Statistical test- Examples

Question 1

Amount of nuts: You are selling choconuts. You are concerned whether your choconuts contain sufficient nuts. You took a sample of 20 packets and find the weiht of nuts in each packet. What statistical test you will use?

• Data
• Sample
• Purpose

Answer 1

• Data: RATIO, weight of nuts, all continous variable are ratio except temperature which is an interval
• Sample : 1 ( 20 packets of choconuts)
• Purpose: to know the MEAN weight of nuts per choconuts

Student T - Test (Test for a mean)

• Data: - Interval /Ratio
• Sample : - 1 sample
• Purpose: - Testing against value

Question 2

Price Tickets: In your promotional campaign to sell your choconuts you calimed that 20% of choconuts should include tickets for free price. You take a sample of 50 packets to find that 7 of them have winnin tickets. What statistical test you should use?

• Data
• Sample
• Purpose

Answer 2

• Data: No tickets or with ticket (nominal/ binary)
• Sample: 1 (50 packets of choconuts)
• Purpose: To know HOW MANY choconuts with olden tickets (proportion)

Point Estimate P = X/n (Test for proportion)

• Data: - Nominal data
• Sample : - 1 sample
• Purpose: - Testing against a value

Question 3

Bar longevity: You think that your choconuts lasts loner than competition chocobars. You et 36 people to eat one of each and record thier eatin times. What Stastical Test you should use?

• Data
• Sample
• Purpose

Answer 3

• Data: Time to eat choconut VS time to eat chcocnuts (RATIO)
• Sample : 1 sample (36 people), 2 measure
• Purpose: Difference of 2 mean (mean time to eat choconut & mean time to eat chocobar)

Paired T Test ( Difference of 2 means , paired)

• Data: - Interval /Ratio
• Sample :- 1 sample, 2 measure
• Purpose: - Testing against value

Question 4

Defective wrapping: You think that there is a difference in performance between the 2 wrapping machines in your factory. You checked 200 bars from one machine and 150 bars from the other. You found that 10 out of 200 bars from the first machine has defective wrapping and 9 out of 150 bars from the second machine has defective wrapping.

• Data
• Sample
• Purpose

Answer 4

• Data: defective wrapping or not NOMINAL
• Sample: 2 samples (2 machines)
• Purpose: difference in 2 PROPORTIONS.

Mann- Whitney U Test (difference of 2 proportions)

• Data: - Nominal data
• Sample : - 2 sample
• Purpose: - Comparing statistics

Question 5

Do stickers help sales: you are thinkin if having free stckers will help your sales. For 13 days you offered free stickers and 10 days you did not. You compute for the sales on those days. What test you will use?

• Data
• Sample
• Purpose

Answer 5

• Data: sales on day with sticker VS without stickers (Ratio)
• Sample: 2 samples (13 days & 10 days)
• Purpose: Difference of 2 MEANS ( not paired - Mean sales on day 13 mean sales on day 10)

Independent T - Test
(Difference of 2 means, independant sample)
- Data: - Interval /Ratio
- Sample : - 2 samples
- Purpose: - Comparing statistics

Question 6

Are sales affected by temperature: You want to see if there are relationship between daily temperature and sales of choconuts . You get a data on sales and temerature and wanted to check difference in sales in 30 days.

• Data
• Sample
• Purpose

Answer 6

• Data: temperature & sales (interval & ratio)
• Sample: 1 (30 days)
• Purpose: RELATIONSHIP between temerature and sales (2 continous variable)
• PEARSON/Spearmans Correlation/// Linear Regression
• Data: - Interval /Ratio
• Sample :- 1 sample, 2 measure
• Purpose: - Looking for relation ship