science does not seek to prove but to disprove * Be suspicious of this claim!

Emotional appeals

* Conclusions should be data-based

* Emotional appeals usually are not data-based

Strong language

* Scientific conclusions should only report what the data supports. * Words should be chosen very carefully to avoid exaggeration or claims

not supported by data.

THE DATA SHOULD CONVINCE YOU, NOT THE WORDS USED!

Appropriate data reported to back conclusions

Are samples and measurements appropriate for the conclusion presented?

Are multiple factors properly accounted for to justify the interpretation of the data?

Data Sources

All organizations produce unbiased data. However, it is important to understand the organization’s motivation to be able to identify potential bias. In some situations, the need to promote special interests or make profits may lead to bias.

1. University Research

2. Corporate Research

3. Government Research

4. Research by Special Interest Groups

grobman, k. h. "confirmation bias." teaching about. developmentalpsychology.org, 2003....

Documents