Reference no: EM133091402
Sampling and Surveys
Aim:
To improve understanding of the issues of bias associated with sampling, and of the principles which underpin good survey design.
Part A:
This first exercise demonstrates the issues of bias which can arise using random sampling if the sample size is small. For this exercise you will need the M&Ms Excel file provided on moodle, and a fun-size pack of M&Ms (if your tutor is unable to provide packs of M&Ms.
Open your pack and count how many of each colour appear in your pack. Enter this data into cells B3 to B8 on the Calculator worksheet of the Excel file. Then hit the Calculate Now button on the Formulas toolbar. The spreadsheet will convert these raw numbers into probabilities, and compare these against the actual distribution of colours as specified by the M&Ms manufacturer. It will convert this into a numeric score where 0 indicates against an exact match between the sample and population distributions, and higher values indicate larger variations.
Compare your result against other members of the class. Is there much variation in scores? What would this mean in terms of the accuracy of a survey carried out on the basis of these samples?
Now add together the data from several (or all) members of the class. Enter that into rows 3 to 8 in a different column of the spreadsheet, and use Calculate Now to calculate its score.
How does this value compare against the individual sample scores? Are most of them higher or lower? What implications does this have for determining an appropriate sample size to maximise the accuracy of a survey?
You may now eat your sample (unless you used the simulator - ingestion of spreadsheets is not recommended).
Part B:
Read the following page on Random Sampling and Biases provided by the Australian Mathematical Sciences Institute (AMSI), and carry out the exercises.
Part C:
Break into small groups of 2 or 3 students. As a group design a suitable sampling regime and survey instrument to investigate the topic "What factors influence the impact of social media use on self-esteem".
Consider what factors might influence someone's use of social media and its impact on their self-esteem, and ensure your survey design gathers information about those factors.
How can self-esteem be measured?
How would you gather a suitable sample to investigate this topic?
When you have completed designing your sample and survey, discuss your design with your tutor.
After the lab class
Think about your research topic. Will a survey be useful for investigating your research question? Is a mixed methods approach required? Are there issues regarding establishing a suitable representative sample?