The “Wisdom of Crowds” concept consists of the belief that if a crowd guesses a value and the average of their guesses is found, it will be somewhat close to the real value. In general terms, the “Wisdom of Crowds” concept explains that the “crowd” collectively has greater accuracy than the average individual. Even though there may be certain individuals with a guess closer to the real value, the chance of those particular individuals being selected at random is very low. Historically, there is a story about a man, Sir Francis Galton, who found that the average of 787 guesses concerning the weight of an ox was only a single pound away from the actual weight. While the “Wisdom of Crowds” theory will show reliably accurate results most of the time, it can still be flawed. This could happen for a number of reasons, but a notable factor is that if the data subjects are not independent of each other, or in other words, some of the results affect the others, this can cause the average result to become significantly less accurate. It is thought that in order for these crowds to be accurate, the crowds must have diverse opinions, and their opinions must not be influenced by the other members in the crowd.
Random Samples
A random sample is a smaller sample taken from a larger group or population that ideally does not contain any bias. The purpose of a random sample is to determine the chance/number of a chosen condition within a given population, without having to survey the entire population. It is vital that the random sample is as random as possible to ensure that it accurately represents the population as a whole. Researchers have to create a random sample to make it as representative of the population as possible. The idea is that from their randomly picked sample every male and female, for example, would have an equal chance of being picked. The drawback is that this process can be costly and time-consuming for researchers. It can also be difficult to gain access to a list of a larger population. There are different types of random sampling. The most common type is simple random sampling. For this type of probability, everyone in the entire target population has an equal chance of being selected, the selection occurs until the desired sample size is achieved. Examples of this include raffles and drawing names from a hat. Stratified random sampling is more controlled than simple random sampling. This method divides the populations into strata. These subgroups may be determined by age, ethnicity, gender or any other dividing characteristic. The required sample size of each subgroup is generally designed to be representative of the known proportions in reality. This does take longer than simple random sampling but the advantage to this is that researchers are able to achieve sufficient sample sizes for key groups of people.
Save your time!
We can take care of your essay
- Proper editing and formatting
- Free revision, title page, and bibliography
- Flexible prices and money-back guarantee
Sources of Error
There are many errors which can occur in random sample research. Unrepresentative samples can negatively affect the reliability of the results. Undercoverage occurs when the sample may not be representative of the population, meaning there may not be the right balance between ages, ethnicities and genders, for the research to be generalised to the wider population. Another significant sample flaw is size, the larger the sample size, the more reliable the results will be. Statistically speaking, an experiment containing 100 people would have much more reliable results than a study with only 10. In a small random sample, outliers have the tendency to drastically change the outcome This is because as the sample size increases the effects of chance and extraneous variables are decreased. Something else that should be noted is that the sample members are the ones who choose to volunteer. This is notable as it gives a bias to the results where the entire sample consists of the type of people who are willing to volunteer, generally these people have certain characteristics such as being more outgoing. Random sampling helps to eliminate the voluntary response bias. Another source of error is participant error. Sample members may not participate correctly in the research or give fake results. These participants can alter the results drastically and prevent the results from being valid.
Sources of Bias
There are many different sources of bias, but as stated before, it is important to make sure that the results do not affect one another. Both the researcher and participant bias can have a big impact on the results. Researchers can ask leading questions, for example, the wording or emphasize parts of the question may have an effect on the participants. Even a tone can alter the participants' decisions to respond differently in a way that benefits the researcher. Participant bias can also be detrimental to a study. The ‘Hawthorne Effect’ is the concept that participants are likely to change their behaviour due to the knowledge that they are in an experiment. Participants are likely to give biased answers in order to impress or satisfy the researcher. Another factor is known as social desirability. This the concept that most people will be reluctant to admit unfavourable opinions as they don’t want to be viewed negatively. This can result in dishonest answers.
Limiting Bias in Random Samples
The two most important ways to remove bias is to make sure that the selection process is truly random, and that the sample is large enough so that irregular results do not heavily affect the average. To ensure that the sample is random, each subject of the group or population has to have an equal chance of being selected. Unfortunately although increasing the sample size is helpful in reducing the sampling error, it doesn't affect sample bias. Single-blind and double-blind experiments can be used to try and eliminate some of the bias variables.
Methodology
To test the concept of the ‘Wisdom of Crowds’ I designed an experiment surveying the year 12’s at Duncraig Senior High School. I selected as many random students as I could. Each student was instructed to guess how many students they believed were in the school. Students were asked to guess the total number of males and the total number of females. These numbers were ultimately combined in the data. It should be noted that students were able to hear each other's answers. This is likely to have caused some sources of error and bias in our experiment.
Discussion of Results
The combined results from the experiment conducted supported the “Wisdom of Crowds” concept. This was proven in the results as the correct number of total students was 1642, and the average number guessed was approximately 1660. Most of the male component of the sample seemed to believe that there was a nearly equal amount of male and female students, as shown in graph 1, but when the female component of the sample was guessing the number of female students, the standard deviation was higher, implying that this aspect was more divisive within the sample.
Stuck on your essay?
