Introduction to Statistics: Investigating and Organizing Data

Section 1.3: Investigating Data

The Statistical Investigation Cycle

The process of statistical investigation is a systematic approach to answering questions using data. This cycle ensures that data analysis is purposeful and scientifically valid.

• Ask Questions: Begin with a clear, focused research question. Good questions guide the entire investigation.

• Consider Data: Identify what data is available or needs to be collected to answer the question. Consider the type, source, and quality of the data.

• Analyze Data: Use appropriate methods to summarize and visualize the data. This may include creating graphs, tables, or calculating summary statistics.

• Interpret Data: Draw conclusions based on the analysis, relating findings back to the original question.

Example: Suppose you have a dataset of nearly 10,000 movies with variables such as title, release date, release_day_of_week, vote_average, and genre. To investigate whether movies released on Fridays receive higher ratings, you would:

• Ask: Do movies released on Friday get higher ratings than those released on other days?

• Consider Data: Use release_day_of_week and vote_average variables.

• Analyze: Compare average ratings by day of release, possibly using graphs or summary statistics.

• Interpret: Decide if there is evidence that Friday releases are rated higher.

Additional info: This cycle is foundational for all statistical studies, ensuring that conclusions are data-driven and reproducible.

Section 1.4: Organizing Categorical Data

Why Organize Data?

Organizing data is essential for making sense of raw information. Proper organization allows for easier analysis, clearer communication of results, and more accurate conclusions.

• Clarity: Well-organized data helps identify patterns and relationships.

• Efficiency: Summarized data is easier to analyze and interpret.

• Comparison: Organized data allows for meaningful comparisons between groups.

Two-Way Tables

A two-way table (also called a contingency table) is a common method for organizing data involving two categorical variables. It displays the frequency of all combinations of the variables.

• Purpose: To summarize and compare the relationship between two categorical variables.

• Structure: Rows represent categories of one variable; columns represent categories of the other.

Example: The following two-way table summarizes responses from the Youth Risk Survey (ages 14-17) about seat belt use by gender:

	Not Always	Always
Male	2	3
Female	3	7

• Variables: Gender (Male, Female) and Seat Belt Use (Always, Not Always).

• Total Males: $2 + 3 = 5$

• Total Females: $3 + 7 = 10$

• Total Not Always: $2 + 3 = 5$

• Total Always: $3 + 7 = 10$

• Sample Size: $2 + 3 + 3 + 7 = 15$

Calculating Percentages from Two-Way Tables

• Percentage of Males: $\frac{5}{15} \times 100\% = 33.3\%$

• Percentage of Females: $\frac{10}{15} \times 100\% = 66.7\%$

• Percentage Not Always Wearing Seat Belt: $\frac{5}{15} \times 100\% = 33.3\%$

• Percentage Always Wearing Seat Belt: $\frac{10}{15} \times 100\% = 66.7\%$

Comparing Groups: To determine if males are more likely than females to not always wear a seat belt, compare the proportions within each gender:

• Males Not Always: $\frac{2}{5} = 40\%$

• Females Not Always: $\frac{3}{10} = 30\%$

Conclusion: In this sample, males are more likely than females to not always wear a seat belt.

Working with Percents

• Finding Number from Percent: If 30% of 300 students do not always wear a seat belt:

$300 \times 0.30 = 90$ students

• Finding Total from Number and Percent: If 43 students represent 20% of a class:

$\text{Total students} = \frac{43}{0.20} = 215$

Another Two-Way Table Example: Gender and Handedness

	Left	Right
Male	10	40
Female	5	35

• Total Students: $10 + 40 + 5 + 35 = 90$

• Left-handed Students: $10 + 5 = 15$

• Percent Left-handed: $\frac{15}{90} \times 100\% = 16.7\%$

• Percent of Males Left-handed: $\frac{10}{50} \times 100\% = 20\%$

Incomplete Data and Comparing Groups

When comparing groups, it is important that the groups are similar in size and context. Percentages or rates are often more informative than raw counts, especially when group sizes differ.

Example: The following table shows the number of injuries by sport:

Sport	Injuries
Baseball	165,842
Basketball	501,251
Bowling	20,878
Football	451,961
Ice hockey	19,035
Soccer	208,214
Softball	121,175
Tennis	23,611
Volleyball	60,159

Note: To determine which sport is most dangerous, you would ideally compare injury rates (injuries per participant), not just total injuries. Raw counts can be misleading if participation rates differ.

Additional info: When data is incomplete, always consider what additional information is needed for a fair comparison (e.g., number of participants in each sport).