Explain how does bias, use of language, ethics, cost, time and timing, privacy, cultural sensitivity influence the collection of data? Give examples.

1. Bias
If a survey or study is done on only one group of people, and all of the people have similar opinions, the statistics are not reliable. For instance, if all of the people surveyed about a laundry detergent were big fans of it, then the laundry detergent would get a good rating, which is what the company wants, but if this information would be put into an advertisement, it would be misleading.

2. Use of Language
The way a question on a survey is put, the way headings are written down on charts, or the way slogans are made can be changed to say many different things with the same words. These things can be positive or negative, whatever is desired. For example, if a company writes for an advertisement (to illustrate 55% of the population), “over half of the population uses and loves this product”, then that gives the company a good image. If they write, “only 55% of the population uses this product”, then that gives them a bad image. Both of these statements are trying to show 55% of the population, but have very different effects on the people reading the advertisement.

3. Ethics
If the whole group of people surveyed or studied had similar beliefs (cultural, religious, etc.) they may all have similar opinions. For instance, if only Muslims were surveyed about their thoughts on Ramadan (the holy month in Islam), they may give similar answers, making the data unreliable. On the contrary, if Muslims, Christians, and Sikhs were surveyed, the data may be more reliable.

4. Cost
Money is a big factor in surveys. If whoever is conducting the study or survey does not have a lot of financial support, they will probably stick with a smaller group of people to survey and round numbers and percentages up to make them sound better (or worse, if that is what’s needed). If a company only surveyed 100 people because they were not able to get financial support to survey 1000, and 45 of these people said one thing and 55 said another, they may round these numbers to 50 and 50 because they do not want to survey more people to get more accurate percentages.

5. Time and timing
Statistics can vary throughout the year, for example, the amount of money spent in clothes and electronics stores rises near the holidays, and the revenue of ice cream shops is greater in the summer months. Also, if the people asked/surveyed were all in the same mood, for example, they were all upset about something that had just happened in their area, the data will be very different than if all of the people asked were happy at the time.

6. Privacy
If there is a survey conducted concerning private matters, then it is very hard to show the results accurately. Either the people surveyed do not want to give this information, or whoever is conducting the survey is not permitted to distribute this information, like in the case of a survey about house addresses, for example. So, privacy can affect the information that is presented as the results of the survey.

7. Cultural sensitivity
Companies need to be aware that cultural differences and similarities exist, otherwise, the collection of data for surveys or studies is affected. For instance, if a survey is done about the supermarkets that the people of one region in Canada visit most often, the results will be very different because there are many different cultures in Canada. Some visit Chinese supermarkets, Russian stores, or Japanese shops. But, if the same survey was done in China, the results would be very similar because there are not as many cultural differences there.

Explain the difference between a population and a sample. Give examples.

1. A population is a group of people or objects from which information is taken (a business, a cultural group, etc.) that has at least one characteristic in common. For example, a survey could be conducted on only girls in a school or only apples in a store.

2. A sample is a subset of a population that is taken from the population if it is not possible or practical to collect information from that whole population. For example, a study could be conducted on only the girls that had at least one sibling or only green apples.

Both inferences and hypotheses can be made about data collected from both a population and a sample. Also, the more specific the characteristic gets, the more unreliable the information is, as the statistics are not taken from a big group of people or objects.

Explain the different types of sampling methods and the benefits of each. (Convenience sample, random sample, stratified sample, systematic sample, and voluntary response sample.) Give examples.

1. Convenience sample
A sample created by choosing individuals from the population who are easy to access.

BENEFITS
– You get to choose who you are surveying
– You control the number of people you are surveying
– Saves time (because of small amount of people)
– Ease of availability

EXAMPLE
A restaurant owner wants to know the favourite pizza topping of customers. He plans to survey every customer who ordered pizza at his restaurant between 5pm and 10pm one evening. This is a convenience sample because it is not random, however the sample does not target customers that will provide enough useful input. These customers easily accessible.

2. Random sample
A sample created by choosing a specific number of individuals randomly from the whole population. Each individual has an equal chance of being chosen, and those who are chosen will represent the whole population. This data can be used to make predictions about the whole population. Stratified sample and Systematic sample are both versions of random samples.

BENEFITS
– Only a small number of people surveyed, not a lot of variety in opinion
– Free from bias and prejudice
– Simple

EXAMPLE
A teacher wants to get feedback from her class about the school dance, she plans to survey 5 students out of her class of 28. She put all of the students names into a box and draws five names. This is a random sample.

3. Stratified sample
A sample created by dividing the whole population into distinct groups, and then crossing the same fraction of members from each group.

BENEFITS
– The sample highly represents the opinion of the people
– Greater precision than a random sample of the same size
– Because of greater precision only need a small sample to survey
– Can guard against an unrepresentative sample (e.g. all male sample from a mixed gender population)

EXAMPLE
a chain store is trying to decide whether to open a store in Camrose, Alberta. The company decides to survey people in Camrose and the nearby towns. The population of each location is as follows: Camrose – 16000, Bashaw – 825, Tofield – 1876, and Daysland- 876. Since the city has more people who use the stores products than the nearby towns they decided to use a stratified sampling method. The surveyed 25% of the population in each town and the city. This is a stratified sampling method since 25% of each group is surveyed, the same proportion of each town is represented in the sample.

4. Systematic sample
A sample created by choosing individuals at fixed intervals from an ordered list of the whole population.

BENEFITS
– Simplicity
– Allows some control over the sample, a degree of system or process to the random selection
– Assures that the population will be evenly sampled
– Time and cost effective

EXAMPLE
A telephone company wants to determine whether a fitness centre would be well used by 3000 employees. The company plans to survey 300 out of their 3000 employees. To ensure that the population is fairly represented they choose to survey every tenth person on the payroll list. This is a systematic sample.

5. Voluntary response sample
A sample created by inviting the whole population to participate.

BENEFITS
– less effort
– No sampling frame required
– Open opinions from whoever wants to speak (sometimes an advantage)
– Very honest report (sometimes a disadvantage)

EXAMPLE
A marketing research company mails surveys to all of the adult residents in town. The survey asks about consumer products. The residents are asked to mail their response in the prepaid envelope. The marketing research company is inviting all residents to participate. This is a voluntary response sample.
Explain how choosing an inappropriate sampling method may bias the data. Give Examples.

– Using an inappropriate sampling method may bias the data because it can give an inaccurate survey of the population. An example of this is if Ellen was to do a survey using a voluntary response sample on her show collecting data on what television talk show people like the best. By doing this you can assume that most people calling in or responding to the survey would say that the Ellen show is the best because that’s what they were watching when they found out about the survey. This would show a complete bias in the data and you wouldn’t be able to get an accurate understanding of the total population’s opinion. Instead you could use a random or stratified sampling method and get a better understanding of the population’s opinions as a whole, not just a specific group’s opinions.

– Another example of how using an inappropriate sampling method can bias data is in Telephone sampling. Telephone sampling is often used in marketing surveys. A random sample method may determine the list of telephone numbers of people in the area being surveyed. This method does involve a random sample, but will not give an accurate reading of the target population’s opinion. This is because it will miss people that don’t have a phone, or whose phone numbers are from different area codes. It will also miss people that do not wish to be surveyed. This type of sampling method excludes certain types of consumers in the area. Overall this method of sampling could show a bias in the data because you are only getting one ‘type’ of person’s opinions, people with phones that wants to be surveyed in a particular area code.
Explain the difference between theoretical and experimental probability. Give examples.

1. Theoretical probability is the expected results of a study, research, or survey. This is not reliable as the actual statistics or results may turn out quite different from what was expected. For example, a clothing store (for women and men) was expecting that 50% of their customers were male and 50% were female. But, when they did a study, they found out that 80% of their customers were female and 20% were male. These results were very different from the theoretical probability.

2. Experimental probability is the actual results of a study, research, or survey. So, in the example above, the fact that 80% of the customers were female and 20% were male is the experimental probability.

As more trials are made, experimental probability becomes closer to statistical probability. So, if the survey above was conducted in many different stores, perhaps the percentages would be much closer to 50 and 50, like 55% and 45%, for instance. Of course, the stores studied would have to be stores that carry clothing for both women and men.

Find 3 examples of misleading statistics used in the media and explain why they are misleading.

1. Book Graph

– Size is not on any particular scale
– Large spaces between the different percentages
– Number of books does not match up with the percentage
They wanted the viewers to believe that a lot more students over the years were earning their high school diplomas than ever before when in fact the the numbers had only increased slightly, and were not as dramatic as the graph portrayed them to be.

2. Abortions Graph

– No y axis scale
– The way the arrows are positioned is presented to believe that 327,000 is a larger number than 935,573
They wanted the viewers to believe that the number of abortions had skyrocketed compared to the cancer screenings and prevention services that were falling when in fact the numbers clearly show otherwise.
3. The Times

– chart does not start at zero on the y axis
– show that The Times is over 50% more popular when in fact it is only about 10%
They wanted viewers to believe that the number of people that liked The Times was substantially greater than those who liked other daily telegraphs.